Multimodal Prototyping for cancer survival prediction
This work addresses computational and interpretability bottlenecks in cancer prognostication, offering a more efficient and explainable approach for medical researchers and clinicians.
The paper tackled the problem of high memory requirements and poor interpretability in multimodal cancer survival prediction by summarizing histology images and transcriptomic profiles with prototypes, achieving over 300x compression and outperforming state-of-the-art methods on six cancer types with less computation.
Multimodal survival methods combining gigapixel histology whole-slide images (WSIs) and transcriptomic profiles are particularly promising for patient prognostication and stratification. Current approaches involve tokenizing the WSIs into smaller patches (>10,000 patches) and transcriptomics into gene groups, which are then integrated using a Transformer for predicting outcomes. However, this process generates many tokens, which leads to high memory requirements for computing attention and complicates post-hoc interpretability analyses. Instead, we hypothesize that we can: (1) effectively summarize the morphological content of a WSI by condensing its constituting tokens using morphological prototypes, achieving more than 300x compression; and (2) accurately characterize cellular functions by encoding the transcriptomic profile with biological pathway prototypes, all in an unsupervised fashion. The resulting multimodal tokens are then processed by a fusion network, either with a Transformer or an optimal transport cross-alignment, which now operates with a small and fixed number of tokens without approximations. Extensive evaluation on six cancer types shows that our framework outperforms state-of-the-art methods with much less computation while unlocking new interpretability analyses.