Survival Prediction on TCGA-COADREAD
72.5C-indexMurreNet
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MurreNetPathological images=true, Genomic profiles=true2025.07 | 72.5 | — | |
| ROUPKTTraining Source=Representation-based knowledge transfer from Ms2026.03 | 71.23 | — | |
| M3IFPathological images=true, Genomic profiles=true2025.07 | 71 | — | |
| Shazam2025.03 | 71 | — | |
| STEPHTraining Source=Model merging-based knowledge transfer from Ms2026.03 | 70.69 | — | |
| ABMILPathological images=true, Genomic profiles=false2025.07 | 70.3 | — | |
| PORPOISEPathological images=true, Genomic profiles=true2025.07 | 70.2 | — | |
| MOMEPathological images=true, Genomic profiles=true2025.07 | 69.2 | — | |
| ABMILFeature Extractor=UNI [7]2025.12 | 68.75 | — | |
| ABMILFeature Extractor=UNI2025.12 | 68.75 | — | |
| DeltaMILFeature Extractor=ResNet-502025.12 | 68.57 | — | |
| DSMILPathological images=true, Genomic profiles=false2025.07 | 68.4 | — | |
| TransMILPathological images=true, Genomic profiles=false2025.07 | 68.4 | — | |
| DeltaMILFeature Extractor=UNI [7]2025.12 | 68.25 | — | |
| DeltaMILFeature Extractor=UNI2025.12 | 68.25 | — | |
| CMTAPathological images=true, Genomic profiles=true2025.07 | 67.8 | — | |
| Fine-tunedTraining Source=Traditional cancer-specific training on Mt2026.03 | 67.53 | — | |
| MLPPathological images=false, Genomic profiles=true2025.07 | 67.5 | — | |
| SURVPATHModality=Multimodal, Magnification=20x2023.04 | 67.3 | — | |
| VanillaTraining Source=Traditional cancer-specific training on Mt2026.03 | 67.25 | — | |
| S4MILFeature Extractor=UNI [7]2025.12 | 66.99 | — | |
| S4MILFeature Extractor=UNI2025.12 | 66.99 | — | |
| Mamba2MILFeature Extractor=ResNet-502025.12 | 66.51 | — | |
| Mean-PoolingFeature Extractor=UNI [7]2025.12 | 66.5 | — | |
| Mean-PoolingFeature Extractor=UNI2025.12 | 66.5 | — | |
| Mamba2MILFeature Extractor=UNI [7]2025.12 | 65.5 | — | |
| Mamba2MILFeature Extractor=UNI2025.12 | 65.5 | — | |
| S4MILFeature Extractor=ResNet-502025.12 | 65.45 | — | |
| DSMILFeature Extractor=UNI [7]2025.12 | 65.18 | — | |
| DSMILFeature Extractor=UNI2025.12 | 65.18 | — | |
| SurvPathPathological images=true, Genomic profiles=true2025.07 | 65 | — | |
| MCATPathological images=true, Genomic profiles=true2025.07 | 64.9 | — | |
| MambaMILFeature Extractor=ResNet-502025.12 | 64.62 | — | |
| H-optimus-12025.03 | 64.6 | — | |
| TIES AMTraining Source=Model merging-based knowledge transfer from Ms2026.03 | 64.44 | — | |
| CLAMPathological images=true, Genomic profiles=false2025.07 | 64.3 | — | |
| MOTCatModality=Multimodal, Fusion Strategy=Early fusion, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 64.1 | — | |
| Fine-tunedTraining Source=Representation-based knowledge transfer from Ms2026.03 | 63.45 | — | |
| MCATModality=Multimodal, Fusion Strategy=Early fusion, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 63.4 | — | |
| DTFD-MILPathological images=true, Genomic profiles=false2025.07 | 63.3 | — | |
| TransMILModality=WSI, Magnification=20x2023.04 | 63.2 | — | |
| ABMILModality=WSI, Magnification=20x2023.04 | 63 | — | |
| CLAMFeature Extractor=UNI [7]2025.12 | 62.76 | — | |
| CLAMFeature Extractor=UNI2025.12 | 62.76 | — | |
