Citation
Novel hemodynamic parameters in pulmonary arterial hypertension

Material Information

Title:
Novel hemodynamic parameters in pulmonary arterial hypertension
Creator:
Schafer, Michal ( author )
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (114 pages). : ;

Subjects

Subjects / Keywords:
Pulmonary artery -- Diseases ( lcsh )
Pulmonary hypertension ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
The diagnostic gold standard for the evaluation of the pulmonary arterial hypertension (PAH) remains to be the right heart catheterization. The emerging possibility of using non-invasive imaging methods for both diagnostic and prognostic clinical purposes is currently frequently investigated field. In this work, three novel hemodynamic parameters, vorticity, wall shear stress (WSS), and helicity are investigated as potential markers of pathophysiologic processes and fluid-tissue interactions. The computed data from the 4D-flow CMR acquisition can reveal important associations with the catheterization and echocardiographic measurements and serve as diagnostical indicators. Two sets of patient groups were investigated in this work (2x n=23). Both cohorts had pre-designed specific research protocols. First group of patients under went same-day CMR and echocardiography, while the second one CMR and catheterization. Specific post-processing protocols with respect to cardiac and vascular domain were created for computation of novel hemodynamic factors. Each novel parameter provided large set of significant correlations with concomitant diagnostic test and presented significant inter-group variations as well. Vorticity computed in cardiac chambers in diastolic phase revealed large number of specific and ventricular interdependent correlations. The WSS provided the strongest correlations with standard clinical determinants mean pulmonary arterial pressure and pulmonary vascular resistance (mPAP: rho = -0.7152, p = 0.0003; PVR = - 0.8031, p < 0.0001). Helicity measured in proximal pulmonary vasculature on the other hand showed significant correlations with the ventricular-vascular coupling ratio Ea/Emax (rho = -0.7696, p = 0.0001). Lastly, the novel hemodynamic markers were considered for the diagnostic power test, using ROC and 3D plot algorithm revealing significant deterministic strength of the data. Every patient in specific cohort was diagnosed properly solely based on the two-method inclusion system. One patient was misdiagnosed due to boundary diagnostic state. The computed hemodynamic indexes can then enable evaluating physician to analyze both cardiac and vascular function as a separate systems, and also as complex coupled system with sensitive description of morphogenesis patterns.
Thesis:
Thesis (M.S.)--University of Colorado Denver. Bioengineering
Bibliography:
Includes bibliographical references.
System Details:
System requirements: Adobe reader.
General Note:
Department of Bioengineering
Statement of Responsibility:
by Michal Schafer.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
913525422 ( OCLC )
ocn913525422

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

NOVEL HEMODYNAMIC PARAMETERS IN PULMONARY ARTERIAL HYPERTENSION by MICHAL SCH€FER B.S c ., Colorado School of Mines, 2013 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Bioengineering 2015

PAGE 2

! ii This thesis for the degree of Master of Science by Michal SchŠfer has been approved for the Bioengineering Program by Kendall S. Hunter Chair J. Kern Buckner Advisor Brett E. Fenster Advisor Michael E. Yeager Vitaly O. Kheyfets April 2 3 2015

PAGE 3

! iii SchŠfer, Michal (M.S ., Bioengineering Program) Novel Hemodynamic Parameters in Pulmonary Arterial Hypertension Thesis directed by Assistant Professor Kendall Hunter ABSTRACT The diagnostic gold standard for the evaluation of the pulmonary arterial hypertension (PAH) remains to be the right heart catheterization. The emerging possibility of using non invasive imaging methods for both diagnostic and prognostic clinical purposes is currently frequently investigated field. In this work, three novel hemodynamic parameters, vorticity, wall shear stress (WSS), and helicity are investigated as potential markers of pathophysiologic processes and fluid tissue interactions. The computed data from the 4D flow CMR acquisition can reveal important associations with the catheterization and echocardiographic measurements and serve as diagnostical indicators. Two sets of patient groups were investigated in this work (2x n=23). Both cohorts had pre designed specific research protocols. First group of patients under went same day CMR and echocardiography, while the second one CMR and catheterization. Specific pos t processing protocols with respect to cardiac and vascular domain were created for computation of novel hemodynamic factors. Each novel parameter provided large set of significant correlations with concomitant diagnostic test and presented significant i nter group variations as well Vorticity computed in cardiac chambers in diastolic phase revealed large number of specific and ventricular interdependent correlations. The WSS provided the strongest correlations with standard clinical determinants mean pul monary arterial pressure and

PAGE 4

! iv pulmonary vascular resistance (mPAP: rho = 0.7152, p = 0.0003; PVR = 0.8031, p < 0.0001). Helicity measured in proximal pulmonary vasculature on the other hand showed significant correlations with the ventricular vascular co upling ratio E a /E max (rho = 0.7696, p = 0.0001). Lastly, the novel hemodynamic markers were considered for the diagnostic power test, using ROC and 3D plot algorithm revealing significant deterministic strength of the data. Every patient in specific coh ort was diagnosed properly solely based on the two method inclusion system. One patient was misdiagnosed due to boundary diagnostic state. The computed hemodynamic indexes can then enable evaluating physician to analyze both cardiac and vascular function a s a separate systems, and also as complex coupled system with sensitive description of morphogenesis patterns. The form and content of this abstract are approved. I recommend its publication. Approved: Kendall Hunter

PAGE 5

! v ACKNOWLEDGEMENTS Anschutz Medical Campus (AMC) gave me a new home and opportunity to conduct so far the most important work in my life. However, I would like to quote one of the greatest minds of human mankind, sir Isaac Newton: If I have done great things, its only becau se I have stood on the shoulder s of giants ". While the great road is ahead of me to maybe one day achieve something great, I have already had greatest honor of my life to meet the academic giants like Dr. Hunter, Dr. Kheyfets, and Dr. Yeager. Dr. Hunter ha s given me chance to study at this university and I will never forget his ever willingness to help me and his passion towa rd being the winner (even in cooking competitions) plus I seriously doubt that I will ever meet a better bioengineer In Dr. Kheyfets I found a new mentor and inspiration, since he was always there for me when I needed to translate science and mathematics into plain language. H e tasted most my clumsiness of all of the advisors, especially during the sport days. The last AMC giant in whom I found great teacher and mentor is Dr. Yeager who taught me one of the most important lesson s in my life, that playing a second scalpel in the world of academic medicine and science means not to play at all. While it is import ant to have great mentors and teachers for guidance through out the early academic steps, it is absolutely vital to have someone who will help you to pointing to the right way and be there for you when the first academic hits come. At AMC I have met Dr. Mc Mahon and I can completely with pure honesty say that without her I would never be able to survive a single week at school. In my class I have met a great friend Chelsea Daniels, who became my closest academic co fighter, and huge thanks to her mainly for helping me with this work and mainly everyday school life. In

PAGE 6

! vi order to understand the adults, one must try to understand children first, and Dr. Truong happens to be the best chaperon in this realm. Dr. Bodine provided necessary directive, support, and pro fessionalism together with important ethical research values which all are becoming more and more scarce. Lastly, there is National Jewish crew who I must thank. Number one, Dr. Fenster who gave me the opportunity of my life to become part of his team, an d who was always there for me to explain the most complicated physiologic phenomena in elegant way. Dr. Fenster had to deal with my impatience, and at the same time it is manly him who is pushing me forward, and if I will ever become a physician I would li ke to be at least as half as good as he is. For the very last I would like to thank Professor Buckner, who is simply the Ace of heart. Professor Buckner is mentoring me for over 4 years already and he never gave up on me. He was there for me at my worst an d best times, and he showed me that if I will understand the human heart, then I will understand myself as well

PAGE 7

! vii TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ......................... 1 Def inition, Epidemiology, Etiology ................................ ............................... 1 Pathophysiology ................................ ................................ ............................. 2 Standar d Diagnostic Imaging Techniques ................................ ..................... 5 A dditional Imaging Diagnostical Measures ................................ ................... 6 Biomarkers ................................ ................................ ................................ ..... 9 Therapy ................................ ................................ ................................ ........ 10 Novel Hemodynamic Indexes ................................ ................................ ...... 12 II. VORTICITY ................................ ................................ ................................ 16 Theory ................................ ................................ ................................ .......... 16 Methods ................................ ................................ ................................ ........ 19 4D CMR + 2D CMR Protocols ................................ ............................. 20 Right Heart Catheterization ................................ ................................ .. 25 Echocardiography ................................ ................................ ................. 26 Statistical Analysis ................................ ................................ ................ 27 Results I Siemens Group ................................ ................................ ........... 28 Results II Butcher Group ................................ ................................ .......... 40 Discussion I Siemens Group ................................ ................................ ..... 41 Discussion II Butcher Group ................................ ................................ .... 50 Conclusion ................................ ................................ ................................ ... 52 III. WALL SHEAR STRESS ................................ ................................ ............. 54

PAGE 8

! viii Theory ................................ ................................ ................................ ......... 54 Methods ................................ ................................ ................................ ....... 57 Results ................................ ................................ ................................ ......... 62 Discussion ................................ ................................ ................................ ... 65 Conclusion ................................ ................................ ................................ .. 68 IV. HELICITY ................................ ................................ ................................ .. 70 Theo ry ................................ ................................ ................................ ......... 70 Methods ................................ ................................ ................................ ....... 73 Results ................................ ................................ ................................ ......... 74 Discussion ................................ ................................ ................................ ... 77 Conclusion ................................ ................................ ................................ .. 79 V. DIAGNOSTIC ALGORIT HM ................................ ................................ ... 81 Introduction ................................ ................................ ................................ 81 Receiver Operating Curves ................................ ................................ ......... 82 Diagnostic Algorithm ................................ ................................ .................. 87 Conclusion ................................ ................................ ................................ .. 91 REFEREN CES ................................ ................................ ................................ ..... 93 LIST OF TABLES

PAGE 9

! ix TABLE 1. Study group characteristics ................................ ................................ .......... 19 2. Abbreviations of the CMR ma rkers ................................ ............................. 25 3. Abb reviations of the RHC markers ................................ .............................. 26 4. Vorti city Measures Siemens Group ................................ .......................... 29 5. RA maximum vorticity ................................ ................................ ................ 31 6. RA minimum vorticity ................................ ................................ ................. 31 7. RA E wave vorticity ................................ ................................ ..................... 31 8. RA A wave vorticity ................................ ................................ .................... 32 9. RV maximum vorticity ................................ ................................ ................ 32 10 RV minimum vorticity ................................ ................................ ................. 33 11. RV E wave ................................ ................................ ................................ ... 33 12. RV A wave ................................ ................................ ................................ ... 3 3 13 LA maximum vorticity ................................ ................................ ................ 34 14 LA minimum vorticity ................................ ................................ ................. 34 1 5. LA E wave vorticity ................................ ................................ ..................... 35 16. LA A wave vorticity ................................ ................................ .................... 35 17. LV minimum vorticity ................................ ................................ ................. 36 1 8. LV E wave vorticity ................................ ................................ ..................... 36 1 9. LV A wave vorticity ................................ ................................ .................... 37 20 LA Ejection fraction ................................ ................................ .................... 38 21. LV Eccentricity ................................ ................................ ............................ 39 22. Hea rt rate multivariate analysis ................................ ................................ ... 39

PAGE 10

! x 23. Vorticity in vasculature ................................ ................................ ................ 40 24. Vascular vorticity ................................ ................................ ......................... 41 25. Wall shear stress results ................................ ................................ ............... 62 26. Wall shear stress multivariate ana lysis ................................ ........................ 65 27. Helicity results ................................ ................................ ............................. 75 28. H elicity multivariate analysis ................................ ................................ ...... 76 LIST OF FIGURES

PAGE 11

! xi FIGURE 1. PAH stage progression ................................ ................................ ............... 3 2. Stress strain curve ................................ ................................ ...................... 4 3. PAH vicious cycle ................................ ................................ ...................... 5 4. P V loops and VVCR ................................ ................................ ................. 8 5. Pharmacologic therapy ................................ ................................ ............. 12 6. Vorticity integrating regions ................................ ................................ .... 22 7. Systolic eccentricity ................................ ................................ ................. 23 8. LA ejection f raction ................................ ................................ ................. 24 9. Echo cardiographic velocity scheme ................................ ........................ 27 10. Minimum vorticity paradigm ................................ ................................ ... 43 11. Vorticity in diastolic phase ................................ ................................ ...... 45 12. Mitral valve vortices ................................ ................................ ................ 47 13. MPA systolic vorticity ................................ ................................ ............. 51 14. Wall Shear Stress computatio n sites ................................ ........................ 58 15. Streamlines guidance system ................................ ................................ ... 59 16. Wall shear stress computation work fl ow ................................ ................ 60 17. In plane point indexing ................................ ................................ ............ 61 18. Wal l shear stress Program output ................................ ............................ 63 19. W all shear stress correlations ................................ ................................ ... 64 20. Helicity football analogy ................................ ................................ ....... 71 21. Helicity integration regions ................................ ................................ ...... 74 22. Helicity correlations ................................ ................................ ................. 77

PAGE 12

! xii 23. Vorticity ROC ................................ ................................ .......................... 84 24. Wall Shear Stress ROC ................................ ................................ ............ 85 25. Helicity ROC ................................ ................................ ........................... 86 LIST OF ABBREVIATIONS

PAGE 13

! xiii 6 MWT 6 Minute Walk Test BNP Brain Natriuretic Peptide CFD Computational Fluid Dynamics CI Cardiac Index cGMP Cyclic Guanine Monophosphate CMR Cardiac Magnetic Resonance CO Cardiac Output CRP C reactive protein CT Computed Tomography DT Deceleration Time ECHO Echocardiography ECM Extracellular matrix EDV End Diastolic Volume GC Gua nylate Cyclase IL 6 Interleukin 6 IRB Institutional Review Board LA Left Atrium LAV Left Atrial Volume LA EF Left Atrial Ejection Fraction LPA Left Pulmonary Artery LRPA Lower Right Pulmonary Artery LV Left Ventricle LV EI Left Ventricular Eccentricity

PAGE 14

! xiv MMP Matrix Metalloproteinase MPA Main Pulmonary Artery mPAP Mean Pulmonary Arterial Pressure MRPA Middle Right Pulmonary Artery MV Mitral Valve NO Nitrous Oxide OSI Oscillatory Shear Index PAH Pulmonary Arterial Hypertension PC MRI Phase Contrast Magnetic Resonance Imaging PCWP Pulmonary Capillary Wedge Pressure PDE 5 5 Phosphodiesterase PVR Pulmonary Vascular Resistance RA Right Atrium RHC Right Heart Catheterization RHF Ri ght Heart Failure ROC Receiver Operating Curves RPA Right Pulmonary Artery RRT Residence Residual Time RV Right Ventricle RVDD Right Ventricular Diastolic Dysfunction RV EF Right Ventricular Ejection Fraction RV SV Right Ventricula r Stroke Volume SV Stroke Volume

PAGE 15

! xv TCV Tricuspid Valve TIMP Tissue Inhibitor of Matrix Metalloproteinase URPA Upper Right Lobe Artery VVCR Ventricular Vascular Coupling Ratio WSS Wall Shear Stress YI Youden Index LIST OF EQUATIONS

PAGE 16

! xvi EQUATION 1. Ventricular Elastance, E max ................................ ................................ ........ 7 2. Arterial Elastance, E a ................................ ................................ ................. 7 3. Efficiency and Morphogenesis ................................ ................................ 13 4. Vorticity, ................................ ................................ .............................. 16 5. Spatial vorticity ................................ ................................ ........................ 17 6. LV Eccentricity ................................ ................................ ........................ 23 7. LA Volume ................................ ................................ .............................. 23 8. LA Ejection Fraction ................................ ................................ ................ 24 9. Distensibility ................................ ................................ ............................ 24 10. Capacitance ................................ ................................ .............................. 24 11. Compliance ................................ ................................ .............................. 24 12. Elastic Modulus ................................ ................................ ....................... 25 13. Stiffness Index ................................ ................................ ...................... 25 14. Plastic ity Morphology Relationship ................................ ..................... 55 15. Wall Shear Stress ................................ ................................ ..................... 56 16. Oscillatory Shear Index ................................ ................................ ............ 57 17. Residual Residance Time ................................ ................................ ......... 57 18. Barras Equation ................................ ................................ ........................ 61 19. Hematocrit Dependent WSS ................................ ................................ .... 61 20. Spatial Helicity ................................ ................................ ......................... 70 21. Total Helicity ................................ ................................ ........................... 73

PAGE 17

! 1 CHAPTER I INTRODUCTION Definition Epidemiology, and Etiology Pulmonary Ar terial Hypertension (PAH) is the first of the five recognized subclasses of Pulmonary Hypertension recognized by the World Health Organization. The majority of the PAH cases in the adult population are idiopathic or familial However the PAH can present it self as a secondary disease to connective tissue d isorders, infectious processes, and congenital structural malformations. Chronic obstructive pulonary diseases develop into mild to moderate PAH in 10% of the cases 1 2 3 A recently reported epidemiologic study shows an occu rrence of 15 cases in the million adult pupulation 4 Interestingly, females are much more prone to acquire PAH, with a predominating overall 3:1 ratio 5 Clinically PAH is defined as an e levated mean pressure in the mai n pulmonary artery (MPA) with a value # 25 m m Hg. This criterion is accompanied with measures of pulmonary capillary wedge pressure (PCWP) $ 15 mm Hg and pulmonary vas cular resistance (PVR) # 3 Wood's units in order to exclude left ventricular failure and other related pulmonary cardiogenic differential diagnoses 4 6 The exact cause of the PAH leading to specific pathophysiologic processes is yet to be determined. The most frequentl y debated mechanisms suggest dysfunctional cellular signaling of the pulmonary endothelium and smooth muscle cells of pulmonary vasculature leading to excessive proliferation and thickening of the medial arterial layer 7 8 9 Genetic studies have revealed a variety of specific gene mutations present in familial patient cohorts 10 The effect of mechanical l y induced stress with subsequent altered cellular modification resulting from

PAGE 18

! 2 endothelial plasticity has b een studied as well 11 Due to the heterogeneous moiety of possible cau ses of PAH, some investigators lean toward the multiple hit' hypothesis leading to development of PAH 8 Pathophysiology Patients presenting with PAH have distinctive biochemical and tissue phenotypic featu res, which result in macro scale tissue observable changes featuring as clinically distinguishable symptoms. From the biochemical/biomarker perspective the most well known feature of PAH is the impaired ratio between prostacyclin and thromboxane A 2 molecules 7 The former is classically responsible for vasodilative vascular response and inhibition of platelet ag gregation, contrary to thromboxane with vasoconstrictive and platelet a ggregation st imulatory effects. Intuitively, both mo lecules serve as counter balancing parts of the vascular homeostasis, which essentially requires precise control. The biochemical analysis revealed that thromboxane levels are elevated in P AH patients, thus resulting in cumulative pulmonary narrowing of the vessels. Additionally, the production of endothelin 1 and nitric oxide (NO), both known as local effectors of physiologic response to blood pressure contr ol is altered as well 12 The conc e n tration of endothelin serving as a vasoconstrictive and cell proliferative agent is increased while synthesis rates of NO, responsible for local vasodilation are lowered. On a larger scale, PAH originates in the distal pulmonary vasculature at the level o f capillary and pre capillary vessels and progresses through the vasculature toward larger more proximal vessels. The stiffening of the arteries is immediately followed by the altered hemodynamic and mechanical response of the entire cardio pulmonary unit 13 (See Figure 1 ) .From the tissue biomechanics point of view this can be assessed by

PAGE 19

! 3 looking at the impedance patterns of higher harmonic values 14 The tissue plasticity of pulmonary vessels and cardiac tissue enables initial compensatory changes in order to deal with increased pressure and resistance, and decreased perfusion 6 These alterations exhibit by overall increased afterload on the right ventricle (RV) which mala dapt s to a hypertrophic state in order to maintain cardiac output (CO) at physiologic levels. In the vascular domain, we observe the loss of the physiolog ic endothelial layer accompanied by constriction and st iffening The two major extracellular matrix (ECM) proteins composing vascular tunica media, elastin and collagen are responsible for the elasticity and stiffness of the vessels respectively Under physiologic conditions their ratio is well balanced, giving a rise to a effectively compliant vessel with anisotropic and viscoelastic properties. Elastin has lower stiffness and Y oung modulus (~0.16 MPa) and is mainly responsible for the elastic Figure 1 The pathologic progression of the PAH via hypertrophied ventricle compensating for the stiffening of vasculature and overall increased PVR. Further progression of PAH eventually leads to RV dilation and right heart fail ure.

