Automated CT-Based Quantification of Pulmonary Veins Shows Greater Central Venous Dilation in Group 2 Pulmonary Hypertension Compared With Group 1 Pulmonary Arterial Hypertension and Control Subjects

Pulmonary hypertension (PH) is a heterogeneous disease that includes pulmonary arterial hypertension (PAH; World Symposium PH group 1); however, PH can also result from left-sided heart disease (pulmonary venous hypertension PVH; World Symposium PH group 2).1Humbert M. Kovacs G. Hoeper M.M. et al.2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension.Eur Heart J. 2022; 43: 3618-3731Crossref PubMed Scopus (1081) Google Scholar Although both groups demonstrate hemodynamic abnormalities on invasive right heart catheterization (RHC), the management approach, particularly pulmonary vasodilator therapy, differs considerably for patients with PVH compared with PAH.1Humbert M. Kovacs G. Hoeper M.M. et al.2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension.Eur Heart J. 2022; 43: 3618-3731Crossref PubMed Scopus (1081) Google Scholar Given that PVH is the most prevalent form of PH and is increasing in incidence,2Thiwanka Wijeratne D. Lajkosz K. et al.Increasing incidence and prevalence of World Health Organization groups 1 to 4 pulmonary hypertension: a population-based cohort study in Ontario, Canada.Circ Cardiovasc Qual Outcomes. 2018; 11e003973PubMed Google Scholar noninvasive methods of differentiating PVH from PAH potentially may improve screening strategies for this large and growing population. CT imaging is increasingly available in groups at risk for PH. Biomarkers based on vascular volumes, which are generated by automated quantitative image processing algorithms, have demonstrated promise as noninvasive indicators of pulmonary vasculopathy.3Synn A.J. De Margerie-Mellon C. Jeong S.Y. et al.Vascular remodeling of the small pulmonary arteries and measures of vascular pruning on computed tomography.Pulm Circ. 2021; 11: 1-9Crossref Scopus (4) Google Scholar Our prior work demonstrated that individuals with PAH had lower volume of the small arteries (vascular “pruning”) and higher volumes of the large central arteries (vascular “engorgement”) on CT scans compared with control participants.4Rahaghi F.N. Nardelli P. Harder E. et al.Quantification of arterial and venous morphologic markers in pulmonary arterial hypertension using CT imaging.Chest. 2021; 160: 2220-2231Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar However, it is unknown whether CT scan-based pulmonary venous volumes differ between patients with PVH and PAH. In this study, we sought to quantify proximal venous engorgement, which is a phenomenon expected in the context of elevated pressures in the venous system, in patients with PVH. Our sample consisted of 113 patients from our institution who underwent RHC between 2011 and 2018 and had thin-sliced chest CT angiography 20 mm Hg, pulmonary artery wedge pressure (PAWP) 20 mm Hg with PAWP ≥ 15 mm Hg. Patients with normal resting/exercise hemodynamics and no cardiopulmonary disease were classified as control participants.4Rahaghi F.N. Nardelli P. Harder E. et al.Quantification of arterial and venous morphologic markers in pulmonary arterial hypertension using CT imaging.Chest. 2021; 160: 2220-2231Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar This study was approved by the Mass General Brigham institutional review board (#2018P000419). Segmentation, identification, and volumetric assessment of intraparenchymal pulmonary veins were performed using the Chest Imaging Platform (www.chestimagingplatform.org) as previously described.4Rahaghi F.N. Nardelli P. Harder E. et al.Quantification of arterial and venous morphologic markers in pulmonary arterial hypertension using CT imaging.Chest. 2021; 160: 2220-2231Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar The method of automated separation of the arteries from the veins has been validated against manual classification.5Nardelli P. Jimenez-Carretero D. Bermejo-Pelaez D. et al.Pulmonary artery-vein classification in CT images using deep learning.IEEE Trans Med Imaging. 