Radiologist- and computer-based chest imaging quantification at 12 months post transplant correlates with baseline...

Area of Science:

• Pulmonary Medicine
• Radiology
• Transplant Surgery

Background:

• Baseline lung allograft dysfunction (BLAD) is defined as spirometry <80% predicted post-lung transplant.
• BLAD is associated with reduced patient survival.
• The CT imaging characteristics of BLAD require further elucidation.

Purpose of the Study:

• To determine the prevalence of CT abnormalities in BLAD.
• To characterize radiologic findings associated with BLAD.
• To identify CT features that predict BLAD.

Main Methods:

• Retrospective cohort study of adult double-lung transplant recipients (12/2017-10/2021).
• Radiologist assessment of CT scans for GGO, reticulation, consolidation, effusion, bronchiectasis, and air-trapping.
• Machine learning-based CT radiomics (CT-measured total lung capacity [CT_TLC], pulmonary vessel volume [PVV]) and parametric response mapping for functional small airways disease.

Main Results:

• BLAD patients (46%) showed longer intubation, hospitalization, and lower donor-to-recipient TLC ratio.
• Radiologist review found more pleural effusions in BLAD; computer-aided CT showed more reticulation, GGO, and parenchymal density.
• Lower CT_TLC and higher PVV were observed in BLAD. PVV was a significant predictor of BLAD (OR=2.65).

Conclusions:

• Computer-aided CT detects structural changes in BLAD missed by standard radiologist review.
• CT_TLC and PVV are strong radiologic predictors of BLAD.
• These findings suggest restrictive physiology and vascular remodeling in BLAD.