Personalized Survival Prediction with Contextual Explanation Networks
This work addresses the need for accurate and interpretable survival predictions in cancer care, offering potential improvements in patient treatment decisions.
The paper tackled the problem of predicting individual cancer survival times with transparency by developing a model that simultaneously learns to predict patient-specific survival distributions and explain its predictions using patient attributes. The model outperformed several baselines on two public datasets.
Accurate and transparent prediction of cancer survival times on the level of individual patients can inform and improve patient care and treatment practices. In this paper, we design a model that concurrently learns to accurately predict patient-specific survival distributions and to explain its predictions in terms of patient attributes such as clinical tests or assessments. Our model is flexible and based on a recurrent network, can handle various modalities of data including temporal measurements, and yet constructs and uses simple explanations in the form of patient- and time-specific linear regression. For analysis, we use two publicly available datasets and show that our networks outperform a number of baselines in prediction while providing a way to inspect the reasons behind each prediction.