2578 - Time-Dependent Interactions between Immune Checkpoint Blockade and Stereotactic Radiosurgery for Brain Metastases

A. Mittal¹, A. Green¹, C. Phillips¹, N. Haghighi¹, N. Plumridge¹, M. Li¹, A. Wirth¹, and J. Sia^1,2; ¹Peter MacCallum Cancer Centre, Melbourne, Australia, ²Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, VIC, Australia

Purpose/Objective(s): Whether immune checkpoint blockade (ICB) and stereotactic radiosurgery (SRS) for brain metastases (BrMs) have a time window for synergistic efficacy and toxicity is unclear. We hypothesized that an anti-tumor interplay between ICB and SRS, if present, would be reflected in an impact of time between an ICB cycle and SRS (tbICB-SRS) on clinical outcomes. To test this, we retrospectively examined a large cohort of SRS+ICB patients for correlation of tbICB-SRS with local control (LC), distant brain control (DBC), overall survival (OS), and radionecrosis (RN) rates. Materials/

Methods: Patients who received ICB within a year of LINAC or a non-invasive stereotactic radiosurgery instrument SRS at a single center from 2018-2023 were identified and screened for the following exclusion criteria: prior whole brain radiotherapy, prior SRS to the same BrM, cavity SRS, and ICB starting >30 days post SRS (to avoid survivor bias). LC (per RANO-BM), DBC (absence of new BrMs or leptomeningeal disease), OS, and any-grade RN were analyzed at per-lesion, per-treatment episode, per-patient, and per-lesion levels respectively by Kaplan Meier, and compared between subgroups by log-rank test. Cox regression was used for uni/multivariate analysis (UVA/MVA).

Results: 125 patients, 161 episodes and 368 BrMs were analyzed. Median age at first SRS episode was 64 years (range 31-91). 45% and 38% of patients had melanoma and non-small cell lung cancer (NSCLC). Median follow-up was 24.8 months. LC and OS curves were significantly separated by tbICB-SRS (=7 vs 8-30 vs >30 days) (p=0.002, p=0.04). On UVA of patients whose tbICB-SRS =30 days, LC and OS were inversely correlated with tbICB-SRS as both continuous and ordinal variables (Table 1), but not with ICB-SRS sequencing. OS was also associated with ECOG status (p=0.02), pre-SRS hemoglobin (HR 0.98, p=0.01), dexamethasone dose (HR 1.10, p=0.001) and small cell lung cancer histology (HR 16.06, p=0.03). tbICB-SRS remained predictive for LC and OS when adjusted for histology and ICB type on MVA (p=0.001, p=0.002). Overall, no factors predicted for DBC, but in patients on dexamethasone <4mg, higher lymphocyte count (HR 0.57, p=0.04) and shorter tbICB-SRS (HR 1.04, p=0.06) correlated with better DBC. RN was not associated with tbICB-SRS, but with BrM size (HR 1.06, p=0.01) and number of ICB cycles received pre-SRS (HR 1.04, p=0.02).

Conclusion: A reduced ICB-SRS interval correlated with improved LC and OS, suggesting a time-sensitive biological interaction. This relationship was not overridden by histology. A weaker correlation with DBC was also seen in patients with minimal dexamethasone use. RN was associated with different risk factors. Abstract 2578 – Table 1

	tbICB-SRS
	Per day increase			8-30 vs =7 days
	HR	95% CI	p	HR	95% CI	p
LC	1.04	1.01, 1.08	0.025	2.83	1.53, 5.21	0.001
DBC	1.02	0.99, 1.06	0.149	1.36	0.79, 2.34	0.271
OS	1.04	1.01, 1.07	0.010	1.78	1.02, 3.11	0.042
RN	0.97	0.90, 1.05	0.481	0.90	0.32, 2.48	0.832