Citation
Modified single-beat method for the estimation of ventricular-arterial coupling ratio in the right ventricle

Material Information

Title:
Modified single-beat method for the estimation of ventricular-arterial coupling ratio in the right ventricle
Creator:
Hobson, Nicholas Evan
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file. : ;

Subjects

Subjects / Keywords:
Heart -- Right ventricle ( lcsh )
Pulmonary hypertension ( lcsh )
Heart -- Right ventricle ( fast )
Pulmonary hypertension ( fast )
Genre:
non-fiction ( marcgt )

Notes

Abstract:
Ventricular performance quantities such as the end-systolic pressure-volume relationship (ESPVR) have been the subject of numerous basic science studies, yet their clinical use remains limited, particularly in the right ventricle (RV). This is primarily due to the difficulty of volume measurements in the small, crescent-shaped RV via catheterization. However, such parameters are a superior indicator of ventricular function compared with other hemodynamic measures used in the prognosis of pulmonary arterial hypertension (PAH), such as pulmonary vascular resistance index (PVRI). Thus, there is clinical interest in methods that estimate ESPVR and related parameters while being minimally invasive. The aim of this study is to examine one such method, a modified single-beat method which estimates the ventricular-vascular coupling ratio (VVCR), or the ratio of end-systolic ventricular elastance (Ees) to arterial elastance (Ea). Within the single-beat elastance framework, the maximum isovolumic pressure (Pmax,iso) and end-systolic pressure are found; based on a novel assumption about the slopes of Ees and Ea, VVCR is then computed using only pressure. A lower coupling ratio is hypothesized to be a good indicator of RV dysfunction and failure, as represented by the World Health Organization Functional Class (WHO-FC). Furthermore, an initial investigation into a non-invasive form of this method was performed, in which pressure data is obtained from the velocity of the tricuspid regurgitant (TR) jet measured by Doppler ultrasound. A total of 98 patients undergoing RV catheterization for PAH were used in this study. The VA coupling ratio was calculated using two definitions for end-systole: 32 ms prior to dP/dtmin; and observed Pmax. Logistic regressions were performed for each coupling ratio, PVRI, and the combination of ratio (Pmax) and PVRI against WHO-FC. The corrected Akaike Information Criteria (AICc) was computed for each regression, and each regression equation was used to predict the WHO-FC for each patient, and correct and incorrect predictions were compared for each method. The AICc for PVRI was slightly lower at 185.12 than the ratio (Pmax) at 187.75 or ratio (Pes) at 188.70, implying a slightly better yet comparable fit to the data by PVRI. Furthermore, ratio (Pmax) was as good or better at predicting WHO-FC II (66.7% for ratio and 41.7% for PVRI) and WHO-FC III (52.4% for each), and the combination of ratio and PVRI was better than PVRI alone at predicting both WHO-FC II (58.3%) and WHO-FC III (61.9%). Therefore, it is likely that the VA coupling ratio contains useful information reflective of RV dysfunction and failure that is not necessarily reflected by PVRI, and shows promise for clinical use in the prognosis of PAH in the future.
Thesis:
Thesis (M.S.)--University of Colorado Denver. Bioengineering
Bibliography:
Includes bibliographic references.
General Note:
Department of Bioengineering
Statement of Responsibility:
by Nicholas Evan Hobson.

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:
861981379 ( OCLC )
ocn861981379

Downloads

This item has the following downloads:


Full Text
MODIFIED SINGLE-BEAT METHOD FOR THE ESTIMATION OF
VENTRICULAR-ARTERIAL COUPLING RATIO IN THE RIGHT VENTRICLE
by
NICHOLAS EVAN HOBSON
B.S., Vanderbilt University, 2009
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 Graduate Program
2012


This thesis for the degree of Master of Science by
Nicholas Evan Hobson
has been approved for the
Bioengineering Graduate Program
by
Robin Shandas, Chair
Kendall Hunter
Dunbar Ivy
November 8, 2012
ii


Hobson, Nicholas Evan (M.S., Bioengineering Program)
Modified Single-Beat Method for the Estimation of Ventricular-Arterial Coupling Ratio in the
Right Ventricle
Thesis directed by Kendall Hunter
ABSTRACT
Ventricular performance quantities such as the end-systolic pressure-volume
relationship (ESPVR) have been the subject of numerous basic science studies, yet their clinical
use remains limited, particularly in the right ventricle (RV). This is primarily due to the
difficulty of volume measurements in the small, crescent-shaped RV via catheterization.
However, such parameters are a superior indicator of ventricular function compared with other
hemodynamic measures used in the prognosis of pulmonary arterial hypertension (PAH), such
as pulmonary vascular resistance index (PVRI). Thus, there is clinical interest in methods that
estimate ESPVR and related parameters while being minimally invasive.
The aim of this study is to examine one such method, a modified single-beat method
which estimates the ventricular-vascular coupling ratio (VVCR), or the ratio of end-systolic
ventricular elastance (Ees) to arterial elastance (Ea). Within the single-beat elastance
framework, the maximum isovolumic pressure (Pmax,iSo) and end-systolic pressure are found;
based on a novel assumption about the slopes of Ees and Ea, VVCR is then computed using only
pressure. A lower coupling ratio is hypothesized to be a good indicator of RV dysfunction and
failure, as represented by the World Health Organization Functional Class (WHO-FC).
Furthermore, an initial investigation into a non-invasive form of this method was performed, in
which pressure data is obtained from the velocity of the tricuspid regurgitant (TR) jet measured
by Doppler ultrasound.


A total of 98 patients undergoing RV catheterization for PAH were used in this study.
The VA coupling ratio was calculated using two definitions for end-systole: 32 ms prior to
dP/dtmin; and observed Pmax. Logistic regressions were performed for each coupling ratio,
PVRI, and the combination of ratio (Pmax) and PVRI against WHO-FC. The corrected Akaike
Information Criteria (AlCc) was computed for each regression, and each regression equation was
used to predict the WHO-FC for each patient, and correct and incorrect predictions were
compared for each method. The AlCc for PVRI was slightly lower at 185.12 than the ratio (Pmax)
at 187.75 or ratio (Pes) at 188.70, implying a slightly better yet comparable fit to the data by
PVRI. Furthermore, ratio (Pmax) was as good or better at predicting WHO-FC II (66.7% for ratio
and 41.7% for PVRI) and WHO-FC III (52.4% for each), and the combination of ratio and PVRI was
better than PVRI alone at predicting both WHO-FC II (58.3%) and WHO-FC III (61.9%).
Therefore, it is likely that the VA coupling ratio contains useful information reflective of RV
dysfunction and failure that is not necessarily reflected by PVRI, and shows promise for clinical
use in the prognosis of PAH in the future.
The form and content of this abstract are approved. I recommend its publication.
Approved: Kendall Hunter
IV


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION.......................................................................1
Background.........................................................................1
The ESPVR Relationship and Elastance...............................................2
Arterial Elastance and Ventricular-Vascular Coupling...............................3
Clinical Studies Utilizing ESPVR and Related Parameters............................4
Single-Beat Methodology............................................................6
Pulmonary Hypertension.............................................................7
Assessment of RV function.........................................................12
II. ANALYSIS OF A NOVEL SINGLE-BEAT METHOD............................................14
Mathematical Basis................................................................14
Data Acquisition..................................................................15
Pressure Data Analysis............................................................15
Statistical Analysis..............................................................16
Results...........................................................................18
Discussion........................................................................27
III. TR JET DOPPLER STUDY..............................................................31
Introduction......................................................................31
Doppler Data Analysis.............................................................32
Results...........................................................................33
Discussion and Future Work........................................................37
IV. CONCLUSION.........................................................................38
REFERENCES.............................................................................41
v


LIST OF ABBREVIATIONS
PV Pressure-volume
sv Stroke volume
LV/RV Left/right ventricle
PA/PAP Pulmonary artery / pulmonary arterial pressure
PVRI Pulmonary vascular resistance index
PWP Pulmonary wedge pressure
CO Cardiac output
RVEF Right ventricular ejection fraction
EDV End-diastolic volume
ESPVR End-systolic pressure-volume relationship
EDPVR End-diastolic pressure-volume relationship
ESPAR End-systolic pressure-area relationship
VVCR Ventricular-vascular coupling ratio
Ees End-systolic elastance
Ea Effective arterial elastance
PH/PAH Pulmonary hypertension / pulmonary arterial hypertension
WHO-FC World Health Organization functional class
6MWT 6-minute walk test
BNP Brain natriuretic peptide
HFnIEF Fleart failure with normal ejection fraction
HTN Hypertension without heart failure
TR Tricuspid regurgitation
RHC Right heart catheterization
ROC curve Receiver-operator characteristic curve
AlCc Corrected Akaike Information Criterion
AUC Area under the curve
VI


LIST OF FIGURES
FIGURE
1. Pressure-volume loop diagram for left ventricle...................................3
2. ROC curves for full dataset for VVC...............................................19
3. ROC curves for "recovery" grouping of data subset.................................24
4. ROC curves for "failure" grouping of data subset..................................25
5. Automatic trace of Doppler US image of TR jet.....................................34
6. Pressure trace and curve fitting of third pulse from Doppler image................34
7. Doppler image yielding poor pressure trace for curve fitting......................35
vii


LIST OF TABLES
TABLE
1. Functional classification of pulmonary hypertension................................11
2. AlCc and related parameters for the three models tested on a dichotomous
categorization of outcome (WHO I & II vs. Ill & IV).......................................20
3. Correct, incorrect, and "close" prognoses for the data set (n = 98) by each of the three
logistic regression models for dichotomous categorization of outcome (WHO I & II vs. Ill & IV). 21
4. AlCc and related parameters for the three models tested on a dichotomous
categorization of outcome (WHO I vs. II, III, & IV).......................................21
5. Correct, incorrect, and "close" prognoses for the data set (n = 98) by each of the three
logistic regression models for dichotomous categorization of outcome (WHO I vs. II, III, & IV).. 22
6. AlCc and related parameters for the three models tested on a trichotomous
categorization of outcome.................................................................22
7. Correct, incorrect, and "close" prognoses for the data set (n = 98) by each of the three
logistic regression models and the combined regression for trichotomous categorization....23
8. Breakdown of correct prognoses by outcome of Ratio (Pmax) and PVRI, and which cases
were correctly predicted by both measures (overlap).......................................23
9. Breakdown of correct prognoses by outcome of Ratio (Pmax) and PVRI together and PVRI
alone, and which cases were correctly predicted by both measures (overlap)................23
10. AlCc and related parameters for the three models tested on the "recovery" dichotomous
categorization of outcome (WHO I vs. II, III, & IV).......................................25
11. AlCc and related parameters for the three models tested on the "failure" dichotomous
categorization of outcome (WHO I & II vs. Ill, & IV)......................................26
12. Coupling ratios for Doppler images with acceptable pressure traces.................35
viii


LIST OF EQUATIONS
EQUATION
1. End-systolic elastance..............................................................15
2. Effective arterial elastance........................................................15
3. Ventricular-vascular coupling ratio.................................................15
4. Bernoulli's principle...............................................................32
IX


CHAPTERI
INTRODUCTION
Background
Characterization of the pressure-volume relationship of the heart as a pump dates back
to an 1898 paper by Otto Frank, in which he presented a P-V diagram of contractions of the
isolated frog ventricle. Frank made two important observations about this diagram: first, peak
systolic activity, whether measured by isovolumic pressure or isobaric ejection volume, is
related to pre-loaded end-diastolic volume (EDV); and second, when the ventricle is in ejecting
mode, the curve produced by connecting the end-systolic points of different PV loops (known as
the end-systolic pressure-volume relationship, or ESPVR) is quite different than the isovolumic
(non-ejecting) curve. These observations were well-received in Europe, and in 1914 Ernest
Starling published a series of papers that led to the now well-known Frank-Starling law of the
heart, which describes the ability of the heart to compensate for varying end-diastolic volumes
by varying its work output accordingly.
These studies left researchers validating these principles until the 1950's, when interest
in pressure-volume relationships of the heart experienced a resurgence. Studies by Ullrichet al
(1954), Hild and Sick (1955), and Hild and Flerz (1956) obtained pressure-volume relationships
for mammalian hearts, which proved to be somewhat different than the previous studies using
the frog ventricle. These studies revealed the ESPVR showed only slight curvilinearity for
physiologic pressure ranges. Monroe et al (1960, 1961,1964) showed similarly that the ESPVR
for the dog ventricle under the control condition was almost linear, contrary to studies of the
frog that showed strong dependence of ESPVR on preload and afterload history.
1


The ESPVR Relationship and Elastance
The near-linearity of the ESPVR curve has been validated in both the left ventricle (Cross
et al 1961, Taylor et al 1969) and right ventricle (Lafontant et al 1962). Perhaps the most
substantial works on the characterization of ESPVR and PV relationships of the ventricle in
general come from the 1970's and 80's in Japan from authors Suga, Sagawa, and Sunagawa.
These serial works on the in vivo canine heart further confirmed and refined the linearity of
ESPVR in the physiologic range and helped advance understanding of a wide array of other
properties of these measurements. Advances in catheter technology and the advent of the
conductance catheter allowed instantaneous volume measurements of the left ventricle in vivo
(Baan et al,1981), meaning ESPVR could be obtained in a cardiac catheter lab by altering preload
(EDV) and obtaining a series of PV loops and connecting their end-systolic points (see Figure 1).
As the end-diastolic volume increases and the loops shift to the right, end-systolic pressure
increases along the ESPVR line up to a maximum pressure for isovolumic contraction (Suga &
Sagawa, 1974). Because ESPVR is relatively insensitive to changes in preload, afterload, and
heart rate, it provides a valuable index of systolic cardiac function. Specifically, the slope of the
ESPVR is a measure of end-systolic elastance (Ees) of the ventricle. Furthermore, a second
useful measure, effective arterial elastance (Ea), can be obtained as the slope of the line from
the pressure-volume loop from the end-systolic PV point (upper-left of the PV box) to the
end-diastolic volume intercept (bottom-right). Arterial elastance was defined as the ratio of
end-systolic pressure to stroke volume, Pes/SV (Fig. 1). This definition, from a 1983 work by
Sunagawa and colleagues about LV-arterial interaction, is based on the assumption that vascular
impedance can be modeled as a three-element Windkessel model. By defining SV in terms of
mean arterial flow over a single cardiac cycle, the Pes-SV relationship in the arteries was
described analytically as a linear relationship when characteristic arterial impedance, peripheral
2


