This multi-center retrospective study evaluated different radiotherapy treatments for brain metastases in SCLC patients. Compared to WBRT alone, WBRT+boost was associated with superior OS and iPFS both in the entire cohort (n = 263) and matched cohort (n = 140). Patients who treated with WBRT+boost also experienced a longer OS than did SRS in matched cohort (n = 74).
Due to the properties of multiple metastases, WBRT was considered the standard treatment for brain metastases from SCLC patients. However, the optimal treatment for brain metastases remained controversial. Some published studies suggested that WBRT plus radiation boost was more suitable for the treatment of brain metastases in SCLC patients than WBRT alone. Sun et al. reported that, compared with WBRT alone, a longer survival was observed in patients who received WBRT plus radiation boost than did WBRT alone (13.4 vs 8.5 months; p = 0.004). To minimize the difference in the number of BMs between the two groups, a subgroup analysis was performed. Among patients with 1–3 brain metastases, WBRT plus radiation boost was also associated with longer OS than WBRT alone (13.4 vs 9.6 months; p = 0.022) [14]. Their findings were similar to ours. Wegner et al. reported that the longer OS was observed in SCLC patients who received WBRT plus SRS than did SRS alone (14 vs 6 months, p = 0.040). Of note, only 6 patients received WBRT plus SRS in their study [21]. Andrews et al. found that WBRT plus SRS group showed an improved KPS score and better OS than WBRT alone group in patients with single brain metastasis (6.5 vs 4.9 months; P = 0.0393). They also found that better control rates at 1 year in the WBRT plus SRS group (82%) vs 71%, P = 0.01). Of note, only 24 (7.2%) SCLC patients were included in their study [16].
Some studies suggested that SRS alone may be appropriate treatment for brain metastases in SCLC patients [13, 22]. Robin et al. reported SCLC patients who treated with upfront SRS was associated with longer OS than did WBRT ± SRS (10.8 vs 7.1 months, p < 0.001). However, in the subgroup analysis, there was no survival difference between SRS alone and WBRT plus SRS (p = 0.601) [22]. Bernhardt et al. also conducted a randomized controlled trial to investigate the treatment response of WBRT or SRS alone in SCLC patients. However, the final conclusions have not yet been published [13]. In a randomized controlled clinical trial published in JAMA, Brown et al. reported that WBRT plus SRS group had no survival difference compared with SRS alone group in patients with 1–3 brain metastases (7.4 vs 10.4 months; P = 0.92). However, the higher intracranial tumor control rates at 3 months was observed in patients who treated with WBRT plus SRS (93.7% vs 75.3%, P < 0.001) [11]. Of note, only 88 (66.7%) patients were lung cancer in their study. Therefore, their findings may not apply to SCLC patients. Aoyama et al. enrolled 132 patients with 1–4 brain metastases, among them, only 88 (66.7%) were lung cancer. They found that patients who treated with WBRT plus SRS group did not improve the survival time than did SRS alone (7.5 vs 8.0 months; P = 0.42), but significantly avoided the risk of intracranial recurrence [23]. Although previous studies were consistent with our findings, most of the above studies did not enroll SCLC patients alone. Combined with the characteristics of multiple metastases from SCLC and the better local control rates of WBRT, adjuvant WBRT combined with focal boost was necessary and indispensable for brain metastases from SCLC.
Our data also showed that the 1–3 brain metastases, and without extracranial metastases were prognostic factor for increased survival, which was similar to previous findings [24, 25]. We also found that gender and systemic treatment were prognostic factors for OS.
Our study also has several limitations. Firstly, the general characteristics of patients are not balanced. Of course, we performed propensity score matching to control the confounding between the WBRT+boost and WBRT or SRS groups. Secondly, RT-related neurotoxicity was not evaluated. Thirdly, the inherent characteristics of retrospective research and patient heterogeneity may bias the results.