Research map / Vision-language and multimodal / Cross-modal learning
LUMIN: A Longitudinal Multi-modal Knowledge Decomposition Network for Predicting Breast Cancer Recurrence
Chunyao Lu, Tianyu Zhang, Xinglong Liang, Yuan Gao, Luyi Han, Xin Wang, Nika Rasoolzadeh, Tao Tan, Ritse Mann
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 contrastive image-text pretraining (ablation) | LUMIN · Breast cancer recurrence (3,924 patients) | AUC: 0.703 -> 0.729 | Improved |
| add CM-DKE (ablation) | LUMIN · Breast cancer recurrence (3,924 patients) | AUC: 0.714 -> 0.729 | Improved |
| add TE-DKE (ablation) | LUMIN · Breast cancer recurrence (3,924 patients) | AUC: 0.721 -> 0.729 | Improved |
| add text-specific knowledge Ktext (ablation) | CM-DKE only · Breast cancer recurrence (3,924 patients) | AUC: 0.691 -> 0.721 | Improved |
| add image-specific knowledge Kimg (ablation) | CM-DKE only · Breast cancer recurrence (3,924 patients) | AUC: 0.706 -> 0.721 | Improved |
| add complementary knowledge Kcomp (ablation) | CM-DKE only · Breast cancer recurrence (3,924 patients) | AUC: 0.711 -> 0.721 | Improved |
| add shared knowledge Kshared (ablation) | CM-DKE only · Breast cancer recurrence (3,924 patients) | AUC: 0.708 -> 0.721 | Improved |
| add time-varying knowledge Kvar (ablation) | TE-DKE only · Breast cancer recurrence (3,924 patients) | AUC: 0.701 -> 0.714 | Improved |
| add Time1-specific knowledge KT1 (ablation) | TE-DKE only · Breast cancer recurrence (3,924 patients) | AUC: 0.695 -> 0.714 | Improved |
| add EHR to image input (single time) | ResNet-50/ST-I · Breast cancer recurrence (3,924 patients) | AUC: 0.605 -> 0.693 | Improved |
| extend single-time fusion to two exams without temporal modeling | ST-B · Breast cancer recurrence (3,924 patients) | AUC: 0.693 -> 0.676 | Hurt |
| use knowledge decomposition instead of naive longitudinal fusion | L-B · Breast cancer recurrence (3,924 patients) | AUC: 0.676 -> 0.729 | Improved |
Closest papers by meaning
- Prototypical Information Bottlenecking and Disentangling for Multimodal Cancer Survival Prediction (ICLR 2024 · similarity 0.86)
- Multimodal Disease Progression Modeling via Spatiotemporal Disentanglement and Multiscale Alignment (NeurIPS 2025 · similarity 0.85)
- CA-MLIF: Cross-Attention and Multimodal Low-Rank Interaction Fusion Framework for Tumor Prognostic Prediction (AAAI 2025 · similarity 0.85)
- Risk Horizons: Structured Hypothesis Spaces for Longitudinal Clinical Prediction (NeurIPS 2026 · similarity 0.85)
- TumorChain: Interleaved Multimodal Chain-of-Thought Reasoning for Traceable Clinical Tumor Analysis (ICLR 2026 · similarity 0.85)
- HEALNet: Multimodal Fusion for Heterogeneous Biomedical Data (NeurIPS 2024 · similarity 0.84)
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.