2018; 37: 2428-2440Crossref PubMed Scopus (127) Google Scholar Small veins are those with cross-sectional area ≤ 5 mm2 and large veins > 5 mm2. Vascular volumes were normalized by radiographic lung volume. Mann-Whitney U tests were used to examine unadjusted differences in small/large venous volumes between PH groups (PAH, PVH, and control). Additionally, we used multivariable linear regression models to investigate the association of PH group and venous volumes with adjustment for age, sex, and smoking status. Because of the known association of systemic sclerosis and pulmonary venous abnormalities (which includes pulmonary veno-occlusive disease) in exploratory models, we performed sensitivity analyses that excluded those participants in the PAH group who had any form of scleroderma (n = 9). Thirty-four of 113 patients (30.1%) in our sample had PVH; 42 patients (37.2%) had PAH, and 37 patients (32.7%) were classified as control participants (Table 1). Median mPAP and pulmonary vascular resistance were higher in PVH and PAH groups vs control subjects; PAWP was higher in PVH patients compared with PAH/control groups. Most cases of PAH were idiopathic or connective tissue disease-associated; most PVH was from heart failure with preserved ejection fraction.Table 1Study Sample CharacteristicsCharacteristicControl Participants (n = 37)Pulmonary Arterial Hypertension: Group 1 (n = 42)Pulmonary Venous Hypertension: Group 2 (n = 34)Age, median (IQR), y53 (20)63.5 (22)69.5 (13)Female sex, No. (%)28 (75.7)35 (83.3)23 (67.7)Ever smoked, No. (%)13 (35.1)19 (45.2)23 (67.7)Mean pulmonary arterial pressure,aWorld Symposium group 1. median (IQR), mm Hg17 (6)42 (24)34.5 (14)Pulmonary artery wedge pressure, median (IQR), mm Hg10 (5)9 (5.5)20 (6)Pulmonary vascular resistance, median (IQR), Wood units1.2 (0.7)7.2 (4.4)2.4 (1.8)CI, median (IQR), L/min/m22.9 (0.7)2.6 (0.7)2.7 (1.1)Radiographic lung volume, median (IQR), L3.7 (1.5)3.1 (1.2)3.2 (1.4)Cause of pulmonary arterial hypertension,a No. (%) Idiopathic. . .20 (47.6). . . Connective tissue disease. . .16 (38.1). . . Congenital heart disease. . .2 (4.8). . . Familial. . .1 (2.4). . . Other. . .3 (7.2). . .Cause of pulmonary venous hypertension,bWorld Symposium group 2. No. (%) Heart failure with preserved ejection fraction. . .. . .20 (58.8) Heart failure with reduced ejection fraction (< 55%). . .. . .2 (5.9) Valvular. . .. . .12 (35.3)Presence of left atrial dilation on echocardiogram, No. (%)3 (8.1)16 (38.1)27 (79.4)Presence of atrial fibrillation, No. (%)0 (0)9 (21.4)17 (50.0)Any use of pulmonary vasodilatory therapy, No. (%)0 (0)19 (45.2)2 (5.9)IQR = interquartile range.a World Symposium group 1.b World Symposium group 2. Open table in a new tab IQR = interquartile range. Examples of vascular reconstructions from each group are shown in Figure 2A and B . In unadjusted comparisons, we found that the PVH group had higher large venous volumes (ie, greater venous engorgement) vs the control group (P = .007) and PAH group (P = .03); no difference was seen between the PAH and control groups (P = .88). A similar pattern was observed in multivariable models (Fig 2C). In contrast, small venous volume (ie, degree of venous pruning) was similar among the PAH/PVH group (P = .66); both of these groups had greater venous pruning when compared with the control participants (P < .0001 for both comparisons). In exploratory analyses, we found no change in the pattern of the primary results when we excluded participants with scleroderma from the PAH group. In this cohort of patients who underwent RHC and CT angiography, we found that patients with PVH and PAH both had lower small venous volumes but that the PVH group had higher large venous volumes compared with the PAH and control groups. Our findings suggest that, although loss of small vessel volume may be a shared feature across these PH subtypes, engorgement of the large central pulmonary veins is more specific to PVH, potentially