Problem
Research questions and friction points this paper is trying to address.
future liver remnant
colorectal liver metastasis
medical image segmentation
surgical planning
AI benchmark
Benchmarking Deep Learning for Future Liver Remnant Segmentation in Colorectal Liver Metastasis
🤖 AI Summary
Accurate segmentation of the future liver remnant (FLR) volume in patients with colorectal liver metastases (CRLM) is critical for preoperative planning but remains challenging due to complex anatomy, diffuse metastatic lesions, and a scarcity of high-quality annotated data. To address this, this study introduces CRLM-CT-Seg—the first open-source, expert-validated benchmark dataset for FLR segmentation—constructed from 197 manually refined CT scans. The work systematically evaluates both cascaded and end-to-end deep learning strategies, demonstrating that a cascaded nnU-Net achieves the best FLR segmentation performance (Dice = 0.767), while a pretrained STU-Net yields superior metastasis segmentation accuracy (Dice = 0.620) and greater robustness to cascaded error propagation. These findings establish a reproducible benchmark and provide methodological guidance for clinical FLR assessment.
📝 Abstract
Accurate segmentation of the future liver remnant (FLR) is critical for surgical planning in colorectal liver metastases (CRLM) to prevent fatal post-hepatectomy liver failure. However, this segmentation task is technically challenging due to complex resection boundaries, convoluted hepatic vasculature and diffuse metastatic lesions. A primary bottleneck in developing automated AI tools has been the lack of high-fidelity, validated data. We address this gap by manually refining all 197 volumes from the public CRLM-CT-Seg dataset, creating the first open-source, validated benchmark for this task. We then establish the first segmentation baselines, comparing cascaded (Liver->CRLM->FLR) and end-to-end (E2E) strategies using nnU-Net, SwinUNETR, and STU-Net. We find a cascaded nnU-Net achieves the best final FLR segmentation Dice (0.767), while the pretrained STU-Net provides superior CRLM segmentation (0.620 Dice) and is significantly more robust to cascaded errors. This work provides the first validated benchmark and a reproducible framework to accelerate research in AI-assisted surgical planning.
Innovation
Methods, ideas, or system contributions that make the work stand out.
future liver remnant segmentation
colorectal liver metastasis
validated benchmark
cascaded vs end-to-end
STU-Net
🔎 Similar Papers
No similar papers found.
