Stereotactic radiosurgery (SRS) for non-small cell lung cancer (NSCLC) in the oncogene-addicted era: A real-world UK tertiary center experience.

Abstract

e14011 Background: 30-50% of NSCLC patients develop brain metastases (BM). Re-evaluation of the role of SRS is required with evolving understanding of molecular subtypes of NSCLC and increasing intracranial efficacy of systemic agents. Methods: We conducted a retrospective analysis of NSCLC patients receiving SRS at our centre from 05/16-11/23. Primary objective was median overall survival (mOS months). Analysis was performed on Prism 10.1.1. Multivariate analysis was conducted using Cox Proportional Hazards Regression, with age, gender, KPS, BM size, total BM volume and oncogene status as predictor variables. Extraction from medical records was undertaken by authors, with local R&D committee approval. Results: 237 patients were evaluable. 59% (N=139) were oncogene-addicted (oNSCLC) and 41% (N=98) non-oncogene-addicted (non-oNSCLC). Median duration of follow up was 43.3m. m OS was 11.9m overall. For oNSCLC mOS was 13.1m vs 8.8m for non-oNSCLC (HR 1.49, P=0.02). mOS was 20.7m for EGFR sensitizing NSCLC (N=48) vs 10.5m without (HR 1.40 P=0.12). mOS for ALK+ (N=22) was 29.3m vs 11.6m for non-ALK+ (HR 1.90, P=0.05). mOS for KRAS+ (N=32) was 11.6m vs 12.0m without (HR 0.91, P=0.68). Significantly inferior mOS was seen for single (N=186 [78%]) vs multiple (N=51[22%]) SRS courses (8.4m vs 38.8m, [HR 0.34 P=<0.0001]). SRS at diagnosis of BM (N=118) had a mOS of 12.3m vs at nadir (N=9, not reached [HR0.44, P=0.18]), progression on 1st line TKI (N=31, 22.1m [HR 1.1, P=0.79]) and at any other time (N=157, 7.8m [HR 1.33, P=0.11 ]). There was no significant difference in survival for patients diagnosed with brain metastases at the diagnosis of metastatic disease compared to those who developed brain metastases during their disease course (12.3m vs 10.6m [HR 1.18, P=0.33]). Survival of those with 1 site of metastatic disease (18.3m) was numerically improved compared to those with 2 (10.7m, HR 1.26, P=0.25) sites, and statistically significant compared to those with 3 (10.1m, HR 1.69, P<0.02) and ≥4 (4.4m, HR 4.41, P<0.0001) sites. Conclusions: Median OS in oNSCLC was superior to non-oNSCLC, driven by numerically superior survival in EGFR+ and ALK+ cohorts. Significantly improved OS was observed in patients with multiple vs single SRS treatments, however this may be confounded by biases. Patients treated with SRS at nadir had numerically superior OS to those treated at other timepoints, however this must be interpreted with caution due to small sample size. Further prospective research is needed to interrogate these findings and guide local therapies in NSCLC with BM. [Table: see text]