Radiomics-based prediction of local control in patients with brain metastases following postoperative stereotactic radiotherapy

Abstract Background Surgical resection is the standard of care for patients with large or symptomatic brain metastases (BMs). Despite improved local control after adjuvant stereotactic radiotherapy, the local failure (LF) risk persists. Therefore, we aimed to develop and externally validate a pre-therapeutic radiomics-based prediction tool to identify patients at high LF risk. Methods Data were collected from A Multicenter Analysis of Stereotactic Radiotherapy to the Resection Cavity of Brain Metastases (AURORA) retrospective study (training cohort: 253 patients (two centers); external test cohort: 99 patients (five centers)). Radiomic features were extracted from the contrast-enhancing BM (T1-CE MRI sequence) and the surrounding edema (FLAIR sequence). Different combinations of radiomic and clinical features were compared. The final models were trained on the entire training cohort with the best parameters previously determined by internal 5-fold cross-validation and tested on the external test set. Results The best performance in the external test was achieved by an elastic net regression model trained with a combination of radiomic and clinical features with a concordance index (CI) of 0.77, outperforming any clinical model (best CI: 0.70). The model effectively stratified patients by LF risk in a Kaplan-Meier analysis (p < 0.001) and demonstrated an incremental net clinical benefit. At 24 months, we found LF in 9% and 74% of the low and high-risk groups, respectively. Conclusions A combination of clinical and radiomic features predicted freedom from LF better than any clinical feature set alone. Patients at high risk for LF may benefit from stricter follow-up routines or intensified therapy. Key points Radiomics can predict the freedom from local failure in brain metastasis patients Clinical and MRI-based radiomic features combined performed better than either alone The proposed model significantly stratifies patients according to their risk Importance of the Study Local failure after treatment of brain metastases has a severe impact on patients, often resulting in additional therapy and loss of quality of life. This multicenter study investigated the possibility of predicting local failure of brain metastases after surgical resection and stereotactic radiotherapy using radiomic features extracted from the contrast-enhancing metastases and the surrounding FLAIR-hyperintense edema. By interpreting this as a survival task rather than a classification task, we were able to predict the freedom from failure probability at different time points and appropriately account for the censoring present in clinical time-to-event data. We found that synergistically combining clinical and imaging data performed better than either alone in the multicenter external test cohort, highlighting the potential of multimodal data analysis in this challenging task. Our results could improve the management of patients with brain metastases by tailoring follow-up and therapy to their individual risk of local failure.