1070 - Genomic Alterations and Intracranial Recurrence Patterns after Melanoma Brain Metastasis Resection
Presenter(s)
J. Lee1, H. Walch2, M. Padilla Mazzeo1, M. I. Parker3, R. R. Patel3, A. Skakodub3, K. Yu4, N. S. Moss4, R. Bou-Nassif1, J. A. Wilcox1, C. Kinslow3, C. B. Jackson3, G. Cederquist3, B. S. Imber1, Y. Yu3, A. T. Ilica5, J. Smithy1, M. A. Postow1, N. Schultz2, and L. R. G. Pike1; 1Memorial Sloan Kettering Cancer Center, New York, NY, 2Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY, 3Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, 4Department of Neurosurgery, Memorial Sloan Kettering Cancer Center, New York, NY, 5Department of Neuroradiology, The University of Miami Miller School of Medicine, Miami, FL
Purpose/Objective(s): Melanoma has a high propensity for brain metastases, and patients who undergo surgical resection remain at significant risk for cavity recurrence and overall intracranial progression. We hypothesized that genomic features present in resected brain metastasis may be associated with CNS-specific outcomes in this population. We present the largest cohort to date examining the molecular determinants of resection cavity–specific and overall intracranial progression following craniotomy for melanoma brain metastases.
Materials/Methods: We included patients with melanoma brain metastases who underwent craniotomy of at least one brain metastasis, with next-generation genomic sequencing (MSK-IMPACT) of that tissue. Co-primary endpoints were time from craniotomy to CNS progression (TTCP) and time from craniotomy to local progression in the surgical cavity (TTLP). Radiographic endpoints were verified through direct inspection of contrast-enhanced MRI by a board-certified neuroradiologist and further independent review by a CNS radiation oncologist. Kaplan-Meier curves were used to estimate survival with cause-specific censoring, and the impact of specific genomic alterations on survival endpoints was evaluated using the log-rank test. Cumulative incidence at 1 and 2 years was estimated for altered and wild-type groups.
Results: A total of 145 patients were identified between 2014 and 2025. The median number of brain metastases was 2 (IQR: 1-3) and the median size of the resected brain metastasis was 3.5cm (IQR: 2.6-4.1). Following surgical intervention, the median TTCP and TTLP were 4.9 month and 8.8 months, respectively. Alterations in the PI3K pathway were associated with shorter TTCP (HR: 1.60; 95% CI: 1.04-2.47; p=0.03), with a 1-year incidence of 69% in the altered group vs 54% in the wild-type group. PIK3CA alterations demonstrated the strongest association (HR: 3.54; 95% CI: 1.28-9.82; p=0.01). B2M alterations were similarly associated with shorter TTCP (HR: 2.28; 95% CI: 1.09-4.79; p=0.02). In contrast, alterations in ATM were specifically associated with shorter TTLP (HR: 4.33, 95% CI 1.29-14.5; p=0.01), with a 1-year incidence of 18% in the wild-type group vs 60% in the wild-type group.
Conclusion: PI3K pathway alterations, driven by PIK3CA, are associated with shorter TTCP, while ATM alterations are implicated in cavity-specific recurrence after melanoma brain metastasis resection. These findings suggest that distant CNS progression and local cavity failure are distinct biologic processes and support the notion that brain metastasis-specific clinical and genomic features may inform risk stratification and adjuvant treatment strategy.