PAGE 20

! 4 behavior of vessels. In the diastolic non stretched phase elastin is condensed in a wavy like shaped structure Under physiological stretch, approximately 90% of all elastin shape changes to straight conformation providing the vessel necessary elasticity. Collagen on the other hand is much stiffer than elastin (~0.75MPa) and only a small fraction is stretched with physiological pressure wave. With increased systolic pressure associated with PAH more collagen fibers exhibit change from wavy to stretched form, giving the vessel its stiff character. Smooth muscle cells due to their relatively low concentration in the proximal pulmonary vasculature show main ly tissue viscous beh avior and do not significantly impact local vascular biomechanics although this behaviour is not observed distal pul monary vasculature compartments However, the vascular remodeling protruding from the distal arteries accompanied by local hemodynamic chan ges, leads to excessive elastin loss, which results in a shift of the stress/strain curve and overall stiffening of the vessel ( See Figure 2 ). The crucial role of Figure 2. Graph representing the non linear elastic modulus of the vasculature. We can see that at the physiologic region, the compliant elastin fibers are completely stretched but only a small portion of stiffer collagen is undergoing shape change. The two tangent lines represent the relative stiffness of elastin and collagen.

PAGE 21

! 5 elastin on biomechanics of pulmonary vasculature and overall physiologic performance has been studied in both animal and clinical settings 11 15 One of the most drastic hemodynamic changes is observed in the decreased wall shear stress (WSS) which has the principle impact on cellular behavior. T he altered fluid tissue interaction inducing different mechano transducti on effects has the ability to change everything from the cellular orientation on the endothelial surface to biochemical pathways and protein synthesis 16 17 Specifically, changes in the WSS promote different local flow conditions and fluid tissue interactions where surface mechano receptors transduce different intracellular responses. The higher WSS is associated with stimulated NO synthesis, tightly regulated smooth muscle cell proliferation, and overall remodeling lead ing toward larger luminal diam eter. On th e other hand, low WSS stimulates cellular proliferation and pla telet aggregation resulting in vascular narrowing. Generally both cardiac and vascular biomechanical/biochemical changes are initially compensated for by tissue remodeling in order to withstand higher pressures and resistance in the pulmonary vasculature. However, as the disease progresses these changes become counter productive, promoting even more rapid development of the PAH 13 This vicious cycle th rows the pulmonary cardiovascular system into a deadly spiral which sequentially lead s to right heart failure (RHF), since the heart cannot anymore adequately compensate for increased afterload (See Figure 3 ) Increased Pressure, PVR Reduced Flow ! Stiffening, Remodeling Decreased WSS ! Figure 3 Vicious cycle scheme in PAH

PAGE 22

! 6 Standard Diagnostic Imaging Techniques I n the 21 st century, the gold and definiti ve standard for diagnosis of PAH remains right heart catheterization (RHC) fo r which patients are referred based on echocardiographic findings. RHC is preferably performed via the jugular vein using the thermo dilution technique with standard Swan Ganz catheter 18 Besides the pressure ins ide the main pulmonary artery (mPAP), PVR, and PCWP, other standard measurement protocol indexes include mean right atrial pressure, pulmonary pressure gradients, and cardiac index. While this method is relatively accurate and represents the most direct mean s to measure the hemodynamic conditions inside the right heart and proximal pulmonary vasculature, the RHC protocols still carry a certain risk of complications due to its invasivity. Echocardiographic evaluation is a standard referral technique which usu ally raises the suspicion of PAH presence based on clinical symptomatics. Physicians can evaluate shape and size of the right heart chambers, which may e xhibit sign s of tissue remodeling, and estimate the size of the MPA 18 The p rogressed phase of the PAH might reflect on impaired RV diastolic function, when elastic recoil following systole provides passive suction force for ventricular filling. The stiffening of the RV limits the optimal recoil affe cting the velocity streams through the tricuspid valve (TCV) detectable via Doppler echocardiography. The accidental finding of pericardial effusion may point to PH disease as well as a manifest of right heart failure 19 The trans thoracic echocardiography serves as a non invasive, read ily available technique, but lacks the necessary specificity toward functional PH class and cannot provide hemodynamic parameters in the pulmonary domain.

PAGE 23

! 7 Addit ional Diagnostic Measures Recently, cardiac magnetic resonance (CMR) and hig h resolution computed to mography (CT) were suggested as possible prognostic and diagnostic techniques 4 18 Both techniques can provide a detailed morphological description of the cardia c tissue and useful tissue dynamic and hemo dynamic parameters. Pressure volume studies on canine models in the early 19 8 0 s introduced a new concept today known as ventriculo arterial or ventriculo vascular coupling ratio (VVCR) describing the relations hip between RV contractility and afterload impo sed by resistive and compliant aspects of the proximal pulmonary vasculature 20 Variation in systemic loading condition s accompanied by in situ hemodynamic measurements provides reflective pressure volume (P V) l oops from which the aforementioned biomechanical indexes are estimated ( See Figure 4 ) The slope line connecting the P V loops represents the end systolic pressure and volume ratio known as the ventricular elastance, typically depicted as E max or E es and is a general measure of the ventricular contractility. Using the physiologic measures it is defined as: !"# !"# !"# ! !"#" !"# (1) where ESP and ESV are end systolic pressure and end systolic volume respectively. Due to close proximity, ESP can be substituted by mean pressure measured in MPA (mPAP). The line connecting the ESP and transi tion point between filling and iso volumetric contraction phase describes the arterial elastance, E a and represents the afterload conditions in pulmonary vasculature. Physiologically E a is defined as: ! !"# !" ! !"#" !" (2) with SV representing the stroke volume in the given cardiac cycle.

PAGE 24

! 8 The ratio between the E a and E max provides insight into the energetic efficiency transfer of the stroke bolus between cardiac and vascular compartments. The optimal transfer is reached when E a /E max is equal to one, with elastance being approximately the same in both compartments. Using CMR, Sanz et al. showed that this ratio is still main tained in the early stage of PAH when cardiac muscle can still compensate for increased after l oad an d the system is said to be coupled 21 However, as the PAH progresses and the afterload increases exacerbated by resistance and reduced compliance in the pulmonary vasculature the ventricle cannot furth er compensate and the system becomes de coupled with E a /E max # 1. Besides VVCR, other novel non invasive analyse s have recently focused on application of echocardiography and CMR with subsequent analysis of hemodynamics in the frequency spectral domain, providing exquisite index of vascular impedance defined as the ratio between pressure and flow 6 14 The data obtained in the impedance spectrum analysis can reveal a significant amount of information about stiffness and resistance of Figure 4. The graphical representation of the VVCR method using variance in pre load conditions. The ratio between ventricular and arterial elastance reflects the efficiency of energy transfer between corresponding compartments.

PAGE 25

! 9 entire pulmonary circuitry. It is now widely accepted that larger inter institutiona l studies may provide important diagnostic and prognostic data for both pediatric and adult PAH populations. However, none of the above mentioned techniques have been yet accepted to the clinical diagnostical protocols and unfortunately remain to serve mor e as research tool s Biomarkers Currently there are only a limited number of generally approved clinically testing biomarkers for PAH among which large number lack necessary sensitivity and specificity. Furthermore, studied biomarkers are not chronologically followed from pediatric to adult populations. Most widely accepted syste mic biomarkers of the PAH are brain natriuretic p eptide (BNP) and its co synthesis byproduct NT proBNP 22 Both molecules are more specific toward RV performance yet they both lack clinically vital cut off values for both diagnostic and prognostic purposes. Additionally, uric and hyaluronic acids were examined to be elevated due to increased systemic organ stress and inflammatory response, respectively 23 Inflammatory markers tend to be less specific due to large spe c trum of potentially associated diseases, and certain classes of cytokines were researched in association with PAH as well. Interleukin 6 (IL 6) and C reactive protein (CRP) are both well recognized non specific inflammatory markers, which have been shwon to be associated with poor prognosis amon g PAH patients. However, both mentioned inflammatory molecular indicators may serve well as a clinical tool for the evaluation of status presence.

PAGE 26

! 10 Increasingly more appealing are ECM modulatory proteins, which are specific indicators of the tissue remodel ing activity. Si milar to prostacyclin thromboxane impairment leading to the cumulative vasoconstriction, the same phenomenon can be observed between the family of matrix metalloproteinases (MMPs) and their inhibitors, tissue inhibitors of metalloproteinase s (TIMPs). The MMPs are secreted by the vascular endothelia l and smooth muscle cells as a result of stimulated proliferative and inflammatory cellular response. However, the syllogism is in this case somewhat different because both MMPs and TIMPs are elevated in the PAH population 22 The concentrations of the former molecule are increased due to excessive prolife ration r equiring ECM remodeling, and latter due to increased stimulation of the negative inhibitory feedback. Additionally, cystatin C, inhibitor of specific lysosomal cysteine proteases is elevated in the course of PAH as well. Recently, the research has been ai med at PAH specific subgroups of circulatory miRNA particles 24 25 Precisely the miRNA 130 family was described to be elevated in the PAH pathophysiology concerned cellular pathways. Accordingly the inhibition of the miRNA 130 stopped progression of induced PAH in animal models. Despite the relative infancy phase of miRNA research in collaboration with the PAH, th is route of investigation appear s to have strong diagnostic and therapeutic potential. Therapy Despite the extens ive research over the last two decades, the medical warfare against PAH is still in the defensive stage, mainly providing palliative therapeutic options. The common aim of all pharmacologic therapy for PAH is induction of

PAGE 27

! 11 vasodilation and restoration of pe rfusion through the pulmonary tissue. The widely prescribed medications are effectors of the cyclic guanine mono adenosine phosphate (cGMP) cellular pathway ( See Figure 5 ). One of the frequently used agents is Sildenafil functioning as the inhibitor of 5 phosphodiesterase (PDE 5) which functions as the hydrolyzing enzyme of the cGMP cGMP is the mediator of NO vasodilatory pathway, thus Sildenafil functions as the promotive molecule via inactivation of the subsequent enzymatic step However, a large multicenter trial revealed that Sildenafil does not improve clinical outcomes of the PAH patient s 26 The qualitative and quantitative measure, the 6 minute walk test (6 MWT) showed no significant difference between treated and placebo groups. The most recently tested med ication, Riociguat serves as a stimulator of the guanylate cyclase (GC), which is crucial for the conversion of the guanine tri phosphate (GTP) to cGMP. In the regulatory pathway perspective, Riociguat functions upstream of Sildenafil, and is capable of st imulation of cGMP synthesis solely with GC independently of NO presence. Indeed, in the first widely reported clinical trials, Riociguat showed promising results, by increasing 6 MWT and improving dyspnoic score, using nearly identical studying protocol as in the Sildenafil trial 27 However, comparative larger volume trials are needed to support these results in order to see wider clinical use. In summary, despite the tremendous pharmacolog ic efforts, there is no adjuvant therapeutic agent for the PAH patients, and physicians must rely on supportive palliative medications accompanied in some case s by oxygen supplement therapy. Additionally, prostacyclins analogs are used to enhance the vasc ular vasodilation and platelet inhibition. Commercially available marks Epoprostenol and Treprostinil

PAGE 28

! 12 were shown to improve the 6 MWT results, however both analogs were associated with adverse outcomes associated with drug administration route 2 8 Novel Hemodynamic Indexes The fluid tissue interaction is sensitive symbiotic relationship possessive of high fidelity patterns and delicate regulatory domain. Any minor disturbance in one compartment is reflected by the negative behavioral pattern in the other. As previously mentioned, endothelial cells in pulmonary vasculature are highly sensitive to viscous drag created by the WSS 29 30 This quiet unique and beautiful dynam ic system has veritable Figure 5. The biochemical representation of the pharmacologic intervention in the guanylate pathway in the PAH. One can see that both drugs tend to promote the stability of cGMP, which subs equently promotes storage of cytosolic calcium leading to a vasodilative effect.

PAGE 29

! 13 physical features detectable by CMR and advanced computational fluid dynamic (CFD) analysis. While insouciant opinion might consider the heart and vasculature as a simple pu mp and pipe' system, the exact opposite nature is true. Se veral forms of energy and forces play vital role in orchestrating the cardiac function. Explicitly, the system is governed by pressure forces, volumetric measures, and bioelectric potentials generated by the cardiac conducting system. However, implicit fac tors exist in this system as well, and their role is vital in order to find the necessary solutions for the entire unit Specifically, the geometrical considerations, elastic potential stored in the tissue, thixotropy of circulating blood, a nd autonomous m etabolic control are all necessary components of the highly ordered morphogenic system. Finally, the all above mentioned properties can be considered as the morphogens of heart and circulation. As rationalized before, the exact mathematical description of the entire unit is considerably formidable, acknowledging the fact that these morphogens are interdependent 31 Either internal or ext ernal physiologic processes can easily alter the efficiency of the system since any of the morphogens can be affected While the immediate changes in cardiac (or vascular) morphology or flow properties are difficult to detect via empirical studies, the thi n interface between both compartments composes a domain where even early sign of tissue rem odeling or minor changes in flow patterns can be detected. !" !" ! ! ! ! !" ! ! ! !" (3) The above depicted relation offers perhaps provocative insigh t into how easily a single change in the cardiovascu lar unit can impact overall dynamic milieu The change in cardiovascular efficiency, can be described by change in morphology (and implicit factors) and by change in fluid dynamic patterns (and explicit factors). As one

PAGE 30

! 14 can see the delicacy of this system is the interdependency. These fluid tissue interactions can then offer a new pool of information which may be used for the prognostic, diagnostic and one day possibly for therapeutic purp oses. The final challenge relies in properly selecting the factor which would adequately measure the fluid tissue interaction. Unfortunately, the vast majority of currently applied clinical diagnostic techniques are not decorous enough to detect fluid ti ssue boundary layer. This critical inters e c tion between fluid and solid tissue domain represents a dynamic system, in which we can observe the effect of mechanical forces on the cellular components. Over the last decade, there has been a substantial increase in using CMR combined with CFD analysis. This engagement produced novel hemodynamic variables which are under current investigation. The earlier mentioned WSS is most likely the most sensitive stress factor describing fluid tissue in teraction. The WSS has been investigated in both systemi c and pulmonary circuitry as well as in pediatric studies 32 33 34 The endothelial cells of tunica intima resp ond variably to different grades of viscous d rag via mechano transduction, which then amplifies specific cellular signaling pathways responsible for remodeling events affecting entir e vessel layers. The plasticity demeanor can be noticed by change s in cyto skeletal rearrangements, synthesis of various inter cellular junctional proteins, proliferation, and attraction of inflammatory cells to outer adventitial layers, i.e. change in vascular morphology. All of these changes carry the caveat of vascular stiffening and of further spread throughout the pulm onary vessels into the cardiac tissue. The second chapter of the presented work introduces to a reader the concept of vorticity, the local spinning nature of the fluid dependent on the directional velocity gradient. It can be observed in variety of instan ces of regular natural phenomena, but it

PAGE 31

! 15 can be also seen in the cardiac chambers and proximal great vessels of pulmonary vasculature. Four dimensional CMR (4D CMR) is an ideal technique, which allows for vorticity computation as described in a subsequent chapter. Phase contrast MRI (PC MRI) and post processing analysis using the Lagrangian method and bicubic interpolation is currently the most frequently used method for measuring the WSS in patients 32 33 Comparatively, the velocity profile combined with the calculation of dynamic viscosity is described in the third chapter of this work Additio nally, complex fluid formations describing local and regional flow shapes are being investigated as a potential diagnostic markers. Additionally, helicity, describing helical formations along the vascular lumen is in the initial stage of clinical investi gation in both in vivo and in vitro models and will be investigated in the fourth chapter In general, novel hemodynamic indexes computed via CFD using data acquired from 4D CMR will be described. Additionally, their diagnostic strength is evidenced by str ong correlations to standard hemodynamic markers currently obtained from RHC and echocardiographic measures. Their powerful potential in desc ribing fluid tissue interaction can reveal pathological changes distant from placid and languor homeostatic equilib rium in healthy cardiovascular physi ology. Finally, the last chapter describes the proposed diagnostical algorithm, which uses as the input the three described novel indexes. The outcomes of this research show s interesting results and may serve as a discovery of a holy grail in PH diagnostics; using non invasive accurate techniques. As stated before, the hope is that after introduction of this material, many aspects will be sufficiently provocative for discuss ion, which is of greatest importance under the current knowledge of PAH disease state.

