DASS Good: Explainable Data Mining of Spatial Cohort Data
This addresses the problem of creating interpretable clinical models with spatial data for oncology researchers and practitioners, though it appears incremental as it builds on existing human-in-the-loop and explainable AI approaches.
The researchers tackled the challenge of developing clinical machine learning models with spatial data by co-designing DASS, a hybrid human-machine system that incorporates visual steering and explainable AI to build predictive models for radiotherapy toxicity in head and neck cancer patients, resulting in two practical clinical stratification models validated with expert feedback.
Developing applicable clinical machine learning models is a difficult task when the data includes spatial information, for example, radiation dose distributions across adjacent organs at risk. We describe the co-design of a modeling system, DASS, to support the hybrid human-machine development and validation of predictive models for estimating long-term toxicities related to radiotherapy doses in head and neck cancer patients. Developed in collaboration with domain experts in oncology and data mining, DASS incorporates human-in-the-loop visual steering, spatial data, and explainable AI to augment domain knowledge with automatic data mining. We demonstrate DASS with the development of two practical clinical stratification models and report feedback from domain experts. Finally, we describe the design lessons learned from this collaborative experience.