Impact of Fractionation and Tumor Size on Radiation Necrosis and Local Control in Patients With Non-Small Cell Lung Cancer Brain Metastases Treated With Stereotactic Radiosurgery and Concurrent Immunotherapy

Document Type

Conference Proceeding

Publication Date

9-2025

Publication Title

International Journal of Radiation Oncology, Biology, Physics

Abstract

Purpose/Objective(s): To evaluate radiation necrosis (RN) and local recurrence (LR) rates in patients with brain metastases (BM) from non-small cell lung cancer (NSCLC) treated with single-fraction (SF) or multi-fraction (MF) stereotactic radiosurgery (SRS) and concurrent immunotherapy. Materials/Methods: We identified 147 NSCLC patients with 836 brain metastases who underwent SRS with concurrent immunotherapy between 2018 and 2024, with a minimum follow-up of 6 months, using a prospectively maintained, multicenter single-institution database. Primary outcomes included RN and LR rates, stratified by fractionation and tumor size. Kaplan-Meier and cumulative incidence analyses assessed outcome differences, while univariable (UVA) and multivariable (MVA) Cox proportional hazards models were constructed to identify clinical and dosimetric predictors of RN and LR. Receiver operating characteristic (ROC) analysis was used to determine tumor volume and diameter thresholds for risk stratification. Results: A total of 784 lesions (93.8%) were treated with SF-SRS. The median (IQR) tumor diameter and volume were 0.6 cm (0.5-1.0) and 0.07 cc (0.03-0.3) in the SF group, compared to 2.8 cm (2.4-4.1) and 6 cc (4.2-12) in the MF group (p< 0.001). Fractionated treatments were either 8-9 Gy x 3 fractions or 6 Gy x 5 fractions while the median marginal dose in the SF arm was 19 Gy, to a median isodose line of 50%. Prior surgical resection was performed in 1% of SF cases and 38% of MF cases (p< 0.001). Immunotherapy regimens included pembrolizumab (78%), durvalumab (10%), and a combination of nivolumab and ipilimumab (5%), with 70% of patients receiving immunotherapy before and after SRS, similarly distributed across fractionation groups. RN occurred in 1.7% of all lesions (1.4% SF vs. 5.8% MF, p=0.051), while LR occurred in 2.5% of lesions (2.6% SF vs. 1.9% MF, p>0.99). On both UVA and MVA, increasing tumor diameter and volume were significantly associated with RN and LR. ROC analysis identified tumor volume >0.67 cc (7.5% vs. 0.3%, p< 0.001, AUC = 0.91) and diameter >1.14 cm (6.0% vs. 0.3%, p< 0.001, AUC = 0.90) as predictors of higher RN risk. Similarly, tumor volume >0.1 cc (4.4% vs. 0.9%, p= 0.002, AUC = 0.72) and diameter >0.68 cm (4.8% vs. 0.5%, p< 0.001, AUC = 0.74) correlated with increased LR risk. At 24 months, cumulative incidence of LR was significantly higher in tumors >2 cm treated with SF-SRS (11%) compared to ≤2 cm SF-SRS (1.8%) and MF-SRS (2.2%) (p< 0.001). Similarly, the cumulative incidence of RN at 24 months was 21% in tumors >2 cm treated with SF-SRS, significantly higher than 0.7% in the ≤2 cm SFSRS and 6.5% in the MF-SRS groups (p< 0.001). Conclusion: SRS demonstrated good efficacy with very low RN and LR rates for small lesions. However, increasing tumor diameter and volume strongly predicted RN and LR. For tumors >2 cm, SF-SRS significantly increased RN and LR rates, which were both improved with the use of MFSRS. More importantly, concurrent immunotherapy did not appear to increase rates of RN above historical standards.

Volume

123

Issue

1 Suppl

First Page

e289

Last Page

e289

Comments

American Society for Radiation Oncology ASTRO 2025: 67th Annual Meeting, September 27 - October 1, 2025, San Francisco, CA

DOI

10.1016/j.ijrobp.2025.06.2169

Link to Full Text

Share

COinS