Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Latent Imaging Biomarkers for Interpretable Microvascular Invasion Prediction in Hepatocellular Carcinoma
Ji Rao, Xinyu Liu, Yong Yi, Ying Xiao, Ye Luo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add mixup augmentation (ablation) | IRCL · HCC-ZSH multi-modality | ACC (%): 79.81 -> 92.78 | Improved |
| add mixup augmentation (ablation) | IRCL · HCC-ZSH multi-modality | F1 (%): 73.16 -> 90.39 | Improved |
| add intra-patient positive samples (ablation) | IRCL contrastive learning · HCC-ZSH multi-modality | ACC (%): 90.37 -> 92.78 | Improved |
| add inter-patient positive samples (ablation) | IRCL contrastive learning · HCC-ZSH multi-modality | ACC (%): 90.55 -> 92.78 | Improved |
| add contrastive loss Lc (ablation) | IRCL Stage I · HCC-ZSH multi-modality | ACC (%): 90.09 -> 92.78 | Improved |
| add reconstruction loss Lr (ablation) | IRCL Stage I · HCC-ZSH multi-modality | ACC (%): 89.35 -> 92.78 | Improved |
| add biomarker prototype attention (ablation) | IRCL Stage II · HCC-ZSH multi-modality | ACC (%): 92.17 -> 92.78 | Improved |
| vary number of biomarkers K and inter-patient positives s (ablation) | IRCL · HCC-ZSH | performance | No clear effect |
| use multi-modality (T1, T1D, T1V) input instead of single modality | IRCL · HCC-ZSH | ACC (%): 88.06 -> 92.78 | Improved |
| use IRCL instead of competing methods on T1D and T1V single modalities | MVI classification · HCC-ZSH (T1D) | AUC (%): 90.55 -> 84.44 | Mixed |
Closest papers by meaning
- Time-to-Event Pretraining for 3D Medical Imaging (ICLR 2025 · similarity 0.83)
- Tumor Micro-environment Interactions Guided Graph Learning for Survival Analysis of Human Cancers from Whole-slide Pathological Images (CVPR 2024 · similarity 0.82)
- Leveraging Tumor Heterogeneity: Heterogeneous Graph Representation Learning for Cancer Survival Prediction in Whole Slide Images (NeurIPS 2024 · similarity 0.82)
- xMIL: Insightful Explanations for Multiple Instance Learning in Histopathology (NeurIPS 2024 · similarity 0.82)
- Modeling Dense Multimodal Interactions Between Biological Pathways and Histology for Survival Prediction (CVPR 2024 · similarity 0.82)
- Symb-xMIL: Symbolic Explanations for Multiple Instance Learning in Digital Pathology (NeurIPS 2026 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
