SAT-097 Assess Biomarkers with a Machine Learning Approach to Develop a Predictive Calculator for First-generation SSA Resistance in Acromegaly

Abstract Disclosure: W. Lin: None. S. Shi: None. Y. Zheng: None. E.R. Laws: None. T. Smith: None. L. Min: None. Context: Acromegaly, caused by excess growth hormone (GH) and insulin-like growth factor-1 (IGF-1) due to pituitary adenomas, often necessitates somatostatin analog (SSA) therapy when surgery is insufficient. However, responses to SSA therapy vary widely among patients. Objective: To develop a machine learning (ML)-based clinical calculator that predicts individual responses to SSA therapy, enabling evidence-based decision-making in acromegaly management. Design: A retrospective study (January 2010–July 2024) utilizing the Research Patient Data Registry (RPDR) to evaluate ten ML algorithms and create a predictive calculator. Setting: Single-center study conducted at Mass General Brigham-affiliated hospitals. Patients: 111 acromegaly patients met inclusion criteria, classified as SSA-responsive (n = 64) or SSA-resistant (n = 47). Interventions: IGF-1 trajectories were analyzed using linear mixed-effects modeling. Ten ML algorithms, including CatBoost and XGBoost, were assessed to predict SSA resistance. SHAP analysis identified key predictors for the development of a web-based clinical calculator. Main Outcome Measures: Model performance was evaluated using AUROC, accuracy, precision, recall, specificity, F1 score, and decision curve analysis. Results: The CatBoost model exhibited optimal performance with AUROC 0.896 (95% CI: 0.751–0.990), accuracy 82.4%, precision 86.7%, specificity 88.2%, and F1 score 81.2%. Key predictors of SSA resistance identified via SHAP analysis included pre-SSA treatment GH, Knosp grade, pre-SSA treatment IGF-1 index, T2-weighted MRI density, and comorbidity burden. The model demonstrated excellent calibration (Brier score 0.131) and clinical utility via decision curve analysis. A web-based calculator was developed for clinical use. Conclusions: The CatBoost-based calculator effectively predicts SSA treatment response in acromegaly patients. Prospective validation is required before clinical implementation. Presentation: Saturday, July 12, 2025