73. PREDICTION OF A SECOND LOCAL RECURRENCE IN SURGICALLY TREATED RECURRENT BRAIN METASTASES PATIENTS

Abstract BACKGROUND Local recurrence is a common occurrence after resection or radiotherapy for brain metastasis (BM). Very little is known about the benefit of (re-)craniotomy in this scenario: does resecting the initial local recurrence (LR1) invariably lead to a second local recurrence (LR2)? This study aimed to analyze occurrence and predictors of LR2 in BM patients undergoing craniotomy for LR1. METHODS Patients were identified from a departmental database at the Brigham and Women’s Hospital, Boston, MA. Multivariable logistic regression and cox regression analysis was performed to identify predictors of binary occurrence of LR2 (yes/no) and time-to-LR2, respectively. Based on predictors, subgroup-specific prevalence of LR2 was explored. RESULTS A total of 188 patients were identified. The median age was 59.5 years and 117 patients (62.2%) were female. Treatment-wise, 64 patients (34.0%) underwent subtotal resection (STR) and 66 (35.1%) received adjuvant radiation. Eighty-one (43.1%) patients experienced LR2 at a median of 7 months after craniotomy. Occurrence of LR2 was significantly associated with STR (OR 6.88, p = 0.0008), surgery as treatment for LR1 (OR = 0.26, p = 0.03), larger tumor volume (OR = 1.14 per 1000 mm3, p = 0.01), and frontal location (OR = 5.23, p = 0.02). Shorter time-to-LR2 was associated with STR (HR = 5.31, p = 0.01) and adjuvant radiation (HR = 2.22, p = 0.03), while temporal (HR = 0.16, p = 0.03) and parietal (0.13, p = 0.03) location were associated with longer time-to-LR2. When stratifying by extent of resection, prevalence of LR2 was 32.8% after gross total resection and 57.1% after STR. CONCLUSION In this population, LR2 occurred in 43.1% of patients. STR was the strongest risk factor for LR2, while tumor size, location, surgical treatment of LR1, and receipt of adjuvant radiation may also influence subsequent recurrence.