PAGE 32

! 16 CHAPTER II VORTICITY Theory Vorticity is one of the newest fluid descriptors introduced to the world of cardiology and cardio thoracic radiology. It precisely describes the local spinning motion generated by gradient differential. While already part of hemodynamic stu dies, its mathematic nature is purely kinematic Both vorticity and its close cousin helicity (described in later chapter) are most heavily used in the mete orological sciences where they are used to describe large scale weather for mations. Nevertheless, despite vex ing complexity their potent ial was recently translated to clinical use. Mathematically, vorticity is defined as: ! ! ! ! ! ! ! !" !" !" ! (4) where represents the del operator, and is a generated velocity vector field. The pulmonary vasculature is represented by a system of high Reynold's numbers (<2000), where the dynamics of viscosity play a mediocre role (not true at the vessel wall boundary layer). Due to the major difference in the velocity filed, which is in the y direction orthogonal to flow and the vessel wall, and vector elementary rules, vorticity tends to copy the surface of the vessel and is restricted to the wall closest the velocity bo undary layer. Perhaps a more picturesque analogy can be provided using a river flow effect. One can easily notice that the stream is the strongest at the middle of the river lumen. If one would place a volleyball c lose to the riverbank, and hold the ball gently temporarily at its center vertical axis, we would see that the ball would start to spin due to

PAGE 33

! 17 velocity differential between the stream further away from the shore and the one juxtaposed to the shore. Tensora l description dictates that vorticity is pri marily generated at solid fluid and fluid fluid interfaces. Additionally, vorticity is independent of dynamic parameters like viscosity and is purely of kinematic nature. Vorticity can be then used to describe th e flow patterns in pulmonary vasculature. However, it can also be used to visualize these patterns in cardiac chambers, where the vorticity vector field is confined to proximity of the endocardium and valvular stru ctures. As described momentarily, the dias tolic filling phase can provide a pool of information regarding the elastic recoil and overall cardiac ability to generate preload. Vorticity can thus be integrated over defined volumetric region of interest and spatially measured: ! ! ! ! !" !" !"# (5) where PAV and BF represent boundary conditions defined by the pulmonary artery valve and th e MPA bifurcation, respectively. The boundary conditions of the above mentioned equation can be easily altered. Per se, one can place the focus on the seco ndar y pulmonary arterial branches, the cardiac ventricles or the atria. However, it is crucial to have these standardized. The concept of vorticity has not been widely used in rigid closed system fluid dynamics due its complex nature and limited potential However, the pulsatile nature of the vascular system together with its compliant properties may provide characteristic vorticity patterns which may differ between healthy individuals and the PAH patients. Specifically, the kinetic gradients will differ v ariably as driving pressure forces vary with the cardiac cycle. Additionally, the directional changes associated with the anatomical

PAGE 34

! 18 structural changes (valve opening, luminal dimensions) will even further enhance the vorticity variability with respect to time and organ system condition. The kinematic and directional nature of the vorticity phenomenon also restricts this measure only to quan titative analysis, which has been neglected in recently introduced qualitative measures of hemodynamic patterns 35 36 Additionally, the recently introduced qualitative investigations of vortex and helix patterns were mostly reserved for healthy subject and serves pre dominantly as a physiologic research tool 37 38 The 4D flow CMR is novel and unique way of measuring complex hemodynamic patterns, and allows for investigation of complex markers like vorticity and helicity. The significant differences between vorticity measures in PAH and control subjects may provide novel sensitive and specific diagnostical clinical tool and spare patients from invasive RHC. The progress of PAH is associated with morphological changes of both ventricles and pulmonary vasculature which may create furth er potential for prognostic and follow up use of this technique 13 39 The vorticity is the first introduced novel hemodynamic index, which was studied in two patient cohorts as described shortly. The principle aim of this study is to assess the difference in vorticity between the healthy controls and PAH patients in peak cardiac cycle conditions. Anatomically, the calculations of vorticity data were also focused on proximal pulmonary vasculature and all cardiac chamber s. Additionally, t he translational potential in measuring vorti city quantitatively may be observed by correlating the CMR derived vorticity with the standardized invasive hemodynamic measurements from RHC and echocardiographic measures.

PAGE 35

! 19 Methods Two patient cohorts we re selected for the specific data acquisition protocols. With the approval of the Institutional Review Board (IRB), patients with suspected and confirmed PAH were referred and enrolled to the specific research cohorts. In the first prospective study, 15 PAH/RVDD (Right Ventricular Diastolic Dysfunction) and 10 controls underwent same d ay ECHO and 4D CMR (Siemens Group) The second cohort of patients was composed from 17 PAH patients and 5 contr ols having the same day RHC and 4D CMR (Butcher Group) One should immediately notice the unique feature of this protocol is in the acquisition o f diagnostical data during single day period, which eliminates major hemodynamic alterations possibly presentable between two imaging testing modalities. Subjects i n both investigated cohorts were 18 years or older without known cardiomyopathy, coronary artery disease (history of myocardial infarction, prior mechanical revascularization including percutaneous coronary intervention or coronary artery bypass surgery, or known coronary artery stenosis > 50%); significant (i.e. > moderate) valvular heart dise ase, or advanced liver disease (cirrhosis, chronic hepatitis B or C infection, or cancer). In addition, control subjects were asymptomatic without parenchymal lung disease, pulmonary hypertension, or smoking history. Siemens Group Butcher Group 13 PAH/RVDD 17 PAH 10 control 5 control Same day 4D CMR and ECHO Same day 4D CMR and RHC Table 1. Study group characteristics

PAGE 36

! 20 4D CMR & 2D CMR Protocol s 4D flow CMR was performed with interleaved 3 directional velocity encoding 4D flow images were acquired using a RF spoiled gradient echo pulse sequence, prospective ECG gating, and respiratory navigators using bellows as previously described 40 Subjects underwent CMR imaging in the supine position using a 1.5 T clinical MRI syst em (MAGNETOM Avanto, Siemens AG Healthcare Sector, Erlangen, Germany) and 8 channel phased array cardiac coil. The cardiac short axis was determined from scout images at the level of the mid ventricle, vertical long axis, and a horizontal long axis. The ba sal short axis image plane was positioned beyond the level of the tricuspid valve plane. A cine steady state free precession technique with retrospective gating was used to image from the base to apex during brief end expiratory breath holds using contigu ous short axis slices in 8 mm increments. The acquisition detection protocol was set for the encoding velocities of 100 cm/s and 150 cm/s for Siemens and Butcher cohorts respectively. The filtered data sets were run through an anti aliasing algorithm in MATLAB and the resulting data was converted to Ensight case file format using the Velomap software package. Two parameters of the anti aliasing algorithm were tuned so that aliasing was corrected while correct velocity data remained untouched. The first p arameter, the pixel velocity threshold for consideration as aliasing was ultimately set at venc/2 (in this case 50 cm/s Siemens group or 75 cm/s Butcher group) and the threshold for a neighboring pixel was also set at venc/2 Increasing the magnitude of these threshold values resulted in the algorithm being unable to identify a significant amount of aliasing in the velocity images.

PAGE 37

! 21 Decreasing these threshold values resulted in false identification an d correction of aliasing The Ensight case files were loaded in the ParaView visualization and quantification software package (Paraview, Kitware, Clifton, NY, USA. The RA and RV regions of interest were manually isolated using a rectangular prism oriented along the RA/RV longitudinal axis. The region of RA w as then more specifically selected using by superimposition of 2D CMR axial and 4 chaber view images. Before spatially integrating the vorticity vectors in the specific regions of interest, vorticity vectors were thresholded to remove regions of low magnit ude vorticity from the integral. This was done to reduce the dependence of the analysis results on the subjective placement of the RA and RV prismatic region of interest. Thresholds boundary values were selected to be 0.03 s 1 as a lower limit and 0.7 s 1 as an upper limit. This range showed to be a good compromise between unwanted noise created by pulmonary interstitium and maximum generated velocity vectors. The generated velocity vector field was combined with the super imposed time dependent 2D CMR ima ges in order to define specific anatomical regions. In the Siemens cohort the defined regions were focused on all cardiac chambers with the boundary defined by endocardial boundaries, and inter chamber valves. Short axis, 4 chamber, long axis, and axial vi ew images were used for the anatomical navigation. The anatomical region covering the main pulmonary artery was defined from the PAV up to bifurcation with the flow parallel boundaries and right pulmonar y artery (RPA) as the segment f r o m the bifurcation up to the first secondary branch right upper lobar branch ( See F igure 6 B ). These regions were selected due to relatively identification process and

PAGE 38

! 22 enabling to use available sofware selection function. However, the patient specific variability in MPA curvature or bifurcation shape might have occurred, but these changes provided m inimal numerical difference. Since the data generation produced variable spatial resolution in voxel size, every calculated vorticity value was then divided by voxel volume. The maximum and minimum vorticity values were recorded for the pulmonary vascular segments with additional E and A diastolic phase vorticities for each cardiac chamber. In order to confirm the morphology dependence and the fluid tissue interaction, the additio nal parameters eccentricity, volumes of cardiac chambers and heart rate dependence were computed as a part of the concomitant study to better understand the vorticity effects The left ventricular systolic eccentricity was computed using short axis Figure 6 Vorticity integrating regions.

PAGE 39

! 23 cine i mages of the LV 5 millimeters below the mitral valve during systolic peak as described previously 41 : !" !" ! ! ! !"#$%& ! ! !"#$%& (6) where l a and l b are maximum length values connecting the endocardial borders parallel and perpendicular to ventricular septum respectively ( See F igure 7 ). The LA volume (V LA ) was computed using the biplane area length method using four chamber and long axis cardiac veiws 42 : !" ! ! !" ! ! ! ! !" ! !"#$%& (7) where A 4CH and A L A represent respective four chamber and long axis view areas of the left atrium. l 4CH is the maximum length between the endocardial borders of the LA from the four chamber view perspective parallel to the atrial septum (See F igure 8 ). The V LA was computed a t the stage of maximum relaxation and contraction of atria. Using the ! ! Figure 7 Systolic eccentricity components.

PAGE 40

! 24 maximum and minimum values, the LA ejection fraction (LA EF) was computed in order to estimate the atrial mechanical performance. !" !" ! !" !"# !" !"# !""# (8) Besides the aforem e ntioned parameters, additional hemodynamic indexes described by Sanz et al. were computed in order to reflect on the possible correlations between fluid behavior and vascular bio mechanics 43 The "Sanz parameters" provide a comprehensive status presents of the vessel condition (see the list of equations and the summary table below). !"#$%&#"'"("$) ! !!"# ! !"# !"# ! !"# !"# !"#" !"#" ! !"# !"# !"" ! ! (9 ) !"#"$%&"'$( ! ! !!"# ! !" !"#" !"#" (10 ) !"#$%&'()* ! ! !!"# ! !"# !"# ! !"# !"# !"#" !"#" (11 ) Figure 8 The computation area of the LA from 4 chamber view of the LA (left) and RA (right) toward the calculation of LA EF.

PAGE 41

! 25 !"#$%&' !"#$%$& !!"# ! !"#" !"#" !"# !"# !"# !"# ! !"# !"# (12 ) !"#$$%&'' !"#$% ! ! !" !"#" !"! !"# !"# ! !"# !"# !"# !"# (13 ) Right Heart Catheterization Catheterization was performed in the Butcher cohort as a part of the diagnostic protocol and assessment of direct hemodynamic parameters in the right heart system. Right heart catheterization (RHC) was performed according to the standard PH diagnostic protocol via right jugular vein using a thermodilution Swan Ganz catheter. Obtained standard hemodynamic measurements were mean arterial pressure (mPAP, mmHg), pulmonary vascular resistance (PVR, Wood's units), pulmonary capillary wedge pressure (PCWP, mmH g), and mean right atrial pressure (mRAP, mmHg) (see table below). The pressure wave records were measured in the MPA. Additional calculated markers were cardiac index (CI), stroke volume (SV), and effective arterial elastance (E a ) which represents the p ulmonary arterial afterload. Hemodynamic Markers CMR Abbreviation Units Vorticity 1/s Eccentricity Index EI Left Atrial Volume LAV mm^3 Left Atrial Ejection Fraction LA EF % Capacitance Ca mm^3/mmHg Compliance Co mm^2/mmHg Distensibility D %/mmHg Elastic Modulus EM mmHg Stiffness Index Beta SI Table 2. List and abbreviations of the CMR markers

PAGE 42

! 26 Echocardiography Patients in the Siemens cohort underwent the same day echocardiography as a part of a completion study to obtain the detailed knowledge of the fluid kinetics and overall ventricular morphology. Using the fact that diastolic phase is a non monotonous event, one can determine the changes in hemodynamic indexes with respect to chronometric, inotropic, and morphologic variables. I mmediately following the ejection phase, isovolumetric relaxation produces necessary suction driving force for the incoming venous blood temporarily stored in the RA and caval system. As the pressure in the RA exceeds that in the RV, the tricuspid valve opens and allows for blood transition between the two compartments powered by a pressure gradient and stored elastic potential in the diastolic recoil. Continual and relatively constant passive filling then adds significantly to the end diastolic volume (EDV). The final atrial contraction contributes to last and minor part of the EDV, dominantly generated by mechanical force of the atria. While healthy el astic cardiac mass will largely assist the initial flow to t he RV, stiff ventricle s will fail to produce sufficient recoil and must concomitantly exaggerate the atrial contraction in order to maintain the physiologic level of the EDV. Intuitively, both flow kinetics and dynamics vary throughout the diastolic phases and can also be used as the comparative tool among the healthy subject and the PAH patient cohort. Hemodynamic Markers RHC Abbreviation Units Mean pulmonary arterial pressure mPAP mmHg Pulmonary vascular resistance PVR WU Pulmonary capillary wedge pres. PCWP mmHg Mean right atrial pressure mRAP mmHg Cardiac Index CI mL/min/m^2 Stroke Volume SV mL Arterial Elastance Ea mmHg/mL Ventricular Elastance Emax mmHg/mL Table 3. List of Abbreviations of RHC markers

PAGE 43

! 27 Using pul sed wave Doppler echocardiography, biventricular diastolic parameters were obtained including mitral valve (MV) and tricuspid valve (TV) early (E) and late (A) filling peak velocities, E/A ratio, and E wave deceleration time (DT). Using pulsed tissue Doppler, earl y (e') and late (a') diastolic velocities were obtained for both the MV (septal and lateral annulus) and TV (lateral only) Above mentioned parameters can be used not only for the assessment of the local kinetic behavior, but maybe used as an indirect meas ure of the juxtaposed tissue stiffness and compliance. The detailed map of these measures is portray ed in the following Figure 9 Statistical Analysis Wilcoxon ranked sums were used as a suitable method for calculation of the inter variable differences. The final obtained data are reported in the form of median value and quartile ranges. The multi parameter regression analysis between the set of vorticit y computed values and RHC data was done using Spearman Rho coefficients. In both cases p < 0.05 was considered as a significant finding. This statistical methodology has been Figure 9. Artistic representation of the cardiac echocardiography measurements assessing the diastolic ventricular performance.

PAGE 44

! 28 conserved for the remainder of this work, and is currently preferred and recogniz ed by a majority of the clinical journals. Results I Siemens Group The computed vorticity for Siemens subjects for individual cardiac and vascular compartments is shown in Table 4 The maximum vorticity in the right heart chambers was higher in the PAH group and lower in control subjects, but exactly the opposite in the left heart where the values were reversed. However the inter group difference was not significant in any of the maximum value measures. While t he maximu m vorticity values did not reveal any significant variation, the minimum vorticity was lower in all cardiac chambers in control subjects. More importantly these findings were significant in the right heart compartment where in the RA minimum was significantly lower in the control group than in PAH patients (1.80 s 1 vs. 5.84 s 1 p = 0.02). The RV exhibited similar functionality with an even more significant inter relationship (1.70 s 1 vs. 3.96 s 1 p = 0.0120). The diastolic vorticity measurem ents revealed the most noteworthy variability. Following the above stated hypothesis of impaired diastolic dynamic phases between early passive filling and final atrial contraction, the computed vorticity in diastole followed the same trend. The E wave vor ticity was higher in all cardiac chambers in control subjects with significant findings in the RA, LA, and LV ( RA: 27.5 s 1 vs. 13.4 s 1 p = 0.0377; LA: 14.8 s 1 vs. 8.8 s 1 p = 0.0324; LV: 17.3 s 1 vs. 10.9 s 1 p = 0.0143). On the c ontrary, the A wave vorticity was higher and statistically significant in all cardiac chambers in PAH patients (RA: 9.3 s 1 vs. 29.6 s 1 p = 0.0048; RV: 8.3 s 1 vs. 43.2 s 1 p = 0.0101; LA: 7.2 s 1 vs. 12.2 s 1 p = 0.0324; LV: 11.0 s 1 vs. 20.9 s 1 p = 0.0377).

PAGE 45

! 29 The LA EF computed from the 2D CMR generated maximum and minimum LA volumes showed a significant difference between PAH and contr ol groups. The LA EF in the control patients was significantly higher than in those with PAH (66% vs. 48%, p = 0.000 3). Similarly, the LV EI computed via LV 2 chamber long axis view images introduced meaningful variability between the two cohorts as well. The LV EI was significantly lower in control group than in PAH subjects (1.14 vs. 1.39, p = 0.0432). Siemens Group Control (n=10) PAH (n=13) P value Right Atrium max (1/s) 50.3 (45.1 51.9) 52.4 (36.1 64.4) 0.8768 min (1/s) 1.80 (0.51 2.32) 5.84 (1.99 8.91) 0.0200 E phase (1/s) 27.5 (14.9 31.1) 13.4 (7.1 21.9) 0.0377 A phase (1/s) 9.3 (4.5 26.8) 29.6 (27.0 44.8) 0.0048 Right V entricle max (1/s) 64.1 (54.4 72.6) 63.1 (50.0 97.5) 0.7330 min (1/s) 1.70 (0.52 2.78) 3.96 (2.57 13.21) 0.0120 E phase (1/s) 37.4 (10.4 48.5) 19.5(6.7 37.4) 0.1450 A phase (1/s) 8.3 (4.7 39.4) 43.2 (31.6 63.3) 0.0101 Left Atrium max (1/s) 25.1 (20.4 26.4) 17.3 (13.6 26.7) 0.2265 min (1/s) 0.73 (0.44 1.77) 2.92 (0.54 3.89) 0.2641 E phase (1/s) 14.8 (9.7 19.7) 8.8 (3.1 12.5) 0.0324 A phase (1/s) 7.2 (3.6 12.7) 12.2(8.8 17.9) 0.0324 LA max Volume (mL) 65.0 (51.7 75.8) 50.9 (38.2 62.9) 0.145 LA min Volume (mL) 21.9 (18.0 26.3) 25.9 (18.0 29.3) 0.3363 LA EF (%) 66 (59 69) 48 (44 55) 0.0003 Left Ventricle max (1/s) 32.1 (30.2 40.5) 27.9 (20.2 42.1) 0.4382 min (1/s) 1.05 (0.14 2.79) 1.81 (1.43 3.61) 0.1137 E phase (1/s) 17.3 (13.1 19.7) 10.9 (3.4 14.0) 0.0143 A phase (1/s) 11.0 (1.9 19.1) 20.9 (12.5 26.1) 0.0377 LV Eccentricity Index 1.14 (1 1.19) 1.39 (1.06 1.62) 0.0432 The inter parameter multivaria te statistical study between cardiac chamber time phase specific calculations and echocardiographic parameters revealed an extensive list Table 4. V orticity measures from 4D CMR and 2D CMR in Siemens Group.

PAGE 46

! 30 of significant correlations (See Tables 1 8 ). In gen eral the minimum computed vorticity generated a substantial number of significant connections, while the amount of correlations with maximum vorticity was limited. Further in the case of the LV maximum vorticity, no statistically significant correlations were produced. One can find below the specific descriptions of correlation study for each cardiac chamber. However, only the highest rho correlations are written in each section due to the massive number of strong statistical relations. The reader is encou raged to review the tables with detailed and complete statistical calculated data. Right Atrium In the case of right heart chambers, the diastolic phase velocities were more sensitive to echocardiographic markers correl ations in the RA than in R V The RA,max positive ly correlates with TCV e'/a' ratio and MV E, but the strongest correlation was associated with LV CO (r ho = 0.6738, p = 0.0004). On the other hand RA,min correlated wi th regular TCV and MV velocities and corresponding ratios, as well as wit h the septal and lateral wall specific velocities. The absolute strongest correlation was with the MV A phase velocity (r ho = 0.7443, p < 0.0001). Both RA E and RA A produced correlations with both echocardiographically (ECHO) measured velocities through both TCV and MV during the respective diastolic phases. While the RA E showed stronger significant correlations with the TCV velocity measures, the most dominant correlation was obtained from the TCV E/A ratio (r ho = 0.6402, p = 0.0010). The RA A vorticity on the other hand, showed the strongest correlation with the lateral wall MV a' velocity (r ho = 0.7672, p < 0.0001).

