Recovering Truncated Chest Wall Anatomy on CT for Downstream Body Composition Analysis: Quantitative Evaluation of an Open-Access Generative Artificial Intelligence Tool

The open-access generative artificial intelligence model CT-Palette may enable more accurate recovery of truncated regions on axial chest CT images at the fifth, eighth, and tenth thoracic vertebral levels compared with S-EFOV, the current state-of-the-art model.