because of elevated left-sided intracardiac filling pressures. Pathologic pulmonary arterial changes in PH have been the focus of considerable research; regardless of PH cause, well-described changes include vascular remodeling with intimal thickening, medial hypertrophy, and eventually loss of pulmonary arterioles.6Tuder R.M. Marecki J.C. Richter A. Fijalkowska I. Flores S. Pathology of pulmonary hypertension.Clin Chest Med. 2007; 28: 23-42Abstract Full Text Full Text PDF PubMed Scopus (276) Google Scholar,7Leopold J.A. Pulmonary venous remodeling in pulmonary hypertension the veins take center stage.Circulation. 2018; 137: 1811-1813Crossref PubMed Scopus (12) Google Scholar Pruning on CT scan is a promising biomarker of pulmonary arteriopathy and has been linked to right ventricular remodeling on cardiac MRI8Rahaghi F.N. Wells J.M. Come C.E. et al.Arterial and venous pulmonary vascular morphology and their relationship to findings in cardiac magnetic resonance imaging in smokers.J Comput Assist Tomogr. 2016; 40: 948-952Crossref PubMed Scopus (17) Google Scholar and poorer clinical outcomes in individuals with COPD,9Washko G.R. Nardelli P. Ash S.Y. et al.Arterial vascular pruning, right ventricular size and clinical outcomes in COPD.Am J Respir Crit Care Med. 2019; 200: 454-461Crossref PubMed Scopus (67) Google Scholar and it can help distinguish those patients with PAH from control patients.4Rahaghi F.N. Nardelli P. Harder E. et al.Quantification of arterial and venous morphologic markers in pulmonary arterial hypertension using CT imaging.Chest. 2021; 160: 2220-2231Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar In contrast, pulmonary venous alterations in PH are not well understood.1Humbert M. Kovacs G. Hoeper M.M. et al.2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension.Eur Heart J. 2022; 43: 3618-3731Crossref PubMed Scopus (1081) Google Scholar,2Thiwanka Wijeratne D. Lajkosz K. et al.Increasing incidence and prevalence of World Health Organization groups 1 to 4 pulmonary hypertension: a population-based cohort study in Ontario, Canada.Circ Cardiovasc Qual Outcomes. 2018; 11e003973PubMed Google Scholar,7Leopold J.A. Pulmonary venous remodeling in pulmonary hypertension the veins take center stage.Circulation. 2018; 137: 1811-1813Crossref PubMed Scopus (12) Google Scholar Recent evidence suggests that patterns of vascular remodeling may differ among PH groups. For example, both arterial and venous remodeling are seen in individuals with PVH, but arterial changes are predominate in those patients with PAH.7Leopold J.A. Pulmonary venous remodeling in pulmonary hypertension the veins take center stage.Circulation. 2018; 137: 1811-1813Crossref PubMed Scopus (12) Google Scholar,10Fayyaz A.U. Edwards W.D. Maleszewski J.J. et al.Global pulmonary vascular remodeling in pulmonary hypertension associated with heart failure and preserved or reduced ejection fraction.Circulation. 2018; 137: 1796-1810Crossref PubMed Scopus (211) Google Scholar A clinical feature of PVH is passive distention of pulmonary veins from elevated hydrostatic pressure, which leads to engorgement of the larger central pulmonary veins.7Leopold J.A. Pulmonary venous remodeling in pulmonary hypertension the veins take center stage.Circulation. 2018; 137: 1811-1813Crossref PubMed Scopus (12) Google Scholar,11Storto M.L. Kee S.T. Golden J.A. Webb W.R. Hydrostatic pulmonary edema: high-resolution CT findings.Am J Roentgenol. 1995; 165: 817-820Crossref PubMed Scopus (103) Google Scholar In this study, this central pulmonary venous engorgement was quantifiable with the use of automated methods applied to noninvasive imaging; the PVH group demonstrated greater venous engorgement compared with those patients with PAH or control patients, although no significant difference was seen in venous engorgement between PAH and control groups. We additionally found evidence of lower small venous volume on CT imaging (ie, venous pruning) in patients with PVH vs control patients and that the degree of venous pruning, as quantified by imaging, was similar between the PAH and PVH groups. Notably, small venous pruning on CT scanning has also been described in individuals with PAH,4Rahaghi F.N. Nardelli P. Harder E. et al.Quantification of arterial and venous morphologic markers in pulmonary arterial hypertension using CT imaging.Chest. 