resistance, compliance, and systolic and diastolic durations are constant. The expression
describing the slope of that relationship involving those arterial parameters was denoted by Ea
(effective arterial elastance). Because both Ees and Ea are given in the same units, their ratio
can be computed as Ees/Ea to provide a measure of ventricular-vascular coupling.
Arterial Elastance and Ventricular-Vascular Coupling
Works by Sunagawa et al (1983, 1985) sought to test the hypothesis that ventricular
external work would be maximized if the ventricular and arterial elastances are equal. They
found the optimal arterial resistance under a number of variable parameters such as
end-diastolic volume, contractility, heart rate, and arterial compliance, and found that optimal
arterial resistance varied only slightly with arterial compliance, while varying widely with
contractility and heart rate. These findings suggest that the ratio of ventricular and arterial
Figure 1 Pressure-volume loop diagram for left ventricle
showing construction of ESPVR and derivation of Ees and Ea
using parameters Pmax, Pes, Ved, and SV.
3


elastance is optimally nearly unity, which gives rise to the idea that the ventricle and vasculature
are coupled from an energetics viewpoint. In other words, optimal flow (and optimal energy
transfer from the ventricle through the vasculature) is achieved at an arterial elastance equal to
ventricular elastance.
Thus, the effects of increased afterload on PV loops and ESPVR should be significant. It
is useful to distinguish between patients with early- and late-stage hypertension as
"compensated" and "failing," in terms of ventricular function. In early-stage pulmonary arterial
hypertension (PAH), as the ventricle must fight greater vascular pressure and impedance, the
end-systolic pressure will increase to compensate, which is achieved by an increase in
ventricular elastance. This progression will also be reflected in a slightly increased Ea as
vascular remodeling gradually stiffens the arterial vasculature, and increased pulse pressure
leads to a greater Pes. Thus, the coupling ratio between the ventricle and arteries, Ees/Ea, may
decrease somewhat but should remain close to unity. Over time however, the heart becomes
overworked, and Pes approaches the maximum isovolumic pressure the ventricle can produce.
A patient transitions from compensated to failing when the ventricle and vasculature become
uncoupled, and the ventricle begins to dilate to meet cardiac output needs. The PV loop shifts
to the right and becomes narrower as end-diastolic volume increases and stroke volume
decreases (Badesch et al, 2009). These patients should have a markedly decreased coupling
ratio, which signifies uncoupling of the ventricle and vasculature.
Clinical Studies Utilizing ESPVR and Related Parameters
Although there has been much work done in the lab characterizing the pressure-volume
relationship of the ventricle, and though these works generally strive to emphasize the potential
clinical usefulness of PV loops, their use in clinical settings has remained extremely limited for a
number of reasons. First, the procedures for obtaining instantaneous pressure and volume
4


measurements require highly invasive catheterization. Techniques for volume measurements
in particular have been difficult and imprecise. Measurements of volume of the right ventricle
are especially troublesome due to its location, smaller size, and non-uniform geometry.
Furthermore, many clinical decisions can be made based on more easily-obtained measures,
such as ejection fraction (Burkhoff, 2005).
However, there have emerged a number of clinical studies in recent years
demonstrating and validating the usefulness of PV-loop-derived parameters. A study in 1992
by Kelly et al tested the use of effective arterial elastance as an index of vascular load in normal
and hypertensive humans. The study found that Ea based on PV loops was nearly identical to
Ea derived from a three-element Windkessel model for normal patients, and that it exceeded
simple resistance by nearly 25% in hypertensive patients, due to decreased compliance and
wave reflection. The findings suggest that even Ea alone, not to speak of full ESPVR or Ees/Ea
coupling, could provide a convenient assessment of arterial impedance and its effects on
ventricular function in hypertensive patients.
The largest population-based study to date assessing diagnostic capabilities of PV loops
comes from Lam et al from Minnesota in 2007. This study compared patients exhibiting heart
failure with a normal ejection fraction (HFnIEF) with patients with hypertension but no heart
failure (HTN), and with patients without cardiovascular disease. The authors characterized left
ventricular volume, effective arterial elastance, left ventricular end-systolic elastance, and left
ventricular diastolic elastance and relaxation noninvasively. Their results showed increased Ees
for hypertensive and HFnIEF patients compared with the control, as well as a shift to the right of
the V0 volume intercept for HFnIEF patients relative to HTN patients, indicating an overall
increase in ventricular volume. They observed similar differences in the EDPVR curves for the
three groups, with hypertensive patients having a greater EDV than HFnIEF patients. Their
5


work showed a probable role of diastolic dysfunction in HFnIEF patients relative to FITN patients.
Furthermore, it shows the potential diagnostic power of ESPVR and EDPVR curves for
distinguishing specific cardiovascular disorders.
Single-Beat Methodology
In 1991, Takeuchi et al proposed a novel method for the calculation of the slope of the
ESPVR (Ees) using a single beat. They realized that the maximum pressure the ventricle can
produce during isovolumic contraction could be estimated from the pressure trace alone, and
the slope of the ESPVR could be obtained by drawing a line from this point tangential to the
upper-left corner of a single PV loop (i.e. near the end-systolic point). Sinusoidal curve fitting is
performed on the pressure trace of the ventricle using data points which correspond to periods
of isovolumic contraction and relaxation in order to estimate Pmax. This is a much easier
method than the previous one for generating the ESPVR, which involved creating at least three
PV loops by altering the end-diastolic volume.
This method has also been validated for the right ventricle by Brimioulle et al in 2003.
A number of assumptions and limitations unique to the RV were addressed in their work. The
triangular shape of typical PV loops in the RV means that end-systole and end-ejection often
occur at different times. The upper-left corner was used as the end-systolic point, which is
valid for the LV but less so for the RV. The assumption that isovolumic beats are sinusoidal,
established by Sunagawa by Fourier analysis for the LV, was verified by comparing estimated
Pmax,iso of ejecting beats with observed Pmaxfor isovolumic beats of the same starting EDV, which
were found to be strongly correlated. Due to the challenges of measuring RV volume in vivo,
they determined RV volume changes by integrating flow measured in the proximal pulmonary
artery with a widely validated ultrasonic method. They also found expected changes in Ees and
Ea under a variety of inotropic states.
6


Furthermore, a 2004 study by Kuehne et al validated the single-beat method for the
right ventricle using cine MRI to obtain RV volume, and also undertook a small clinical study
comparing 6 control patients to 6 patients with chronic pulmonary hypertension. They
obtained indexes of cardiac pump function, measured as cardiac index or the ratio of cardiac
output to body surface area, Ees indexed to myocardial mass, and ventricular-arterial coupling
(Ees/Ea). In patients with PAH, RV pump function was decreased, myocardial contractility was
enhanced, and VVC was inefficient compared with the control group. These results agree with
the expected physiologic impact of PAH on RV function discussed above, and are encouraging
for the minimally-invasive prognosis of PAH.
Pulmonary Hypertension
Pulmonary hypertension was first classified in 1973 in a meeting organized by the World
Health Organization. PH is defined as a hemodynamic and pathological condition in which a
mean pulmonary arterial pressure (PAP) of 25 mmHg or greater at rest is measured by right
heart catheterization. Normal PAP is 14 3 mmHg with an upper limit of about 20 mmHg.
The significance of a mean PAP of 21 to 24 mmHg is unclear.
The clinical classification of PH has undergone some changes in international
conferences since the initial 1973 WHO classification, but the basic differentiation of primary
and secondary PH is still apparent. The broad categories of the most recent classification
agreed upon at the fourth World Symposium for PH at Dana Point, California in 2008 are as
follows:
1. Pulmonary arterial hypertension (PAH)
1'. Pulmonary veno-occlusive disease and/or pulmonary capillary haemangiomatosis
2. Pulmonary hypertension due to left heart disease
3. Pulmonary hypertension due to lung diseases and/or hypoxia
7


4. Chronic thromboembolic pulmonary hypertension
5. PH with unclear or multifactorial mechanisms
The focus of this paper is primary hypertension (group 1, PAH), or PH not caused by other health
problems.
Pathology
The initiation of pathological changes seen in PAH is not well understood, although the
condition is characterized by numerous pathologies in a variety of biochemical pathways and
cell types. An increase in pulmonary vascular resistance (PVR) is caused by vasoconstriction of
the arteries, inflammation, thrombosis, and remodeling of the vessel walls. Vasoconstriction
has been linked to abnormal function of potassium channels in smooth muscle cells as well as
endothelial cell dysfunction leading to impaired production of vasodilators such as nitric oxide
and prostacyclin, and overproduction of vasoconstrictors such as thromboxane-A2 and
endothelin-1. Over time, vascular remodeling leads to stiffening and fibrosis by increased
production of collagen and elastin in the adventitia (Galie et al, 2009).
The increase in PAP and PVR represents an increased workload applied to the ventricle.
This can eventually lead to pathological hypertrophy and ventricular stiffening and dilation.
This is the point at which ventricular-arterial decoupling occurs. In the sense that the heart is a
pump, it will eventually be unable to compensate for the increased arterial load, leading to
heart failure and death.
Diagnosis
Diagnosis of PH is a process of exclusion, so often many different diagnostic tools and
methods need to be employed to eliminate all other possible causes. In the case of PAH, this
means ruling out left heart disease, veno-occlusive disease, lung disease, and chronic
8


thromboembolism.
Symptoms of PAH are often mild and present mainly during exercise, and can include
shortness of breath, fatigue, weakness, angina, syncope, and abdominal distension. Physical
signs that may aide diagnosis are primarily heart sounds: tricuspid valve regurgitation murmur,
a diastolic murmur of pulmonary insufficiency, and an RV third sound. Patients in an advanced
disease state may present with signs such as ascites, peripheral edema, and cool extremities.
However, these physical signs vary widely and are often not present at all, so they are of limited
use.
The electrocardiogram may provide diagnostic value in showing the presence of RV
hypertrophy and right atrial dilation, though the absence of these findings should not be taken
as a negative diagnosis. ECG is limited to electrical information, so it cannot give any direct
indication about the underlying mechanical state of the ventricle and vasculature.
A chest radiogram is abnormal in 90% of patients presenting with PAH. Central PA
dilation contrasted with loss of capillaries is suggestive of PAH. The chest radiogram can also
show RV enlargement in advanced cases, and serve to eliminate lung disease and venous
hypertension due to left heart disease as possible causes. However, the radiogram cannot give
a detailed idea of the status of the ventricle and vasculature except in advanced states.
Echocardiography can provide an estimation of PAP by obtaining velocity measurements
of regurgitant tricuspid valve jets, which occur in more than 70% of patients with PAH (Yock and
Popp 1984). Velocity of TR jets is measured using Doppler ultrasound. By using a simplified
form of Bernoulli's equation, mean PAP can be calculated as
4 x (tricuspid regurgitant velocity)2 + (estimated right atrial pressure). However,
echocardiography is limited by the noise inherent in ultrasound imaging, which can lead to faint
or inaccurate signals.
9


The gold standard for the diagnosis of PAH is right heart catheterization (RHC), and it is
necessary for confirming such a diagnosis. The following variables must be obtained: PAP,
right atrial pressure, RV pressure, pulmonary wedge pressure (PWP), and cardiac output (CO) by
either thermodilution or the Fick method. A mean PAP of 25 mmHg or greater with a PWP of
15 mmHg or less in the absence of other causes confirms a diagnosis of PAH (Galie et al, 2009).
Although RHC is a relatively safe procedure, it is invasive, and therefore carries some risk to the
patient.
Prognosis
One of the most important prognostic measure of hypertension is the WHO functional
class devised in 1998 by the New York Heart Association, which is a relatively good predictor of
mortality, outlined in Table 1. In untreated patients with idiopathic PAH or heritable PAH,
historical data showed a median survival of 6 months for WHO-FC IV, 2.5 years for WHO-FC III,
and 6 years for WHO-FC I and II.
10


Table 1 Functional classification of pulmonary hypertension modified after the New York Heart
Association functional classification according to the WHO 1998
Class 1 Patients with pulmonary hypertension but without resulting limitation of physical activity. Ordinary physical activity does not cause undue dyspnoea or fatigue, chest pain, or near syncope.
Class II Patients with pulmonary hypertension resulting in slight limitation of physical activity. They are comfortable at rest. Ordinary physical activity causes undue dyspnoea or fatigue, chest pain, or near syncope.
Class III Patients with pulmonary hypertension resulting in marked limitation of physical activity. They are comfortable at rest. Less than ordinary activity causes undue dyspnoea or fatigue, chest pain, or near syncope.
Class IV Patients with pulmonary hypertension with inability to carry out any physical activity without symptoms. These patients manifest signs of right heart failure. Dyspnoea and/or fatigue may even be present at rest. Discomfort is increased by any physical activity.
Prognostic indices yielded by echocardiography include pericardial effusion (Eysmann et
al, 1989), indexed right atrium area, LV eccentricity index, and the RV Doppler index. Systolic
PAP estimated from TR jet velocity is not prognostic (Raymond et al, 2002).
Hemodynamic measurements taken by RHC at rest are also prognostic. These include
PA oxygen saturation, right atrial pressure, CO, PVR, and a marked vasoreactivity response.
Exercise capacity as measured by the 6-minute walking test (6MWT) is simple and
well-standardized.
Finally, biochemical markers have been gaining appeal as a non-invasive means to
11


prognosing PH in the last decade. These primarily describe RV function in some way. Serum
uric acid is an indicator of impaired oxidative metabolism of ischaemic peripheral tissue, and
high uric acid levels were shown to correlate with poor survival in patients with PAH. Another
biochemical marker that has shown a correlation with survival in PAH is brain natriuretic
peptide, a low-molecular weight protein that acts as a vasodilator, present in greater amounts in
patients with poor survival (Galie et al, 2009). The initiation of the synthesis of BNP results
from wall stress due to volume expansion or pressure overload (Bernus et al, 2009). A 2009
study by Bernus et al found no strong correlation between BNP and commonly used
hemodynamic and echocardiographic parameters, but that the change over time of BNP did
correlate with the change of those parameters, and that patients with a BNP greater than 180
pg/mL had a decreased survival rate.
Assessment of RV function
A number of methods for the assessment of RV function have been proposed and
validated over the past 30 years, varying widely in invasiveness as well as efficacy and accuracy.
In 1984, Kimchi et al became the first group to evaluate RV ejection fraction (RVEF) using
radionuclide ventriculography. They defined RV failure as RVEF less than 38%. In 1989,
Vincent et al obtained RVEF using a modified PA catheter with a fast-response thermistor,
utilizing the thermodilution method. However, these methods are not ideal both because they
are either nuclear or invasive, and because RVEF varies widely in normal patients from 30-60%.
Cardiac magnetic resonance (CMR) imaging is the gold standard for evaluating RV
function. It is the only tool which can reliably quantify RV volume and RVEF. Delayed
contrast-enhanced CMR has also been shown to detect fibrotic tissue in the RV wall, but can
only detect regional changes in myocardial tissue. Contrast-enhanced Tl-weighted CMR allows
12


the detection of more global fibrosis (Mertens & Friedberg, 2010). However, CMR is expensive
and availability is often limited.
Echocardiography is currently the most practical tool for assessing RV function at the
bedside. A failed right ventricle is severely dilated, which can be assessed by comparing the area
of the chamber with that of the left ventricle. A ratio of RV:LV area less than 0.6 is considered
normal, and greater than 1 is considered failing. In 1995, Masahiro, et al demonstrated the
efficacy of an automated method for obtaining a correlate of the ESPVR based on the area of
the chamber from echocardiography, rather than volume measured by catheterization. They
showed that the ESPAR responded similarly to ESPVR as RV function declined. Another sign
readable from echocardiography is paradoxical septal motion during systole, which indicates RV
systolic overload. In Doppler mode, the ejection flow can show pulmonary hypertension by a
shortening of the acceleration time of flow (<100 ms). A related parameter proposed by Tei et
al in 1995 takes the ratio of combined isovolumic contraction and relaxation times compared
with ejection time, termed the myocardial performance index (MPI). Longer periods of
isovolumic contraction and relaxation yield a higher MPI, and indicate worse RV function.
Interestingly, this measure bears a certain similarity to VVCR, because longer isovolumic periods
will inevitably yield lower estimated maximum isovolumic pressures.
Advances in 3D echocardiography in the past 5 years have made the non-invasive
measurement of RV volume possible. There are two methods for calculating RV volumes from
echocardiographic images. The first involves the manual tracking of RV area through a series
of slices, while the other uses border-detection algorithms to automate this process. These
measurements have been shown to correlate well with volume measurements obtained from
CMR, but consistently predict volumes 20-34% smaller than those obtained by CMR (Mertens &
13