PAGE 47

! 31 Siemens Group Correlations RA max Vorticity R ho value P Value TCV e'/a' 0.4740 0.0223 MV E 0.5223 0.0126 LV CO 0.6738 0.0004 Siemens Group Correlations RA min Vorticity R ho value P Value TCV e'/a' TDI 0.4248 0.0434 TCV e'/a' 0.4379 0.0366 TCV A vel 0.5303 0.0092 TCV E/e' vel 0.6278 0.0013 MV A vel 0.7443 <0.0001 MV E/A vel 0.5866 0.0041 Sep MV e' 0.4998 0.0152 Sep MV e'/a' 0.5395 0.0079 Lat MV a' 0.5577 0.0057 Lat MV e'/a' 0.4535 0.0297 Lat MV E/e' 0.4833 0.0211 Siemens Group Correlations RA E Wave Vorticity R ho value P Value TCV e' TDI 0.5197 0.0110 TCV E vel 0.5437 0.0073 TCV E/A vel 0.6402 0.0010 TCV e' VVi 0.5571 0.0058 MV E vel 0.6534 0.0015 MV E/A 0.4568 0.0326 Sep MV e' 0.4937 0.0167 Sep MV e'/a' 0.5969 0.0026 T able 5 Multi Parameter Regression for RA max vorticity T able 6 Multi Parameter Regression for RA min vorticity T able 7 Multi Parameter Regression for RA E vorticity

PAGE 48

! 32 Right Ventricle The RV,max revealed s imilar but stronger correlations as the RA minimum vorticity with the variance in correlation with the septal MV e'/a' velocity displacing the LV CO correlation. The RV,min provided an extensive list of correlations with ECHO velocities through both TCV and MV. Additionally, lateral and septal wall aspects of ECHO velocities showed significant correlations. Interestingly, the strongest correlation between the RV,min and ECHO indexes wa s with the MV A wave velocity (r ho = 0.7783, <0.0001). Both RV E and RV A produced fewer numbers of significant correlations than corresponding RA diastolic vo rticities. The RV E produced the strongest correlation with MV E velocity while the RV A with the lateral wall MV a' velocity (r ho = 0.5506, p = 0.0079; r ho = 0.7024, p = 0.0002). However, both phases showed strong s ignificant correlations with their respective and rationally comparable indexes, TCV E and TCV A velocities. Siemens Group Correlations RV max Vorticity R ho value P Value TCV E/A vel 0.5337 0.0084 MV E vel 0.5506 0.0079 Sep MV e'/a' 0.4557 0.0281 Siemens Group Correlations RA A Wave Vorticity Rho value P Value TCV A vel 0.6525 0.0007 TCV E/A vel 0.5337 0.0087 TCV E/e' vel 0.4549 0.0292 Sep MV e' 0.4408 0.0353 Sep MV e'/a' 0.5889 0.0031 Lat MV a' 0.7672 <0.0001 LV SV 0.4720 0.0230 T able 9 Multi Parameter Regression for RV max vorticity T able 8 Multi Parameter Regression for RA A wave vorticity

PAGE 49

! 33 Siemens Group Correlations RV min Vorticity R ho value P Value TCV A vel 0.4834 0.0194 TCV E/e' vel 0.6685 0.0005 MV A vel 0.7783 <0.0001 MV E/A 0.4295 0.0461 Sep MV e' 0.4334 0.0388 Sep MV e'/a' 0.5004 0.015 Sep MV E/e' 0.5902 0.0038 Lat MV e' 0.4246 0.0434 Lat MV a' 0.5541 0.0061 Lat MV e'/a' 0.5265 0.0099 Lat MV E/e' 0.5836 0.0044 LV CO 0.4668 0.0247 Siemens Group Correlations RV E Wave Vorticity R ho value P Value TCV E vel 0.4854 0.0189 TCV E/A 0.5357 0.0084 MV E vel 0.5506 0.0079 Sep MV e'/a' 0.4557 0.0281 Siemens Group Correlations RV A Wave Vorticity R ho value P Value TCV A vel 0.6540 0.0007 TCV E/A 0.5327 0.0089 TCV E/e' vel 0.4554 0.0290 MV A vel 0.4632 0.0299 Sep MV e'/a' 0.4814 0.0200 Lat MV a' 0.7024 0.0002 T able 10 Multi Parameter Regression for RV min vorticity T able 11 Multi Parameter Regression for RV E wave vorticity T able 12 Multi Parameter Regression for RV A wave vorticity

PAGE 50

! 34 Left Atrium The analysis of the left heart chamber vorticities offered perhaps unforeseen while interesting results. As expected LA provided generally weaker but still significant correlations than it s right side counter part However, the LV showed in some instances more numerous and stronger association s with the ECH O markers than the RV. Similar to the right heart comput ed vorticities, the LA,max showed only a limited number of correlations with the strongest one to be with the septal MV e' velocity (r ho = 0.4996, p = 0.0152). The most significant LA,min correlation was with the septal MV a' velocity (r ho = 0.5369, p = 0.0083). Both diastolic measured vorticities LA E and LA A provided extensive and multispectral sets of correlations with TCV, MV, and location specific velocities. The LA E indicated the strongest correlation with the LV SV (r ho = 0.5725, p = 0.0043), while the LA A was with the lateral wall MV a' velocity (r ho = 0.6777, p = 0.0004). Siemens Group Correlations LA max Vorticity R ho value P Value Sep MV e' 0.4996 0.0152 Lat MV e' 0.4859 0.0187 LV SV 0.4275 0.0419 Siemens Group Correlations LA min Vorticity R ho value P Value TCV a' TDI 0.459 0.0276 TCV A vel 0.455 0.0291 TCV E/A vel 0.4472 0.0324 Sep MV a' 0.5369 0.0083 Lat MV a' 0.4661 0.0250 T able 13. Multi Parameter Regression for LA max vorticity T able 14 Multi Parameter Regression for LA min vorticity

PAGE 51

! 35 Siemens Group Correlations LA E Wave Vorticity R ho value P Value TCV e'/a' 0.4424 0.0345 TCV e' VVi 0.4315 0.0398 MV A vel 0.4361 0.0425 MV E/A 0.5734 0.0053 Sep MV e' 0.4206 0.0457 Sep MV e'/a' 0.4623 0.0264 Lat MV a' 0.4167 0.0479 LV SV 0.5725 0.0043 Siemens Group Correlations LA A Vorticity R ho value P Value TCV e' TDI 0.4197 0.0462 TCV a' TDI 0.4306 0.0403 TCV A vel 0.5586 0.0056 TCV E/A vel 0.4965 0.0160 TCV E/e' vel 0.4806 0.0203 TCV a' VVi 0.4316 0.0398 MV A vel 0.4819 0.0231 Sep MV e' 0.4888 0.0180 Sep MV a' 0.4199 0.0461 Sep MV e'/a' 0.5984 0.0026 Lat MV a' 0.6777 0.0004 Left Ventricle The LV,max did not reveal any major correlations with the ECHO indexes. On the c ontrary, the LV,min provided a large list of the ECHO marker associations with both TCV and MV measured velocities. The strongest correlation of LV,min was found to be with the septal MV a' velocity. The LV diastolic vorticities showed a considerable amount of correlations with ECHO m arkers measuring both left and right heart diastolic performance. Additionally, both LV E and LV A showed LV regional velocity specific association as well. The most significant LV E was with the septal MV e' velocity (r ho = T able 15 Multi Parameter Regression for LA E wave vorticity T able 16 Multi Parameter Regression for LA A wave vorticity

PAGE 52

! 36 0.6617, p = 0.0006 ). The LV A followed in this case a logical trend and provided the strongest correlation with the lateral wall MV a' velocity (r ho =0.6189, p = 0.0016). Siemens Group Correlations LV min Vorticity R ho value P Value TCV A vel 0.5763 0.004 0 TCV E/A vel 0.5318 0.0090 MV A vel 0.5165 0.0138 MV E/A 0.5078 0.0158 Sep MV e' 0.4404 0.0354 Sep MV a' 0.6344 0.0011 Sep MV e'/a' 0.5413 0.0076 Lat MV a' 0.5773 0.0039 Siemens Group Correlations LV E wave Vorticity R ho value P Value TCV e' TDI 0.4786 0.0209 TCV e'/a' 0.4557 0.0289 TCV E/A vel 0.5912 0.003 0 TCV e' VVi 0.65 0.0008 MV E vel 0.5687 0.0058 MV A vel 0.4372 0.0419 MV E/A 0.6035 0.0029 Sep MV e' 0.6617 0.0006 Sep MV e'/a' 0.6421 0.001 0 Lat MV e' 0.6298 0.0013 Lat MV a' 0.4607 0.0269 Lat MV e'/a' 0.5419 0.0076 LV SV 0.6596 0.0006 LV CO 0.4936 0.0167 T able 17 Multi Parameter Regression for LV min wave vorticity T able 18 Multi Parameter Regression for LV E wave vorticity

PAGE 53

! 37 Siemens Group Correlations LV A wave Vorticity R ho value P Value TCV e'/a' 0.426 0.0427 TCV A vel 0.4726 0.0228 TCV E/e' vel 0.4302 0.0405 MV A vel 0.4836 0.0266 MV E/A 0.4346 0.0433 Sep MV a' 0.5223 0.0106 Sep MV e'/a' 0.5376 0.0081 Lat MV a' 0.6189 0.0016 Concomitant Study I. Left Atrial Ejection Fraction The parallel calculations of the LA EF from the 2D CMR images introduced correlation with both vorticity measures and standard ECHO markers. The LA EF revealed associations with vorticity variants in different time points of the cardiac cycle at each cardi ac chamber. The diastolic phase measured correlations were predominant (See Table 20 on the next page ). The most significant correlation showed a negative relationship with the LA A (rho = 0.6370, p = 0.0011). The similar number of right and left heart E CHO markers also provided physiologically specific correlations. The most significant correlations in both cardiac compartments were with the e' velocity (RV e': rho = 0.7168, p = 0.0001; Sep MV e': rho = 0.7153, p = 0.0001). T able 19 Multi Parameter Regression for LV A wave vorticity

PAGE 54

! 38 Siemens Group Correlations Left Atrial Ejection Fraction R ho value P Value LA A 0.637 0 0.0011 LV E 0.466 0 0.025 0 LV A 0.5159 0.0117 RV A 0.5392 0.0079 RA min 0.4132 0.05 00 RA E 0.4887 0.0317 RA A 0.5797 0.0037 ECHO MARKERS R ho value P Value RV e' TDI 0.7168 0.0001 RV e' VVI 0.6957 0.0002 RV e'/a' 0.5585 0.0056 RV A vel 0.5117 0.0126 RV E/A 0.4905 0.0175 RV E/e' 0.6592 0.0006 MV A 0.4809 0.0235 MV E/A 0.5329 0.0107 Sep MV e' 0.7153 0.0001 Sep MV e'/a' 0.6918 0.0003 Sep MV E/e' 0.6093 0.0026 Lat MV e' 0.5785 0.0038 Lat MV a' 0.5204 0.0109 Lat MV e'/a' 0.4356 0.0377 Lat MV E/e' 0.4335 0.0439 Concomitant Study II. Left Ventricular Eccentricity The second additional measure from 2D CMR provided eccentricity of the LV measured at peak systole (See Table 21 on the following page) As expected, the correlations were predominantly with the left heart chambers, specifically with the maximum and E phas e vorticity in both LA and LV. However, an additional unexpected correlation was revealed with the RA A The most significant corr elation from LV EI study was with the LA max (rho = 0.6815, p = 0.0003). T able 20 Multi Parameter Regression for LA EF

PAGE 55

! 39 Siemens Group Correlations LV Eccentricity Rho value P Value LA max 0.6815 0.0003 LA E 0.4332 0.0389 LV max 0.4644 0.0256 LV E 0.4932 0.0168 RA A 0.5145 0.012 Concomitant Study III. Heart Rate (Frequency) The heart rate measured values from the ECHO studies were translated to the specific vorticity calculations. The minimum generated vorticity showed significant correlations in the LV, RV and RA, with the most significant finding being association with the LV min (rho = 0.7244, p = <0.0001). Additional A phase vorticity calculated in the LV and RA provided st rong positive correlations as well (See Table 22 ) One can notice that correlations from concomitant study II. exploring the effect of LV EI, provided inverse correlations to those observed in this study. While the LV EI showed predominantly strong associations with the maximum and E phase vorticities, heart rate frequency analysis showed correlations exclusively c onnected only with the minimum and A phase vorticities. Siemens Group Correlations Heart Rate Rho value P Value LV min 0.7244 <0.0001 LV A 0.4333 0.0389 RV min 0.5679 0.0047 RA min 0.6816 0.0003 RA A 0.5075 0.0134 T able 21 Multi Parameter Regression for LV EI Table 22 Multi Parameter Regression for HR

PAGE 56

! 40 Results II Butcher Group The Butcher cohort was merely focused on the evaluation of novel hemodynamic parameters in the proximal pulmonary vasculature. In both arterial segments, the vorticity was reduced in the PAH patients. Vorticity computed in the MPA and RPA segments showed edge significance and major significance in the former and latter arteries respectively (MP A: 65.3 vs. 43.9, p = 0.0601;RPA: 4.27 vs. 2.33, p = 0.0077) Butcher Group Control (n=5) PAH (n=17) P value MPA (1/s) 10 2 65.3 (50.0 80.5) 43.9 ( 35.2 53.8) 0.0601 RPA (1/s) 10 2 4.27 (3.53 9.34) 2.33 (1.42 3.23) 0.0077 mPAP (mmHG) 21 (16 22) 33 (27 38) 0.0019 PVR (WU) 2.2 (1.2 2.9) 5.7 (3.3 7.6) 0.0054 PCWP (mmHg) 9 (7 14) 11 (9 14) 0.4307 mRAP (mmHg) 4 (3 7) 8 (5 9) 0.1343 CI (mL/mi/m ^2) 3.28 (2.51 3.84) 2.80 (2.27 3.33) 0.3470 SV (mL) 81.0 (68.4 110.7) 64.6 (49.9 81.0) 0.1171 Ca pacitance 4.09 (4.04 7.08) 1.98 (1.45 2.41) 0.0149 Co mpliance 8.08 (6.77 11.22) 6.03 (2.53 8.36) 0.2036 D istensibility 1.44 (1.16 1.74) 0.61 (0.27 1.13) 0.0262 Elastiv Modulus 69.5 (57.4 86.5) 163.4 (88.2 434) 0.0262 Stiffness Index Beta 3.67 (2.81 3.98) 4.29 (3.26 12.7) 0.2662 E a /E max 0.45 (0.25 0.84) 1.02 (0.62 1.79) 0.0404 The correlation analysis with the RHC hemodynamic makers provided significant correlations in both MPA and RPA, with the RPA showing stronger statistical dominance by providing larger and stronger correlations. Both MPA and RPA showed negative significant associations with the standard diagnostic index mPAP (MPA: r ho = 0.6461, p = 0.0012; RPA: r ho = 0.8163, p < 0.0001) and with the additional standard PVR index (MPA: rho = 0.6136, p = 0.0024; RPA: r ho = 0.8163, p < 0.0001). Lastly, the third common marker showing significant correlation was the E a /E max VVCR ratio also T able 23 Vorticity analysis in proximal pulmonary vasculature and Sanz indexes

PAGE 57

! 41 relating the MPA and RPA in a negative fashion (MPA: rho = 0.5842, p = 0.0086; RPA: rho = 0.5053, p = 0.0273). Additionally, the RPA provided significant negative relations with the PCWP mRAP, and two positive wi th the SV and capacitance (See T able 24 below). Butcher Group Correlation MPA RPA R ho value P Value R ho value P Value mPAP 0.6461 0.0012 0.8163 <0.0001 PVR 0.6136 0.0024 0.8328 <0.0001 PCWP 0.1971 0.3794 0.4725 0.0264 mRAP 0.2689 0.2516 0.5697 0.0087 CI 0.3701 0.0986 0.0818 0.7244 SV 0.2761 0.2136 0.4466 0.0372 E a /E max 0.5842 0.0086 0.5053 0.0273 Capacita nce 0.2977 0.2023 0.7489 0.0001 Compliance 0.1324 0.5780 0.4190 0.0660 Distensibility 0.0572 0.8108 0.4129 0.0704 Elastic Modulus 0.0617 0.7962 0.4180 0.0666 Stiffness Index Beta 0.0587 0.8059 0.2572 0.2736 Discussion I Siemens Group Hemodynamic vorticity is by its nature dependent on the velocity, directional gradient, and fluid solid interface. Thus vorticity measures the effect of vessel geometry, jet velocity through the MPA valve and branching orifices, and dynamics of pooling blood in the ventricles during diastole. Since the maximum vorticity values were increased in all vascular compartments in healthy subjects wi th significant variation, and sensitive diastolic phases showed corresponding alterations in vorticity at the PAH pathological state, it is reasonable to argue that both morphological and dynamic aspects are altered in the organ system effect by PAH. With regard to the heart as a singular functional unit, the maximum computed vorticity numbers did not tend to reveal any T able 24 Multi Pa rameter Regression Butcher Group

PAGE 58

! 42 significant correlations toward the ECHO markers. This may account for the fact that maximum vorticity was observed in the large majority o f the cases during the systolic stroke. One can assume that ventricular systolic compensatory performance in PAH patients would be more altered in the later stages of the disease. Additionally, the systolic vorticity is a substantially more complex functio n for the computation with larger number of variable s playing a deterministic role such as the morphology of the RV outflow tract, MPA valve, and the possibility of the TCV regurgitation. The Minimum Vorticity in Cardiac Chambers T he minimum cardiac chamber vorticity numbers revealed a considerable pool of interesting correlations. The extravagant question then arises asking why minimal generated vorticity would provide an extensive list of relation s with ECHO markers? One possible explanation may deal with the increased stiffness in cardiac tissue, which restricts the diastolic recoil. Assuming the com promised ventricular elastance i s a sign of the PAH progression to the heart, neither the atria nor the ventricles cannot fully expan d into maximum relaxed diastolic dimensions and provide necessary negative gradient suction potential However, s ince the vorticity is dependent on the fluid solid interaction, the lower diastolic surface area in healthy size d ventricles restricts the fiel d for vorticity generation ( See Figure 10 ). In this case when the gradient is minimal due to passive filling phase (or isovolemic relaxation) the surface area aspect dominates over velocity, resulting in smaller minimum vorticity in control subjects. Thus increased minimum vorticity is a potential sign of the increased RV stiffness (and of the remaining cardiac unit) and potential presence of the PAH. This proposed explanation is further supported

PAGE 59

! 43 by the significant inter group variance in the right he art chambers where the minimum vorticity differential was significant between the PAH and control groups (RA: 1.80 CT vs. 5.84, p = 0.0200; RV: 1.70 CT vs. 3.96, p = 0.0120). It should be expected that RV revealed more solid variability, assuming its juxta po sition and connectivity to the pulmonary vasculature from which PAH originate s While the trend of the increased minimum vorticity was also observed in the left heart chambers, t he significa nce was not present, possibly indicating the slower spread of dysfunctionality to the left heart. The l ast piece of evidence supporting the minimum produced vorticity i s by closer review of the inter variable correlation study, where A and a' velocities were positively associated with the minimum vorticities, while the E and e' velocities presented exactly the opposite nature. Figure 10 Artistic representation of the minimum generated vorticity phenomenon. While the velocity gradient is comparatively the same during the passive filling phase (or isovolumetric relaxation phase), the overall ventricular cavity area in the progressed PAH stage is larger than in a normal, healthy heart.

