Predictors of neurological outcomes after resection of brain metastases: a cohort study and classification for patient selection
Chia-En Wong1,2, Yu Chang1, Chih-Hao Tien1
11Division of Neurosurgery, Department of Surgery, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan.
Objective:
Neurological outcomes after surgery for brain metastasis (BM) are critical for patient survival. Patient selection is vital to avoid neurological deterioration and preserve function. The aim of this study was to investigate outcome predictors following BM surgery and to propose a simple classification system to guide patient selection for BM surgery.
Methods:
This retrospective single-center study included 500 consecutive patients who underwent resection of a single BM between January 2013 and June 2023. Patients who underwent resection of a single BM between July 2023 and June 2024 were included in the validation cohort. Neurological outcomes were assessed as changes in modified Rankin Scale scores within 6 months. Patient and radiographic predictors of neurological outcomes were investigated. A BM score consisting of tumor eloquence, depth, and hemorrhage status at presentation was constructed to predict neurological outcomes after BM surgery.
Results:
Neurological deterioration was observed in 122 of 500 patients (24.4%). Independent predictors of neurological deterioration included eloquence (p < 0.001), near eloquence (p = 0.037), tumor depth ≥ 1 cm (p < 0.001), absence of tumor hemorrhage (p = 0.009), and a recursive partitioning analysis (RPA) class of III (p = 0.004). Neurological deterioration occurred in 8%, 13%, 28%, 51%, and 87% of patients with BM scores of 1, 2, 3, 4, and 5, respectively. Receiver operating characteristic (ROC) analyses demonstrated that the BM score (area under the curve [AUC] = 0.739) predicted neurological outcomes better than the RPA class (AUC = 0.642). In the validation cohort, ROC analyses showed AUCs of 0.707, 0.867, and 0.875 for the RPA class, BM score, and RPA class and BM score combined, respectively.
Conclusions:
Tumor eloquence, deep location, and hemorrhagic presentation were significant independent predictors of neurological outcomes following BM surgery. The BM score can predict neurological outcomes and refine patient selection for BM surgery.
