2657 - Predicted vs. Observed Survival of Patients Undergoing Surgical Resection for Brain Metastases

Purpose/Objective(s): Surgery is important for palliation of large, symptomatic brain metastases. However, accurate prognostication is critical for patient selection. The updated diagnosis-specific graded prognostic assessment (DS-GPA) is an accepted tool for estimating survival of brain metastasis patients. However, its applicability to patients requiring surgery is uncertain. Here we compared survival outcomes of brain metastasis patients who underwent surgery and post-operative radiosurgery at a single institution to DS-GPA estimated predictions. We hypothesized that 1-year survival would be worse than DS-GPA predicted survival given the need for surgery.

Materials/Methods: Patients who underwent surgery and post-operative radiosurgery for a brain metastasis from 01/2018-06/2024 were included in the analysis. Only lung, breast, renal, melanoma, and gastrointestinal (GI) primary cancers (which permit DS-GPA calculation) were included. For each patient, GPA and eligibility quotient (EQ), the probability of surviving 1 year from the date of surgery, was calculated using the online calculator. Expected 1-year survival was calculated and compared with Kaplan-Meyer survival curves. Results were considered significant if expected 1-year survival fell outside the 95% confidence interval (CI) of observed 1-year survival.

Results: Thirty-nine patients were included. Median GPA was 2 (range 0.5-4). For the full cohort, 1-year survival was significantly greater than predicted (Table 1). Subset analysis by primary site showed trends toward better-than-predicted survival for breast, renal, and GI. Subset analysis by number of metastases showed significantly better-than-predicted survival for patients with solitary metastases but not for patients with >1 metastasis.

Conclusion: Surprisingly, observed survival was significantly better for patients undergoing brain metastasis surgical resection than DS-GPA predicted survival. This finding was not strongly associated with a specific primary site, but it was driven by patients with solitary brain metastases. DS-GPA was highly accurate for patients with >1 metastasis. These findings may help guide surgical decision making. Further, they suggest that differential improvements in prognosis for patients with solitary vs. multiple brain metastases have been made since the data informing DS-GPA was collected, possibly reflecting improvements in care for patients with oligometastatic cancer.