Radiomic modeling to predict risk of vertebral compression fracture after stereotactic body radiation therapy for spinal metastases

OBJECTIVE In the treatment of spinal metastases with stereotactic body radiation therapy (SBRT), vertebral compression fracture (VCF) is a common and potentially morbid complication. Better methods to identify patients at high risk radiation-induced VCF are needed evaluate prophylactic measures. Radiomic features from pretreatment imaging may be employed more accurately predict VCF. The objective this study was develop machine learning model based on clinical characteristics radiomic after SBRT for metastases. METHODS Vertebral levels C2 through L5 containing treated were included if they naive prior surgery or therapy, target delineation consensus guidelines, 1-year follow-up data available. Clinical features, including patient, disease, treatment, obtained chart review. extracted planning volume (PTV) CT T1-weighted MRI. selected by least absolute shrinkage selection operator (LASSO) regression in random forest classification models, which trained within 1 year SBRT. Model performance assessed leave-one-out cross-validation. RESULTS Within SBRT, 15 95 analysis demonstrated new progressive Selected BMI, status, total prescription dose, dose 99% PTV, lumbar location, 2 components Spine Instability Neoplastic Score (SINS): lytic tumor character misalignment. 5 3 best-performing model, derived combination sensitivity 0.844, specificity 0.800, area under receiver operating characteristic (ROC) curve (AUC) 0.878. This significantly accurate than alternative models only (AUC = 0.795, p 0.048) SINS 0.579, < 0.0001). CONCLUSIONS incorporating both predicted excellent specificity, outperforming developed alone. If validated, these findings allow judicious interventions.