DITTO: A Visual Digital Twin for Interventions and Temporal Treatment Outcomes in Head and Neck Cancer

TL;DR

DITTO is a digital twin system utilizing deep reinforcement learning and visual explainability to assist Head and Neck Cancer clinicians in personalized treatment decision-making by analyzing complex risk profiles.

Contribution

The paper introduces DITTO, a novel visual digital twin system that combines deep reinforcement learning and explainability to support personalized treatment planning in HNC.

Findings

DITTO accurately predicts patient-specific risks for outcomes and toxicities.

Clinicians found DITTO's visual explanations helpful for trust and decision-making.

Quantitative and qualitative evaluations demonstrate DITTO's effectiveness.

Abstract

Digital twin models are of high interest to Head and Neck Cancer (HNC) oncologists, who have to navigate a series of complex treatment decisions that weigh the efficacy of tumor control against toxicity and mortality risks. Evaluating individual risk profiles necessitates a deeper understanding of the interplay between different factors such as patient health, spatial tumor location and spread, and risk of subsequent toxicities that can not be adequately captured through simple heuristics. To support clinicians in better understanding tradeoffs when deciding on treatment courses, we developed DITTO, a digital-twin and visual computing system that allows clinicians to analyze detailed risk profiles for each patient, and decide on a treatment plan. DITTO relies on a sequential Deep Reinforcement Learning digital twin (DT) to deliver personalized risk of both long-term and short-term disease outcome and toxicity risk for HNC patients. Based on a participatory collaborative design alongside oncologists, we also implement several visual explainability methods to promote clinical trust and encourage healthy skepticism when using our system. We evaluate the efficacy of DITTO through quantitative evaluation of performance and case studies with qualitative feedback. Finally, we discuss design lessons for developing clinical visual XAI applications for clinical end users.