Survival Prediction on TCGA-COAD
0.694C-indexCrossFusion
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CrossFusionScale=Multi Scale, Backbone=ResNet502025.03 | 0.694 | — | |
| SCMILScale=Single Scale, Backbone=ResNet502025.03 | 0.677 | — | |
| PathOmicsPretrain data modality=image+mRNA, Finetune data modality=image+mRNA2023.07 | 0.6732 | — | |
| PathOmicsPretrain data modality=image+Methy, Finetune data modality=image+Methy2023.07 | 0.6722 | — | |
| CrossFusion w/o CPScale=Multi Scale, Backbone=ResNet502025.03 | 0.669 | — | |
| ProtoSurvScale=Single Scale, Backbone=ResNet502025.03 | 0.668 | — | |
| AMILScale=Single Scale, Backbone=ResNet502025.03 | 0.662 | — | |
| DeepSetsPretrain data modality=image+Methy, Finetune data modality=image+Methy2023.07 | 0.6561 | — | |
| TransMILScale=Single Scale, Backbone=ResNet502025.03 | 0.656 | — | |
| PORPOI-SEPretrain data modality=image+mRNA, Finetune data modality=image+mRNA2023.07 | 0.6531 | — | |
| HIPTScale=Multi Scale, Backbone=ResNet502025.03 | 0.651 | — | |
| MCATPretrain data modality=image+mRNA, Finetune data modality=image+mRNA2023.07 | 0.6502 | — | |
| MCATPretrain data modality=image+CNA, Finetune data modality=image+CNA2023.07 | 0.6466 | — | |
| ZoomMILScale=Multi Scale, Backbone=ResNet502025.03 | 0.642 | — | |
| MuSTMILScale=Multi Scale, Backbone=ResNet502025.03 | 0.64 | — | |
| PathOmicsPretrain data modality=image+mRNA, Finetune data modality=image2023.07 | 0.6378 | — | |
| CSMILScale=Multi Scale, Backbone=ResNet502025.03 | 0.636 | — | |
| CrossFusion w/o F&CScale=Multi Scale, Backbone=ResNet502025.03 | 0.631 | — | |
| PORPOI-SEPretrain data modality=image+Methy, Finetune data modality=image+Methy2023.07 | 0.6184 | — | |
| PathOmicsPretrain data modality=image+Methy, Finetune data modality=Methy2023.07 | 0.6156 | — | |
| PatchGCNScale=Single Scale, Backbone=ResNet502025.03 | 0.612 | — | |
| PathOmicsPretrain data modality=image+CNA, Finetune data modality=image+CNA2023.07 | 0.6119 | — | |
| DSMILScale=Single Scale, Backbone=ResNet502025.03 | 0.61 | — | |
| MCATPretrain data modality=image+Methy, Finetune data modality=image+Methy2023.07 | 0.6098 | — | |
| PathOmicsPretrain data modality=image+mRNA, Finetune data modality=mRNA2023.07 | 0.6076 | — | |
| PathOmicsPretrain data modality=image+Methy, Finetune data modality=image2023.07 | 0.6043 | — | |
| TransMILPretrain data modality=image+CNA, Finetune data modality=image+CNA2023.07 | 0.598 | — | |
| DeepGraphSurvScale=Single Scale, Backbone=ResNet502025.03 | 0.591 | — | |
| DeepSetsPretrain data modality=image+mRNA, Finetune data modality=image+mRNA2023.07 | 0.587 | — | |
| PathOmicsPretrain data modality=image+CNA, Finetune data modality=image2023.07 | 0.5806 | — | |
| PORPOI-SEPretrain data modality=image+CNA, Finetune data modality=image+CNA2023.07 | 0.5732 | — | |
| MSNEDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 0.573 | 1 | |
| PathOmicsPretrain data modality=image+CNA, Finetune data modality=CNA2023.07 | 0.5643 | — | |
| IntegrAODownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 0.553 | 2 | |
| AB-MILPretrain data modality=image+CNA, Finetune data modality=image+CNA2023.07 | 0.5468 | — | |
| TransMILPretrain data modality=image+mRNA, Finetune data modality=image+mRNA2023.07 | 0.5415 | — | |
| AB-MILPretrain data modality=image+mRNA, Finetune data modality=image+mRNA2023.07 | 0.5412 | — | |
| LWRDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 0.541 | 3 | |
| MINDDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 0.54 | 4 | |
| TransMILPretrain data modality=image+Methy, Finetune data modality=image+Methy2023.07 | 0.5335 | — | |
| JASMINEDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 0.531 | 5 | |
| DeepSetsPretrain data modality=image+CNA, Finetune data modality=image+CNA2023.07 | 0.515 | — | |
| AB-MILPretrain data modality=image+Methy, Finetune data modality=image+Methy2023.07 | 0.4966 | — |