Time to Surgery and Tumor Size: Predictors of LMD in Preoperative SRS

Summary: Historically, the standard treatment for large, resectable brain metastases has been surgery followed by whole brain radiotherapy (WBRT). Stereotactic radiosurgery (SRS) has replaced WBRT as the standard of care, but this approach has been linked to increased rates of leptomeningeal disease (LMD), potentially due to tumor seeding during resection without coverage of the meninges with radiation. Preoperative SRS may reduce LMD rates by devitalizing tumor cells before surgery. Our institution has collected data on patients treated with preoperative SRS to evaluate factors associated with LMD and local failure. Patients treated with preoperative SRS, receiving 15 Gy in a single fraction, between January 2011 and August 2023, including those from a phase 1 dose-escalation study were evaluated. Patient charts were reviewed for dates of treatment, follow-up and clinical outcomes. Cox proportional hazards models were used to analyze LMD and local failure with respect to potential risk factors, including time to surgery after SRS, tumor histology, size and location of the index metastasis, number of intracranial metastases, dose, and gross total resection (GTR). In total, 96 patients were included, with a median follow-up time of 9 months. Eleven patients developed LMD, and 14 experienced local failure. Six patients received 12 Gy as part of a dose escalation phase I trial. In the LMD analysis, increased time to surgery (HR 1.19, 95% CI 1.07-1.33, p=0.001) and index metastasis size (HR 3.26, 95% CI 1.43-7.43, p=0.005) were significantly associated with LMD (Table 1). No factors reached statistical significance for local failure, though GTR (HR 3.78, 95% CI 0.98-14.64, p=0.054) and histology of primary tumor (breast, NSCLC, melanoma, etc.) (HR 0.32, 95% CI 0.09-1.18, p=0.088) approached significance (Table 2). In this retrospective cohort, increasing tumor size and time from SRS to surgery were significantly associated with LMD development. These findings suggest the potential for a repopulation phenomenon, warranting further investigation, as no randomized data currently assess time to surgery as a predictor of LMD. We hypothesize that increasing dose of preoperative SRS may help to overcome this observation. While previous studies have identified histology, multiple metastases, and tumor location as predictors of LMD in postoperative SRS, none were significant in our analysis. GTR may be a protective factor for local failure, though further research is needed to clarify this relationship John Hoyle, MD (Presenting Author) - University of Alabama at Birmingham; James Markert, MD (Co-Author) - UAB Neurosurgery; Kristen Riley, MD (Co-Author) - UAB Neurosurgery; Samuel Marcrom, MD (Co-Author) - UAB Radiation Oncology; Christopher Willey, MD, PhD (Co-Author) - UAB Radiation Oncology; Richard Popple, PhD (Co-Author) - UAB Radiation Oncology; Markus Bredel, MD, PhD (Co-Author) - University of Miami Radiation Oncology; John Fiveash, MD (Co-Author) - UAB Radiation Oncology