A comparative study of CT-based volumetric assessment methods for total lung capacity with the development of an adjustment factor: incorporating VR imaging for improved accuracy

Physiological methods for measuring total lung capacity (TLC), including body-box plethysmography (BBP), are costly and require specialized expertise. Computed tomography (CT)-based TLC assessment is essential in clinical practice for candidates of lung transplantation and those unable to undergo standard lung function testing. While CT-based algorithms were studied to estimate TLC, their accuracy should be further evaluated. This study aimed to compare the BBP measurement of TLC (TLCpleth) with three CT-based methods for measuring TLC, one of them is an innovative virtual reality (VR)-based method. Additionally, we aimed to develop an adjustment factor that will allow a new, non-invasive, cost-effective estimation of the TLCpleth. TLC was calculated for 24 adult patients using three different CT-based volumetric assessment methods: an older region-growing algorithm (TLCrg), a more recent convolutional neural network-based algorithm (TLCcnn), and a VR-based method (TLCvr). Agreement between each method and TLCpleth was evaluated, and an adjustment factor was developed using linear regression. The correlation between the three CT-based methods and TLCpleth ranged from 0.91 to 0.92 (p < 0.001). TLCvr measurements were 80.13% (CI:75.08–85.18%, P < 0.001) of TLCpleth measures, whereas TLCcnn and TLCrg estimates were 71.3% and 77.1% of TLCpleth, respectively. An adjustment factor is proposed to estimate TLCpleth based on the three CT-based methods. This study is the first to evaluate the correlation between BBP, VR volumetric analysis, and two iterations of CT volumetric software for measuring total lung capacity (TLC). After being corrected by an adjustment factor, VR- and CT-based assessments provide accurate estimates of TLCpleth.