Friedberg, 2010). Thus, 3D echocardiography requires further refinement and validation, but
could become the new gold standard in the future.
14


CHAPTER II
ANALYSIS OF A NOVEL SINGLE-BEAT METHOD
Mathematical Basis
The basis for this work is a modification of the single-beat method that estimates the VA
coupling ratio while eliminating the need for volume measurement. Re-examining Figure 1, it
can be seen that:
F
ues
SV
and given the definition
p £es
a sv
(1)
(2).
Inserting equations (1) and (2) into the expression for the VA coupling ratio, we have:
Ees ^ Pm ax Pes ^ _ Pmax Pes
E~ SV P,~
es r es
= Dnax_1 (3).
Pes
Thus, from (3) we obtain an expression for VA coupling as a simple ratio of maximum isovolumic
pressure to end-systolic pressure. The single-beat method can be applied to pressure traces
from the right ventricle, estimating Pmax by sinusoidal curve fitting to periods of isovolumic
contraction and relaxation, and comparing this value to the observed end-systolic pressure to
obtain an estimate of VVCR. It is hypothesized that VVCR alone is prognostic of pulmonary
hypertension.
This model makes a number of assumptions. First, we assume that isovolumic
contraction can be defined by the period between the time at which dP/dt exceeds 200 mmFIg
and dP/dtmax, and similarly for isovolumic relaxation. It is assumed that isovolumic contraction
and relaxation are sinusoidal in their pressure development, and that maximum isovolumic
pressure can be estimated as the maximum of this sinusoid. Finally, we assume that the
15


catheter gives sufficient temporal resolution for an accurate calculation of the first derivative of
pressure.
Data Acquisition
All clinical studies had IRB approval. With informed consent and/or assent when
applicable, we studied 98 patients undergoing evaluation of reactivity using oxygen and/ or
nitric oxide in the cardiac catheterization laboratory at The Children's Hospital in Denver, CO.
Patients ranged in age from 1 month to 26 years (mean = 9.136.29 yrs, 55 males). Pressure
was obtained within the RPA using 5- or 6-fr Swan-Ganz catheters (Transpac IV, Abbott Critical
Care Systems, Abbott Park, IL).
Pressure Data Analysis
All data analysis was performed in MATLAB. Pressure traces from the right ventricle
were divided into single beats by ECG gating. Pmax was determined by fitting the equation
P = a + b sin (c t + d), where P is pressure and t is time, to pressure values from the
end-diastolic point up to the point of the maximal value of the first derivative, corresponding to
isovolumic contraction, and from the time of the minimal value of the first derivative to the
same pressure value as the end-diastolic point, corresponding to isovolumic relaxation. The
end-diastolic point was defined as the time at which dP/dt exceeds 200 mmHg/s (Takeuchi
1991). Curve fitting was performed using the Levenberg-Marquardt algorithm for non-linear
least squares. Each pulse was examined individually to determine whether the fit was
acceptable for analysis. In some instances, there was insufficient data (three or fewer data
points) in the determined period of either isovolumic contraction or relaxation, and this fits
were discarded. Other discarded pulses had fits with inappropriate amplitudes or periods,
16


where the pressure trace only occupied the upper portion of the sinusoid, or the sinusoid had a
frequency that doubled that of the pulse.
p
The coupling ratio Ees/Ea was determined as 1, using equation (3) from
Pes
above. The definition used for end-systole was 32 ms prior to the minimum value of dP/dt.
Alyono et al (1985) found that 30 ms before dP/dtmin to be the best definition for end-systole
using pressure data alone, compared with definitions using pressure as well as volume.
Because the pressure data was obtained with a sampling frequency of 250 Hz, data points were
separated by 4 ms, so end-systole was defined as 32 ms, or 8 data points, prior to dP/dtmin
rather than 30 ms prior.
Statistical Analysis
The computed VVCR for 98 patients was tested for prognostic capability against the
WHO-FC outcome of each patient at follow-up. In each analysis, the coupling ratio was
compared with pulmonary vascular resistance index (PVRI) as the hemodynamic measure with
prognostic value already established. Furthermore, the coupling ratio was calculated using
both the 32 ms prior to dP/dtmin definition for end-systole as well as simply using Pmax for a
given pulse as end-systole, and each method was compared. Outcomes were grouped for
analysis in three different ways. WHO-FC I and II and WHO-FC III and IV formed two groups for
the first analysis, which allowed the construction of receiver-operator characteristic curves (ROC
curves). This grouping was chosen because WHO-FC I and II are less severe compared with
WHO-FC III and IV, which represent the onset and progression of RV dysfunction and failure.
The area under each ROC curve was calculated using trapezoidal integration. The analysis for
the second grouping was performed similarly, but by grouping WHO-FC II and lll/IV together.
This dichotomous categorization would represent the ability of the three measures to
17


distinguish between mild hypertension and both early and late stages of its progression.
In the third grouping, only WHO-FC III and IV were grouped together, leaving three
outcomes to be predicted: WHO-FC I, WHO-FC II, and WHO-FC lll/IV. Multiple logistic
regressions were performed on the coupling ratio and PVRI, and the predictive power and
goodness of fit were assessed by the AlCc and c-values. Furthermore, a multiple logistic
regression was performed using the combination of both VVCR and PVRI as the predictors.
Finally, the logistic models were applied to the data and the number of correct prognoses for
each measure were compared. Finally, we examined the number of "close" prognoses for each
measure, which was defined as a probability within 10% of the correct prognosis. For example,
if a case is predicted to be FC II with 45% likelihood and FC I with 40% likelihood, and the
observed outcome was FC I, this would be considered "close." This is justified in the context of
the clinic, where a measure that is statistically on the cusp between predicted outcomes would
warrant further investigation to confirm prognosis, and therefore not represent a missed
prognosis in the strictest sense.
Finally, a subset of the data was studied in which only patients with an outcome of FC II
or greater at initial catheterization were included (n = 54). Because patients with a WHO-FC of
I are essentially asymptomatic, this allowed a study of the prognostic capability of each measure
for predicting whether sick patients will recover (WHO-FC I at follow-up) or worsen (WHO-FC III
or IV at follow-up). Two dichotomous categorizations were used for this analysis. The first
compared WHO-FC I (better than at initial) with WHO-FC II, III, and IV, and constituted the
"recovery" study. The second compared WHO-FC I and II (better or the same as initial) with
WHO-FC III and IV, and constituted the "failure" study.
18


Results
ROC Curves Full Dataset (n = 98)
ROC curves for VA coupling ratio and for PVRI are shown in Figures 2. The area under
the curve (AUC) for the coupling ratio using estimated end-systole was 0.8158, meaning the
measure had an 81.58% probability of correctly distinguishing between WHO-FC l/ll and
WFIO-FC lll/IV. In comparison, the AUC for coupling ratio using max pulse pressure was 0.8164,
while the AUC for PVRI was 0.8642.
ROC Curve: Ratio (end-systole) ROC Curve: Ratio (maximum pressure)
>
(7)
c
<1>
CO
ROC Curve: PVRI
AUC = 0.8642
0.6 0.8 1
1-Specificity
Figure 2 ROC curves for full dataset for VVC (end-systole, top left, AUC = 0.8158), VVC
(maximum pressure, top right, AUC = 0.8164), and PVRI (bottom left, AUC = 0.8642).
19


Logistic Regression Dichotomous Categorization (WHO I & II vs. Ill & IV)
In addition to constructing simple ROC curves for the dichotomous case, logistic
regressions were created, and the second-order Akaike Information Criterion (AlCc) was
obtained for each of the three models. The AlCc for each model, as well as the AAIQ (A|),
Akaike weight (W|,
eV2
Sf=1e-V2
), and evidence ratio (w/wj, where Wj is the Akaike weight for the
model with the smallest AlCc value) are shown in Table 2, in order of ascending AlCc values.
Table 2 AlCc and related parameters for the three models tested on a dichotomous
categorization of outcome (WHO I & II vs. Ill & IV)
Model AlCc £i Akaike weight Evidence ratio
iWjl (Wi/Wr)
PVRI 74.17 0.00 0.9918 1.00
Ratio (Pmax) 84.83 10.66 0.0048 206.62
Ratio 85.53 11.36 0.0034 291.71
(end-systole)
Finally, the logistic models created were applied to the data and the number of correct,
incorrect, and "close" prognoses were recorded. These results are shown in Table 3.
20


Table 3 Correct, incorrect, and "close" prognoses for the data set (n = 98) by each of the three
logistic regression models for dichotomous categorization of outcome (WHO I & II vs. Ill & IV)
Model Correct (n/%) Incorrect (n/%) Correct and Close
(nZ%l
PVRI 84/85.71% 15/15.30% 85/86.73%
Ratio (Pmax) 85/86.73% 14/14.28% 86/87.77%
Ratio (end-systole) 82/86.37% 17/17.35% 88/89.79%
Logistic Regression Dichotomous Categorization (WHO I vs. II, III, & IV)
Regressions and analysis were performed in the same way for this dichotomous
grouping as in the previous section. Information loss analysis and observed predictions are
shown in Tables 4 and 5, respectively.
Table 4 AlCc and related parameters for the three models tested on a dichotomous
categorization of outcome (WHO I vs. II, III, & IV)
Model AlCc £i Akaike weight Evidence ratio
Ml (Wi/Wr)
Ratio (Pmax) 117.93 0.00 0.5299 1.00
Ratio 118.20 0.27 0.4629 1.14
(end-systole)
PVRI 126.53 8.60 0.0072 73.60
21


Table 5 Correct, incorrect, and "close" prognoses for the data set (n = 98) by each of the three
logistic regression models for dichotomous categorization of outcome (WHO I vs. II, III, & IV)
Model Correct (n/%) Incorrect (n/%) Correct and Close
in/%1
PVRI 64/65.31% 34/34.69% 79/80.61%
Ratio (Pmax) 69/70.41% 27/27.55% 73/74.49%
Ratio (end-systole) 68/69.39% 30/30.61% 73/74.49%
Logistic Regression Trichotomous Categorization
Logistic regressions were created for trichotomous categorization of outcomes in the
same manner as for the dichotomous case, with WHO-FC I, WHO-FC II, and WHO-FC lll/IV
comprising the three categories. The AlCc and related parameters are shown in Table 6.
Correct, incorrect, and "close" prognoses counts for each model are shown in Table 7. Finally,
Table 8 characterizes performance of Ratio (Pmax) and PVRI by outcome and shows the overlap
between these measures.
Table 6 AlCc and related parameters for the three models tested on a trichotomous
categorization of outcome
Model AlCc £i Akaike weight Evidence ratio
Ml (Wi/Wr)
PVRI 185.12 0.00 0.6967 1.00
Ratio (Pmax) 187.75 2.63 0.1870 3.73
Ratio 188.70 3.58 0.1163 5.99
(end-systole)
22


Table 7 Correct, incorrect, and "close" prognoses for the data set (n = 98) by each of the three
logistic regression models and the combined regression for trichotomous categorization of
outcome
Model Correct (n/%) Incorrect (n/%) Correct and Close ln/%1
PVRI 54 (55.10%) 45 (45.92%) 70 (71.43%)
Ratio (Pmax) 57 (58.18%) 41 (41.84%) 67 (68.37%)
Ratio (end-systole) 53 (54.08%) 46 (46.94%) 70 (71.43%)
PVRI + Ratio (Pmax) 56 (57.14%) 42 (42.86%) 65 (66.32%)
Table 8 Breakdown of correct prognoses by outcome of Ratio (Pmax) and PVRI, and which
cases were correctly predicted by both measures (overlap)
# Correct OC 3 (n = 21) # Correct OC 2 (n = 36) # Correct OC 1 (n = 41) Total
Ratio (Pmax) 11 (52.4%) 24 (66.7%) 22 (53.7%) 57
PVRI 11 (52.4%) 15 (41.7%) 28 (68.3%) 54
Overlap 5 (23.8%) 10 (27.8%) 14 (34.1%)
Table 9 Breakdown of correct prognoses by outcome of Ratio (Pmax) and PVRI together and
PVRI alone, and which cases were correctly predicted by both measures (overlap)
# Correct OC 3 (n = 21) # Correct OC 2 (n = 36) # Correct OC 1 (n = 41) Total
PVRI + Ratio 13 (61.9%) 21 (58.3%) 22 (53.7%) 56
(Pmax)
PVRI 11 (52.4%) 15 (41.7%) 28 (68.3%) 54
Overlap 11 (52.4%) 12 (33.3%) 18 (43.9%)
23


ROC Curves Partial dataset (n = 54)
ROC curves for the "recovery" grouping and the "failure" grouping are shown in Figures
3 and 4, respectively. The AUCs for the "recovery" grouping were: VVCR (end-systole) =
0.8818, VVCR (maximum pressure) = 0.8886, PVRI = 0.5750. For the "failure" grouping, the
AUCs were: VVCR (end-systole) = 0.8338, VVCR (maximum pressure) = 0.8279, PVRI = 0.8500.
ROC Curve: Ratio (end-systole)
ROC Curve: Ratio (maximum pressure)
1-Specificity 1-Specificity
ROC Curve: PVRI
1-Specificity
Figure 3 ROC curves for "recovery" grouping of data subset (sick at initial catheterization, n =
54). VVC (end-systole, top left, AUC = 0.8818), VVC (maximum pressure, top right, AUC =
0.8886), and PVRI (bottom left, AUC = 0.5750).
24


