AI-augmented quantitative analysis of lungs photon-counting computed tomography in heart failure patients

Abstract Introduction Heart failure (HF) is a syndrome associated with poor prognosis, impacting lung structure/function and resulting in dyspnea and type 2 pulmonary hypertension (PH). PH develops in up to 80% of patients with HF and is linked to worse survival rates. Purpose To assess and compare quantitative computed tomography (qCT) pulmonary findings in HF patients and control subjects. Methods The study involved patients with HF and non-HF control subjects. No participant had significant lung disease. Participants underwent laboratory testing, echocardiography, right heart catheterization, and native chest imaging (in full inspiration and expiration) by Naeotom Alpha photon-counting CT scanner (Siemens). Automated AI-based analysis (VIDA) was employed to evaluate lung volumes, densities, vessel volumes, and tissue texture. Image series were used in mediastinal and lung window reconstruction with slice thickness 0,8 mm, increment 0,5 mm, matrix 512 x 512, mediastinal kernel Br36 and lung kernel Bl60, in cases of metallic implants presence also with iMAR reconstructions. FOV was based on the width of patients to cover whole lungs in x and y axis. Results A total of 229 subjects were analyzed: 186 HF and 43 control subjects. Mean patient age was 57.8 ± 11.0 years, 19.7 % of them were women. Among the HF patients, 158 (85 %) exhibited PH (mean pulmonary artery pressure mPAP 20 mmHg). HF patients demonstrated comparable inspiratory lung volumes to controls, but had reduced air volume and increased lung tissue volume. Mean lung density was significantly higher in the HF group (-787 Hounsfield units HU, IQR: -855 to -787) compared to controls (-833 HU, IQR: -855 to -833; ; p 0.0001). There was no significant difference in air trapping between groups. HF patients exhibited a lower volume fraction of vessels 1 mm² and 1-2 mm² compared to controls, but a higher proportion of vessels 2 mm². (Picture 1) Ground-glass opacity (GGO) mass was higher in HF patients (48.8 g vs 9.1 g; p 0.001), as was reticulation (42.2 g vs. 8.4 g; p 0.001) and consolidation (0.5 g vs 0.2 g; p=0.0011), with no significant difference in the extent of honeycombing. GGO mass showed a stronger correlation with mean pulmonary artery pressure (mPAP) and mean pulmonary capillary wedge pressure (PCWP) than reticulations (r=0.31, p0.001 for mPAP and r=0.27, p=0.002 for PCWP). Picture 2 shows comparison of subpleural lung images between control (A) and HF patient (B) Conclusion qCT distinguished patients with HF vs. controls, with lung density, the fraction of the smallest vessels demonstrating significant differences between groups. Mass of ground-glass opacities and reticulations were significantly correlated to mPAP and PCWP.Comparison of vessel volume fractions Subpleural lung CT reconstructions