Selective salvage radiotherapy could provide favorable outcomes in patients with refractory or relapsed primary central nervous system lymphoma.

Introduction: To date, there is no relevant data supporting the role of salvage radiotherapy (sRT) in patients with refractory or relapsed primary central nervous system lymphoma (PCNSL). Herein, we aimed to investigate the impact of sRT in patients with refractory or relapsed PCNSL following upfront HD-MTX.

Methods: We retrospectively reviewed 89 patients who had refractory (n = 16) or recurrent disease after an initial favorable response (n = 73); among them, 41 were treated with sRT and 48 were treated without sRT (nRT). Event-free survival (rEFS) and overall survival (rOS) after first recurrence were considered from the date of recurrence to date of each event.

Results: Overall, the first failure was diagnosed at a median of 11.0 months [interquartile range (IQR), 5.6-26.4] after first treatment. More than half of the patients had recurrent disease involving initial tumor bed (n = 47), deep structure (n = 67), and multiple lesions (n = 58). Among 19 patients who were initially treated with 23.4 Gy of whole brain RT, 10 patients received sRT as a re-irradiation; other 31 patients in sRT group were RT naïve patients. There was no significant difference in tumor characteristics between sRT and nRT group. Overall and complete response after salvage treatment were 80% and 48%, respectively; sRT provided higher overall response rate than nRT (93% vs. 69%, p = 0.011). With a median follow-up of 14.3 months (IQR, 7.9-31.4), 2-year rEFS and rOS rates were 27% and 57%, respectively. There were no differences in rEFS and rOS according to sRT (sRT vs. nRT, 26% vs. 28%, p = 0.730; 63% vs. 50%, p = 0.690). Poor performance, recurrence interval < 8 months, and unfavorable response following salvage treatment were associated with inferior rEFS and rOS. Additionally, sRT and stem cell transplantation improved response rate independently after multivariate analysis for complete/partial response.

Conclusions: We found favorable response rate and comparable survival outcomes following sRT compared with non-local treatments for patients with refractory/relapsed PCNSL. Further studies of patient selection could stratify patients who can benefit from sRT.