PAGE 60

! 44 Diastolic Phase Vorticity in Cardiac Chambers The diastolic phase chrono logical measurements indicated significant inter group variability in both E and A phase in each cardiac chamber with the exception of the E phase variance in the RV. As expected, the E wave generated vorticity dominates over A phase in healthy control sub jects, and the phenomenon is exactly reversed in the PAH patients ( See Figure 11 on the next page ) The diastolic vorticity measurements also provided widespread number of correlations with anatomically and physiologically respective ECHO markers. However, several cross septal associations between ECHO and vorticity measurements exposed another proof of the debated concept of ventricular interdependency. The interdependency spectacle may be considered as a relatively fresh concept, which has been in the vas t majority of the 20 th century insou ciantly overlooked. Currently this aspect is most heavily studied from the left heart failure perspective, when severe pathologic alterations in the LV translate into the RV performance as well 44 Importantly, generated data and subsequent correlation analysis of the present work shows that interdependency can also be observed in cardiac atria.

PAGE 61

! 45 Both RA and LA showed rational strong correlations with diastolic markers describing the velocity through the respective tr icuspid and mitral valves. The RA E correlated with the TCV E, e' and E/A ratio and the RA A with the TCV A, E/A, E/e' indexes. The LA showed similar trends in t he respective phases with more location specific markers for lateral and septal LV walls. The LA E correlated with the septal MV e', e'/a' and lateral wall a' velocity. Chronologically later measured LA A was strongly associated with al l of the septal wall parameters: MV A, and lateral MV a'. However both RA and LA showed large numbers of correlations with velocity markers measured in the opposite heart compartment One can argue that due to synchronized nature of the !"#$%"&' ()' !"#$%"&' ()' Figure 11 Representative images of vorticity (green glyphs) and velocity (white glyphs) vectors in the LV (gray mask) of a control (top row) and PH (bottom row) subject during progressive phases of diastole viewed from a cranial view (bottom panel). The LV is oriented with the MV annulus on the top, apex o n the bottom, septal wall on the left, and lateral wall on the right. Note the lesser density of vorticity glyphs in the PH subject during peak E phase compared to the control subject, particularly in the septal region.

PAGE 62

! 46 cardiac cycle, this similarity sho uld be expected. Taking into account that these physiologically logical associations hold reciprocally for both atria and in the pathophysiological shift toward the A wave diastolic dominance in PAH, the vorticity represents a new differentiable variable w hich can be even signified if normalized by the respective cardiac chamber size and volume. Interestingly enough, the most significant correlation in t he atrial diastolic analysis was between RA A and the lateral MV a' velocity parameter (r = 0.7672, p < 0.0001). The plentiful number of ventricular diastolic vorticity correlations further fortified the hypothesis about vorticity possessing biomechanical and clinical evaluation capacity, from both a local chamber and interdependency perspective. The natur e of the correlations in the ventricular domain has a repetitive analogy with the atrial correlations. Both RV and LV provided phase and location respectful correlations as well as inter septal associations. The reader is encouraged to review the list of s ignificant correlations in the Method s section. One should notice that the LV provided a large number of correlations than the RV, which enhances the interdependency phenomenon considering the right heart pathophysiological nature of the PAH. The spread of the stiffness and loss of contractility to the LV tissue has been reported as the secondary sign of the PAH 44 45 46 The quantitative portion of this work also correlates with the qualitative description of the complex fluid formations in the ventricular (or vascular) spaces. It has been already described that reduction in the inter cavital surface area limits the generation of the specific vorticity vector field. Additionally, the complex vortex rings formation is suppressed as well. It is important to highlight the difference between the two terms. While vorticity results from quantitati ve analysis of the velocity vector field

PAGE 63

! 47 and its gradient in spatial domain, vortex formatio ns are purely dependent on velocity glyphs and their present directionality. Nevertheless, even the vortex formations are suppressed in the progressed PAH stage, when the septal shift and overall reduced LV volume limits the complete generation of vortices Interestingly the sliding fluid shear levels produce visible vortices on the both surface aspects of the mitral valve leaflets ( See F igure 12 ) The anterior leaflet vortex is more susceptible to deformation (~reduction) in compromised ventricular function. Interesting theory can be proposed when one looks at this phenomenon from the momentum conservation point of view. Stronger vortices should conserve more ene rgy in the fluid and assist' in the ventricular ejection phase. However, this analysis r epresents pure ly the qualitative nature of the in situ conditions and its comparative variability still represents a challenging aspect of hemodynamics. Nevertheless, this finding provides an additional example of t he fluid tissue interaction and how impor tant it is to study the overall morphogenesis in both tissue and fluid domains. ! ! ! Figure 12 Vortices generated around the mitral valve leaflets. The long axis two chamber view (right) shows the posterior leaflet on the left side of the picture. The left picture d epicts the same case from the 4 chamber view. Both vortices form at the peak E diastolic phase.

PAGE 64

! 48 Concomitant Studies The LA EF was studied in parallel to the primary hemodynamic marker vorticity in order to assess the hypothesis of the ventricular interdependency and quantify the compensatory effect of the atrial ejection phase during diastolic filling in the PAH. The impaired relaxation in the RV compromises the filli ng properties of the LV and due to complex 4 chamber coupling also reflects on the LA function. The studies involving LA EF are usually limited to associations with the left heart failure and consequent hospitalization and normalization research 47 48 49 However, the association of the LA EF with the complex hemodynamic markers and PAH has not yet been determined. The LA EF showed to be significantly decreased in the p atients suffering from PAH (66% vs. 42%, p = 0.0003). The only known similar type of study investigated the role LA EF in diabetic patients showing the analogous trend of decreased LA EF in the diseased group 50 Furthermore, the LA EF showed correlations with both vorticity calculations and ECHO markers. The most significant connection with the vorticity was with the LA A (rho = 0.6370, p = 0.0011). While the exact nature of the decreased LA EF in the PAH patients is not known yet, one of the rudimentary explanations may still reside in the compromised LV elastic tissue properties. The stroke volume generated by the LA at the end of the diastolic phase work s against already distended LV chamber, which is already under considerable p reload stress. At this point the slope of compliance curve is steep, and greater force generated by pressure from the contracted LA is required to add the last contribution to the final EDV. Since the stiff LV will possess limited distensibility range at this point cardiac cycle, the LA EF will consequently decrease as shown in the presented data.

PAGE 65

! 49 The LV EI computed at the peak of systolic phase has been shown to suggest impaired bi ventricular properties in the PAH 41 In the context of this LV EI is shown to correlate with the maximum and early E diastolic phase vorticity in the left heart chambers. As described earlier, the regional magnitude of produce d vorticity is dependent on the interaction of the solid fluid boundary, thus is proportional to the surface area of cardiac structures. In the maximum systolic contracted state should be the transversal profile of the LV in circular shape with the septal and its perpendicular diagonal of the same length. With the progression of the PAH and RV pathologic remodeling, the inter ventricular septum tends to bow in to the LV cavity. This effect reduces the inner surface area of the LV thus limiting the generation of vorticity. As seen from the produced correlations the maximum vorticity in both LA and LV showed negative relationship with measured LV EI. Similarly, the grade of E wave velocity is dependent on elastic recoil and suction prop erties of the LV. The reduced stiffness then impairs this driving force for the E wave generation and consequently reduces the vorticity concurrently with the reduced LV inter cavity area. In the last concomitant study the direct effect of he art rate on vorticity wasd investigated The found correlations found were of a reciprocal nature to those found in LV EI study. In this case, minimum and A wave vorticity in both right and left heart compartments showed significant positive correlations w ith the heart rate. Regarding the minimum vorticity, as a heart rate frequency increases, the ESV increases as well together with the minimum intra cavi ty area, with the final result being larger minimum vorticity. Analogically, it is a known physiologic p henomenon, that with increasing heart rate, passive filling does not have time for inflow from RA and reservoir in pulmonary

PAGE 66

! 50 veins, and diastole must rely on atrial kick' to meet the sufficient requirements for EDV and to obey Frank Starling law. In other words, with higher frequency, diastole shortens more rapidly than systole, and final atrial contraction becomes a vital component of the EDV. This physiologic vignette is indirectly supported by the positive significant correlations between A phase vortic ity and heart rate. Discussion II Butcher Group Since the heart and circulatory system is one symbiotically connected unit, the study of vorticity excluded to the cardiac chambers would surmise the overall complete cardiac vessel coupling system philosophy. This is e specially true in the case of vortici ty study which is fortunately dependent on the minimal manipulation of the velocity vector field and gross spatial geometry of heart and large vessels. The direct study of currently accepted RCH hemodynamic parameters with parallel investigation of newly p roposed hemodynamic factor computed from 4D CMR, represents the unique set of data, especially considering the acquisition within 1 day period. The careful selection of the computation fields enabled working with specific segments of the proximal pulmonary vessels (See Figure 13 ) Both MPA and RPA d emonstrated that vorticity is lower in PAH patients than in healthy controls. However, only RPA provided significant variance between the two cohorts (4.27 vs. 2.33 (x10 2 s 1 ), p = 0.0033). It should be mentioned that MPA vorticity variability significance was border line to be considered significant. Nevertheless, the RPA was in summary more sensitive for vorticity computation and subs equent correlation analysis than MPA. Several numbers of explanations might be offered for this occurrence. Firstly, the RHC measurements are most often conducted in

PAGE 67

! 51 the RPA, very close to first superior lobar branch. The catheter tends passively to swim into this region due to catheter properties and flow distribution, despite the m ore continual luminal profile with the LPA. Secondly, the MPA is due to its direct anatomical connection via cardiac skeleton to the RV, prone to geometrical changes generated by instantaneous movement of cardiac muscle. The RV outflow tract also might be variable in shape creating non uniform jet streams among individuals. The RPA is more embedded, between the aortic arch and pulmonary veins. Lastly, the presence of the pulmonary valve and upstream bifurcation introduces another complex morphological and mechanical variability among the patients. Figure 13 Generated vorticity field in the MPA through out the systolic phase. The flow profile curves are depicted to enable further interpretation association with the specific point in the cardiac cycl e. The lower right image portarays the vorticity streamlinesat peak systole.

PAGE 68

! 52 Despite the complex anatomical geometry, the maximum MPA vorticity generated at the peak of systole still produced significant correlations with the mPAP and PVR (mPAP: rho = 0.6461, p = 0.0012, PVR: rho = 0.6136, p = 0.0024). The RPA correlations with the same parameters indicated stronger relationships (mPAP: rho = 0.8163; p < 0.0001;PVR: rho = 0.8328, p <0.0001). These powerful correlations indicate the strong associat ion between complex he modynamic factors, and regular standard measures in PAH diagnostic protocol. The RPA also provided strong correlation with vascular capacitance (rho = 0.7489, p = 0.0001) reflecting the direct relationship between the tissue propertie s and hemodynamic flow parameter. Importantly, both MPA and RPA provided negative correlations with the E a /E max ratio n representing VVCR phenomenon, implementing an even more complex relationship between cardiovascular coupling and hemodynamics. The negati ve relationship agrees with the already known effect of VVCR when E a /E max $ 1, and ventricular and vascular compartments are decoupled. This finding introduces a question, if vorticity might be used to assess the ventricular vascular functional unity and consequently the severity of PAH. Vorticity proved to be sensitive to bot h ventricular and vascular functions, and showed significant variance between PAH and healthy subjects. Additionally, the ventricular interdependency resulted in altered left heart compartment diastolic properties when affected by PAH. Conclusion In summ ary, the vorticity has a potential as a diagnostical tool for the evaluation of the status presence in suspected PAH. Additionally, the vorticity computation can be applied to assess the severity of the PAH progression and potentially serve as the

PAGE 69

! 53 prognost ic variable assuming multistep chronologic evaluations and further propagation of stiffness and loss of vital tissue elasticity. Vorticity as a novel and relativel y complex hemodynamic parameter can be computed from non invasively acquired 4D flow CMR, and analyzed in both cardiac and vascular sections. While the vorticity carries relatively strong diagnostic potential, the calculations and correlations may fail in very early stage of PAH when the pathophysiologic mor phogens have not yet reached' the RV. The more precise role of PAH in deterministic diagnosis is discussed in last chapter, when the overall diagnostic algorithm is introduced. It should be pointed out that this study carries severa l pioneering methodolog y steps and differs vehemently from the vortex formation research presented in earlier publications which is dominantly relying on qualitative flow properties 38 36 35 The introduced methodology in this and subsequent chapters relies almost exclusively on quantitative approach and requires precise selection of the anatomical regions fo r volumetri c integration. Finally, it is suggested to give attention to the fact that vorticity measures are dependent on solid fluid interaction, which may provide new and better insight into fluid tissue interaction and overall understanding of undesired PAH induce d morphogenesis. Vorticity can be relatively easily measured using non invasive 4D flow MRI and can provide fluid specific information about function of vascular and ventricular compartments. This technique could be translated for further prognostic meas ures and evaluation of prescribed PAH therapy per individual patient bases. The elimination of the invasive RHC would also enable to perform this anlysis in relatively short period of time and reduce hospital in patient monitoring.

PAGE 70

! 54 CHAPTER III WALL SHEAR STRESS Theory As already mentioned in the previous chapters, WSS is currently the best known parameter describing the fluid tissue interaction from both macro and micro environmental perspective. The endothelial cells function in this system as high fideli ty transistors, which possess the ability to control morphological change of the vessel anywhere from proliferative, stiff, constricted stage to lumen with compliant and dilated character. The scientific inquiry regarding WSS has been heavily investigated over last decade from many perspectives. The hierarchical concepts considered complex fluid formations producing specific WSS patter n s and subsequently WSS directing the cellular and extracellular control in both systemic and pulmonary vasculature 16 One can notice that WSS can be then used in variety of situations simultaneously as both input and output of higher order morphogenesis function. Recall equation (3) where complex fluid tissue interaction is described as differential of a state in specific compartment describing overall cardiovascular performance with a vitally present inter dependency !" !" ! ! ! ! !" ! ! ! !" (3.1) While this perhaps uncertain process may appear somewhat ambiguous the WSS could serve in this case as an explicit illustration of the concept. As said earlier, WSS can play bilateral input output inter pla yer role in vascular physiology. For example the flow pattern property function which takes as input more physically intuitive properties like pressure, lumen area, or flow velocity provides the WSS as an output product intermediate. WSS then manifests the tonus of vasculature, which can induc e over

PAGE 71

! 55 chronic period of lasting plasticity changes morphology change, essentially translating the tissue properties. The proposed reasoning can be summarized in following diagram: ! ! ! ! ! !" !" !" ! ! ! !" !" (14) where the (flow property function) takes as in put variables as pressure, luminal area, or flow velocity and demeans the behavior of tissue on multiple scale level giving a rise to variable chemical morphogens concentrations C i and at the same effecting macro scale histologic ( Hx ) and geometric properties ( Gx ). The WSS ! then plays a role of inter playing function, which can relate both differentiable aspects of the cardiovascular system. The most important property of this sys tem is the functional interdependency and reversibility Based on this theory, one can appreciate the role of the WSS, and see that the abilit y to quantitatively assess its value may serve to discovery of a deeper and essential concepts in PAH pathophysio logic processes. While this system can be recognized as idealized, the large amount of basic and clinical research work indicates that this mechano transduction principle should not be ignored 16 17 51 The endothelial monolayer and SMCs in media are under influence of stretching forces variable in their magnitudes and periodicities. In instances of higher or maintained level of W SS, the vessels tend to preserve their dilated state and non proliferative behavior. Inversely, low WSS leads to decrease in luminal diameter and overall thickening of the vessel wall. From the clinical perspective, the assessment of WSS has been already explored in terms of genesis and predisposition to atherosclerotic lesions, systemic hypertension, diabetes, and pulmonary hypertension as well. The primary non invasive imaging methods for the WSS require high spatial resolution and ability of full set

PAGE 72

! 56 ch aracterization of vessel at the computation site. These conditions come from the most concrete definition of the WSS: ! ! !" !" (15) where is dynamic viscosity and !" !" represents the strain term. The WSS can be measured via more than one method. Lagrangian and tangential techniques were already proven to show promising results 33 In the association with the PAH, the WSS was investigated in proximal pulmonary vasculature using combination of segmentation analysis and Phase Contrast MRI (PC MRI) with promising results 34 Nevertheless, this study introduces the WSS computed based on patient specific dynamic viscosity and generated velocity profile curves throug h the pulmonary blood vessels. Importantly, the ability of specific filtration methodology using a stream lines patterns enabled the computation of the WSS in dista l lobar arteries of right lung. Assuming that the WSS will in parallel associate with trends of standard diagnostic markers obtained by RHC, it can be asserted that the WSS can be used as the direct diagnostic tool for PAH. As of now the very exact mechanism of the WSS transduction and role of morphogen is unknown and one may even consider the WS S alteration to a primary starting point of PAH pathophysiology. However, the evocation for the non invasive diag nostic tool has not been yet called for and the WSS certainly is the best candidate. Relatively less known associated descriptors of the WSS w ere described in the past with the main focus on the systemic vasculature 29 52 These indexes take into consideration the WSS oscillatory and frequency characteristics which may possess specific variability throughout the cardiac cycle. Importantly, the frequency analysis of th e WSS should not be avoided in discretization of the WSS impact on the endothelial

PAGE 73

! 57 cell mechano transduction. Among theses WSS closely related derivatives are oscillatory shear index (OSI) and residual resident time (RRT) both previously described in the a ssociation with endothelial morphology and atherosclerotic studies. The former mentioned marker is defined as: !"# ! ! ! !"" ! ! !" ! !"" ! ! !" ! ! ! !"# ! ! (16) and is used for the identification of the vascular regions undergoing highly oscillatory stress leading to progressive hyperplastic changes in vessel wall. Values approaching zero are then associated with the luminal locations with limited flow disruptions The RRT can be related to OSI via following relation: !!" ! ! ! !"# !"" !" ! !"" ! ! ! !" ! ! !"" !" ! (17) where WSS TA represents the time averaged WSS described later in the method section. Methods In this particular study was the WSS examined in the Butcher patient cohort (introduced in previous chapter), which possesses the unique feature of the same day RHC and 4D/2D CMR. The concept of the WSS truly describes the local fluid tissue interface. Thu s the comparison and correlation with in situ measured RHC hemodynamic factors if of great value. The data pre processing algorithm was indeed the same as described in previous chapter, while the subsequent post processing algorithm in this differed signif icantly. Intuitively, the localization and geometric consideration of the proximal pulmonary vasculature is done with relative ease. With the analytical progression toward the distal branches of pulmonary ar terial tree, the luminal shapes

PAGE 74

! 58 tracking and over all spatial resolution decreases due to anatomical and technical reasons. The present character of given datasets enabled accurate computation of the WSS in the MPA and primary main branches RPA and LPA. The acquisition field view then additionally provide d sufficient information for computation of the WSS character in three branching arteries originating directly from the RPA lumen composing anatomically speci fic lobar arteries ( Figure 14 ) In the figure portrayed above can reader visualize the specific computation planes. The normalization protocol for the computation of the WSS in the MPA was determined to be one cm distally from the pulmonary valve. The R PA and LPA planar segments were set to be approximately two cm f rom the main bifurcation orifice. The upper right lobe artery (URPA) plane was defined to be 2 cm from the main branching ostium, and similarly the segments of middle and lower right pulmonary arteries (MRPA and LRPA) Figure 14 Artistic representation of the WSS sites measurements in the pulmonary vasulature.

