2640 - Evaluating Neurologic Causes of Death in Melanoma Patients with Brain Metastasis Treated with SRS and Immunotherapy: A Retrospective Cohort Study

Purpose/Objective(s): Melanoma brain metastases have traditionally had a high rate of death from neurologic causes due to hemorrhage, leptomeningeal spread, or local and distant brain failure. The advent of immunotherapy (I/O) has improved survival and decreased neurologic death in metastatic melanoma patients, including patients with brain metastases. It remains unknown, however, what are the causes of death in melanoma brain metastasis patients in the I/O era.

Materials/Methods: A single institution retrospective review was conducted to assess patients with brain metastases from melanoma for causes of death. Eligible patients were identified using our departmental radiosurgery (SRS) database. Patient demographics, tumor characteristics, and clinical outcomes were extracted from the EMRs. Neurologic death was defined as patients dying with progressive neurologic decline or patients with severe neurologic dysfunction dying from intercurrent disease evidenced on MRI or CT scans. Neurologic deaths were then subclassified into the following causes: leptomeningeal disease, local failure, distant failure, and hemorrhage. Baseline characteristics were compared using Pearson’s chi-squared test for categorical variables and Kruskal-Wallis test for continuous variables. Incidence of neurologic death and subclassifications of neurologic death between patients treated with I/O within 3 months after first SRS treatment and those not treated with I/O were compared using the log-rank test.

Results: Between 2000 and 2019, a total of 140 patients with melanoma brain metastases were treated with SRS. Of these patients, 24 (17.1%) were treated with I/O. The mean duration of follow-up from SRS was 13 months. The mean age of patients was 58 years old. At initial SRS, patients had a mean KPS of 79 and a mean of 3 brain metastases. The I/O group had a significantly higher number of brain metastases (5 vs 2.5, p<0.001) and proportion of patients alive at end of study follow-up (29.2% vs 3.4%, p<0.001) compared to patients without I/O. Cumulative incidence of neurologic death for all patients in the series was 20.7%, 32.9% and 39.3% at 6, 12 and 24 months. Cumulative incidence of neurologic death was 16.7%, 16.7% and 20.8% at 6, 12 and 24 months for patients receiving I/O and 21.6%, 36.2% and 42.2% at 6, 12 and 24 months for patients not receiving I/O (p=0.01). The proportion of patients not receiving I/O who died of leptomeningeal disease, local failure, distant failure, and hemorrhage compared to patients receiving I/O was not significantly different (13%, 4%, 44% and 39% vs 14%, 0%, 43% and 43%, p=0.43).

Conclusion: I/O has decreased the incidence of neurologic death in patients with melanoma brain metastases treated with SRS. This decrease in neurologic death is not driven by a particular etiology of disease progression. Despite the higher burden of brain metastasis at initial SRS in patients receiving I/O in our study, we observed improved clinical outcomes without a change in the distribution of neurologic causes of death.