Prognostic Value of Pulmonary Transit Time and Pulmonary Blood Volume Estimation Using Myocardial Perfusion CMR

Objectives The purpose of this study was to explore the prognostic significance of PTT and PBVi using an automated, inline method of estimation using CMR. Background Pulmonary transit time (PTT) and pulmonary blood volume index (PBVi) (the product of PTT and cardiac index), are quantitative biomarkers of cardiopulmonary status. The development of cardiovascular magnetic resonance (CMR) quantitative perfusion mapping permits their automated derivation, facilitating clinical adoption. Methods In this retrospective 2-center study of patients referred for clinical myocardial perfusion assessment using CMR, analysis of right and left ventricular cavity arterial input function curves from first pass perfusion was performed automatically (incorporating artificial intelligence techniques), allowing estimation of PTT and subsequent derivation of PBVi. Association with major adverse cardiovascular events (MACE) and all-cause mortality were evaluated using Cox proportional hazard models, after adjusting for comorbidities and CMR parameters. Results A total of 985 patients (67% men, median age 62 years [interquartile range (IQR): 52 to 71 years]) were included, with median left ventricular ejection fraction (LVEF) of 62% (IQR: 54% to 69%). PTT increased with age, male sex, atrial fibrillation, and left atrial area, and reduced with LVEF, heart rate, diabetes, and hypertension (model r2 = 0.57). Over a median follow-up period of 28.6 months (IQR: 22.6 to 35.7 months), MACE occurred in 61 (6.2%) patients. After adjusting for prognostic factors, both PTT and PBVi independently predicted MACE, but not all-cause mortality. There was no association between cardiac index and MACE. For every 1 × SD (2.39-s) increase in PTT, the adjusted hazard ratio for MACE was 1.43 (95% confidence interval [CI]: 1.10 to 1.85; p = 0.007). The adjusted hazard ratio for 1 × SD (118 ml/m2) increase in PBVi was 1.42 (95% CI: 1.13 to 1.78; p = 0.002). Conclusions Pulmonary transit time (and its derived parameter pulmonary blood volume index), measured automatically without user interaction as part of CMR perfusion mapping, independently predicted adverse cardiovascular outcomes. These biomarkers may offer additional insights into cardiopulmonary function beyond conventional predictors including ejection fraction.

T he pulmonary circulation is inextricably linked with cardiac physiology, but our understanding of the cardiopulmonary axis in various disease states is limited. Use of noninvasive imaging biomarkers as surrogate indicators of cardiopulmonary status may facilitate risk stratification and outcome prediction, potentially contributing to personalized clinical care.
Pulmonary transit time (PTT) and pulmonary blood volume (PBV) are physiological parameters reflective of cardiopulmonary hemodynamics (1). Both are known to be altered in various disease states, including heart failure (2,3), pulmonary hypertension (4,5), and chronic lung disease (6), and to correlate with structural, functional, and biochemical parameters of pulmonary (7) and cardiac function (8). Pulmonary transit time, defined as the time interval for a contrast bolus to pass from the right-to left-sided circulation, and PBV (the product of PTT and cardiac output), correlate with established prognostic biomarkers, including right ventricular (RV) and left ventricular (LV) ejection fraction (9), markers of LV diastolic function (10), brain natriuretic peptide levels (9), and pulmonary vascular resistance (4). Importantly, a small number of studies suggested an independent prognostic utility of PTT and PBV in specific disease models (2,5,11).
Despite extensive research supporting a clinical utility of PTT and PBV, at-scale analysis and clinical adoption have been hindered by challenges in data acquisition, requiring either invasive catheterization (1) or manual segmentation and data extraction from noninvasive tests (2,(5)(6)(7)(8)(9)(10). Recent  been used to explore the prognostic effect of myocardial blood flow (MBF) and myocardial perfusion reserve (MPR), and has been previously described (12). In brief, consecutive adult patients referred for a myocardial perfusion scan were included. Patients with congenital heart disease, known intracardiac shunts (known to affect methods based on the indicator dilution principles [13]), and inherited or infiltrative cardiomyopathies (hypertrophic cardiomyopathy and cardiac amyloid) were excluded.
The primary outcome was the incidence of major adverse nonfatal cardiovascular events (defined as myocardial infarction, stroke, heart failure admis-    Stress myocardial perfusion was performed, using adenosine as pharmacological stressor according to guidelines (14). The myocardial perfusion sequence is a single-bolus, dual sequence described previously (15). Basal, midventricular, and apical short-axis perfusion images were acquired at both stress and rest. Image acquisition was performed over 60 to 90 heartbeats and a bolus of 0.05 mmol/kg gadoterate meglumine (Dotarem, Guerbet, Paris, France) was administered at 4 ml/s during both maximal hyperemia and subsequently at rest (for estimation of stress and rest MBF respectively). MPR was defined as the ratio of stress MBF over rest MBF. PTT data was calculated from perfusion imaging, and PBV was estimated utilizing resting cardiac output measurement from cardiac volumes obtained from short-axis stack cine images.  Figure 1). The use of centroids of the AIF curves was previously shown to be superior to peakto-peak methods for PBV estimation (19). Pulmonary transit time normalized for heart rate (PTTn) was estimated by dividing PTT with the duration of the cardiac cycle (R-R interval, in seconds) as performed in previous studies (8,9): Pulmonary blood volume was estimated as the product of PTT and cardiac output as originally described from indicator dilution methods (20): This was indexed to body surface area (BSA), allowing calculation of pulmonary blood volume index (PBVi): The LV stroke volume was estimated using steady state free precession (SSFP) cine images from manual planimetry of a full short-axis stack in end-diastole and -systole, and the patient's heart rate at rest was used to derive cardiac output (cardiac output ¼ stroke volume Â heart rate). Rest PTT was used for the primary analysis, including estimation of PBV as cardiac output data was only available during rest. Associations between stress PTT and outcomes were performed as a secondary exploratory analysis.  Prognostic Value of Automated CMR-Derived PTT  Log-rank p = 0.021 Event-free survival curves for major adverse cardiovascular events (heart failure hospitalization, myocardial infarction, stroke and ventricular tachycardia/implantable cardioverter-defibrillator treatment) according to mean PTT (8.05 s) (A) and mean PBVi (414 ml/m 2 ) (B). Longer PTT and higher PBVi were associated with higher rates of major adverse cardiovascular events (log-rank p ¼ 0.043 and p ¼ 0.021, respectively).
Abbreviations as in Figure 1.