PAGE 75

! 59 were selected to be 2 cm passed the distal RPA bifurcation fork. While the selection of the particular compuation location may introduce a certain degree of variability, the chosen vascular segments compose relatively conserved an atomical region with minimal functional change. The velocity vector field generated by 4D flow CMR and pre processed in Paraview software kit, provided initial visualization of the velocity vector glyphs. Every computational plane was positioned orthogonally to the direction of the flow. In the case of distal arteries, the filtered stream lines were used an assistive guidance tool for localization of lobar arteries (See Figure 15 ) The four velocity profile curves were generated in the proximal pulmonary vessels, with careful placement of the line edges Figure 15. Generated streamlines from the velocity vector glyphs for the visualization of the proximal and distal pulmonary vasculature.

PAGE 76

! 60 within the endothelial boundary. This set up provided of total eight points for the WSS measurement with special designation names (MPA1 8, RPA1 8, LPA1 8). The reader is encouraged to explore the F igure s 1 6 and 1 7 (next page) in order to see the computational methodology and localization of the specific computation points. D istal pulmonary arteries were analyzed with two velocity profile curves providing four calculation points. The principal quantitative assessment of the WSS was based on the equation 15, as product of the patient specific dynamic viscosity and strain rate a t the juxta position to the vessel wall. The dynamic viscosity was computed using the logarhytmic relation between hematocrit and viscosity proposed by Barras 52 : Figure 16 The work flow process of generating the velocity profile curves. a) visualization of the vector field b) Superimposed scout image of the pulmonary artery c) filtration and localization using streamlines method d) final drawing of the velocity profile lines.

PAGE 77

! 61 ! !" ! !"#!$ !"# ! !"## (18 ) where the Hct represent the percentile form of hematocrit. The final version of the WSS equation was then formed into: ! !"# !" !" (19 ) with hematocrit dependent d ynamic viscosity and point specific strain du/dy. The dimensional configuration was set to provide the units in the standard SI form of N/m 2 The in plane WSS (WSS i p ) was computed as an average value from all point locations in selected plane. Additionally, the time averaged WSS (WSS TA ) was computed for one complete cardiac cycle. The full post processing method and WSS calculation in every artery was standardized by using specifically designed Matlab program providing the full scale information regarding the WSS and flow analysis per patient basis. Figure 17 Indexing of the specific points in the plane of computation for WSS analysis. The same trend was maintained throughout the entire vasculature.

PAGE 78

! 62 The RHC indexes were collected and organized in the very same manner as in vorticity studies in chapter 2. The subsequent statistical analysis for inter group variability for the helicity calculation was then assessed using Wilcoxon ranked sums method and reported in the form of median value and quartile ranges. The multivariate analysis between the set of helicity computed values and RHC data was done using Spearman Rho coefficients. In both cases p < 0.05 was considered as a significant finding. Results The simultaneous complex analysis of the WSS in the entire pulmonary circuitry was enabled using aforementioned specific Matlab script providing the complete set of information. The reader can view the sample script output in the Figure 18 (next page) The overview of poin t and vessel specific WSS i p calculations accompanied by the flow analysis is provided. The filtration process using specific streamline patterns enables clear identification of the proximal and distal vessels in all subjects. However, the further analysis in more distal tertiary segmental arteries would be subjected to high anatomical variab ility and insufficient resolution. As expected and following the already established trend, the WSS i p measured in all proximal pulmonary arteries was larger in control subjects than in PAH patients ( See Table 25 ) WSS (N/m 2 ) Control PAH p value Upper Right Lobe A. 0.56 (0.38 0.58) 0.31 (0.21 0.38) 0.0523' Middle Right Lobe A. 0.11 (0.05 0.16) 0.09 (0.06 0.11) 0.5357 Lower Right Lobe A. 0.58 (0.33 0.66) 0.24 (0.13 0.38) 0.0231* Main Pulmonary A. 0.13 (0.11 0.18) 0.09 (0.08 0.13) 0.0431* Right Pulmonary A. 0.40 (0.34 0.59) 0.17 (0.13 0.28) 0.0044* Left Pulmonary A. 0.13 (0.08 0.20) 0.10 (0.06 0.13) 0.3000 T able 25 Wall shear stress analysis

PAGE 79

! 63 The significant difference between control and PAH group wa s in L RPA, MPA and RPA. However, the URPA showed nearly significant variability. Nevertheless, the most si gnificant finding was observed in the RPA (0.40 vs. 0.17 N/m 2 p = 0.0044). The multivariate analysis explored the association of the WSS i p with RHC parameters. In general, every artery except URPA, revealed multiple number of significant correlations (See Table 26 ). Every correlating vessel show ed negative strong significant correlation with the PVR (See Figure 19 ) The strongest correlations overall provided RPA. T he Figure 18 The Matlab script output, here only shown for the distal vasculature. The bar graphs on the left side depict the WSS at specific in plane point.

PAGE 80

! 64 sta ndard PAH diagnostic indexes mPAP and PVR, provided the most significant correlati on in the RPA specific analysis (mPAP: rho = 0.7152, p = 0.0003; PVR = 0.8031, p < 0.0001). However, there were no correlations with the third diagnostic determinant PCWP. Complex cardiovascular system indicator E a /E max correlated in convincing negative fashion with the two primary branches RPA and LPA, and also with two distal arteries MRPA and LRPA. Cardiac functional and physiologic indicators RV EF and RV SV correlated positively with the same set of arteries, with RV SV additionally providing the association with the MPA. Interestingly, relatively distant RA showed significant negative correlations via mRAP with the MRPA and LRPA. The reader is enc ouraged to explore the details of the WSS i p c omplete multivariate analysis Table 26 shown on the next page Figure 19. The WSS correlations with the RHC diagnostic determinants, mPAP (left) and PVR (right ).

PAGE 81

! 65 U R P A MRPA LRPA RHC rho p value rho p value rho p value mPAP 0.3706 0.0982 0.4278 0.0531 PVR 0.3801 0.0892 0.5783 0.0060* 0.3992 0.0730* PCWP 0.1097 0.6360 0.3258 0.1495 0.4417 0.0450* mRAP 0.2386 0.3252 0.4684 0.0431* 0.5515 0.0144* RV CI 0.2496 0.2885 0.4331 0.0565 0.3521 0.1278 RV SV 0.2519 0.2706 0.439 0.0465* 0.5289 0.0137* E a /E max 0.0795 0.754 0.7606 0.0002* 0.5382 0.0212* RV EF 0.0083 0.9724 0.5062 0.0228* 0.2781 0.2351 MPA RPA LPA RHC rho p value rho p value rho p value mPAP 0.4298 0.0518 0.7152 0.0003* 0.6192 0.0028* PVR 0.5796 0.0059 0.8031 <0.0001 0.6555 0.0013 PCWP 0.0261 0.9105 0.2129 0.3542 0.0624 0.7882 mRAP 0.0539 0.8265 0.3014 0.2099 0.1343 0.5835 RV CI 0.3353 0.1484 0.4075 0.0745 0.4287 0.0593 RV SV 0.6195 0.0027* 0.6610 0.0011* 0.5788 0.0060* E a /E max 0.4221 0.0810 0.6698 0.0024* 0.6061 0.0077* RV EF 0.337 0.1463 0.5724 0.0084* 0.5546 0.0112* Unlike the WSS i p the WSS TA did not reveal any significant inter group variability and major multivariate correlations. However, the overall trend was kept intact, when control group exhibited larger WSS TA than the PAH in every proximal and distal vessel. Discussion The role of the WSS as pathogenic mediator in vascular diseases is currently one of the most discussed topics, which may convey the known pathophysiologic changes caused by still considered idiopathic nature of the PAH. It is of great importance that the modified profiles of WSS in the vasculature should be investigated when taking into acc ount the already explored vari ations in pediatric and adult subpopulations on both T able 26 Wall Shear Stress Multi Parameter Regression Analysis

PAGE 82

! 66 systemic and pulmonary side. The mechano transduction, while envision as a micro environmental phenomenon is dependent on the basic physiologic principles. The complex system of cells hierchially and log istically categorized into large skeletal musc l e organ can be can be considered as parallel model to vascular wall and its various tunicas The mechanical stress on this organ can be considered as an act of exercise with the variability in the load and short term and long term frequency. When the stress (i.e. WSS) is applied in the controlled amount muscle will maintain its strength, shape, and will b e able to respond to sudden stress (metabolic or mechanic). When the load and tension are applied imprope rly over pro longed period of time the muscle will undergo morphological changes. Comparatively, pulmonary vascular wall is placed under wall shear throughout the entire life period, and when the applied stress is decreased the vessels undergo histological and geometrical changes, similarly as a skeletal muscle when offloaded from mechanical stress produced over a longer period of time. While this theory is only an approximation to the complexity of vessel wall and neighboring pulmonary tissue and intersti tium, the already carried investigation done via MRI technology, shows that the WSS can be considered as an important mediator and marker of the present collective morphogenesis in the PAH. Decreased WSS had been observed in all proximal pulmonary vessels in both adults and children. (These effects has been thoroughly investigated by A. Barker & U. Truong) 34 53 Considering the further investigations carried in the systemic side of circulation, it would be tremendously ambitious to ignore or not think that the WSS physiology is not causative (if not the primary in sult) in many vascular diseases. In the previously discussed studies, the authors used single plane PC MRI measurements in specific location in the proximal pulmonary

PAGE 83

! 67 circuitry. Their findings are corresponding to those produced in this study using 4D flow CMR and flow profile curve methodology. The decreased WSS was indeed observed in all investigated vessels with two proximal vessels, RPA and MPA, providing the significant difference. It was somewhat expected and exciting to observe that two of the distal lobar arteries, LRLA and UR LA, indicated significant inter variability as well (URLA showed near significance with p = 0.0523). Perhaps the most important finding were large number of strong associations between measured maximum WSS i p and RHC in dexes standardly used for definitive diagnosis of the PAH. The RPA showed the most significant associations following already observed trend. From the clinical translational perspective the most valuable correlations are between the WSS measured in RPA and mPAP and PVR (mPAP: rho = 0.7152, p = 0.0003; PVR = 0.8031, p < 0.0001) Retrospectively, perhaps one possible advantage of this method over those used in previously mentioned work s might have been in the single CMR acquisition and tracing of the flow into the distal lobar pulmonary arteries which exhibit the decreased WSS in the PAH patients as well. When considering the PVR and mPAP measured via catheterization to be also the mark of the PAH severity we can see that WSS has a great potential for the p rognostic measurements if periodical imaging examination would be considered and tested. Interestingly the WSS TA did not reveal any significant variations while the median inter group values remained to be higher in the control group for every considered a rtery. Regarding the WSS one must consider the gross anatomical variations among the patients and the direct effect on the WSS computation. For example, the LPA is from the luminal perspective more continuous with the MPA and tends to be of larger diamete r

PAGE 84

! 68 than RPA. On the other hand, RPA deviates from the MPA under sharper angle, when vessel in approximately 2 cm segment change path under relatively large acute angle, which can be in some individuals close to 90 degrees. It is then interesting to see that the most significant changes in WSS i p appear to happen in the mid RPA section. The flow at this location reflects a prominent pulsatile fashion, when streak of the blood is redistributed from the posterior wall of the MPA bifurcation point. Even more variability is present among the distal vessels when bifurcating ostia of lobar and further segmental arteries deviate at various sides with non uniform anatomical pattern creating an obstacl e to large scale studies. Consequently, future studies should investigate the detailed anatomy geometrical effect on the WSS in order to take into account fully morphological aspect of the in situ vascular condition. Conslussion The WSS remains to be the most sensitive marker of the PAH assessed by non invasive imaging technique. As shown in this experimental set up, several routes toward the computation may be taken. The 4D flow technique definitely benefits from the additional localization of the distal arteries and in situ visualization of the flow (interpolated in the form of velocity vector field). The technique would further benefit from the higher degree of fidelity regarding the viscosity measurement. While the logarhytmic approximation based on he matocrit provides relatively easy method to assess the dynamic viscosity, this assumption is valid solely on the rheological and thixotropic properties prescribed by the majority content of RBCs. The further studies may consider the use of viscosity meter as a more accurate quantitative method for assessment of the

PAGE 85

! 69 vital part of the WS computation. However, to the current knowledge of the author, no previous studies has used patient specific determinant for the viscosity calculation. Additionally, this stud y was limited by relatively small sample size, and future work would benefit from the integrated multi center study investigating the WSS computation from the 4D flow CMR. The selection of analysis location introduced another variability in the calculation and should be acknowledged since the WSS is not measured along the entire vessel length. However, it is expected that these changes would introduce minimal error since the depicted vessel segments represent relatively conserved anatomical region. The ana lysis of the WSS in the frequency domain might provide another information with regar d to the nature of the disease, since it is a valid assumption to expect that oscillatory nature and specific directionality of the shear stress on endothelial cells may play both intrinsic and extrinsic role in the PAH propagation and stiffness development. The biochemical assays from mammalian models provide some initial pioneering insights and may further evocate inquiries leading to detail pathogenesis of the PAH and effect of the WSS as of principal mediator. Nevertheless, it is hoped that after review of this work, further experiments will focus on the WSS investigations mediated by CMR technolo gy as this coupled research con tinuous providing promising results.

PAGE 86

! 70 CHAPTER IV HELICITY Theory Helicity is the third novel hemodynamic parameter introduced in this work, and perhaps the least intuitive one. The theoretical groundwork surrounding helicity has relatively simple roots, while the application of this highly dynamic feature is reserved mainly for hydrodynamic studies in meteorological disciplines. Perhaps the scala r nature of this measure introduces less intuitive interpret ation in the close d vessel system. Despite the problem of discrete visualization, the helicity is alr eady used as a descriptive flow pattern tool in the aorta, in association with diseases prominent in the proximal systemic circulation like aortic dissecti ons, aneurysms, or valvular stenosis The leading and seminal work toward the quantification of helicity done by Morbiducci introduced several higher spatial derivatives related to helicity computation 37 Simplified mathematical definition of helicity provides the first hint for potential visual representation of the helical flow effect: ! ! ! ! ! !" (20) where one can recognize ! ! as the spatially dependent vorticity. The dot product of the velocity field and respective vorticity generated field composes the scalar product termed as a spatially dependent helicity. However, the above mentioned equation may be easily converted into the time dependent variable as well. The degree of alignment between two vector fields is from the kinetic energy p erspective a crucial indicator of flow stability. At this point analogous real life exa mple may provide a better understanding of helicity. From the aerodynamic perspective, synchronization of pitch and torsion is the crucial

PAGE 87

! 71 element in order to generate efficient flight pattern. For example, the properly thrown football has a straight veloc ity trajectory accompanied by long axis spin. If one would then mark the point at the widest radius of the football and create a kinematic plot, perfect spiral or helix would appear (See Figure 20 ) Contrary the ball with twirling motion in the main veloci ty trajectory and/or imperfect axial spin would have diminished flight properties. If one would translate this analogy to the stroke bolus of a fluid the pattern in the closed vessel system can be observed Assuming same amount of input energy is given to the ejecting stroke the bolus of blood propagating in the former described manner will conserve more amounts of kinetic translational energy and propulse' further in the vessel. On the other hand, when the velocity fiel d (directionality) is misaligned with the Figure 20 The artistic representation of the football flight analogy. The top flight trajectory represent efficiently thrown football with direct trajectory and correct spin. The bottom version is analogous to the bolus of fluid propagated in the dys functional system.

PAGE 88

! 72 vorticity one, the stroke bolus will dissipate the kinetic energy in to the elastic one vanished in the vascular wall compliant components. In summary, helicity represents the level of alignment between velocity an d vorticity. Better alignment allows for more laminar efficient flow and we observe less dissipating energy along the vessel path. Contrary, the misalignment will produce turbulent flow patterns with induced energy loss. This chapter informs about an in vestigative attempt to identify if PAH driven by morphological changes worsening the local and complex flow patterns consequently decreases helicity in proximal pulmonary arteries. Additionally the more detailed qualitative analysis can be assessed to diff erentiate between levorotatory and dextrorotatory helices in various segments and patient groups. However, these directional changes can be spatially visualized only using the velocity vector field, since the scalar nature of the helicity will limit itself from the quantitative point of by change of sign. Several Eularian descriptors of helicity were already introduced and used for the flow and shear assessment in carotid arteries 54 In this context, the work flow conducted in this experiment was designed to calculate helicity in the proximal pulmonary circulation, using the spatial integration at the peak vascular systole. Similarly as in vorticity, helicity can provide information about the quality of flow using non invasive 4D flow CMR tech nique. This approach can spare patients of invasive RHC, NO challenge, and additional exercise induced stress measuring techniques performed in catheterization labs. The principle aim of this study is to assess the difference in helicity in vascular compar tments between PAH and control subjects. The reduced helicity in proximal pulmonary conduit in PAH patients, would support the

PAGE 89

! 73 mechanical/mathematical theory of reduced flow efficiency in PAH vessels. As stated before, the non invasive aspect of this study would allow for more frequent follow up studies and monitoring of the patient specific disease state and response to therapy. Methods Similarly to the vorticity calculation protocol in the pulmonary arteries, the Butcher patient cohort was selected as a suitable dataset for the high value same day RHC measurements. For reminder, the Butcher cohort consisted from 17 PAH patients and 5 age matched controls. The same RHC protocol was conducted as described in previous chapters. The post processing involved thorough selection of the anatomical region for the spatial integration. The helicity was then computed at the peak vascular systole identified ba sed on the vector glyph pattern using following equation serving as an input to the Paraview programed pipelin e: ! ! ! ! (21) where the H + represents the left handed helix component (+ sign of helicity component) and H indicates right handed helix component ( sign of helicity component). The exact definition for the MPA luminal integration region was set be from the pulmonic valve to the far end of the bifurcation wall with the wall parallel boundaries of the main lumen ( See Figure 21 ). The RPA region was this time defined from the edge of the parallel line (from MPA boundary) to the distal bifurcation of the right middle and right lower lobar arteries. The proximal ostium giving a rise to the upper lobar segment artery was then considered as part of the calculation region coefficients

PAGE 90

! 74 Additionally, both MPA and RPA total helicity values were combined into one region in order to encounter for possible discrepancy between the definitive boundaries within the considered vascular segments. Inter group variability for the helicity calculatio n was then assessed using Wilcoxon ranked sums method and reported in t he f o r m of me dian value and quartile ranges. The multivariate analysis between the set of helicity computed values and RHC data was done using Spearman Rho coefficients. In both cases p < 0.05 was considered as a significant finding. Results The helicity was and its scalar components were calculated in each patient. Two patients who were on the PAH therapy, were excluded from the inter group analysis, but were part of multivariate cor relation assessment. The reason for the exclusion is contemplated from the possibility of the vascular flow, which would improve the overall conduction and final helicity. The total helicity was larger in both MPA and RPA in Figure 21 Helicity integration regions for MPA and RPA.

