Concurrent Immunotherapy With Stereotactic Radiation for Brain Metastases is Not Associated With Increased Rates of Radionecrosis.

PURPOSE/OBJECTIVE(S) Early concerns were raised about higher rates of radionecrosis (RN) with concurrent use of immune checkpoint inhibition (ICI) and stereotactic radiotherapy (SRT) for brain metastases (BM). We sought to clarify if this same concern translated to linear accelerator-based SRT for BM. MATERIALS/METHODS We retrospectively reviewed cases of SRT delivered at a single institution for BM (2004-2019). In total, 251 patients and 723 courses of treatment were analyzed on a per-treatment basis for any grade RN. RN was graded by CTCAE v5 and defined as an enlarging lesion after SRT that resolved spontaneously, resolved with steroids, or was surgically resected and found to have no viable tumor. Differences were assessed with Chi-squared, Kaplan-Meier, and univariable/multivariable Cox regression analyses with candidate variables including timing of ICI with SRT (neoadjuvant, concurrent defined as within 3 months, or adjuvant), primary histology, and symptomatic vs not at time of BM diagnosis, among others. RESULTS In total, 251 patients received SRT for BM: 53 without ICI, 58 neoadjuvant ICI, 102 concurrent ICI, and 38 adjuvant ICI. Median follow up for the surviving patients who received any ICI was 24.0 months (IQR 7.6-41.4) and not different from those without ICI at 26.8 months (IQR 8.8-45.6) (P = 0.22) with 38% of patients alive at last follow up. Analyzed by timing of ICI at the patient level, those not receiving ICI were more likely to have a diagnosis of breast or other cancers as compared to melanoma, lung, and kidney primaries that were more common with ICI (P < 0.01). This is likely due to the longer historical use of ICI in these malignancies. The overall rates of Grade 1-3 RN were 4.0%, 4.6%, and 2.2% (no Grade 4+ events). Median time to RN was 11.4 months (IQR 6.0-22.5). Analyzed by timing of ICI on a per-treatment basis, there was no difference among rates of any grade RN for no (13.5%), neoadjuvant (9.8%), concurrent (8.5%), and adjuvant ICI (14.1%) (P = 0.052). On univariable analysis, primary histology, performance status, SRT fractionation, and metastasis diameter were significantly associated with time to RN (all P < 0.05). Time to RN did not vary among treatment courses with neoadjuvant, concurrent, adjuvant, or no ICI (P = 0.72). On multivariable analysis, once histology and SRT fractionation were included, only a performance status of ECOG 0 (HR 0.43 CI 0.23-0.81) and increasing diameter (HR 1.58 CI 1.24-2.03) remained significant (both P < 0.01). CONCLUSION In this large single-institution study of linear accelerator-based SRT, consistent with previous reports, we found that lesion diameter positively correlated with risk of RN. However, there was not an increased risk of RN in patients who received ICI. On subset analysis, there was no association of RN with ICI timing - including with concurrent administration. These data suggest that linear accelerator-based SRT is safe to combine with ICI. AUTHOR DISCLOSURE S.G. Allen: Employee; University of Michigan Mott Children's Hospital. J.J. Waninger: None. S.R. Birer: None. S. Journey: None. J. Skvarce: None. D.R. Wahl: Research Grant; Agios Inc, Innocrin Inc, American Cancer Society, NIH. Stock Options; Lycera Inc. Advisory Board Member; Agios Inc. T.S. Lawrence: None. J.A. Hayman: Research Grant; Blue Cross Blue Shield of Michigan. M.M. Kim: Research Grant; Blue Earth Diagnostics.; Red Journal. Site PI of ongoing CCTG CE.7, N0577, A071801, and NCI 9979 trials; NCTN (Alliance, NRG, CCTG) and ETCTN.M. Green: None.