RESULTS
COHORT DESCRIPTION AND BASELINE CHARAC-TERISTICS. A total of 1,049 patients with CMR myocardial perfusion imaging data were available for inclusion as previously described (12). Of these, 4 (0.4%) had confirmed intracardiac shunts and were therefore excluded, in addition to 60 (5.7%) patients with incomplete or erroneous rest perfusion data (including incorrect automated blood pool identification, incorrect timing of contrast administration, and poor AIF signal of either the RV or LV). A total of 985 patients with available rest PTT data were therefore included in the main analysis.
Median age of the patients was 62 years (IQR: 52 to 71 years) and 660 (67%) were men. There were 281 (28.6%) patients with diabetes mellitus, and 306 (31%) patients had a prior history of either PCI or CABG. The median LVEF across the cohort was 62% (IQR: 54% to 69%). Baseline characteristics are summarized in Table 1.   Table 1.  PBVi are presented in Figure 3.
All-cause mortality data was available over a me- PTT was also extracted during adenosine stress first Values are median (interquartile range) or n (%).
BSA ¼ body surface area; MACE ¼ major adverse cardiac events (myocardial infarction, stroke, heart failure admission, and ventricular tachycardia or appropriate implantable cardioverter-defibrillator treatment [including implantable cardioverter-defibrillator shock and/or antitachycardia pacing]); PBVi ¼ pulmonary blood volume index; PTTn ¼ pulmonary transit time normalized for heart rate; TIA ¼ transient ischemic attack; other abbreviations as in Table 1. pass perfusion, and an exploratory analysis between stress PTT and outcomes was performed. A total of 963 cases with stress PTT data were available for analysis, following exclusion of cases with incomplete or erroneous stress perfusion data. As expected, median stress PTT was shorter than rest PTT  Table  S5, Supplemental Figure S4).

DISCUSSION
This study investigated the prognostic power of pul- Very few studies previously investigated the association of PTT parameters and outcomes, and these studies were focused on specific disease entities.  (2,3,9) and to be associated with markers of diastolic function in patients with hypertrophic cardiomyopathy (10). A number of studies investigating the relation between PTT and cardiac volumes or biomarkers used a normalized PTT by adjusting for heart rate. The method of correction of PTT varied between studies (10,23), but given the association between heart rate and PTT also shown in our data ( Figure 2, Table 2), we performed a further analysis using PTT normalized for heart rate (PTTn). PTTn was also predictive of MACE, and similarly to PTT and PBVi, was not predictive of all-cause mortality (Supplemental Table 4, Supplemental Figure S2). As the estimation of PBVi incorporates the use of cardiac output at rest, the impact of resting heart rate is incorporated in this metric.
Stress PTT extracted during adenosine stress perfusion was also found to be independently asso-  showing event-free survival for major adverse cardiovascular events.
varied over a much larger dynamic range than cardiac output. Furthermore, the study was designed primarily to assess the prognostic value of biomarkers (PTT and PBVi) that could be automatically derived from CMR sequences obtained as part of routine clinical imaging protocols. Although all first-pass perfusion studies rely on the indicator dilution principles, there are important variations between different methods of PTT estimation. Different sampling locations have been described, including the pulmonary trunk to left atrium (19), the RV to the left atrium, as well as the RV to LV (8,9,23). Evidently, the estimation of PTT and PBVi will vary depending on the anatomic landmarks selected. In our study, the RV and LV cavities were used for sampling as these can easily be sampled during the perfusion sequence, eliminating the need for additional planning and image acquisition. Patients had been clinically referred for myocardial perfusion CMR, and therefore the cohort predominantly included patients with known or suspected coronary artery disease. This may have introduced bias in terms of the association of PTT metrics. However, our analysis was adjusted for a number of cardiovascular risk factors as well as myocardial perfusion reserve, previously shown to independently predict adverse events within this patient cohort (12).

CONCLUSIONS
PTT and PBV, measures of the cardiopulmonary system, can now be derived automatically without user input from latest-generation CMR perfusion mapping studies. Here, we show that these metrics are independently associated with adverse cardiovascular events over and above conventional factors, potentially providing clinically feasible imaging biomarkers of cardiopulmonary physiology.