PAGE 91

! 75 control groups (MPA: 32.0 vs. 1 0.3, p = 0.0251; RPA: 10.2 vs. 1.6, p = 0.0041). Additionally, there was significant variability found in the H component in also in both vessel segments (MPA: 25.5 s 2 vs 6.0 s 2 p = 0.0137; RPA: 5.9 s 2 vs. 0.4 s 2 p = 0.0036). Precisely, the H provided stronger correlations than total helicity. The H + did not reveal any significant finding. Furthermore, the RPA was again more sensitive to the overall computation of give novel hemodynamic index. The r eader is advised to review the T able 27 with t he comprehensive analytical summary. Helicity ( m 2 s 2 ) Control (n=5) PAH (n=16) P value MPA H(+) 1.6 (0.3 8.5) 1.9 (0.3 10.3) 0.8228 MPA H( ) 25.5 ( 32.6 | 12.5) 6.0 ( 10.5 | 0.99) 0.0137 MPA H total 32.0 ( 15.2 33.9) 10.3 (5.1 19.7 0.0251 RPA H(+) 4.5 (0 .1 9.9) 0.4 (1.9 3.9) 0.7201 RPA H( ) 5.9 ( 6.9 | 3.7) 0.4 ( 1.3 | 0.2) 0.0036 RPA H total 10.2 (6.5 14.4) 1.6 (3.2 4.4) 0.0041 RPA + MPA 42.91(21.79 47.68) 12.43 (7.64 25.2) 0.0156 Multivariate analysis showed numerous correlations with RHC indexes with helicity data sets form both arterial regions. The standard diagnostic measure mPAP indicated strong negative correlations with total helicity in both MPA and RPA (MPA: rho = 0.5574, p = 0.0074; RPA: rho 0.6279, p = 0.0018). As the rational dictates, the same correlation trend was observed with the PVR (MPA: 0.5442, p = 0.0088; RPA: 0.6064, p = 0.0028). However, in the scope of analysis regardi ng the helix directionality and its separate components, it was exciting to find that H revealed strong positive correlations with mPAP and PVR, with the significance superior to the total computed helicity in MPA (mPAP: rho = 0.6517, p = 0.0010; PVR: rh o = 0.6019, p = 0.0030) In the RPA, the H computations did not exceed from the significance point view the total T able 27 Helicity analysis in proximal pulmonary vasculature

PAGE 92

! 76 helicity measures. However, the positive correlation with mPAP and PVR were observed here as well (mPAP: rho = 0.5578, p = 0.0070; PVR: rho = 0.5499, p = 0.0080). Additionally, the summed total helicity for both MPA and RPA regions showed significant correlation with every RHC marker besides PCWP and mRAP. The strongest associations of this study were found between total MPA +RPA helicity and E a /E max (rho = 0.7696, p = 0.0001) and with the systolic performance markers RV EF and RV CI (RV EF: rho = 0.6790, p = 0.0007; RV CI: rho = 0.6548, p = 0.0013). The complete multivariate analysis is presented in the Table 28 on the next page [=] m 2 .s 2 MPA H(+) MPA H( ) MPA H Total MPA + RPA RHC index rho p value rho p value rho p value rho p value mPAP 0.1971 0.3794 0.6517 0.0010 0.5544 0.0074 0.5737 0.0052 PVR 0.2244 0.3155 0.6019 0.0030 0.5442 0.0088 0.5657 0.0061 PCWP 0.0318 0.8882 0.3618 0.0980 0.2301 0.3030 0.2709 0.2226 mRAP 0.0266 0.9114 0.4559 0.0433 0.3328 0.1516 0.3190 0.1974 RV CI 0.5599 0.0083 0.2871 0.2070 0.6262 0.0024 0.6548 0.0013 RV SV 0.1276 0.5713 0.3095 0.1610 0.4948 0.0192 0.5298 0.0112 Ea/Emax 0.0430 0.8612 0.5678 0.0112 0.7398 0.0003 0.7696 0.0001 RV EF 0.2307 0.3144 0.4224 0.0565 0.6855 0.0006 0.6790 0.0007 [=] m 2 .s 2 RPA H(+) RPA H( ) RPA H Total RHC index rho p value rho p value rho p value mPAP 0.1824 0.4166 0.5578 0.0070 0.6279 0.0018 PVR 0.2294 0.3034 0.5499 0.0080 0.6064 0.0028 PCWP 0.0988 0.6617 0.3789 0.0821 0.3414 0.1200 mRAP 0.0456 0.8486 0.4385 0.0531 0.4559 0.0433 RV CI 0.2105 0.3598 0.2832 0.2135 0.4898 0.0242 RV SV 0.1356 0.5475 0.5795 0.0047 0.5750 0.0051 E a /E max 0.3273 0.1713 0.5818 0.0090 0.6994 0.0009 RV EF 0.2703 0.2360 0.3710 0.0978 0.4971 0.0219 Interestingly, t he most significant correlations in the entire experiment were observed in both MPA and RPA total helicity with the E a /E max (MPA: rho = 0.7398, p = T able 28 Helicity Multi Parameter Regression Analysis

PAGE 93

! 77 0.003, RPA: rho = 0.6994, p = 0.0009). Additional strong associations for both vessels and total helicity measures were with C I (+), RVSV (+), RVEF (+), and mRAP ( ) as an extra observed correlation in RPA. The H + provided only one major correlation with the RV CI (rho = 0.5599, p = 0.0083). Discussion To the best knowledge of author s this is the first time assessment of helicity in the pulmonary circulation and its integration to the study of the pulmonary hypertension. While the concept itself may be at the first notion bewildering, the association of helicity with the real life phe nomenon (e.g. football flight efficiency) may help to understand the propulsion of the stroke volume throughout the vasculature. It is important to carry in mind, that helicity and helix formation are two different concepts, similar a s the vorticity is to vortices However, they symbiotically a ffect each other into complete flow description of particular order in quantitative and qualitative fashion. Helicity then represents the scalar field similarly as the electric field, but in this case tells in what de gree is the vorticity and velocity field aligned at specific spot in the space. "#"! $#"! %"#"! %$#"! &"#"! &$#"! '"#"! "! $! %"! %$! &"! ()*!+,--./0!12345! 6783934:!+; & #0 <& 5! "! %"! &"! '"! ="! $"! >"! ?"! "! $! %"! %$! &"! ;(@(!+;;6A5! 6783934:!+; & #0 <& 5! *B-!C! < "#>">= D!C!"#""&E *B-!C! < "#>&?F D!C!"#""%E Figure 22 RPA computed helicity and correlations with the RHC diagnostic indexes PVR (left) and mPAP (right).

PAGE 94

! 78 The data collected in this study confirmed that helicity is decreased in the PAH, and that the diseased effected morphological changes in the vasculature conduct flow in less efficient way. However additional measurements directly assessing the ventricular wor k and related energetics would be necessary to underline this hypothesis. It was interesting to find that H showed more significant inter group variation than overall total helicity in both analyzed vessels. At the same time the H showed positive correlations with mPAP, PVR, and E a /E max Contrary, the total helicity correlated in positive fashion with the above mentioned parameters (See Figure 22 ) While this implementation might appear at the first view confusing the transposition of th e graphical quadrants and realization that the larger negative values are associated with the lower mPAP, PVR, and E a /E max values enables better understanding of the resulting concept. The presence of H + is less clear, and its quantitative presence, could suggest that fluid particles are following the helical pathway in opposite direction. However, it cannot be said with a certainty that this effect is beneficial or energetically withdrawing from the coupled system, since the inter group variability was not significant in neither vessel segments. The fact that VVCR ratio showed the most significant association suggests that helicity might be novel and sensitive marker to assess the cardio vascular system as a one functional unit. After all, the ventricle generates and propels the bolus stroke in specific manner and vessel conducts the given bolus distally and this translational part has its own properties as well which can be easily altered. Thus, considering again the football flight analogy, initial magn itude and nature of the throw and spin play important role together with the distal flight conditions (wind, drag) both effect the flight and its efficiency. The former aspect can be then applied to the ventricular performance, while the latter one to

PAGE 95

! 79 the vascular Perhaps this demonstration might appear highly grotesque, however, it is hoped that after presentation of the calculated data and provided analogy, the concept of helicity would become less enigmatical. Conclusion From all novel hemodynamic descriptors introduced in this work, helicity provided the strongest association with the E a /E max This finding suggests, that helicity can be very indicative of the actual mechanical state of the couple (or decoupled) system where and at the same as indic ator of the PAH presence. Scrupulous analysis of the collected data provided novel insight into actual helicity parameters with respect to multi unit functionality performance, whereas most of the previous work secluded helicity and its associated metrics to single vessel segment. The propagation of the flow through pulmonary vasculature in the unremitting fashion is crucial for entire cardio vascular insicitium and since the effectiveness of this complex motion further implies to tissue morphogen esis towar d the undesired state, and resulting pathogenesis of the PAH as we know it. It is important to notice that helicity might not be the best diagnostic route of testing in the case of early stage of the PAH where the cardio vascular system is still coupled and does not aggravate distinct histological changes in the RV. The relatively small total patient cohort also limited this analysis. Additionally, the limited temporal resolution and the non exact non uniform voxel volume may introduce certain level of error. Further studies may take advantage of the multi center co operation and benefit from larger number of control subjects. Eventual computation of the helicity trough out

PAGE 96

! 80 the entire cardiac cycle may provide new insights and set of correlative data as well. Nevertheless, this analysis provided possibly new perspective on the car diovascular system, and hopefully introduced new set of questions for further discussion about the PAH mechanism.

PAGE 97

! 81 CHAPTER V DIAGNOSTIC ALGOR ITHM Introduction This last part of this work will be written in slightly futuristic fashion and it is hoped that its contents will inspire research and clinical investigations toward more aggressive approach in PAH diagnostics. As stated in the introductory chapter, the only and limited method for definitive diagnosis is via RHC. While this metho d provides sensitive and valuable data accompanied by additional information regarding the cardiac systolic performance, it still represents the certain degree of sensitivity. It is also important to mention that RHC is in some cases accompanied with the v asoreactive testing such as nitrous oxide challenge, and in some instances angiography is done when thromboembolic disease or other PH functional class is suspected. Additionally, the therapeutic effects are assessed via RHC as well. In the large multicen ter retrospective and prospective study, Hoeper et al investigated mortality and morbidity associated with the RHC 55 From 7,218 recorded procedures done in 4 year p eriod, 1.1% of the cases resulted in seriou s complications and less than 0.1% of fatal complications. The nature of the complications varied from induced hematomas, vasovagal reactions, induced pneumothorax, or in initiation of the venous access. The vast majority of the patients also under went the vasoreactive challenge. While these numbers are not alarming at the first site, one has to consider the fragility of some of the PAH patients as well as the recurring number of RHCs for monitoring and therapeutic evaluation purposes.

PAGE 98

! 82 The non invasive imag ing testing represents one of the best modalities for the assessment of the PAH. While the PAH is one of the five sub functional classes of the PAH, it can be safely diagnosed ba sed on proposed guidelines. The PAH is precisely distinguished from other categories, by elevated mPAP (# 25 mmHg) and PVR (# 3 WU) with PCWP maintained at physiologic levels ($ 15 mmHg). These criteria are sensitive to arterial pulmonary tree and at the same exclude the possibility of the left heart ongoing pathology. As of now the WSS represents the strongest potential for diagnosing PAH, with a great correlation studies with both mPAP and PVR. Vorticity and helicity computed in the proximal pulmonary arteries revealed significant variability between the diseased and control g roups as well. Sanz et al proposed using the pulsatility and additional measured indexes of pulmonary stiffness in the MPA measured from the same day CMR, with a relatively good diagnostic potential defended by the AUC of 0.91 (n=94) 43 The hemodynamic data acquired for the purposes of this work were obtained form the 4D flow CMR and compared to the same day RHC measures as well. Indeed, the patient population is smaller (n=23), but the proposed hemodynamic indexes are more sensitive t o the fluid tissue interaction and their data can be also used to study indirectly VVCR. At the same time all hemodynamic indexes can be computed from the single CMR acquisition with appropriate 4D flow and 2D scout image generation protocols. This study pro vided same day CMR study in tandem with either and RHC or ECHO and provided exclusive relationships and insights between the intrinsic mechanistic properties of the cardiovascular system and more extrinsic flow measures.

PAGE 99

! 83 Receiver Operating Curves Beside s the generated data, large amount of significant correlations, and overall analysis of morphologically driven hemodynamic changes, the calculated data sets were used to design a novel diagnostic algorithm. This algorithm could serve as a potential startin g point for developing ultimate universal diagnostic method, which would be universal and differentiable toward any PAH subtype and would be based solely on single CMR test. The standard evaluation method for newly proposed diagnostic technique is via rece iver operating curves (ROC ), which are composed from iterative calculations of sensitivity and specificity, based on obtained data sets. Generated ROCs can then provide intuitive sense of diagnostic efficiency and at the same time provide necessary cut off values for final diagnostic purposes. The widely accepted method for finding the desired cut off value in given ROC is v ia derivation of Youden index (YI ), which is defined as: ! !"#$%&%'%&( ! !"#$%&%$%'( ! (22) where i index specifies the given iteration in ROC calculation 56 The maximum YI value is then used to find the corresponding cut off value. All three described hemodynamic parameters were considered for the generation of specific ROC calculated using specifically designed Matlab program From the voriticity cal culations, the RPA maximum systolic vorticity revealed the strongest set of multivariate correlations with the PVR and mPAP, an were t hen used for ROC generation. The given range of values was 43.9 to 1072 .32 s 1 suggesting using 5 s 1 as a suitable itera tive step, and 1100 s 1 as a maximum iterative cut off point. The generated ROC curve for vorticity data is shown in the following figure.

PAGE 100

! 84 The computed ROC for the maximum systolic vorticity in RPA was 0.9400 (n=20). The calculated maximum YI was J = 0.7059, with a final cut off value 305.00 s 1 The input patient characteristics included 5 controls and 15 PAH confirmed patients. Two patients with PAH on medical therapy were excluded from the computation. Their vorticity values would place them into non PAH group and reducing the quality of the ROC to 0.9000. Three patients with vorticity values 485.61 s 2 318.12 s 2 and 367.26 s 2 would be improperly diagnosed. The patient with vorticity 485.61 s 1 had a boundary diagnostic measures with mPAP = 25 mmHg and PVR = 2.9 WU. Other two vorticity based misdiagnosed patients had exceedingly high mPAP and PVR. The in plane WSS generated also showed strong associations with RHC measured hemodynamics and was considered for the ROC analysis. Again, the RPA indicated strongest diagnostic potential based significant negative correlation with both mPAP and PVR. The measured WSS values value range was 0.083 N.m 2 to 0.755 N.m 2 The Figure 22 ROC curve for RPA systolic maximum vorticity.

PAGE 101

! 85 chosen iteration step was determined to be 0.005 N.m 2 with maximum iterative cut off point at 0.8000 N.m 2 The generated WSS ROC is displayed below: The computed ROC for WSS was 0.9412 (n=20). The calculated maximum YI was 0.8824 corresponding to cut off value 0.2700 N.m 2 The input characteristics are same as in the previous ROC analysis with two patients excluded from already described reasons. However, the two patients on therapy did not change the computed ROC value. Three be patients would improperly diagnosed with WSS values 0.298 N.m 2 0.492 N.m 2 and 0.279 N.m 2 The first patient is the one previously mentioned with boundary RHC deterministics. Patient with high outlying WSS was value had an exceeding viscosity value (% = 5.2863 cP). The third patient had a regular RHC data indicating PAH diagnosis (mPAP = 33, PVR = 3.6). Lastly, the spatially computed helicity in the RPA was used for the ROC a s well. The total maximum systolic helicity was used for the computation. The data range was Figure 23 ROC for measured in plane WSS in RPA.

PAGE 102

! 86 between 0.32 m 2 .s 2 and 14.86 m 2 .s 2 and iteration steps were chosen to be 0.05 m 2 .s 2 and the maximum iterative cut off point 15.00 m 2 .s 2 The generated ROC for total RPA helicity is shown below. The total helicity showed the highest ROC = 0.9889. The maximum YI was 0.9444 with corresponding cut off value 5.900 m 2 .s 2 The input characteristics were in this case different with 18 confirmed PAH patients and 5 controls. Two patients on therapy were included in this analysis not changing the ROC curve. Additional PAH patient (Butcher # 40) was just added to the study as t he latest protocoled subject. Single patient was misdiagnosed by this method with standard RHC hemodynamics indicating PAH (mPAP = 37, PVR = 8.5). Figure 24 ROC generated for total helicity in RPA.

PAGE 103

! 87 Diagnostic algorithm In summary, all generated ROCs provided diagnostic accuracy above 90%. Furthermore most of the misdiagnosed patients were somewhat controversive. For example, patients would be mismatched due to irregularly high viscosity, therapeutic effects, or they had boundary standard RHC diagnostic values. At this point it is important to mention that all of the proposed novel hemodynamic measures did not correlate with PCWP, which is important co determinant toward PAH diagnosis in order to exclude left heart chamber or valvular disease. At the same time strong correlation with the mPAP and PVR, qualifies the data for index specific and combinatory potential diagnostic analysis. The careful analysis of all outliers indicates that some technique, which did not properly match certain outliers, did on the other hand correctly diagnosed patients impr operly diagnosed in other sets. This suggests that combination of more than one of the novel techniques could serve as the input into diagnostic algorithm. At the point when technique may fail, the other can overpower the false validity and correctly diagn ose the patient. At the same time, the proposed new hemodynamic markers are direct descriptors o f the fluid tissue interactions, with each providing slightly different view of the cardio vascular system. While helicity can elegantly assess the VVCR, vortic ity can directly describe the flow at the solid fluid boundary, and WSS direct fluid layer inducing stress on endothelial cells. Unlike mPAP and PVR which both represent somewhat intrinsic description of the system in the form potential and resistance of f low. Resulting spectrum of the pathologic morphogenesis driven sensitive hemodynamic data can then provide physician the more accurate and state of the

PAGE 104

! 88 cardiovascular system and perhaps better formulate and dictate th e direction of the therapy. Following scenario might occur, when the patient suspected for PAH undergoes clinical evaluation and obtained CMR data indicate low WSS, elevated vorticity, and boundary level helicity. Rational explanation may have in this case multiple solutions, but closer evalua tion of specific data indicate that pulmonary vasculature and vascular pulmonary interstitium space is already compromised by disease progression, flow in the proximal v essels is irregular and that heart vessel system is still coupled but heart is already in the compensatory state. Similar analogies could be offered, but the same underlying effect remains: two out of three hemodynamic indexes correctly diagnosed a patient and together with the third marker they describe more precisely the state of the syste m This logically driven system would diagnose correctly the entire Butcher cohort except one patient who as mentioned before had a boundary RHC characteristics for PAH diagnosis. These observations can be summarized into inter marker plots allowing the d irect visualization effects of the diagnostic power The fact that all introduced hemodynamic indexes provided same nature of negative correlations with the RHC markers enabled generation of the 3 dimensional scheme of the entire patient cohort with depend ency on solely based on the calculated hemodynamics. The only present obstacle was an appropriate scaling factor among all indexes. The entire Butcher data set was then subjected to 2D and 3D specific diagnostical placement system, which revealed interesti ng and data corresponding patterns. The Figure 25 portrayed on the next page shows the specific 2D plots and comprehensive 3D plot for specific study enrolled subjects.

