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

TL;DR

The TriALS challenge introduces a new benchmark for liver lesion segmentation in non-contrast CT, highlighting the difficulty of NCCT segmentation and the impact of training data scale and pre-training strategies.

Contribution

This work provides a multi-centre dataset, a benchmark for NCCT liver lesion segmentation, and insights into factors affecting algorithm performance.

Findings

Top method achieved Dice of 0.754 on venous-phase images.

Performance dropped to 0.57 on NCCT, showing challenges in contrast-limited settings.

Scaling pre-training alone does not overcome perceptual barriers on NCCT.

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.