CORRELATION OF QUANTITATIVE BLOOD VOLUME ANALYSES FROM CT IMAGING WITH HEMODYNAMICS IN PATIENTS WITH PULMONARY HYPERTENSION

Abstract

SESSION TITLE: Technological Innovations in ImagingSESSION TYPE: Original InvestigationsPRESENTED ON: 10/17/22 1:30 PM - 2:30 PMPURPOSE: The characterization of pulmonary vasculature via imaging may be a noninvasive diagnostic method to characterize pulmonary hypertension (PH). Limited studies have assessed pulmonary blood volume (PBV) using CT imaging. This study’s purpose was to make an integrated framework to analyze and quantitate PBV on CT imaging and correlate it to the hemodynamic variables from right heart catheterization (RHC).METHODS: This study was approved by the Institutional Review Board at the University of Virginia. Demographics, RHC parameters, pulmonary function tests (PFT), and 6-minute-walk-test (6MWT), were retrospectively collected using the PH registry. CT pulmonary arteriogram (CTPA) images were analyzed using the open-source Advanced Normalization Tools (ANTs). Whole lung segmentation was carried out using a deep learning network. From the whole lung mask, lobar masks were estimated using a separate network permitting tabulation of lobe-specific CT values. PBV was reported according to lobar distribution. Spearman’s correlations were assessed by the quantitative PBV in the whole lung, left vs. right lung, and upper vs. lower lobes, RHC parameters, PFT, and 6MWT.RESULTS: Among 47 patients (70% female/30% male, median age was 67 [IQR 56-77]), mean pulmonary pressure (mPAP) was 39 mmHg (IQR 30-47), median pulmonary vascular resistance 5.73 WU (IQR 2.9-9.8), median right atrial pressure 8 mmHg (IQR 5-13), and pulmonary arterial saturation 64% (IQR 57-69). Mean PBV for whole lung was 2.80x105 mm3 (IQR 1.23-5.18x105), left lung 1.13x105 mm3 (IQR 0.44-2.46x105), right lung 1.52x105 mm3 (IQR 0.49-2.92x105), upper lobes 1.47x105 mm3 (IQR 0.37-3.03x105), and lower lobes 1.07x105 mm3 (IQR 0.34-2.83x105). PBV was positively associated with increased PVR and PA saturation. mPAP was positively correlated with PBV in the right lung but not in the left lung. PBV did not correlate with cardiac output but negatively correlated with pulmonary capillary wedge pressure.CONCLUSIONS: Our exploratory study shows a positively correlating PBV and PVR and PA saturation and divergent PBV between right and left lungs. These patterns of correlations between PBV from CTPA and RHC hemodynamics suggest that PBV may increase as a potential compensatory mechanism related to pre-capillary pulmonary vascular resistance.CLINICAL IMPLICATIONS: Quantitative image analysis of CTPA has the potential to phenotype PH and related vascular physiology and will require additional research to validate these findings. This technique may aid in understanding the intricate pressure-volume relationship in the pulmonary circulation during disease states.DISCLOSURES: No relevant relationships by Carissa Harnish-CruzNo relevant relationships by Jhosep HuaromoNo relevant relationships by Jaime MataNo relevant relationships by Sula MazimbaNo relevant relationships by Andrew Mihalek, value=Grant/Research SupportRemoved 04/13/2022 by Andrew MihalekNo relevant relationships by Kun QingNo relevant relationships by Prerna SharmaNo relevant relationships by Y ShimNo relevant relationships by Nicholas Tustison SESSION TITLE: Technological Innovations in Imaging SESSION TYPE: Original Investigations PRESENTED ON: 10/17/22 1:30 PM - 2:30 PM PURPOSE: The characterization of pulmonary vasculature via imaging may be a noninvasive diagnostic method to characterize pulmonary hypertension (PH). Limited studies have assessed pulmonary blood volume (PBV) using CT imaging. This study’s purpose was to make an integrated framework to analyze and quantitate PBV on CT imaging and correlate it to the hemodynamic variables from right heart catheterization (RHC). METHODS: This study was approved by the Institutional Review Board at the University of Virginia. Demographics, RHC parameters, pulmonary function tests (PFT), and 6-minute-walk-test (6MWT), were retrospectively collected using the PH registry. CT pulmonary arteriogram (CTPA) images were analyzed using the open-source Advanced Normalization Tools (ANTs). Whole lung segmentation was carried out using a deep learning network. From the whole lung mask, lobar masks were estimated using a separate network permitting tabulation of lobe-specific CT values. PBV was reported according to lobar distribution. Spearman’s correlations were assessed by the quantitative PBV in the whole lung, left vs. right lung, and upper vs. lower lobes, RHC parameters, PFT, and 6MWT. RESULTS: Among 47 patients (70% female/30% male, median age was 67 [IQR 56-77]), mean pulmonary pressure (mPAP) was 39 mmHg (IQR 30-47), median pulmonary vascular resistance 5.73 WU (IQR 2.9-9.8), median right atrial pressure 8 mmHg (IQR 5-13), and pulmonary arterial saturation 64% (IQR 57-69). Mean PBV for whole lung was 2.80x105 mm3 (IQR 1.23-5.18x105), left lung 1.13x105 mm3 (IQR 0.44-2.46x105), right lung 1.52x105 mm3 (IQR 0.49-2.92x105), upper lobes 1.47x105 mm3 (IQR 0.37-3.03x105), and lower lobes 1.07x105 mm3 (IQR 0.34-2.83x105). PBV was positively associated with increased PVR and PA saturation. mPAP was positively correlated with PBV in the right lung but not in the left lung. PBV did not correlate with cardiac output but negatively correlated with pulmonary capillary wedge pressure. CONCLUSIONS: Our exploratory study shows a positively correlating PBV and PVR and PA saturation and divergent PBV between right and left lungs. These patterns of correlations between PBV from CTPA and RHC hemodynamics suggest that PBV may increase as a potential compensatory mechanism related to pre-capillary pulmonary vascular resistance. CLINICAL IMPLICATIONS: Quantitative image analysis of CTPA has the potential to phenotype PH and related vascular physiology and will require additional research to validate these findings. This technique may aid in understanding the intricate pressure-volume relationship in the pulmonary circulation during disease states. DISCLOSURES: No relevant relationships by Carissa Harnish-Cruz No relevant relationships by Jhosep Huaromo No relevant relationships by Jaime Mata No relevant relationships by Sula Mazimba No relevant relationships by Andrew Mihalek, value=Grant/Research Support Removed 04/13/2022 by Andrew Mihalek No relevant relationships by Kun Qing No relevant relationships by Prerna Sharma No relevant relationships by Y Shim No relevant relationships by Nicholas Tustison