Reproduce “ModalSurv: Investigating opportunities and limitations of multimodal deep survival learning in prostate and bladder cancer” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Noorul Wahab, Ethar Alzaid, Jiaqi Lv, Fayyaz Minhas, Adam Shephard, Shan E Ahmed Raza
ModalSurv predicts cancer recurrence over time by projecting available data types into shared representations. Cross-attention lets each modality draw on the others before a survival head estimates risk across time intervals. Prostate discovery testing favored cross-attention over simple concatenation, but external performance was weaker. Bladder test score was 0.5740, below a clinical-only challenge entry’s 0.6828; this comparison cannot isolate fusion.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.