2021; 160: 2220-2231Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar exercise-induced PH,4Rahaghi F.N. Nardelli P. Harder E. et al.Quantification of arterial and venous morphologic markers in pulmonary arterial hypertension using CT imaging.Chest. 2021; 160: 2220-2231Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar and individuals with COPD who smoke.8Rahaghi F.N. Wells J.M. Come C.E. et al.Arterial and venous pulmonary vascular morphology and their relationship to findings in cardiac magnetic resonance imaging in smokers.J Comput Assist Tomogr. 2016; 40: 948-952Crossref PubMed Scopus (17) Google Scholar Taken in context, our findings suggest that small venous alterations may be a shared feature across different causes of pulmonary vascular disease, and therefore less specific than differences in the large veins in distinguishing PVH from PAH. We note that, although “pruning” of the small pulmonary veins may appear similar with this imaging technique, it is possible that the mechanism of those changes differs between PVH and PAH. For example, a prior study found that the character of small venous remodeling is distinct in PVH compared with another form of PH (pulmonary veno-occlusive disease), and that this remodeling may be reactive in the setting of pulmonary edema caused by congestive heart failure.10Fayyaz A.U. Edwards W.D. Maleszewski J.J. et al.Global pulmonary vascular remodeling in pulmonary hypertension associated with heart failure and preserved or reduced ejection fraction.Circulation. 2018; 137: 1796-1810Crossref PubMed Scopus (211) Google Scholar Although our study was not designed to elucidate the precise mechanism of the small venous changes, we hypothesize that this finding may be due to redistribution of small vessel blood volume into the dilated large veins, small vein remodeling, and/or poor circulation/reduced contrast penetration of the small veins caused by resistance and poor cardiac output in the setting of PH. Interpretation of our findings is limited by our small single-center sample and the retrospective nature of our study. In particular, certain inclusions/exclusions for the PH groups in our study (Fig 1) were adjudicated by PH specialist review that was based on the available retrospective, and thus potentially limited, data. However, to account for this, each case was reviewed by two independent clinical PH specialists, and all participants met the objective hemodynamic criteria for their respective PH group. In addition, CT scans and RHC were not performed contemporaneously, and clinical status (eg, volume status) may have differed between these time points. Finally, given that the data for our study were obtained during a clinical evaluation, participants may not have been necessarily in an optimized stable state during their CT scan, although none of the participants in our sample were on vasopressors or inotropes at the time of imaging. In conclusion, the results of our study suggest that large vessel engorgement on CT scanning is a noninvasive feature of PVH that may help to differentiate PVH from other types of PH. Additional research is necessary to replicate these preliminary findings in larger prospective studies, which would include those individuals with simultaneous imaging and catheterization, and in individuals with mixed pre- and postcapillary PH. If confirmed, our results may help support the application of CT scanning-based quantification of venous engorgement, in conjunction with existing invasive and noninvasive clinical tools, to improve screening and phenotyping of this highly morbid condition. This work was supported by grants to F. N. R. from NHLBI (K23HL136905, 1R01HL164717) and A. J. S. (K23HL164976).