ROC Curve: Ratio (end-systole)
1-Specificity
ROC Curve: PVRI
1-Specificity
ROC Curve: Ratio (maximum pressure)
0 0.5 1
1-Specificity
Figure 4 ROC curves for "failure" grouping of data subset (sick at initial catheterization, n = 54).
VVC (end-systole, top left, AUC = 0.8338), VVC (maximum pressure, top right, AUC = 0.8279),
and PVRI (bottom left, AUC = 0.8500).
Logistic Regressions Partial Dataset (n = 54)
Logistic regressions were performed for the partial dataset in the same manner as the
previous dichotomous analyses, with groupings of outcomes at follow-up. The AlCc analysis
for the "recovery" grouping and "failure" grouping are shown in Figures 9 and 10, respectively.
Table 10 AlCc and related parameters for the three models tested on the "recovery"
dichotomous categorization of outcome (WHO I vs. II, III, & IV)
Model AlCc At Akaike weight (Wil Evidence ratio (Wi/W|)
Ratio (Pmax) 35.50 0.00 0.9055 1.00
Ratio (end-systole) 40.02 4.52 0.0945 9.58
PVRI 55.05 19.55 5.148E-5 1.76E4
25


Table 11 AlCc and related parameters for the three models tested on the "failure
dichotomous categorization of outcome (WHO I & II vs. Ill, & IV)
Model AlCc £i Akaike weight iWjl Evidence ratio (Wj/Wi)
Ratio (Pmax) 53.50 0.00 0.7412 1.00
Ratio (end-systole) 56.77 3.27 0.1444 5.13
PVRI 57.24 3.74 0.1142 6.49
Discussion
This study sought to determine the efficacy of a novel method for estimating the
p p
ventricular-arterial coupling ratio as max'lS0 and max'lS0. These measures were obtained by
^es Pmax
sinusoidal curve fitting to pressure traces from catheterization of the right ventricle, and their
ability to predict the WHO functional class was compared with PVRI. The major findings were
presented in the previous section, and will be discussed in detail below.
First, it is interesting to note that the VA coupling ratio using Pmax performed better
across the board than the ratio using Pes defined by 32 ms prior to dP/dtmin. This could be for a
number of reasons. The definition of end-systole as 32 ms prior to minimum dP/dt is
established for the left ventricle, not the right. The typical pressure-volume loop for the right
ventricle is more triangular than for the left ventricle, meaning that the upper-left corner of the
PV loop actually coincides closely with maximum pressure. Also, because the data comes from
patients with varying states of RV dysfunction, their length of isovolumic contraction and
relaxation periods will not necessarily be consistent. This could ultimately yield more variance
in the estimated end-systolic point than in the maximum pressure point. Furthermore, the age
of the patients used could affect the timing of end-systole. This study consisted primarily of
children, while the study defining end-systole as 30 ms prior to dP/dtmin used dogs, so this
definition may not have been appropriate for the purposes of this study. Because the
26


estimation of coupling ratio used in this paper is a relative measure, it makes sense that the
more consistent measure would be the more prognostically valuable one. Finally, maximum
pressure is frequently used an as estimation of end-systole in the right ventricle by clinicians
(source?). Therefore, the coupling ratio discussed in this section will be limited to the ratio
derived from Pmax.
For the dichotomous categorization regression, the AlCc showed that the PVRI model
minimizes the information loss from the data far better than did the coupling ratio. The Akaike
weight shows that PVRI is more than 99% likely to minimize information loss. This is somewhat
surprising considering the coupling ratio is based on RV function and should be especially adept
at distinguishing RV failure. Flowever, PVRI can often increase to as much as ten times the
normal range of values in patients considered OC 3. Thus, statistically there is much greater
separation between patients with a PVRI of 1-5 mmFIg and patients with a PVRI of 15-30 mmFIg.
Still, although the AUC of the ROC curve for PVRI was greater than that for the coupling ratio, an
AUC value greater than 0.80 does imply significant prognostic ability of the coupling ratio to
distinguish between mild and severe PH. Furthermore, the regression for the coupling ratio
actually correctly predicted outcome in one case more than PVRI. So, although the PVRI model
is better in the goodness-of-fit sense, it's apparent that there is still significant prognostic value
in the coupling ratio. Also, while AlCc suggests which model minimizes information loss from
the given data, it is a relative measure, and says nothing about the goodness of fit of either
model in an absolute sense.
The AlCc values for the trichotomous categorization were much closer together,
suggesting a more comparable performance of the models. Akaike weights of about 0.70 for
PVRI and 0.19 for coupling ratio suggest that PVRI is still likely the better model, but that the
ratio is certainly worth considering. The number of cases correctly or closely predicted by each
27


regression were again fairly comparable, with the ratio correctly predicting four more cases than
PVRI, but PVRI getting two more cases than ratio with the close calls included. Table 6 offers
further insight into the relative performance of the two measures. PVRI is somewhat better at
predicting OC 1, while coupling ratio is better at predicting OC 2, and they are equal in OC 3.
However, there is relatively little overlap between these predictions (50% or less), which
suggests that the coupling ratio predicts some aspect of RV function that PVRI does not.
One of the biggest issues in prognosticating PH is the variability of its pathology and
progression among individuals. In particular, the manner and degree to which the RV
compensates for the elevated PA pressure it pumps against affect the symptoms present as well
as pathological remodeling of the vasculature and ventricle. In the compensated state,
hypertrophy of the ventricle increases its contractility, and arteries stiffen to maintain cardiac
output. There is a mild increase in PVR, and hemodynamics are usually minimally affected.
Because cardiac output is normal, these patients likely predominately show few or no symptoms
and are in the OC 1 group. This means that there may be substantial overlap of observed PVRI
values between OC 1 and 2, causing OC 2 to be less distinguishable by PVRI. On the other
hand, coupling ratio may be slightly increased in the compensated state relative to normal if it is
affected at all, so OC 1 and 2 may be better differentiated. More importantly, OC 2 represents
the onset of noticeable symptoms, and therefore also represents the transition from the
compensated toward the failing state. In this sense, the improvement in prediction by the
coupling ratio over PVRI observed for OC 2 could be interpreted as both logical and promising.
Whether a patient's disease will continue or worsen is more vital information to a clinician than
whether it is improving. Furthermore, both the coupling ratio and PVRI correctly predicted 11
of the OC 3 cases, but with an overlap of only five. This implies that the coupling ratio is
detecting some aspect of RV dysfunction that PVRI isn't. Furthermore, a regression of coupling
28


ratio to PVRI gave an r2 value of 0.08, which suggests a fairly weak fit. This further
corroborates the assertion that the coupling ratio is giving us information that PVRI is not.
This is strongly corroborated by the results of the "recovery" study on the partial
dataset. By isolating patients with a WHO-FC II or greater at initial catheterization, we could
determine the strength of VVCR and PVRI at predicting which patients would improve at
follow-up. The ROC curves for VVCR were quite strong, while the one for PVRI was in this case
barely better than random selection. The AlCc analysis also showed that VVCR modeled the
data much better than PVRI. This means that in patients who improved between the initial
catheterization and follow-up, PVRI was no different than in those that showed no
improvement, whereas VVCR was significantly higher in improved patients. It is reasonable to
speculate that this is because VVCR is a reflection of ventricular performance, which is a primary
factor in a patient's capacity for recovery. More importantly, these results highlight the
potential prognostic ability of VVCR in assessing RV function. The results from the "failure"
grouping were similar to the other dichotomous groupings, with PVRI performing slightly better
in ROC curves and AlCc. This is neither surprising nor discouraging, because of the sick patients
used in this analysis, only one actually worsened, making this analysis nearly the same as the
second dichotomous grouping for the full dataset.
There are a number of limitations associated with this model. Because VVCR is a ratio
of elastances, it can only provide an indirect measurement of ventricular performance as it
relates to the state of the vasculature. Thus, parameters involving direct measurement of RV
volume, such as ESPVR, will likely always remain better indicators of RV function.
Furthermore, although it the measurement of pressure is much simpler than the measurement
of pressure and volume together, VVCR still requires catheterization, meaning it offers no large
advantage in terms of invasiveness. It also suffers from having a fairly small range compared
29


to its variance among individuals, although this is alleviated by considering it alongside other
measures, such as PVRI. Additionally, although it has been shown here to be prognostic, this
method for obtaining VVCR requires validation. This should be accomplished by comparing the
estimated VVCR against VVCR obtained from PV loops, in both animals and humans.
30


CHAPTER III
TR JET DOPPLER STUDY
Introduction
The regurgitant tricuspid valve is fairly common in patients with PH, present in 70% or
more cases with significantly heightened pulmonary artery pressure (Yock and Popp 1984), and
results in a backflow known as a TR jet. Changes in the geometry of the ventricle due to
hypertrophy or dilation, as well as the excess pressure present in the ventricle in PH often
causes the tricuspid valve to function improperly, not completely closing. The velocity of the
TR jet can be measured in real time non-invasively using continuous-wave Doppler ultrasound.
In 1984, Yock and Popp published a method for obtaining an estimate of peak ventricular
systolic pressure by using a simplified form of Bernoulli's principle.
In its general form, Bernoulli's principle describes the pressure gradient across an
obstruction (in this case, the tricuspid valve) is given as
Pt-P2= \p{Vi Vi2) + p £ § ds + R(V) (4)
where the terms on the right side of the equation describe convective acceleration, flow
acceleration, and viscous friction, respectively. In this scenario, the flow acceleration and
viscous friction terms are negligible compared to the convective acceleration term, and Vi is
small compared to V2, so we obtain a modified form of Bernoulli's principle as AP
4V2, where V is the maximal velocity of the TR jet, and the constant reflects the density of blood
as it factors into the equation. Peak ventricular systolic pressure is then obtained by adding
right atrial pressure estimated clinically using jugular vein pressure.
By using this simplified form of Bernoulli's principle, we can take the complete velocity
envelope from the TR jet and obtain a continuous pressure trace, allowing the modified
single-beat method to be used non-invasively. In essence, it is not even necessary to translate
31


from velocity to pressure in this case, because VVCR is calculated as a simple ratio, except that
the definition used for the onset of isovolumic contraction involves the first derivative of
pressure. However, this definition could be put in terms of velocity by use of the simplified
form of Bernoulli's principle discussed above.
Doppler Data Analysis
Doppler ultrasound data was visualized and processed in Matlab. Noise reduction was
applied using the discrete wavelet transform method described by Zhang, et al in 2001. The
wavelet used was the 8th-order symlet wavelet with a coarsest resolution of six. The threshold
value was computed automatically based on the noise level, soft thresholding was applied to
the detail coefficients, and the image was reconstructed.
Traces of TR jet velocity were performed by an automatic edge detection routine using
an adjustable threshold magnitude. Because Doppler ultrasound data contains spectral
velocity data over time, edge detection was performed for each column of data by marking the
first pixel to exceed the threshold value from the top of the image (maximal negative velocity).
The resulting trace was smoothed using cubic splines.
The smoothed velocity trace was transformed into approximate right-ventricular
pressure by simplification of Bernoulli's principle as P = 4V2. This data was subsequently
analyzed using the techniques described in Section 2.1 above. Having already performed
better in the catheterization cases, the maximal pressure was used in computing the coupling
ratio, rather than the end-systolic definition of 32 ms prior to dP/dtmin.
The primary assumption and potential limitation of this method is that the velocity of
the TR jet corresponds to ventricular pressure for the entire systolic envelope, which may not be
the case and likely depends somewhat on individual physiology. Using TR jet velocity as a
means of estimating ventricular pressure has been validated only for peak systolic pressure, not
32


pressure development over time. Furthermore, atrial filling begins during isovolumic
relaxation of the ventricle, which may affect the velocity of the TR jet, and therefore the
estimated ventricular pressure. This may account for the slope being steeper during
contraction than relaxation in some of the Doppler images that were examined.
Results
Attempts at obtaining coupling ratios from Doppler images had mixed results. Of the
32 images analyzed, 19 yielded pressure traces with at least one good pulse for curve fitting.
Noise was generally not a problem after applying the noise reduction algorithm and smoothing
the raw trace obtained by thresholding. An example of an image yielding good results is shown
in Figure 5 along with the corresponding pressure trace and curve fitting in Figure 6. A
frequent problem encountered is shown in Figure 7 the portion of the image corresponding to
isovolumic contraction is much steeper than the portion corresponding to isovolumic relaxation,
yielding a poor fit. This shallow portion on the back side of the pulse is likely due to an
unwanted signal from the Doppler image, perhaps the fluttering of the tricuspid valve. Mean
coupling ratio and standard deviation for the patients yielding good traces and fits are given in
Table 7. Patients with multiple images were averaged together.
33


Figure 5 Automatic trace of Doppler US image of TR jet. Red raw trace; Blue smoothed trace
Pressure (mmHg) p , ,
Figure 6 Pressure trace and curve fitting of third pulse from Doppler image in Figure 5 above
34


Table 12 Coupling ratios for Doppler images with acceptable pressure traces
Patient ID N (# of pulses) Mean Ratio Standard Deviation
PHT0251 7 1.646 0.535
PHTR001 6 1.455 0.145
PHTR002 2 1.452 0.096
PHTR003 4 2.214 0.572
PHTR010 3 2.480 0.979
PHTR012 4 1.169 0.117
PHTR013 3 1.194 0.061
PHTR014 1 1.499 0
PHTR015 3 1.140 0.101
PHTR022 2 2.444 0.283
PHTR023 5 1.216 0.163
35