| Model Avg.Training Source=Model merging-based knowledge transfer from Ms2026.03 | 62.6 | — | |
| MLPModality=Omics, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 62.5 | — | |
| Virchow22025.03 | 62.5 | — | |
| MOTCatPathological images=true, Genomic profiles=true2025.07 | 61.8 | — | |
| MambaMILFeature Extractor=UNI [7]2025.12 | 61.74 | — | |
| MambaMILFeature Extractor=UNI2025.12 | 61.74 | — | |
| Max-PoolingFeature Extractor=UNI [7]2025.12 | 61.53 | — | |
| Max-PoolingFeature Extractor=UNI2025.12 | 61.53 | — | |
| MorphDistill2026.04 | 61.51 | — | |
| Surgery AMTraining Source=Model merging-based knowledge transfer from Ms2026.03 | 61.37 | — | |
| DSMILFeature Extractor=ResNet-502025.12 | 60.87 | — | |
| LWRDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 60.8 | 1 | |
| CLAMFeature Extractor=ResNet-502025.12 | 60.79 | — | |
| SNNPathological images=false, Genomic profiles=true2025.07 | 60.6 | — | |
| ABMIL2026.04 | 60.34 | — | |
| IntegrAODownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 60.2 | 2 | |
| TransMILFeature Extractor=UNI [7]2025.12 | 60.11 | — | |
| TransMILFeature Extractor=UNI2025.12 | 60.11 | — | |
| CoxnnetPathological images=false, Genomic profiles=true2025.07 | 60.1 | — | |
| RRT-MIL2026.04 | 59.91 | — | |
| AdaMergingTraining Source=Model merging-based knowledge transfer from Ms2026.03 | 59.71 | — | |
| TransMILFeature Extractor=ResNet-502025.12 | 59.64 | — | |
| UNI 22025.03 | 59.6 | — | |
| Prov-Gigapath2025.03 | 59.5 | — | |
| ABMIL (Cat)Modality=Multimodal, Fusion Strategy=Concatenation, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 59.2 | — | |
| ABMIL (KP)Modality=Multimodal, Fusion Strategy=Kronecker product, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 58.4 | — | |
| PANTHER2026.04 | 58.32 | — | |
| S-MLPModality=Omics, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 58.1 | — | |
| Mean-PoolingFeature Extractor=ResNet-502025.12 | 57.81 | — | |
| Phikon-v22025.03 | 57.4 | — | |
| AMISL (KP)Modality=Multimodal, Fusion Strategy=Kronecker product, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 56.7 | — | |
| TransMIL (KP)Modality=Multimodal, Fusion Strategy=Kronecker product, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 56.6 | — | |
| MINDDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 56.4 | 3 | |
| Max-PoolingFeature Extractor=ResNet-502025.12 | 56.38 | — | |
| ABMILFeature Extractor=ResNet-502025.12 | 55.96 | — | |
| MSNEDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 55.8 | 4 | |
| Iso-C AMTraining Source=Model merging-based knowledge transfer from Ms2026.03 | 55.32 | — | |
| JASMINEDownstream predictor=penalized Cox's proportional hazards models, Evaluation protocol=five-fold cross-validation, Input features=patient information and embeddings2026.06 | 54 | 5 | |
| TransMIL (Cat)Modality=Multimodal, Fusion Strategy=Concatenation, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 53.9 | — | |
| SNNModality=Omics, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 52.1 | — | |
| AMISL (Cat)Modality=Multimodal, Fusion Strategy=Concatenation, Magnification=20x, Pathway Sets=Reactome and Hallmark2023.04 | 51 | — | |
| AMISLModality=WSI, Magnification=20x2023.04 | 50 | — | |
| DSMIL2026.04 | 50 | — |