Automated Cancer Subtyping via Vector Quantization Mutual Information Maximization
This addresses the challenge of medically controversial labels in cancer subtyping for improved therapy, though it appears incremental as it builds on existing clustering approaches.
The paper tackles the problem of automated cancer subtyping from high-dimensional genetic expression profiles by proposing an unsupervised clustering method that maximizes mutual information to determine the number of subtypes, and it demonstrates that the refined labels correlate highly with cancer survival rates.
Cancer subtyping is crucial for understanding the nature of tumors and providing suitable therapy. However, existing labelling methods are medically controversial, and have driven the process of subtyping away from teaching signals. Moreover, cancer genetic expression profiles are high-dimensional, scarce, and have complicated dependence, thereby posing a serious challenge to existing subtyping models for outputting sensible clustering. In this study, we propose a novel clustering method for exploiting genetic expression profiles and distinguishing subtypes in an unsupervised manner. The proposed method adaptively learns categorical correspondence from latent representations of expression profiles to the subtypes output by the model. By maximizing the problem -- agnostic mutual information between input expression profiles and output subtypes, our method can automatically decide a suitable number of subtypes. Through experiments, we demonstrate that our proposed method can refine existing controversial labels, and, by further medical analysis, this refinement is proven to have a high correlation with cancer survival rates.