TriALS: Triphasic-Aided Liver Lesion Segmentation Benchmark in Non-Contrast CT

2026-05-15

AI Summary

Automatic segmentation of liver lesions in non-contrast CT remains a significant challenge in resource-limited settings due to the absence of annotated benchmarks. This study addresses this gap by constructing the first publicly available, multicenter, four-phase dataset specifically designed for non-contrast CT and proposes a triphasic auxiliary strategy that integrates large-scale pretraining with data-driven methodologies within a systematic algorithm evaluation framework. The proposed approach achieves a Dice score of 0.57 on non-contrast CT and outperforms existing models by 28% in external validation. Notably, it attains near-human performance in the venous phase (Dice: 0.754), demonstrating that reliance solely on pretraining is insufficient to overcome perceptual limitations and underscoring the critical importance of multiphase information fusion.

Abstract

Automated segmentation of liver lesions on non-contrast computed tomography (NCCT) is clinically important but fundamentally challenging, particularly in low-resource settings across Africa and Asia where contrast agents are frequently unavailable. Progress has been limited by the absence of annotated NCCT benchmarks. Here we describe the TriALS challenge for automated liver lesion segmentation under contrast-limited conditions, supported by a multi-centre dataset of 150 cases with four-phase CT acquisitions (600 volumes) from Egyptian and Chinese institutions. Algorithms were evaluated on 70 cases from three institutions, including an independent external cohort. The top-performing method achieved a mean venous-phase Dice of 0.754, consistent with human-level performance, yet dropped to 0.57 on NCCT. On external validation, the leading method outperformed off-the-shelf models by up to 28% in Dice on NCCT. Algorithm performance was most strongly predicted by training data scale and pre-training strategy. A cross-year comparison exposed a persistent perceptual barrier on NCCT that scaling pre-training alone cannot overcome. Data, annotations, and code are available at https://github.com/xmed-lab/TriALS.

Problem

• liver lesion segmentation
• non-contrast CT
• medical image analysis
• benchmark
• automated segmentation

Innovation

• non-contrast CT
• liver lesion segmentation
• multi-phase CT benchmark
• automated medical image analysis
• external validation