HFSRT + WBRT showed good results in terms of efficacy, toxicity and survival time.
Some studies suggest that SRS can be considered a standard treatment that is a less toxic alternative to SRS + WBRT[4, 24]. However, WBRT reduction of brain tumor recurrence rate may translate to improved survival in patients with intracranial tumor progression; thus, it seems reasonable to suggest that the addition of WBRT may affect survival outcomes in selected patients[25, 26]. Wegner RE et al.[27] reviewed 36 patients who were treated with HFSRT alone (24Gy in2-5F), with 6- and 12-month LC rates of 73% and 63%, respectively. Kim KH et al.[14] reviewed 46 patients who were treated with HFSRT alone. Patients were randomized to receive 24, 27, or 30Gy in 3 fractions, with 12-month LC rates of 65%, 80% and 75%, respectively. The intracranial LC rate in this study was superior to the two studies mentioned above, further studies are needed to investigate the most suitable population for HFSRT + WBRT.
Our results suggest that HFSRT + WBRT achieved equivalent or even better survival outcomes compared to those of SRS + WBRT in previous reports. Three secondary analyses of previous randomized clinical trials described survival outcome in patients with BMs who were treated with SRS + WBRT, and the MST was from 5.7 to 7.9 months[26, 28, 29]. Our better survival rate compared to previous studies may be due to the fact that the patients in the abovementioned three studies were enrolled from 1996 to 2013, during which time they did not have access to the improved new targeted therapy and immunotherapy regimens currently available. Similar results were observed in some recent studies. Sallabanda et al.[30] retrospectively reviewed 200 patients who were treated with SRS or HFSRT from 2010 − 2016. The median OS was 8 months, and the 1- and 2-year actuarial OS were 40% and 24.5%, respectively.
In our study, the HFSRT group received a smaller total dose and number of fractions than those reported in the literature. We reviewed recent studies on patients with BMs who were treated with HFSRT or HFSRT + WBRT; the most common HFSRT dose was 27Gy/3F (range, 24 − 41Gy/2 − 6F), the OS rates ranged from 13.0 − 69.0% at 12 months, and the MST ranged from 12.2 − 16.2 months[9, 14, 27, 31–34]. Our results are also comparable to those of several studies on HFSRT for BMs in the surgical cavity. Two studies assessed the efficacy and safety of postoperative HFSRT in patients with BMs. The 12-month OS was 62% and 58%, respectively. The radiation necrosis rate was 5.1% and 8.9%, respectively[13, 35].
PI are beneficial to clinical and therapeutic decision-making. However, GPA classes were not statistically significant in our analysis. Conversely, many previous studies concluded that GPA is a reliable, simplistic, and powerful tool for predicting survival[18, 36]. The possible reasons for this difference include the following. First, GPA includes the number of metastases and does not involve metastasis volume. We noted that the number of BMs in the patients and the GTV size were not proportional. The median GTV was greater in patients with 1 BM than in patients with 2 − 3 BMs (9.45 and 7.25 cm3, respectively). Some studies concluded that brain tumor volume had a significant association with OS[37, 38]. Second, the primary tumors in our study included lung cancers, breast cancers, gastrointestinal cancers, and gynecologic tumors. There was marked heterogeneity in outcomes among patients with BMs and the differences in outcome were not only related to the diagnosis but also to diagnosis-specific prognostic factors[36, 39]. Further, the dissimilar proportions of patients within the prognostic classes could be another reason.
In addition, we found that the PIs in our study were not uniformly recommended in different studies, and there was a large discrepancy between the expected and actual survival[18, 40–42]. This finding supports the need for a better prognostic tool or index.
In this study, we used two instruments for efficacy evaluation. There was no significant difference in efficacy outcomes between the two criteria; this was possibly due to: 1) the small number of patients included in the efficacy analysis, 2) of all patients treated with HFSRT, only one patient showed asynchronous changes in the intracranial lesions, which may have attenuated the difference in the number of target lesions between the two criteria, and 3) although corticosteroid use is not included in the RECIST 1.1 criteria, no patient who met the imaging criteria was ineligible for corticosteroid use, thus eliminating the potential discrepancy from this definition. The RANO-BM criteria, may provide a more comprehensive assessment of patient outcomes than the RECIST 1.1 criteria. However, in our practical application of the RANO-BM criteria we found that the criteria complicate the assessment of patients with BMs in clinical trials.
We believe the results of the present study are important for several reasons. First, we used the same HFSRT and WBRT scheme for the entire study, reduced treatment-derived differences in the analysis of treatment efficacy. Second, to our knowledge, our study involved the lowest total radiation dose and number of fractions, but we observed desirable survival outcomes. Third, we analyzed survival outcomes and symptom and KPS score improvement in patients with a poor KPS post-treatment. Our findings demonstrate that patients with KPS < 70 may not be unfavorable candidates for SRT. Lastly, to our knowledge, this is the only study that has employed both RECIST 1.1 and RANO-BM criteria to evaluate the HFSRT + WBRT efficacy for BM.
This study was limited by its retrospective design. During the radiotherapy and follow-up periods, patients also received systemic treatments, which might influence their survival and local control. In addition, we did not avoid the hippocampal regions during irradiation in WBRT. We also did not evaluate the neurocognitive function of patients, post-treatment using scores.