Discussion and Future Work
The goal of this portion of the study was to apply the modified single-beat method
developed in this study to Doppler ultrasound images of tricuspid valve regurgitation jets by
using a simplification of Bernoulli's principle in order to obtain approximate pressure over time
in the right ventricle from the velocity of the jet. Although there is much work to be done to
validate and improve this method, it shows great promise in the form presented here. In
theory, this method can provide the same information as the modified single-beat method used
on catheter pressure traces that formed the main portion of this study, yet it has the enormous
advantages of being completely non-invasive, cheap, and quick and easy to perform. Thus,
even if it proves to be somewhat less accurate than catheterization, it could still prove to be an
extremely powerful tool for clinicians.
One of the first things that can be done to improve the method is to implement a tool
for manual edge definition. This was performed by a simple algorithm for the purposes of this
paper, which resulted in some images with potentially valid pressure information, such as the
one shown in Figure 7, to be impossible to obtain a proper isovolumic sinusoidal fit due to
unwanted signals. These signals could simply be ignored using a manual trace, which could
increase the number of usable images substantially.
Next, this method needs to be validated. First, the ratios obtained with Doppler
ultrasound should be compared with ratios obtained from the same patients by RV
catheterization. If there is reasonably good agreement between the invasive and non-invasive
measurements, it will go a long way toward supporting the use of Doppler ultrasound in the
assessment of RV function. Furthermore, there should be a statistical analysis of the ratios
obtained from this method with WFIO-FC in the same manner as the catheterization method in
this paper. Again, it should not be expected that this method will be as accurate as
36


catheterization, but if it proves to be reasonably good at predicting clinical outcome, it could
prove extremely useful as a preliminary prognostic tool.
Another thing that could be considered from a statistical standpoint would be the
clinical outcome (WHO-FC) of patients presenting with TR jet compared with those who have no
tricuspid regurgitation. It would be expected that patients in a more advanced state of PH and
RV dysfunction would be more likely to present with tricuspid regurgitation. This tool could
also be used to assess the severity of tricuspid regurgitation compared with WHO-FC.
37


CHAPTER IV
CONCLUSION
The single-beat methodology for estimating the ventricular-arterial coupling ratio
examined in this paper has a number of advantages over other commonly-used hemodynamic
measures, and shows promise for clinical use in the future. Perhaps most importantly, the
VVCR reflects the state of both arterial compliance and ventricular contractility at once. It
indicates the status of the pulmonary cardiovascular circuit as a unit, whereas a measure such as
PVRI only indicates properties of the vasculature. Thus, although it was found that PVRI may
predict clinical outcome as well or slightly better than VA coupling ratio, PVRI can only indicate
heart failure indirectly as vascular resistance rises.
Furthermore, this method has the advantage over all other current methods of
measuring ventricular properties that no volume measurement is necessary. This makes it
particularly desirable in the right ventricle, where the small size and uneven shape make
volumetric measurement problematic. It also has the potential to become totally non-invasive
with the use of Doppler ultrasound utilizing the regurgitant tricuspid valve velocity. This is a
nearly trivial extension of the current method of estimating peak systolic pressure using the
peak velocity of the TR jet, and could prove extremely useful depending on additional validation.
VVCR can also be used in conjunction with other hemodynamic measures to strengthen
prognosis, as was shown in this paper using PVRI. Because the measure partially reflects
ventricular properties, it contains unique information not available from other measures
without volume measurements. This also means it can be compared with PVRI in studying the
effects of treatments in order to differentiate between effects on the vasculature and the
myocardium (as most treatments used for PH are known to affect both in some way).
Comprehensive clinical trials to this end could yield interesting information both about the
38


particular clinical usefulness of the VVCR and about the specific effects of some of the
treatments commonly used for dealing with PH. Another useful clinical trial that could be
performed is determining whether the coupling ratio can accurately predict transplant and
death. Although this was performed to a small extent in this paper, the small number of
transplants and deaths in the pediatric patients comprising this study limited its usefulness.
Finally, although PVRI was found to be a slightly better indicator of failure, it was clearly
demonstrated that VVCR is superior to PVRI in predicting recovery in sick patients, which holds
immense clinical value. Taken with its relative simplicity compared with other means of
assessing RV function, these factors make VVCR estimated by the modified single-beat method
presented in this paper an effective prognostic tool for the clinic.
The next step remains validation. This method can be validated for both the left and
right ventricle by comparing the estimated VVCR with VVCR obtained from PV loops, first in
animal models, likely the rat and cow, then in humans. Because volume measurements by
conductance catheters are problematic in the RV, MRI volume measurements should be
obtained as well. The Doppler ultrasound TR jet method should undergo similar validation.
Finally, more comprehensive clinical trials should be performed, involving mortality and
transplant rates, adult patients, and additional follow-ups with patients.
39


REFERENCES
Abel, L. (1981). Maximal negative of end of systole as an indicator. Society.
Antonio, S. (1985). Defining End Systole for End-Systolic. October, 350, 344-350.
Baan, Jan, Tjong T Aouw Jong, Peter L M Kerkhof, Rudolf J Moene, Arjan D Van Dijk, E.
T. V. D. V. and J. K. (1981). Continuous stroke volume and cardiac output from
intra-ventricular dimensions obtained with impedance catheter. Cardiovascular
Research, 15(6), 328-334. Retrieved from
http ://cardiovascres. oxfordj oumal s. org / content/15/6/328. short
Badesch, D. B., Champion, H. C., Sanchez, M. A. G., Hoeper, M. M., Loyd, J. E., Manes,
A., ... & Torbicki, A. (2009). Diagnosis and assessment of pulmonary arterial
hypertension. Journal of the American College of Cardiology,54(\sl), S55-S66.
Bemus, A., Wagner, B. D., Accurso, F., Doran, A., Kaess, H., & Ivy, D. D. (2009). Brain
natriuretic peptide levels in managing pediatric patients with pulmonary arterial
hypertension. Chest, 135(3), 745-51. doi:10.1378/chest.08-0187
Brimioulle, S., January, F., Tabima, D. M., Hacker, T. A., Chesler, N. C., Wauthy, P.,
Ewalenko, P., et al. (2003). Single-beat estimation of right ventricular end-systolic
pressure-volume relationship Single-beat estimation of right ventricular end-systolic
pressure-volume relationship. American Journal Of Physiology, (January 2003).
doi:10.1152/ajpheart.01023.2002
Brown, K. a., & Ditchey, R. V. (1988). Human right ventricular end-systolic
pressure-volume relation defined by maximal elastance. Circulation, 75(1), 81-91.
doi: 10.1161/01. CIR. 78.1.81
40


Burkhoff, D., Mirsky, I., & Suga, H. (2005). Assessment of systolic and diastolic
ventricular properties via pressure-volume analysis: a guide for clinical, translational,
and basic researchers. American journal ofphysiology. Heart and circulatory
physiology, 289(2), H501-12. doi:10.1152/ajpheart.00138.2005
Chen, C. H., Fetics, B., Nevo, E., Rochitte, C. E., Chiou, K. R., Ding, P. a, Kawaguchi, M.,
et al. (2001). Noninvasive single-beat determination of left ventricular end-systolic
elastance in humans. Journal of the American College of Cardiology, 38(1), 2028-34.
Retrieved from http://www.ncbi.nlm.nih.gOv/pubmed/l 1738311
Desk, R., Williams, L., & Health, K. (1973). Ventricle and Effects of Epinephrine and
Heart Rate on the Ratio Load Independence of the Instantaneous Pressure-Volume
Ratio of the Canine Left Ventricle and Effects of Epinephrine and Heart Rate on the
Ratio. Circulation Research.
Eysmann SB, Palevsky HI, ReichekN, Hackney K, Douglas PS. Two-dimensional and
Doppler-echocardiographic and cardiac catheterization correlates of survival in
primary pulmonary hypertension. Circulation 1989;80:353-360.
Galie, N., Hoeper, M. M., Humbert, M., Torbicki, A., Vachiery, J.-L., Barbera, J. A.,
Beghetti, M., et al. (2009). 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. European heart journal, 30(20),
2493-537. doi:10.1093/eurheartj/ehp297
41


Kelly, R. P., Ting, C.-tai, Yang, T.-ming, Liu, C.-peng, Maughan, W. L., Chang, M.-song,
& Kass, D. A. (1992). Effective Arterial Elastance as Index of Arterial Vascular Load
in Humans. Circulation, 513-521. doi:10.1161/01.CIR.86.2.513
Kuehne, T., Yilmaz, S., Steendijk, P., Moore, P., Groenink, M., Saaed, M., Weber, O., et
al. (2004). Magnetic resonance imaging analysis of right ventricular pressure-volume
loops: in vivo validation and clinical application in patients with pulmonary
hypertension. Circulation, 770(14), 2010-6.
doi: 10.1161/01.CIR.0000143138.02493 .DD
Lam, C. S. P., Roger, V. L., Rodeheffer, R. J., Bursi, F., Borlaug, B. a, Ommen, S. R., Kass,
D. a, et al. (2007). Cardiac structure and ventricular-vascular function in persons with
heart failure and preserved ejection fraction from Olmsted County, Minnesota.
Circulation, 775(15), 1982-90. doi: 10.1161/CIRCULATIONAHA. 106.659763
Mertens, L. L. & Friedberg, M. K. (2010). Imaging the right ventricle current state of the
art. Nat. Rev. Cardiol. 7, 551-563; published online 10 August 2010;
doi:10.1038/nrcardio.2010.118
Naeije, R., & Torbicki, A. (1995). More on the non invasive diagnosis of pulmonary
hypertension: Doppler echocardiography revisited. European Respiratory Journal,
1445-1449. doi: 10.1183/09031936.95.08091445
Nichols, W. W., & Edwards, D. G. (2001). Arterial Elastance and Wave Reflection
Augmentation of Systolic Blood Pressure: Deleterious Effects and Implications for
Therapy. Journal of Cardiovascular Pharmacology and Therapeutics, 6(1), 5-21.
doi: 10.1177/107424840100600102
42


Raymond RJ, Hinderliter AL, Willis PW, Ralph D, Caldwell EJ, Williams W, Ettinger NA,
Hill NS, Summer WR, de Boisblanc B, Schwartz T, Koch G, Clayton LM, Jobsis
MM, Crow JW, Long W. Echocardiographic predictors of adverse outcomes in
primary pulmonary hypertension. JAm Coll Cardiol 2002; 39:1214-1219.
Suga, H., & Sagawa, K. (1974). Instantaneous pressure-volume relationships and their
ratio in the excised, supported canine left ventricle. Circulation research, 35(1),
117-26. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/4841253
Sunagawa, K., Maughan, W. L., Burkhoff, D., & Sagawa, K. (1983). Left ventricular
interaction with arterial load studied in isolated canine ventricle. The American
journal of physiology, 245(5 Pt 1), H773-80. Retrieved from
http://www.ncbi.nlm.nih.gov/pubmed/6638199
Sunagawa, K., Maughan, W. L., & Sagawa, K. (1985). Optimal arterial resistance for the
maximal stroke work studied in isolated canine left ventricle. Circulation research,
56(4), 586-95. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3978773
Takeuchi, M., Igarashi, Y., Tomimoto, S., Odake, M., Hayashi, T., Tsukamoto, T., Hata,
K., et al. (1991). Single-beat estimation of the slope of the end-systolic
pressure-volume relation in the human left ventricle. Circulation, 53(1), 202-12.
Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1898642
Yock, P. G., & Popp, R. L. (1984). Noninvasive estimation of right ventricular systolic
pressure by Doppler ultrasound in patients with tricuspid regurgitation. Circulation,
70(4), 657-662. doi: 10.1161/01.CIR.70.4.657
Zhang, Y., Wang, Y., Wang, W., & Liu, B. (2001). Based on Wavelet Frames, 48(3).
43


Full Text

PAGE 1

nr nn r r !"#!$%%& !'###()!!!*!' +( !*,!'n(!+'** *,!' "#!*,* **-! ,( )! *,!'.()!#,*!'/*, #!*,++ r*//n(!*/) $%0$

PAGE 2

'#!'##,*!'/*,#!*,++ +'* #"*#* '#--*",*!' r*//n(!*/) *'#' 1 (! (" *")2$%0$

PAGE 3

*#*+'* #"3r*//*/) 4 *,/ r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

PAGE 4

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

PAGE 5

" nn n 0 r+:/*( 0 !*#'#!+ $ #!+!+( #+( *(/ = + !(#! 6/ !) !# C / r!!'** */ A ( )*-!#* @ ####)!*,,(+!* 0$ Dnr 0C !')!+ r## 0C !+.(#!* 0? ##(! ## 0? !!#!+ ## 0A #( !# 02 #+(##* $@ ED =0 !*(+!* =0 *-! ## =$ #( !# == #+(##*(!(9*: =@ =2 C0

PAGE 6

" nr ##("* ()!*:"* ()>,!>/'!"!+ >( )*!>-( )*! ##( ( )*"#+( ##!+89( )*7/-##(+*(!-(!/'!"!+( 5+!*,+!*#!* +"* ()##!* +-##("* () !*#'-#!* +-##("* () !*#'##!* +-##( !*#'-!+( "#+( +*(/!*###!* + #!+ ,,+!"! #!+ >( )*'-!#*>-( )*! '-!#* 99* !'/6!*,(+!* + ## A9A)(!7 :!#!rr!(!+--! !, (7!'*) 5+!*,+! -!#*7!'*(!'!, (+(#-/(/!!*/'!'!+!'!6!*+("+"*-!*+'+!#!++("+*+!::,*)!*!*(!'+("

PAGE 7

" rn n0##("* () **-/),* ,!"!+ = $+("#,*,( !#!,* 0& =+("#,*F+*"F/*(-/*,!#( #! $C C+("#,*F, (F/*(-/*,!#( #! $? ?(!*)!+!+*,*-)/*,5! =C A##(!++(",!!/*,!'-( #,*)*-)/ =C @*-)/ /-**-##(!+,* +(",!!/ =?

PAGE 8

" rn n0(+!* + ##,+!**,-( )*'-! #* 00 $+ !-)!#,*!'!')* #!#!*+'*!*)*(# +!/*6!**,*(!+*)39G"#G4 $% =*+!+*+!F+ *#F-*/*##,* !'!#!3H&24+'*,!'!' */#!+/##*)* #,*+'*!*)*(#+!/* 6!**,*(!+*)39G"#G4$0 C+ !-)!#,*!'!')* #!#!*+'*!*)*(# +!/*6!**,*(!+*)39"#G4 $0 ?*+!+*+!F+ *#F-*/*##,* !'!#!3H&24+'*,!'!' */#!+/##*)* #,*+'*!*)*(#+!/* 6!**,*(!+*)39"#G4$$ A+ !-)!#,*!'!')* #!#!*!+'*!*)*(# +!/*6!**,*(!+*) $$ @*+!+*+!F+ *#F-*/*##,* !'!#!3H&24+'*,!'!' */#!+/##*)* #!'+*)/## *,*!+'*!*)*(#+!/*6!*$= 2r:*7*,+*+!-*/*##*(!+*)*, !*3)847'+'+## 7+*+! -+!*!')#(#3*" -4 $= &r:*7*,+*+!-*/*##*(!+*)*, !*3)84!*/!' *7'+'+##7+*+! -+! *!')#(#3*" -4 $= 0%+ !-)!#,*!'!')* #!#!*!'F+*"F+'*!*)*(# +!/*6!**,*(!+*)39"#G4 $? 00+ !-)!#,*!'!')* #!#!*!'F, (F+'*!*)*(# +!/*6!**,*(!+*)39G"#G 4 $A 0$*(/!*#,**-)/#7!'++! -##(!+# =?