PAGE 105

! 89 The red areas, bounded by hemodynamic index specific values, represent in every case the region where patients have with 100% sensit ivity PAH. On the other hand blue regions represent region where patients does not have a PAH with 100% certainty. Remaining non shaded regions represent controversial patients, which are correctly diagnosed according to only hemodynamic index per specific plot. The important question then arises inquiring if this methodology would be applicable to a clinical setting and what would be necessary statistical power to approve this test. Figure 25 The diagnostic patient specific diagram. The top left figure shows all three data sets combined into 3D plot. All hemodynamic measures were appropriately scaled.

PAGE 106

! 90 We could answer the raised questions in the systematic fashion considerin g the every clinical aspect of the disease and the current diagnostic method For approximately three decades have been PAH and its closely related conditions diagnosed via RHC providing mPAP, PVR, PCWP, and flow dependent measures as a concrete data sets. If one would consider complet e and definitive RHC replacement by CMR study, the CO, SV, and PCWP data would be lost. Arguably, the flow conditions could be calculated from the 4D flow CMR despite currently sufficient accuracy of the rmodilution and Fick's method for CO computation. However, the term sufficient accuracy should not be used in any clinical terminology. While brilliant and revolutionary ideas gave a rise to thermodilution and Fick's method, the instrumental aspect of the measure acquisition lac ks the necessary precision. Most of the experienced medical centers with catheterization labs nowadays prefer thermodilution as a gold standard for diagnostic testing for PAH, since direct Fick's method is more demanding from the equipment point of view. However, the studies revealed, that simultaneous measurements of CO by the two methods can reveal significant discrepancy exceeding 20% differentiation between two techniques 57 Obviously, the diagnostic determination and implementation in prognostic or therapeutic assessments may be considerably compromised. Additionally, the thermodilution itself is thought to have a approximately 20% error rate as well, since the initial value inputs of body and room temperatures are highly prone to inaccuracy 58 As we can see, the quality of the RHC generated measures can be easily altered by many factors. The reader is at this point encouraged to reconsider th e allegoric situation introduced shortly above. While the CMR acquisition places patient under

PAGE 107

! 91 claustrophobic stress and general discomfort, the risk of iatrogenic injury is essentially zero, and requires patient to actively co operate for roughly 45 minut es. The computed hemodynamic indexes can then enable evaluating physician to analyze both cardiac and vascular function as a separate systems, and also as complex coupled system with sensiti ve description of morphogenesis patterns. This morphogenesis aspec t include s pathophysiologic nature of flow, altered vascular biomechanics, and gross overview of tissue remodeling. Conclusion There are still many physiologic and PAH induced phenomena yet to be determined in order to fully underst and the nature of th e disease. The palliation remains to be unfortunately the only possible route of treatm ent despite the extensive research on many fronts. The presented work investigated the PAH form slightly different angle where the discrete physiological processes were studied in conjuncture with the mechanical properties of bodily tissues. Several interesting correlations between the hemodynamic flow characteristics and standard clinical markers were discovered. Most importantly, these markers show strong potential for direct diagnostical purposes. While the investigated data sets were limited by small number of patients, it is hoped that the results of this work will propel further research in order to find further valuable determinants in the PAH and strengthen the cli nical translation. Comparative analysis of the novel hemodynamic parameters may be correlated with the impedance spectra and disease specific local tissue biomarkers like miRNA or certain types of matrix proteases. Even more prospectively, the evaluation of the system

PAGE 108

! 92 energe tics and specific tissue kinetics may provide new insights into the disease mechanism. At this moment, the invasive therapeutic approach seems very distant, mostly due to lack of the large group chronological retrospective data. Thus, the prognostic studie s need to be considered using the proposed hemodynamic factors in order to retain necessary research momentum. Finally, it is hoped that continuum of this research work will be applied also to remaining functional groups of PH which would benefit from the more accurate disease descriptors as well. The cardiovascular system is obviously a vital physiologic system. It is complex systems driven by many morphogens and perhaps one day it will be fully understand. The sincere and designated research of the engin e of the life may also provide new exciting foundations of classical medicine. This route may be possible via understanding of the PAH disease pathology. After all it would be great benefit to all patients, clinicians and the entire world population to un derstand human hearts.

PAGE 109

! 93 REFERENCES 1. Chaouat a, Chaouat a, Naeije R, Naeije R, Weitzenblum E, Weitzenblum E. Pulmonary hypertension in COPD. Eur Respir J Off J Eur Soc Clin Respir Physiol 2008;32(5 ):1371 85. doi:10.1183/09031936.00015608. 2. Thabut G, Dauriat G, Stern JB, et al. Pulmonary hemodynamics in advanced COPD candidates for lung volume reduction surgery or lung transplantation. Chest 2005;127(5):1531 1536. doi:10.1378/chest.127.5.1531. 3. Huber LC, Soltermann A, Fischler M, et al. Caveolin 1 Expression and Hemodynamics in COPD Patients. Open Respir Med J 2009;3:73 78. doi:10.2174/1874306400903010073. 4. Gali N, Hoeper MM, Humbert M, et al. Guidelines for the diagnosis and treatment of pulmonary hypertension: the Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS), endorsed by the Internat. Eur Heart J 2009;30(20):2493 537. doi:10.109 3/eurheartj/ehp297. 5. Fuster V, Walsch RA, Harrington R a. Hurst's The Heart 13th ed. The McGraw Hill; 2013. 6. Champion HC, Michelakis ED, Hassoun PM. Comprehensive invasive and noninvasive approach to the right ventricle pulmonary circulation unit: s tate of the art and clinical and research implications. Circulation 2009;120(11):992 1007. doi:10.1161/CIRCULATIONAHA.106.674028. 7. Farber HW, Loscalzo J. Pulmonary arterial hypertension. N Engl J Med 2004;351(16):1655 65. doi:10.1056/NEJMra035488. 8. McLaughlin V V, McGoon MD. Pulmonary arterial hypertension. Circulation 2006;114(13):1417 31. doi:10.1161/CIRCULATIONAHA.104.503540. 9. Humbert M, Sitbon O, Simonneau G. Treatment of pulmonary arterial hypertension. N Engl J Med 2004;351(14):1425 36. doi:10.1056/NEJMra040291. 10. Ma L, Roman Campos D, Austin ED, et al. A novel channelopathy in pulmonary arterial hypertension. N Engl J Med 2013;369(4):351 61. doi:10.1056/NEJMoa1211097. 11. Lammers SR, Kao PH, Qi HJ, et al. Changes in the structure fu nction relationship of elastin and its impact on the proximal pulmonary arterial mechanics of hypertensive calves. Am J Physiol Heart Circ Physiol 2008;295(4):H1451 9. doi:10.1152/ajpheart.00127.2008.

PAGE 110

! 94 12. Rabinovitch M. Molecular pathogenesis of pulmonar y arterial hypertension. J Clin Invest 2012;122(12):4306 13. doi:10.1172/JCI60658. 13. Vonk Noordegraaf A, Haddad F, Chin KM, et al. Right heart adaptation to pulmonary arterial hypertension: physiology and pathobiology. J Am Coll Cardiol 2013;62(25 Suppl):D22 33. doi:10.1016/j.jacc.2013.10.027. 14. Hunter KS, Lee P, Lanning CJ, et al. with Pulmonary Hypertension. 2011;155(1):166 174. doi:10.1016/j.ahj.2007.08.014.Pulmonary. 15. Stenmark KR, Fagan K a, Frid MG. Hypoxia induced pulmonary vascular rem odeling: cellular and molecular mechanisms. Circ Res 2006;99(7):675 91. doi:10.1161/01.RES.0000243584.45145.3f. 16. Malek AM, Alper SL. and Its Role in Atherosclerosis. 1999;282(21):2035 2042. 17. Barakat AI. Author s personal copy Comptes Rendus Phys ique Blood flow and arterial endothelial dysfunction&: Mechanisms and implications Author s personal copy. 2013;14:479 496. 18. McLaughlin V V, Archer SL, Badesch DB, et al. ACCF/AHA 2009 expert consensus document on pulmonary hypertension: a report of the American College of Cardiology Foundation Task Force on Expert Consensus Documents and the American Heart Association: developed in collaboration with the American College Circulation 2009;119(16):2250 94. doi:10.1161/CIRCULATIONAHA.109.192230. 19. Raymond RJ, Hinderliter AL, Willis PW, et al. Echocardiographic predictors of adverse outcomes in primary pulmonary hypertension. J Am Coll Cardiol 2002;39(7):1214 1219. doi:10.1016/S0735 1097(02)01744 8. 20. Sunagawa K, Maughan W. Left ventricular inte raction with arterial load studied in isolated canine ventricle. Am J Physiol 1983. Available at: http://ajpheart.physiology.org/content/ajpheart/245/5/H773.full.pdf. Accessed December 28, 2014. 21. Sanz J, Garc’a Alvarez A, Fern‡ndez Friera L, et al. Ri ght ventriculo arterial coupling in pulmonary hypertension: a magnetic resonance study. Heart 2012;98(3):238 43. doi:10.1136/heartjnl 2011 300462. 22. Foris V, Kovacs G, Tscherner M, Olschewski A, Olschewski H. Biomarkers in pulmonary hypertension: what do we know? Chest 2013;144(1):274 83. doi:10.1378/chest.12 1246. 23. Warwick G, Thomas PS, Yates DH. Biomarkers in pulmonary hypertension. Eur Respir J 2008;32(2):503 512. doi:10.1183/09031936.00160307.

PAGE 111

! 95 24. Bertero T, Lu Y, Annis S, et al. Systems leve l regulation of microRNA networks by miR 130 / 301 promotes pulmonary hypertension. 2014;124(8). doi:10.1172/JCI74773DS1. 25. Wei C, Henderson H, Spradley C, et al. Circulating miRNAs as potential marker for pulmonary hypertension. PLoS One 2013;8(5):e64 396. doi:10.1371/journal.pone.0064396. 26. Zisman D a, Schwarz M, Anstrom KJ, Collard HR, Flaherty KR, Hunninghake GW. A controlled trial of sildenafil in advanced idiopathic pulmonary fibrosis. N Engl J Med 2010;363(7):620 8. doi:10.1056/NEJMoa1002110. 27. Ghofrani H A, Gali N, Grimminger F, et al. Riociguat for the treatment of pulmonary arterial hypertension. N Engl J Med 2013;369(4):330 40. doi:10.1056/NEJMoa1209655. 28. Duarte JD, Hanson RL, Machado RF. Pharmacologic treatments for pulmonary hype rtension: exploring pharmacogenomics. Future Cardiol 2013;9(3):335 49. doi:10.2217/fca.13.6. 29. Himburg H a, Grzybowski DM, Hazel AL, LaMack J a, Li X M, Friedman MH. Spatial comparison between wall shear stress measures and porcine arterial endothelial permeability. Am J Physiol Heart Circ Physiol 2004;286:H1916 H1922. doi:10.1152/ajpheart.00897.2003. 30. Silber H a, Bluemke D a, Ouyang P, Du YP, Post WS, Lima J a. The relationship between vascular wall shear stress and flow mediated dilation: endot helial function assessed by phase contrast magnetic resonance angiography. J Am Coll Cardiol 2001;38(7):1859 1865. doi:10.1016/S0735 1097(01)01649 7. 31. Turing AM, Aug N, Turing BYAM. The Chemical Basis of Morphogenesis THE CHEMICAL BASIS OF MOKPHOGENES IS. 2007;237(641):37 72. 32. Barker AJ, Lanning C, Shandas R. Quantification of hemodynamic wall shear stress in patients with bicuspid aortic valve using phase contrast MRI. Ann Biomed Eng 2010;38(3):788 800. doi:10.1007/s10439 009 9854 3. 33. Cheng CP Parker D, Taylor C a. Quantification of Wall Shear Stress in Large Blood Vessels Using Lagrangian Interpolation Functions with Cine Phase Contrast Magnetic Resonance Imaging. Ann Biomed Eng 2002;30(8):1020 1032. doi:10.1114/1.1511239. 34. Barker AJ, Ro ld‡n Alzate A, Entezari P, et al. Four dimensional flow assessment of pulmonary artery flow and wall shear stress in adult pulmonary arterial hypertension: Results from two institutions. Magn Reson Med 2014;00. doi:10.1002/mrm.25326.

PAGE 112

! 96 35. Pedrizzetti G, L a Canna G, Alfieri O, Tonti G. The vortex -an early predictor of cardiovascular outcome? Nat Rev Cardiol 2014;11(9):545 53. doi:10.1038/nrcardio.2014.75. 36. Reiter U, Reiter G, Kovacs G, et al. Evaluation of elevated mean pulmonary arterial pressure bas ed on magnetic resonance 4D velocity mapping: Comparison of visualization techniques. PLoS One 2013;8(12):1 9. doi:10.1371/journal.pone.0082212. 37. Morbiducci U, Ponzini R, Rizzo G, et al. In vivo quantification of helical blood flow in human aorta by t ime resolved three dimensional cine phase contrast magnetic resonance imaging. Ann Biomed Eng 2009;37(3):516 31. doi:10.1007/s10439 008 9609 6. 38. Reiter G, Reiter U, Kovacs G, et al. Magnetic resonance derived 3 dimensional blood flow patterns in the m ain pulmonary artery as a marker of pulmonary hypertension and a measure of elevated mean pulmonary arterial pressure. Circ Cardiovasc Imaging 2008;1:23 30. doi:10.1161/CIRCIMAGING.108.780247. 39. Shimoda L, Laurie S. Vascular Remodeling in Pulmonary Hyp ertension. J Mol Med 2013;91(3):297 309. doi:10.1007/s00109 013 0998 0.Vascular. 40. Hudsmith  L, Petersen  S, Francis J, Robson M, Neubauer S. Normal Human Left and Right Ventricular and Left Atrial Dimensions Using Steady State Free Precession Magnetic Resonance Imaging. J Cardiovasc Magn Reson 2005;7(5):775 782. doi:10.1080/10976640500295516. 41. Harouni AA, Skrok J, Boyce D, Lechtzin N, Mathai SC, Vogel claussen J. Regional and Global Biventricular Function in Pulmonary Arterial Hypertension&: 2013;266(1). doi:10.1148/radiol.12111599/ /DC1. 42. Nacif MS, Barranhas AD, TŸrkbey E, et al. Left atrial volume quantification using cardiac MRI in atrial fibrillation: comparison of the Simpson's method with biplane area length, ellipse, and three dimen sional methods. Diagn Interv Radiol 2012;19(3):213 20. doi:10.5152/dir.2012.002. 43. Sanz J, Kariisa M, Dellegrottaglie S, et al. Evaluation of pulmonary artery stiffness in pulmonary hypertension with cardiac magnetic resonance. JACC Cardiovasc Imaging 2009;2(3):286 95. doi:10.1016/j.jcmg.2008.08.007. 44. Voelkel NF, Quaife R a., Leinwand L a., et al. Right ventricular function and failure: Report of a National Heart, Lung, and Blood Institute working group on cellular and molecular mechanisms of right heart failure. Circulation 2006;114:1883 1891. doi:10.1161/CIRCULATIONAHA.106.632208.

PAGE 113

! 97 45. Chang S M, Lin C C, Hsiao S H, et al. Pulmonary hypertension and left heart function: insights from tissue Doppler imaging and myocardial performance index. Echoca rdiography 2007;24(4):366 373. doi:10.1111/j.1540 8175.2007.00405.x. 46. Guazzi M, Borlaug B a. Pulmonary hypertension due to left heart disease. Circulation 2012;126:975 990. doi:10.1161/CIRCULATIONAHA.111.085761. 47. Appleton CP, Kov‡cs SJ. The role of left atrial function in diastolic heart failure. Circ Cardiovasc Imaging 2009;2:6 9. doi:10.1161/CIRCIMAGING.108.845503. 48. Aune E, Baekkevar M, Roislien J, Rodevand O, Otterstad JE. Normal reference ranges for left and right atrial volume indexes an d ejection fractions obtained with real time three dimensional echocardiography. Eur J Echocardiogr 2009;10:738 744. doi:10.1093/ejechocard/jep054. 49. Welles CC, Ku I a., Kwan DM, Whooley M a., Schiller NB, Turakhia MP. Left atrial function predicts hea rt failure hospitalization in subjects with preserved ejection fraction and coronary heart disease: Longitudinal data from the heart and soul study. J Am Coll Cardiol 2012;59(7):673 680. doi:10.1016/j.jacc.2011.11.012. 50. Graa B, Ferreira MJ, Donato P, Gomes L, Castelo Branco M, Caseiro Alves F. Left atrial dysfunction in type 2 diabetes mellitus: insights from cardiac MRI. Eur Radiol 2014;24(11):2669 76. doi:10.1007/s00330 014 3299 2. 51. Li Y SJ, Haga JH, Chien S. Molecular basis of the effects of s hear stress on vascular endothelial cells. J Biomech 2005;38:1949 1971. doi:10.1016/j.jbiomech.2004.09.030. 52. Reinke W, Gaehtgens P, Johnson PC. Blood viscosity in small tubes: effect of shear rate, aggregation, and sedimentation. Am J Physiol 1987;25 3(29):H540 H547. 53. Truong U, Fonseca B, Dunning J, et al. Wall shear stress measured by phase contrast cardiovascular magnetic resonance in children and adolescents with pulmonary arterial hypertension. J Cardiovasc Magn Reson 2013;15(1):81. doi:10.118 6/1532 429X 15 81. 54. Gallo D, Steinman D a., Bijari PB, Morbiducci U. Helical flow in carotid bifurcation as surrogate marker of exposure to disturbed shear. J Biomech 2012;45(14):2398 2404. doi:10.1016/j.jbiomech.2012.07.007. 55. Hoeper MM, Lee SH, V oswinckel R, et al. Complications of right heart catheterization procedures in patients with pulmonary hypertension in experienced centers. J Am Coll Cardiol 2006;48(12):2546 52. doi:10.1016/j.jacc.2006.07.061.

PAGE 114

! 98 56. Hajian Tilaki K. Receiver operating cha racteristic (ROC) curve analysis for medical diagnostic test evaluation. Casp J Intern Med 2013;4:627 635. 57. Fares WH, Blanchard SK, Stouffer G a., et al. Thermodilution and Fick cardiac outputs differ: Impact on pulmonary hypertension evaluation. Can Respir J 2012;19(4):261 266. 58. Lehmann KG, Platt MS. Improved accuracy and precision of thermodilution cardiac output measurement using a dual thermistor catheter system. J Am Coll Cardiol 1999;33(3):883 891.