For a growing number of cancer survivors, ringing the bell at the end of primary treatment marks a transition into a complex phase of care that is often less structured and harder to predict. Even after therapy concludes, patients may experience lingering physical symptoms, emotional distress or other unexpected medical needs that lead to emergency department or urgent care visits or even hospitalizations, along with worsening symptom burden.
A new multidisciplinary study from Sylvester Comprehensive Cancer Center, part of the University of Miami Miller School of Medicine, suggests that the key to anticipating those outcomes may lie in closely examining electronic health records and patient-reported data systematically, using novel artificial intelligence (AI) technologies.
Read the full article at Inventum.
Congratulations to IDSC Authors Jerry Bonnell, Mitsunori Ogihara, and Ravi Vadapalli, and their co-authors on the publication of “Machine Learning Risk Stratification Approach Using Patient-Reported Outcomes for Forecasting Unplanned Health Care Use and Symptom Burden in Cancer Survivors” in the JCO Clincial Cancer Informatics Journal.
Abstract
Purpose
Effective risk stratification in cancer survivorship requires handling longitudinal data characterized by multimodal inputs, irregular follow-up, and recurrent clinical events. This study evaluated the incremental value of integrating patient-reported outcomes (PROs) with electronic health record (EHR) data and identified optimal windowing strategies for machine learning–based prediction of adverse survivorship outcomes.
Patients and Methods
This study used a cohort of 25,592 cancer survivors followed for 36 months. Data from four domains were integrated: baseline measures, treatments, PROs, and health care utilization (emergency room visits and hospitalizations). Two classification models, LASSO and CATBOOST, were applied across modality combinations and five temporal representations of patient history: static early-phase (0-6 months), cumulative history, sliding windows (4- and 12-month), and a most-recent baseline. Performance was evaluated for predicting monthly health care utilization and patient-reported symptom burden using average precision (AP). SHapley Additive exPlanations (SHAP) analysis identified key predictors and characterized their evolving influence.
Results
For health care utilization, CATBOOST models trained on the full multimodal data set with time-windowed predictors achieved strong discrimination (AP = 0.207), outperforming static baselines by 27%. SHAP analyses emphasized dynamic contributions from recent utilization and treatment toxicity. For symptom burden, PRO integration was crucial, nearly doubling clinical-only performance (AP = 0.132 v 0.071), with longer historical context improving characterization of progressive functional decline and symptom severity. Flagging the top 10% of patients by predicted risk captured 51.7% of health care utilizations and 46.7% of symptom burden events.
Conclusion
Adverse survivorship risk is dynamic and outcome-specific: acute health care utilization is best predicted by recent clinical momentum, while longitudinal patient-reported trends drive symptom burden. Implementing decoupled, dynamic windows provides a flexible framework for risk stratification and risk prediction beyond standard clinical heuristics, facilitating proactive, precision-based survivorship care.
__________
Akina Natori, Jerry R. Bonnell, Vasileios Stathias, Sara E. Fleszar-Pavlovic, Mitsunori Ogihara, Andrew Wang, Ravi Vadapalli, Blanca Silvia Noriega Esquives, Tracy E. Crane, Frank J. Penedo. Machine Learning Risk Stratification Approach Using Patient-Reported Outcomes for Forecasting Unplanned Health Care Use and Symptom Burden in Cancer Survivors. JCO Clin Cancer Inform 10, e2500389(2026).
DOI:10.1200/CCI-25-00389