PAGE 9

8 rnr nr0##!* + #!+ 0? $,,+!"! #!+ 0? =!+( "#+( +*(/!* 0? Cr*( I#-+=$

PAGE 10

0 r nn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

PAGE 11

$ r "! !*,!'+("'#" !*!'!' ,!"!+ 3*## 0&A0 *! 0&A&4/'!"!+ 3,*!!! 0&A$4'-#!')*#! #(#!! 7*:#*!'+'+!6!**, !*#'-#*,!'"!+ / +*),*)!'0&@%J#2%J#E-,*) (!'*#(//7(/7 '## 7*:#*!'""*+'!,(! '+*,),!' !*, !'-'#* */+/' -"+( #!/*,7*,*!' -*-!#*,!'#)#()!#"+#+!' !!+'* */!'"!*,!' +*(+!++!'! *7#!!*(#"* ()) #()!#*,!' ,!"!+ ""* 3r! 0&204)/+*( *! +++!'! !/* 34*!/##*, **-#+*+ !/!'##!* +-*!#3#/(04 #!'#!* +"* ()+##!' **-# #',!!*!'/'!##!* +-##( +## */!' (-!*)8)()-# #(,*#*"* ()++*!+!*3(/G /70&@C4r+(## !" ## !"!*+'/#*,! '!!!-*"#" ( 8*,##!* +++,(+!*-+,+ !'# *-*,!' #)#(*,##!* + #!+3#4*,!'"!+ (!')*#+* (#,( )#(,,+!"! #!+34+*!#!'# *-*,!' ,*) !'-##("* () **-,*)!'##!* +*!3(-,!*,!'*84!*!' #!* +"* ()!+-!3*!!*)/'!4! #!+7#,#!'!**, ##!* +-##(!*#!*:"* ()#>3/04'#,!*,*)0&2=7*: (/7+* /(#*(!! !+! *##*!'##()-!*!'!"#+( )-++)* #!' )!9:# # )* r,/!)#*, )! *7*"#/ ++++ !' # !*#'-!'!#7# #+ !+ # !*#'-7' +'+!#!+! )-+--'

PAGE 12

= ##!++*)+##!* +#!* + (!*#+*#!!'8-##* #+/!'# *-*,!'! !*#'-"* "/ !'*#! -)!#7#*! 3,,+!"! #!+4r+(#*!'#/"!'#)(!#!'!* ++*)-(!##>!*-*")#(*,"!+( "#+( +*(/ #!" 9*:#(/7! 30&2=0&2?4#*(/'!!*!# !!''-*!'##!'!"!+( 8! 7*:7*( )8)6,!'"!+( #!+#.( ,*(!'*-!) ##!+(() *," -)!##(+'# #!* +"* ()+*!+! !'!! +*)+,*(!'!*-!) ##!+"* # /'! 7!'! +*)+7' "/7 7!' +*!+! !'!!'#,/##(// #!!'!!'!**,"!+( ##( "* () **-/),* ,!"!+ #'*7/+*#!(+!**,"!**,# (#/-)!#)8#

PAGE 13

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r#+'! $%%&4'#-!!##' *( '"): +#+*(/ !*7'+'#/,#(+*(/*,!'"!+ "#+( !( $ rr% !'*(/'!''#)(+'7*:*!' +'+!6/!'-##("* () !*#'-*,!'"!+ !'*(/'!'#7* :#/ #!"!*)-'#6!'-*!! + + (#,( ##*, **-#!'(#+ + #!!/#'#)8!) )!,* ()*,#*##!!'-*+(#,**! /#!!*(#-##("* ()

PAGE 14

? )#()!#.('/' "#"+!'!6!* +'.(#,*"* ())#()!# -!+( '",,+( !)-+# #()!#*,"* ()*,!'/'!"!+ #-+ !*( #*)(!*!# *+!*#) #6*(,*)/*)! (!')*)+ + +#*#+) #*)*# *!)#(# #(+'#5+!*,+!*3r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

PAGE 15

A 7*:#'*7-* *,#!* +#,(+!* -!!# !"!*-!!# (!')*!#'*7#!'-*!! /*#!+-*7 *,+("#,* #!/(#'/#-+,++*"#+( #*# #&' 0&&0:(+'! -*-*#*" )!'*,* !'+ +( !**,!'# *-*,!' 3#4(#/#/ !' 6!'!!')8 )()-##(!'"!+ + -*(+(/#*"* ()++*!+!*+*( #! )!,*)!'-##(!+ !'# *-*,!'+*( *!7/ ,*)!'#-*!!/! !*!' (-,!+**,#/ **-3!' ##!* +-*!4(#* +(",!!/# -,*)*!'-##(!+*,!'"!+ (# /!-*!#7'+'+*#-*!*-*# *,#*"* ()++*!+!* 8!**! *#!)!)8'##)(+'# )!'*!'!'-"*(#*,*/!/!' 7'+'"* "+!/! #!!' **-# !/!'#!* +"* ()'#)!'*'# #*" !,*!'/'! "!+ r)*( $%%= ()*,##()-!*# )!!*#(.(!*! '7##!'7*:' !/( #'-*,!-+ **-#!') #!'!##!* 5+!**,! *++(!,,!!)#'(-,!+*7 #(##!'##!* +-*!7'+'# ,*!'(! ###*,*!''##() -!*!'!#*"* ()+!##(#* #! #'(/7*( ##,*!' 7#",+*)-/#!)! )8#**,5+!/!#7!'*#")8,*#*"* ()+!#*,!'#)#!!/7' +' 7,*(!*#!*/ +* !(!*!'+ /#*,)#(/"* ()""* !'!)"* ()+'/#!/!/, *7)#(!'-*8) -( )* !7!'7 !( !#*+)!'* #*,*(8-+!+'/## ("!*,*!*-+#!!#

PAGE 16

@ (!')*$%%C#!(1('! !'#/ !)!'*,*!' /'!"!+ (#/+!**!"* () #*(!**:#) + + #!( +*)-/A+*!* -!!#!*A-!!#7!'+' *+-( )*'-!#*' *!8#*,++-()-,(+!*)#( #++8*!'!**,++ *(!-(!!**#(,+#8!*)*+ )##"!+( +*(/ 3#>4-!!#7!'-()-,(+!*7#+ #)*+ +*!+! !7# '+7#,,+!+*)-7!'!' +*!* /*(-'##( !#/7!' !'8-+!-'#* */+)-+!*,*,(+! *#+(##*"+*(// ,*!')) "#"-*/*##*, r%''"! ( )*'-!#*7#,#!+ ##,0&@ =)!/*/6!'9* !'/6!*#,#')*) +-!'* */+ +*!*7'+' )-( )*! -##(34*,$?))/* /!!#!#)#(/'! '!+!'!6!**) #0CK=))/7 !'(-)!*,*(!$%))/ '#/,++*,)*,$0!*$C))/# (+ '+ + + ##,+!**,'#(/*#* )+'/#!!* +*,+##+!'! 0&@=9+ ##,+! *(!!'#+,,!!**,-) #+*##! --!'*+! /*#*,!')*#!+!+ ##,+!* /(-*!!',*(!'9* )-*#(),*! *! ,*$%%2# ,* *7#<0( )*! '-!#*340I( )*"**++ (#"##>*-( )* +')/*)!*## $( )*'-!#*(!* ,!'!# # =( )*'-!#*(!* (/### >*'-*8

PAGE 17

2 C'*+!'*)*)* +-( )*'-!#*?7!'(+ *)( !,+!* )+'#)#',*+(#*,!'#--#-)'-!#*3/ *(-04**!+(#*!'' !' -* )#r''!!**,-!'* */+ +'/## #*!7 (#!** !'*(/'!' +*!*#+'+!6()*(#-!'* */# "!*,*+')+ -!'7# + !-#+#-( )*"#+( # #!+34#+(#"#*+*#!+!**, !'!#, ))!*!'*)*##)* /*,!'"## 7 ##*+*#!+!* '# :!**) ,(+!**,-*!##()+ ##)**!')(#+ + ##7 # *!' + #,(+!* /!*)-*(+!**,"#* !*##(+'#!+*8 -*#!++ *"-*(+!**,"#*+*#! +!*##(+'#!'*)*8$ *!' 0"!)"#+( )* / # !*#!,,/,*##+# -*(+!**,+* / #!!'"!! 3n $%%&4 '+#-#!#+# 7*: *-!*!'"!+ '#+"!( !*-!'* */+ '-!*-' "!+( #!,,/ !* '##!'-*!!7'+'"!+( + *(/*++(#!'##!'!!''!# -()-!7 "!( ( !*+*)-#!,* !'+#! /!* '!, (!' !/*##*,#-*+##*,8+ (#*#**,! ),,!/*#!+!** # )!'*#!*)*!* )! *!' -*## +(##!'+#*,!'# )#( /*(! ,!'!##"**++ (#" ## (/##+'*+

PAGE 18

& !'*)*)* #))-!*)#*,*,!) -#!) (/8+#++ ( #'*!##*,!',!/(7:##/# +*-*) #!#*'#+ #/#!'!)/*##-) '!# *(#
PAGE 19

0% '/* #!,*!'/*##*,#/' !'!+!'!6!*34!# +##,*+*,)/#(+'/*##',* *7/" #)(#!*!< /'!! -##(-##(-( )*7/ -##(394++*(!-(!34 !'!')* (!**!'+:)!'*) *,$?))/*/!7!'9*, 0?))/* ##!'#+*,*!'+(##+*, )#/*##*,3n $%%&4 !'*(/'# !" #,-*+(!# "#"!',*+##*)#:!*!' -!!r! !*,!')*#!)-*!!-*/*#!+)#(*,' -!#*#!'9,(+!* + ##"#0&&2!'7D*:!##*+ !*7'+'# !" /**-+!**, )*! !*(! 0(!!-! !#7!'*-!'+*'! '#!*+ !#'*7)#("" *,A)*!' #,*9$?#,*9 A#,*9

PAGE 20

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

PAGE 21

0$ -*/*#/!' #!+'#-) #+,(+!*#*)7() (++#+!**,)-*8!")! #)*,#+')+--' !##( '/'(++ #7#'*7!*+* !7!' -**#("" -!!#7!'*!' *+')+ ):!'!'##'*7+* !*7! '#("" #!(!+ --! *7)* +( 7/'!-*!!'!+!# #"#* !*-#!/!)*(!# -!!#7!'-**#("" 3n $%%&4 '!!**,!'#!'##*,r#( !# ,*)7 #!##(!*"* ()8-#**-##( *" *3r(#! $%%&4$%%& #!(r(#! ,*(*#!*/+* !* !7r+*))* (# ')*)++'*+*/-'+-)!#(!! '!!'+'/*"!)*,r +* !7!'!'+'/*,!'*#-)!# !'!-!!#7!'r/!!'02% -/>)'+##("" !!!!%(( ()*,)!'*#,*!'####)!*,,(+! *'"-*-*# !*"!'-#!=%#"/7 "#"###7 #,,++++(+ 0&2C1)+'! +)!',#!/*(-!*" (!5+!*,+!*34(#/ *(+ "!+( */-'',, (# ##!'=2B0&2& +!! *!(#/)*,+! '!7!',#!#-*#!')#!* (! 6/!'!')* (!*)!'**7"!'# )!'*#*! *!'+(#!' !'(+ *"#"+(#" #7 *) -!!#,*)=%A%B +)/!+#*+34)//#!'/* #!,*" (!/ ,(+!*!#!'* !** 7'+'+ (!,"* () +*!#!'+'# #*#'*7!*!+! ,*!+!##(!'7 (!+ !+!/* +'/#)*+ !##( *!#!'+07/'! *7#

PAGE 22

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

PAGE 23

0C /$%0%4'(#=+'*+*/-'.( #,(!',)!" !*(! +*( +*)!'7/* #!!',(!(

PAGE 24

0? r n)nn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

PAGE 25

0A +!'!/"##(,,+!!)-* #* (!*,* ++(!+ +( !**,!',#!"!"*, -##( + + + #!(#'r--*" 9!',*) +*#!>*##!7' -+ 7#!(&2-!!#(/*/" ( !**,+!"!(#/*8/>* !+*8!'+++!'!6!* *! *!'' J#*#-! !!#//,*)0)*!'!*$A#3) H&0=KA$&#??) #4##( 7#*!7!'!'(#/?*A,7n 6+!'!#3#-+*!!!+ #!)#*!!:4 r!!'! ! ##7#-,*)r##( !+#,*)!'/'!"!+ 7"!*#/ !#n/!/)87#!),!!/!'.(!* -(.*/01* 7' #-##( #!)!*-##(" (#,*)!' #!* +-*!(-!*!'-*!*,!')8) (*,!',#!"!"+*#-*/!* #*"* ()++*!+!*,*)!'!)*,!') ) (*,!',#!"!"!*!' #)-##(" (#!'#!* +-*!+* #-*/!*#*"* ()+ 8!*' #!* +-*!7#,#!'!)!7'+ -2 8+#$%%))/>#3:(+' 0&&04(",!!/7#-,*)(#/!'" /.(! /*!'),** #!#.(#+'-( #7#8)"( !*!)7'!'!',!7# ++-! ,* ###*)#!+#!' 7##(,,+!!3!'*,7! -*!#4!'!)-**,!'#*"* () ++*!+!** 8!*!'#,!# 7#+!'#+-( ##',!#7 !'--*-!)!(#*-*#

PAGE 26

0@ 7'!'-##(!+* *++(-!'(---* !**,!'#(#**!'#(#*' ,.(+!'!*( !'!*,!'-( #'+*(/!* 2'7#!)# !#$ +, (#/.(!*3=4,*) *"',!*(#,*##!* 7#=$ )#-*!*!'))()" (*, -2 **! 30&2?4,*(!'!=%)#,* -2345!*!'#!,!*,*##!* (#/-##(! *+*)-7!',!* #(#/-##(#7 #"* () r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

PAGE 27

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r+(#-!!#7!'9*, ##! #)-!*)!+!'# *7#!( *,!'-*/*#!++!*,+')#( ,*-+!/7'!'#+:-!!#7 +*"3 9!,* *7(-4*7*#39 *!,* *7(-47*+'*!*)*(#+!/*6!* #7(#,*!'# ##',#! +*)-93!!!'!! 47!'9 +*#!!(!!' F+*"F#!('#+*+*)-9 3!!*!'#)#! 47!' 9+*#!!(!!'F, (F#! (

