Radionecrosis after repeated courses of radiotherapy under stereotactic conditions for brain metastases: Analysis of clinical and dosimetric data from a retrospective cohort of 184 patients

Abstract

Purpose: Between 10 and 40% of patients with cancer will develop one or more brain metastases (BMs). Stereotactic radiotherapy (SRT) is part of the therapeutic arsenal for the treatment of de novo or recurrent BM. Its main interest is to delay whole brain radiation therapy (WBRT), which may cause cognitive toxicity. However, SRT is not exempt from long-term toxicity, and the most widely known SRT is radionecrosis (RN). The objective of this study was to analyze the occurrence of RN per BM and per patient.

Material and methods: Between 2010 and 2020, data from 184 patients treated for 915 BMs by two to six SRT sessions for local or distant brain recurrence without previous or intercurrent WBRT were retrospectively reviewed. RN was examined on trimestral follow-up MRI and potentially confirmed by surgery or nuclear medicine. For each BM and SRT session plan, summation V_12Gy, V_14Gy, V_21Gy and V_23Gy isodoses were collected. Volumes of intersections were created between the 12Gy isodose at the first SRT and the 18Gy isodose of the following SRT (V_18-12Gy).

Results: At the end of follow-up, 23.0% of patients presented RN, and 6.3% of BM presented RN. Median follow-up of BM was 13.3 months (95%CI 18.3-20.8). The median interval between BM irradiation and RN was 8.7 months (95% CI 9.2-14.7). Six-, 12- and 24-month RN-free survival rates per BM were 75%, 54% and 29%, respectively. The median RN-free survival per patient was 15.3 months (95% CI 13.6-18.1). In multivariate analysis, the occurrence of RN per BM was statistically associated with local reirradiation (P<0.001) and the number of SRTs (P<0.001). In univariate analysis, the occurrence of RN per patient was statistically associated with the sum of all V_18-12Gy (P=0.02). No statistical association was found in multivariate analysis. A sum of all V_18-12Gy of less than 1.5ml was associated with a 14.6% risk of RN, compared with 35.6% when the sum of all V_18-12Gy was superior to 1.5ml. The sum of all V_18-12Gy larger than 1.5ml was associated with a 74% specificity and 53% sensitivity of RN (P<0.001).

Conclusion: Based on these results, a small number of BMs show RN during repeated SRT for local or distant recurrent BMs. Local reirradiation was the most predictive factor of brain RN. A V_18-12Gy larger than 7.6ml in the case of local reirradiation or larger than 1.5ml in proximity reirradiation were prognostic factors of RN. The more BM patients need radiation therapy, and the longer they survive after irradiation, the higher their individual risk of developing RN.

Keywords: Brain metastases; Métastases cérébrales; Radionecrosis; Radionécrose; Repeated radiosurgery; Réirradiation; Salvage radiation; Stereotactic radiosurgery; Stéréotaxie; Toxicity; Toxicité; Traitement de rattrapage.