SRDA-02 DEVELOPMENT OF A SCORING SYSTEM TO PREDICT LOCAL RECURRENCE IN BRAIN METASTASES FOLLOWING COMPLETE RESECTION AND OBSERVATION

Abstract BACKGROUND Postoperative stereotactic radiosurgery (SRS) to the resection cavity in patients with brain metastases is guideline-recommended therapy. However, Japanese Clinical Oncology Group 0504 study showed that postoperative observation could be a therapeutic option in patients with completed resected brain metastases. However, clear criteria for choosing between SRS and observation remain undefined. This study aimed to identify factors associated with local recurrence following observation alone and to develop a scoring system for predicting recurrence risk. METHODS We retrospectively analyzed 53 patients (54 lesions) who had no prior radiation, underwent complete resection, and were managed with observation between January 2016 and December 2021. Variables assessed included age, primary cancer type, preoperative Karnofsky Performance Status (KPS), primary tumor control, extracranial disease status, postoperative systemic therapy, tumor size, and resection technique. RESULTS The median tumor diameter was 34 mm, and 72.2% underwent piecemeal resection. Local recurrence occurred in 37 of 54 tumors (68.5%), with a median time to local recurrence of 5.1 months (range: 0.5-34.4 months). Univariate and multivariate analyses revealed that non-lung adenocarcinoma, infratentorial tumors, and no postoperative systemic therapy were identified as risk factors for local recurrence (non-lung adenocarcinoma, p = 0.035; infratentorial tumors, p = 0.044; and no postoperative systemic therapy, p = 0.0069). Assigning one point to each factor, patients with scores of 2–3 had a significantly shorter median time to recurrence (2.1 months) compared to those with scores of 0–1 (30.8 months; p=0.0002). CONCLUSION Local recurrence is common after complete resection followed by observation. Non-lung adenocarcinoma, infratentorial tumors, and no postoperative systemic therapy were risk factors for local recurrence. Our scoring system can predict local recurrence, thus potentially guiding treatment decisions.