PAGE 28

0& !! ,!#!-./01+("#,*+*(/!*,* #'*7/(#$'( !'+("34,*!'+*(/!*(#/#!) !##!* 7#%20?2)/!' )#('20?2B-* !*,+*+! #! /(#'/!79> 9>+*)-#*!',*+*(/ !*(#/)8-( #-##(7#%20AC 7' !',*7#%2AC$ +("#,*,( !#!,*3##!* *,!H%20?24 3)8)()-##(!*-/'!H%20AC4 3*!!*) ,!H%2AC$4 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1-SpecificitySensitivityROC Curve: Ratio (end-systole) AUC = 0.8158 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1-SpecificitySensitivityROC Curve: Ratio (maximum pressure) AUC = 0.8164 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1-SpecificitySensitivityROC Curve: PVRI AUC = 0.8642

PAGE 29

$% !! # %! $ -2 3,!431 !*!*+*#!(+!/#)+("#,* !'+'*!*)*(#+# */#!+ /##*#7+!!'#+**: :,*)!*!*3+47# *!,*+'*,!'!')* #'+,* +')* #7 #!'M3M4 ::7/'!3778 9 2:; 78 9 2: <9=> 4"+!*375>77'75#!'::7/'!,*!' )* 7!'!'#) #!+" (4#'*7 $**,#+/+" (# + !-)!#,*!'!')* #! #!*+'*!*)*(# +!/*6!**,*(!+*)39G"#G4 !" # $ %%&'( )& n+",#(! )& .& @C0@ %%% %&&02 0%% !*3 )8 4 2C2= 0%AA %%%C2 $%AA$ !* 3##!* 4 2??= 00=A %%%=C $&0@0 !' */#!+)* #+!7-!*!'!!'()*,+*+! +*+!F+ *#F-*/*##7+*' ##( !##'*7 =

PAGE 30

$0 / *+!+*+!F+ *#F-*/*##,*!' !#!3H&24+'*,!'!' */#!+/##*)* #,*+'*!*)*(#+!/* 6!**,*(!+*)39G"#G4 !" !#(),.0* ,#!#(),.0* !#(,"! ),.0* 2C>2?@0B 0?>0?=%B 2?>2A@=B !*3 )8 4 2?>2A@=B 0C>0C$2B 2A>2@@@B !* 3 ##!* 4 2$>2A=@B 0@>0@=?B 22>2&@&B !! # %! $ -2 ,!45531 /##*# ##7-,*)!'#) 7,*!'#+'*!*)*(# /*(-/#!'-"*(##+!*,*)!* *## ##*#"-+!*# #'*7 #C?#-+!" 1 + !-)!#,*!'!')* #! #!*+'*!*)*(# +!/*6!**,*(!+*)39"#G4 !" # $ %% &'( )& n+",#(! )& .& !*3 )8 4 00@&= %%% %?$&& 0%% !* 3##!* 4 002$% %$@ %CA$& 00C 0$A?= 2A% %%% @$ @=A%

PAGE 31

$$ 2 *+!+*+!F+ *#F-*/*##,*!' !#!3H&24+'*,!'!' */#!+/##*)* #,*+'*!*)*(#+!/* 6!**,*(!+*)39"#G4 !" !#(),.0* ,#!#(),.0* !#(,"! ),.0* AC>A?=0B =C>=CA&B @&>2%A0B !*3 )8 4 A&>@%C0B $@>$@??B @=>@CC&B !*3 ##!* 4 A2>A&=&B =%>=%A0B @=>@CC&B !! # %! $ */#!+/##*#7+!,*!+'*!*)*(# +!/*6!**,*(!+*)#!' #))#,*!'+'*!*)*(#+#7!'9 99> +*)-#/!'!'+!/*#'+ !-)!##'*7 A *+!+*+!F+ *#F-*/*##+*(!#,* +')* #'*7 @ 2+'+!6#-,*)+*,!*3)84*(!+*)#'*7#!'*" !7!'#)#(#3 + !-)!#,*!'!')* #! #!*!+'*!*)*(# +!/*6!**,*(!+*) !" # $ %%&'( )& n+",#(! )& .& 02?0$ %%% %A&A@ 0%% !*3 )8 4 02@@? $A= %02@% =@= !* 3##!* 4 022@% =?2 %00A= ?&&

PAGE 32

$= 4 *+!+*+!F+ *#F-*/*##,*!' !#!3H&24+'*,!'!' */#!+/##*)* #!'+*)/## *,*!+'*!*)*(#+!/*6!**, *(!+*) !" !#(),.0* ,#!#(),.0* !#(,"! ),.0* ?C 3 ??0%B 4 C?3 C?&$B 4 @%3 @0C=B 4 !*3 )8 4 ?@3 ?2 02B 4 C03 C02CB 4 A@3 A2=@ B 4 !*3 ##!* 4 ?=3 ?C%2B 4 CA3 CA&CB 4 @%3 @0C=B 4 N!*3 )8 4 ?A3?@0CB4 C$3C$2AB4 A?3 AA=$B4 5 r:*7*,+*+!-*/*##*(!+*)*,!* 3)847'+' +##7+*+! -+!*!')#(#3*" -4 6!#(r / ),7* 6!#(r ),7/3* 6!#(r ),71* !( (!) 89 00 3?$CB4 $C 3AA@B4 $$ 3?=@B4 ?@ 00 3?$CB4 0? 3C0@B4 $2 3A2=B4 ?C r+: ? 3$=2B4 0% 3$@2B4 0C 3=C0B4 ; r:*7*,+*+!-*/*##*(!+*)*,!* 3)84!*/!' *7'+'+##7+*+! -+! *!')#(#3*" -4 6!#(r/ ),7* 6!#(r ),7/3* 6!#(r ),71* !( <(! )89* 0=3A0&B4 $03?2=B4 $$3?=@B4 ?A 00 3?$CB4 0? 3C0@B4 $2 3A2=B4 ?C r+: 003?$CB4 0$3===B4 02 3C=&B4

PAGE 33

$C ,!#r !-.671+("#,*!'F+*"F/*(-/!'F, (F/*(-/#'*7/(# =C#-+!" '#,*!'F+*"F /*(-/7<3##!* 4H %22023)8)()-##(4H%222AH% ?@?%*!'F, (F/*(-/!' #7<3##!* 4H%2==23)8 )()-##(4H%2$@&H%2?%% / +("#,*F+*"F/*(-/*,!#(#!3 #+:!! +!'!6!*H ?C43##!* !*,!H%22024 3)8)()-##(!*-/'!H %222A43*!!*) ,!H%?@?%4 0 0.5 1 0 0.5 1 1-SpecificitySensitivityROC Curve: Ratio (end-systole) AUC = 0.8818 0 0.5 1 0 0.5 1 1-SpecificitySensitivityROC Curve: Ratio (maximum pressure) AUC = 0.8886 0 0.5 1 0 0.5 1 1-SpecificitySensitivityROC Curve: PVRI AUC = 0.5750

PAGE 34

$? 1 +("#,*F, (F/*(-/*,!#(#!3# +:!! +!'!6!*H?C4 3##!* !*,!H%2==243)8 )()-##(!*-/'!H%2$@&4 3*!!*) ,!H%2?%%4 !! !#r !-.671*/#!+/##*#7-,*),*!'-! !#!!'#))#!' -"*(#+'*!*)*(# ##7!'/*(-/#*,*( !+*)#!,* *7(-'+ ## ,*!'F+*"F/*(-/F, (F/*(-/ #'*7/(#&0%#-+!" = + !-)!#,*!'!')* #! #!*!'F+*"F +'*!*)*(#+!/*6!**,*(!+*)39"# G4 !" # $ %%&'( )& n+",#(! )& .& !*3 )8 4 =??% %%% %&%?? 0%% !* 3##!* 4 C%%$ C?$ %%&C? &?2 ??%? 0&?? ?0C2 ? 0@AC 0 0.5 1 0 0.5 1 1-SpecificitySensitivityROC Curve: Ratio (end-systole) AUC = 0.8338 0 0.5 1 0 0.5 1 1-SpecificitySensitivityROC Curve: Ratio (maximum pressure) AUC = 0.8279 0 0.5 1 0 0.5 1 1-SpecificitySensitivityROC Curve: PVRI AUC = 0.8500

PAGE 35

$A + !-)!#,*!'!')* #! #!*!'F, (F +'*!*)*(#+!/*6!**,*(!+*)39G" #G4 !" # $ %%&'( )& n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r+(#!'

PAGE 36

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

PAGE 37

$2 /##*7/, +*)7!'!' !*+*+! -+!/,*()*+##!' (!/!!/!7*)*+##!'!*7 !'!'+ *#+ #+ ( A*,,# ,(!'#/'!!*!' !"-,*)+*,!' !7*)#(###*)7'!!!! -+!/07' +*(/!*#!!! -+!/$!'.( = *7"!'# !" !! *" -!7 !'#-+!*#3?%B* ##47'+' #(//#!#!'!!'+*(/!*-+!##*)#+!*,,(+!*!'!*#*! *,!'//#!##(#-*/*#!+!/# !'" !*,!#-!'* */ -*/##*)*/"( #-!+( !' )/!*7'+'!' +*)-#!#,*!' "!-##(!-()-# /#!,,+!!'#)-!*)#-#!#7 #-!'* */+ )* /*,!'"#+( !(" !+ !'+*)-#!#!! '-!*-'*,!'"!+ +##!#+*!+! !!##!,,!*)!++ *(!-(!'#) +#')* )+#(#( )) ,,+! r+(#++*(!-(!#*) !'#-!!# : -*)! #'*7,7**#)-!*)# !'0/*(-'#)#!'!!') #(#!! *" -*,*#" (#!70$+(#/$!* ## #!/(#' !'*!' '+*(/!*)# /'! +#! '+*)-#!#!! !"!**) ,!# ,,+!! #*0$)!!,, !!*)-*!! $-#!# !'*#!*,*!+ #)-!*)#!',* # *-#!#!'!#!*,*)!' +*)-#!!*7!', /#!!!'### !')-*")!-+!*!' +*(/!**"*#",*$+*( !-!#*!' */+ -*)#/ 9'!'-!!I###7 +*!(*7*# #)*"! ,*)!*!*+ +!' 7'!'!#)-*"/(!')**!'!'+*( /!*+*+! -+!00 *,!'=+##(!7!'*" -*,* ," '#)#!'!!'+*(/!*# !+!/#*)#-+!*,#,(+!*!'! #I!(!')*/##**,+*(/

PAGE 38

$& !*!*/"$" (*,%%27'+'#(//#!#, 7:,! '#,(!' +***!#!'##!*!'!!'+*(/!* #/"/(#,*)!*!'!#*! '###!*/ +***!!'#( !#*,!' F+*"F#!(*!'-! !#!r#* !/-!!#7!'9* /!!! +!'!6!*7+*( !)!'#!/!'*,!-+! /7'+'-!!#7*( )-*"! ,* *7(-'+("#,*7.(!#!* /7' !'*,*7#!'#+# !!!'*)# +!*'+ ## #*#'*7!'!)* !' !)(+'!!!''#)#!'!-! !#7'*)-*"!7!'! +!'!6!*,* *7(-7#*,, !!'!'*#!'!#'*7* )-*")!7'#7##/,+! '/' )-*"-!!#!##* !* #-+( !!'!!'##+(##, +!* *,"!+( -,*)+7'+'#-) ,+!*-!!I#+-+!,*+*"* )-*!! !'##( !#'/' /'!!' -*!! -*/*#!+ !*,####/ ,(+!*'#( !#,*)!'F, (F /*(-/7#) !*!'*!'+'*!*)*(#/*( -/#7!'-,*)/# /'! !! +("#+'##!'#(-#/ *#+*(//+(#*,!'#+:-!!# (#!'# ##* *+!( 7*# ):/!'# ## !'#)#!' #+*+'*!*)*(#/*(-/,*!',( !#!'()*, )!!*###*+!7!' !'#)* r+(##!* *, #!+#!+* -*"+!)# ()!*,"!+( -,*)+#! !#!*!'#!!*,!'"#+( !('(#)!#"* "/+!)#()!*, "* ()#(+'#7 : 7#) !!+!*#*,,(+!* (!')* !'*(/'!!')#()!*,-##( #)(+'#)!'!')#()! *,-##("* ()!*/!'#! .( #+!'!6!*)/!*,,#* / "!/!)#*,"#"##! #*#(,, #,*)'"/, #) /+*)-

PAGE 39

=% !*!#"+)*/"( # !'*(/'!'## "!+*#/! */#*!' )#(##(+'#!* !'*(/'! '##'*7'!*-*/*#!+!'# )!'*,**!/.(#" !*' ##'*( ++*)#'+*)-/!' #!)!/#!*!,*) **-# *!') #'()#

PAGE 40

=0 r 8rr) '/(/!!!+(#-" "#, +*))* -!!#7!'-#!@%B* )*+##7!'#/,+! '/'!-( )* !-##(3D*+:*--0&2C4 #( !#+:, *7:*7#5!'/# !'/*)!*,!'"!+ (!* '-!*-'* !*#7 #!'8+##-# #(-#!!'"!+ *,! +(##!'!+(#-" "!*,(+!*)-**!+*)! + *#/'" *+!*,!' 5!+)#( !)*"#" (#/+*!(*(#7"*-( !#*( 0&2CD*+:*---( #')!'*,**! /#!)!*,-:"!+( ##!* +-##((#/#),,*)*,r *( I#-+!#/ ,*)r*( I#-+#+ #!'-##(/!+*## *#!(+!*3!'#+#!'!+(#-" "4# /"# -A+-(A B C +CA B DE& F F G EH GIC F G A3C4 7'!'!)#*!'/'!#*,!'.(!* #++*"+!"++ !*, *7 ++ !*"#+*(#,+!*#-+!" !'##+*!', *7++ !* "#+*(#,+!*!)#/ / +*)-!*! '+*"+!"++ !*!)0# #) +*)-!*$#*7*!)*,,*)*,r*( I#-+ # 8-J C? 7'#!')8) *+!*,!'5! !'+*#!!, +!#!'#!*, ** #!,+!*#!*!'.(!*:"!+( ##!* +-##(#!'*!/ /'!! -##(#!)!+ + (#/5( /( "-##( r(#/!'##),,*)*,r*( I#-+ 7+!:!'+*)!" *+! *-,*)!'5!*!+*!(*(###(!+ *7/!')*, #/ !)!'*!*(#*"#" ##+!#*!"+##!*!#

PAGE 41

=$ ,*)" *+!!*-##(!'#+#+(# #+ +( !##)!*8+-!!'! !',!*(#,*!'*#!*,#*"* ()++* !+!*"* "#!',#!"!"*, -##(*7"!'#,!*+*( -(! !)#*," *+!(#*,!'#), ,*)*,r*( I#-+#+(##*" ""'! *-( !#*(!7#"#( 6-*+## *#(+!*7# -(#/!'#+!7" !!#,*))!'* #+P'/! $%%0' 7" !(#7#!'2!'*#) !7" !7!' +*##!#* (!**,#8'!'#'* (7#+*)-(!(!*)!+ #*!'*# #*,!!'#'* /7#-!* !'! +*,,+!#!')/7#+*#! (+! +#*,5!" *+!7-,*)(! *)!+/!+!**(!(#/ 5(#! !'#'* )/!(r+(#*-( !#*(!+*!##-+! *+!!*"!)/!+!*7#-,*) ,*+'+* ()*,!):/!' ,#!-8 !*8+!'!'#'* (,*)!' !*-*,!')/3)8) /!"" *+!4 '#( !/!+7##)**!'(#/+(+## '#)**!'" *+!!+7#!#,*)!* --*8)!/'!"!+( -##(#),+!**,r*( I#-+# -(C'#!7##(#.(! 6(#/!'!+'.(##++!* $0*""/ -,*) !!!'+!'!6!*+##!')8) ##(7#(#+*)-(!/!'+*(/ !*!'!'!'##!* +,!**,= $)#-*!*>!) '-)##()-!*-*!! )!!**, !'#)!'*#!'!!'" *+!*, !'5!+*#-*#!*"!+( -##(,* !'!##!* +" *-7'+')*! !'+# : -##*)7'!*"( -'#* */#/5!" *+!# )#*,#!)!/"!+( -##('#" !* ,*-:##!* +-##(*!

