Integrating Multi-Modal Data for Predicting Post-Surgical Pain Outcomes: Towards Personalized Pain Management

PROJECT SUMMARY/ABSTRACT Post-surgical pain (P-SP) affects up to 30% of U.S. patients and up to 50% for high-risk surgeries like thoracotomies and amputations, with global incidence ranging from 5% to 85%. P-SP imposes an economic burden of $560-$635 billion annually, including direct medical expenses and indirect costs like lost productivity and disability claims. Current treatment relies on generalized pain management protocols, leading to poor outcomes, higher chronic pain risk, and increased opioid use. A more personalized approach is urgently needed to improve treatment effectiveness by accounting for individual pain sensitivity and psychological factors. To elucidate the neural and psychological mechanisms underlying P-SP and to facilitate early identification of patients at risk for chronic pain, enabling more targeted interventions, the proposed secondary analysis project will develop an interpretable and predictive model for P-SP outcomes by 1) identifying the most predictive features of P-SP outcomes using multi-modal data (functional neuroimaging, psychological, behavioral, and medical history); 2) applying Shapley value analysis to the output of supervised machine learning models to interpret the contributions of each feature to the predictions, providing clinicians with patient-specific insights to improve pain management and support personalized treatment planning; 3) developing a time-series model to predict P-SP outcomes and 4) applying Shapley value analysis to identify key temporal patterns that influence these predictions. The expected outcomes of this highly innovative project include the identification of reliable biomarkers for P-SP, which will serve as a foundation for developing individualized pain management strategies. The findings have the potential to transform post-surgical care by improving the early detection of P-SP, optimizing treatment protocols, and reducing opioid use. Ultimately, this project aligns with the broader movement toward personalized medicine and precision healthcare, offering a comprehensive solution for enhancing patient outcomes and alleviating the long-term burden of chronic pain on the healthcare system. This directly supports one of the primary missions of the National Institute of Biomedical Imaging and Bioengineering (NIBIB). 2