Deep Relevance Regularization: Interpretable and Robust Tumor Typing of Imaging Mass Spectrometry Data
This work addresses the need for more reliable and interpretable AI tools in medical diagnostics, particularly for tumor typing across different laboratories, though it appears incremental as it builds on existing neural network methods.
The paper tackled the problem of neural network performance degradation due to confounding factors in multi-laboratory tumor typing using imaging mass spectrometry data, and introduced Deep Relevance Regularization to improve classification robustness and interpretability, demonstrating effectiveness on a challenging inter-lab dataset of breast and ovarian carcinoma.
Neural networks have recently been established as a viable classification method for imaging mass spectrometry data for tumor typing. For multi-laboratory scenarios however, certain confounding factors may strongly impede their performance. In this work, we introduce Deep Relevance Regularization, a method of restricting what the neural network can focus on during classification, in order to improve the classification performance. We demonstrate how Deep Relevance Regularization robustifies neural networks against confounding factors on a challenging inter-lab dataset consisting of breast and ovarian carcinoma. We further show that this makes the relevance map -- a way of visualizing the discriminative parts of the mass spectrum -- sparser, thereby making the classifier easier to interpret