PAGE 42

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

PAGE 43

=C 2 (!*)!+!+*,*-)/*,5! 7!+;r (#)**!'!+ 3 ##(!++(",!!/*,!'-( #,* )*-)//(?*"

PAGE 44

=? 4 *-)/ /-**-##(!+,*+(" ,!!/ *(/!*#,**-)/#7!'++-! -##(!+# (,(>)6!?:* ,(!(,"">+ (!, r24@313=@2/2 ==3@122=@12 ==@12=@=;3 ==/1@1=@24 ==/@15==@;4; =1@3;=@4 =//@;1=@=3 =1@1;;= =2/@1==@= =@111=@5/ =/2@3=@3/

PAGE 45

=A !!! 2 '/* *,!'#-*!**,!'#!(7#!*-!')*,#/ !)!'* *-!'##!(!**-( !#*()/ #*,!+(#-" "/(/!!*5!# (#/#),+!**,r*( I#-+*!**!--*8)!-##(*"!) !'/'!"!+ ,*)!'" *+!*,!'5! !'*(/'!'#)(+'7*:!**!* !)-*"!'#)!'*!#'*7#/!*)#!',*)-#!' !'*!'#)!'*+-*"!'#),*)!* #!')*,#/ !)!'*(# *+!'!-##(!+#!'!,*)!')-* !**,!'##!(!!'#!'*)*(# "!/#*,/+*)! *"#"+'.(+:#!*-,*)'(# ",!-*"#!*#*)7'! ##++(!!' +!'!6!*!+*( #! -*"!* 8!) -*7,( !** ,*+ +#*,!',#!!'/#!'!+*!*)-*" !')!'*#!*))!!** ,*)( /,!*'#7#-,*) #)/*!'),*!'-(-*##*,!'# --7'+'#( !#*))/#7!'-*!! -##(,*)!*#(+'#!' *#'*7/(@!*)-*## !**! -*-#*"* ()+#(#* ,!(!* (7!#/ #'##/ #+*( #)/ *(#/)( !+7'+'+*( +#!'()*,(# )/##(#!! 8!!'#)!'*#!*" !#!! '!*#*!7!'*-( !#*(#'*( +*)-7!'!*#*! ,*)!'#)-!!# +!'!6!*,!'##* /**/ )!!7!'"#"*"#" )#()!#!7 /* */7!*7#(--*! /!'(#*,*-( !#*(!' ####)!*,,(+!*(!')*!'#'*( #!!#!+ ##*,!'!*# *!,*)!'#)!'*7!'9!'#)) #!'+!'!6!*)!'* !'#--/!#'*( *!8-+!!'! !'#)!'*7 #++(!#

PAGE 46

=@ +!'!6!*(!,!-*"#!*#* /**!-+!/+ + *(!+*)!+*( -*"8!) (#,( #)-*/*#!+ !** *!'!'/!'!+*( +*#,*)#!! #!+ #!-*!7*( !' + + *(!+*)394*,-!!#-#!/7 !'5!+*)-7!'!'*#7'*'"* !+(#-/(/!!*!7*( 8-+!!'! -!!#)*"+#!!*, #,(+!*7*( )* : !*-#!7!' !+(#-/(/!!*'#!** +*( #*(#!*####!'#"!*,!+(#/(/!!*+*)-7!'9

PAGE 47

=2 r nn '#/ !)!'** */,*#!)!/!'" !+( +*(/!* 8)!'#--'#()*,"!/#* "*!'+*))* (#')*)+ )#(##'*7#-*)#,*+ + (#!' ,(!('-#)*#!)-*!! !' +!#!'#!!*,*!'! +*)+ "!+( +*!+! !!*+! +!#!'#!!(#*,!'-( )*+*"#+( ++(!#(!7'#)#(#(+'# +!#-*-!#*,!'"#+( !( '(# !'*(/'!7#,*(!'!) -+!+ + *(!+*)#7 *# /'! !! !'+*(/!*+* +! '!, (+! #"#+( ##!+# # (!')*!'#)!'*'#!'"!/*" *!'+(!)!'*#*, )#(/"!+( -*-!#!'!*"* ()) #()!#+##'#):#! -!+( # !'/'!"!+ 7' !'#) #6("#'-): "* ()!+)#()!-* )!+! #*'#!' -*!! !*+*)!*! *"#" 7!'!'(#*,*-( !#*((! 6/!' /(/!!!+(#-" "" *+!'## !" 8!#**,!'+(!)!'**, #!)!/-:##!* +-##((#/!' -:" *+!*,!'5!+*( -*"8!) (#,( -/*!* !* + #*(#+*5(+!*7!'*!'' )*)+)#(#!*#!/!' -*/*###7##'*7!'#--(#/ r+(#!')#(-! +!# "!+( -*-!#!+*!#(.(,*)! **!" ,*)*!')#(# 7!'*(!"* ())#()!#'# #*)#!+ +*)-7!'#!(/!' ,,+!#*,!!)!#*!*,,!!! 7,,+!#*!'"#+( !(!' )*+()3#)*#!!!)!#(#,*:*7 !*,,+!*!'#*)74 *)-'#"+ + #!*!'#+*( !#!/,*)!**!'*(!!'

PAGE 48

=& -!+( + + (#,( ##*,!'*( !!'#-+,+,,+!#*,#*)*,!' !!)!#+*))* (#,* /7!'*! '(#,( + + !'!+*( -,*)#!)/7'!'!'+*(/!* +++(! -+!!#! !' !'*(/'!'#7#-,*)!*#) 8! !!'#--!'#) ()*, !#!#!'#!'-!+-!!#+* )-#/!'##!( )!!#(#,( ## !'*(/'7#,*(!*# /'! !!+!**,, (!7#+ )*#!!!'!##(-*!*+!/+*"#+:-!!#7'+''* # ))#+ + (:7!'!# !"# )+!+*)-7!'*!')#*, ####/,(+!*!'#,+!*#):#! )!!')*,#/ !)!'* -#!!'#--,,+!"-*/*#!+!** ,*!'+ + '8!#!-)#" !*'#)!'*+ !,**!'!' ,! /'!"!+ +*)-/!'#!)!7! '*!,*) **-#,#! ) )* # : !'!+*7!''() #r+(#"* ())#()!# +*(+!++!'!#-* )!+!' "* ())#()!##'*( *!#7 '*-( !#*(5!) !'*#'*( (/*#) !* )*+*)-'#"+ + ##'*( -,*)"* "/)*! !#!!#( !-!!#!* ,* *7(-#7!'-!!#

PAGE 49

C% nAbel, L. (1981). Maximal negative of end of systole as an indicator. Society Antonio, S. (1985). Defining End Systole for End-Sy stolic. October 350 344-350. Baan, Jan, Tjong T Aouw Jong, Peter L M Kerkhof, Ru dolf J Moene, Arjan D Van Dijk, E. T. V. D. V. and J. K. (1981). Continuous stroke vol ume and cardiac output from intra-ventricular dimensions obtained with impedanc e catheter. Cardiovascular Research 15 (6), 328-334. Retrieved from http://cardiovascres.oxfordjournals.org/content/15/ 6/328.short Badesch, D. B., Champion, H. C., Sanchez, M. A. G., Hoeper, M. M., Loyd, J. E., Manes, A., ... & Torbicki, A. (2009). Diagnosis and assess ment of pulmonary arterial hypertension. Journal of the American College of Cardiology 54 (1s1), S55-S66. Bernus, A., Wagner, B. D., Accurso, F., Doran, A., Kaess, H., & Ivy, D. D. (2009). Brain natriuretic peptide levels in managing pediatric pa tients with pulmonary arterial hypertension. Chest 135 (3), 745-51. doi:10.1378/chest.08-0187 Brimioulle, S., January, F., Tabima, D. M., Hacker, T. A., Chesler, N. C., Wauthy, P., Ewalenko, P., et al. (2003). Single-beat estimation of right ventricular end-systolic pressure-volume relationship Single-beat estimation of right ventricular end-systolic pressure-volume relationship. American Journal Of Physiology (January 2003). doi:10.1152/ajpheart.01023.2002 Brown, K. a., & Ditchey, R. V. (1988). Human right ventricular end-systolic pressure-volume relation defined by maximal elastan ce. Circulation 78 (1), 81-91. doi:10.1161/01.CIR.78.1.81

PAGE 50

C0 Burkhoff, D., Mirsky, I., & Suga, H. (2005). Assess ment of systolic and diastolic ventricular properties via pressure-volume analysis : a guide for clinical, translational, and basic researchers. American journal of physiology. Heart and circulato ry physiology 289 (2), H501-12. doi:10.1152/ajpheart.00138.2005 Chen, C. H., Fetics, B., Nevo, E., Rochitte, C. E., Chiou, K. R., Ding, P. a, Kawaguchi, M., et al. (2001). Noninvasive single-beat determinatio n of left ventricular end-systolic elastance in humans. Journal of the American College of Cardiology 38 (7), 2028-34. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1 1738311 Desk, R., Williams, L., & Health, K. (1973). Ventri cle and Effects of Epinephrine and Heart Rate on the Ratio Load Independence of the In stantaneous Pressure-Volume Ratio of the Canine Left Ventricle and Effects of E pinephrine and Heart Rate on the Ratio. Circulation Research Eysmann SB, Palevsky HI, Reichek N, Hackney K, Doug las PS. Two-dimensional and Doppler-echocardiographic and cardiac catheterizati on correlates of survival in primary pulmonary hypertension. Circulation 1989;80:353–360. Gali, N., Hoeper, M. M., Humbert, M., Torbicki, A. Vachiery, J.-L., Barbera, J. A., Beghetti, M., et al. (2009). Guidelines for the dia gnosis 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 European heart journal 30 (20), 2493-537. doi:10.1093/eurheartj/ehp297

PAGE 51

C$ Kelly, R. P., Ting, C.-tai, Yang, T.-ming, Liu, C.peng, Maughan, W. L., Chang, M.-song, & Kass, D. A. (1992). Effective Arterial Elastance as Index of Arterial Vascular Load in Humans. Circulation 513-521. doi:10.1161/01.CIR.86.2.513 Kuehne, T., Yilmaz, S., Steendijk, P., Moore, P., G roenink, M., Saaed, M., Weber, O., et al. (2004). Magnetic resonance imaging analysis of right ventricular pressure-volume loops: in vivo validation and clinical application in patients with pulmonary hypertension. Circulation 110 (14), 2010-6. doi:10.1161/01.CIR.0000143138.02493.DD Lam, C. S. P., Roger, V. L., Rodeheffer, R. J., Bur si, F., Borlaug, B. a, Ommen, S. R., Kass, D. a, et al. (2007). Cardiac structure and ventricu lar-vascular function in persons with heart failure and preserved ejection fraction from Olmsted County, Minnesota. Circulation 115 (15), 1982-90. doi:10.1161/CIRCULATIONAHA.106.65976 3 Mertens, L. L. & Friedberg, M. K. (2010). Imaging t he right ventricle current state of the art. Nat. Rev. Cardiol. 7, 551–563; published onli ne 10 August 2010; doi:10.1038/nrcardio.2010.118 Naeije, R., & Torbicki, A. (1995). More on the non invasive diagnosis of pulmonary hypertension: Doppler echocardiography revisited. European Respiratory Journal 1445-1449. doi:10.1183/09031936.95.08091445 Nichols, W. W., & Edwards, D. G. (2001). Arterial E lastance and Wave Reflection Augmentation of Systolic Blood Pressure: Deleteriou s Effects and Implications for Therapy. Journal of Cardiovascular Pharmacology and Therapeu tics 6 (1), 5-21. doi:10.1177/107424840100600102

PAGE 52

C= Raymond RJ, Hinderliter AL, Willis PW, Ralph D, Cal dwell EJ, Williams W, Ettinger NA, Hill NS, Summer WR, de Boisblanc B, Schwartz T, Koc h G, Clayton LM, Jobsis MM, Crow JW, Long W. Echocardiographic predictors o f adverse outcomes in primary pulmonary hypertension. J Am Coll Cardiol 2002; 39:1214–1219. Suga, H., & Sagawa, K. (1974). Instantaneous pressu re-volume relationships and their ratio in the excised, supported canine left ventric le. Circulation research 35 (1), 117-26. Retrieved from http://www.ncbi.nlm.nih.gov/ pubmed/4841253 Sunagawa, K., Maughan, W. L., Burkhoff, D., & Sagaw a, K. (1983). Left ventricular interaction with arterial load studied in isolated canine ventricle. The American journal of physiology 245 (5 Pt 1), H773-80. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/6638199 Sunagawa, K., Maughan, W. L., & Sagawa, K. (1985). Optimal arterial resistance for the maximal stroke work studied in isolated canine left ventricle. Circulation research 56 (4), 586-95. Retrieved from http://www.ncbi.nlm.nih .gov/pubmed/3978773 Takeuchi, M., Igarashi, Y., Tomimoto, S., Odake, M. Hayashi, T., Tsukamoto, T., Hata, K., et al. (1991). Single-beat estimation of the sl ope of the end-systolic pressure-volume relation in the human left ventricl e. Circulation 83 (1), 202-12. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1 898642 Yock, P. G., & Popp, R. L. (1984). Noninvasive esti mation of right ventricular systolic pressure by Doppler ultrasound in patients with tri cuspid regurgitation. Circulation 70 (4), 657-662. doi:10.1161/01.CIR.70.4.657 Zhang, Y., Wang, Y., Wang, W., & Liu, B. (2001). Ba sed on Wavelet Frames, 48 (3).