Analysis of the efficacy and prognostic factors of linear accelerator-based hypofractionated stereotactic radiotherapy for brain metastasis of non-small-cell lung cancer

Objective To investigate the efficacy and prognostic factors of treatment of brain metastases (BMs) of non-small-cell lung cancer (NSCLC) patients with hypofractionated stereotactic radiotherapy (HSRT) based on linear accelerators. Methods A total of 80 NSCLC patients with BMs were treated with HSRT under the image guidance of tomotherapy (TOMO) or volumetric modulated arc therapy (VMAT) in Yunnan Cancer Hospital from Jan 1 st , 2017 to Aug 7 th , 2019. The outcome measures included overall survival (OS), intracranial local progression-free survival (iLPFS) and intracranial regional progression-free survival (iRPFS). The related prognostic factors were analyzed. Results The median OS, iLPFS and iRPFS in 80 patients were 29.2 months, 19.3 months and 20.5 months, respectively. Using HSRT to treat BMs, whether combined with platinum-containing chemotherapy, targeted therapy, or both chemotherapy and targeted therapy, could improve OS ( P =0.027), iLPFS ( P =0.050) and iRPFS ( P =0.124). Among these patients, there were no significant differences in median OS (28.5 months vs 28.3 months, P =0.785), 6-month iLPFS rates (81.3% vs 72.9%, P =0.998) and 6-month iRPFS rates (92.9% vs 74.8%, P =0.974) between patients with 4-10 BMs and those with 1-3 BMs. Conclusion The use of HSRT to treat BMs of NSCLC based on linear accelerators is safe and effective. Use of HSRT in combination with other antitumor therapies to treat BMs is a favorable prognostic factor that can affect OS and iLPFS. HSRT treatment of 4-10 BMs or 1-3 BMs resulted in similar OS, and there were no significant differences in 6-month iLPFS rates and 6-month iRPFS rates.

by 3 mm external expansion of GTV. Prescription dosages ensured at least 95% coverage of PTV and 100% coverage of GTV area.
All patients underwent clinical follow-up, including a craniofacial MRI or enhanced CT scan 1 month after the end of radiotherapy and every 3 months thereafter.

Statistical analysis
Statistical analysis was carried out using SPSS 22.0 software. Survival analysis was performed with Kaplan-Meier method. The log-rank test was used for single variable analysis of prognostic factors. The variables with P<0.1 in the single variable analysis are included in the Cox scale risk regression model for multivariate analysis to evaluate the independent prognostic factors associated with OS, iLPFS, and iRPFS. P<0.05 was statistically significant.

Survival and intracranial control
In our study, the median follow-up period was 22.4 months, ranging from 1-33 months to the last follow-up date of October 20 th , 2019, in which 23 patients (28.8%) died and 31 patients (38.8%) had progression of intracranial lesions. Twenty-two patients (27.5%) were actively treated for intracranial progression.
The median OS from HSRT combined with other antitumor therapies reached 31.6 months, but the difference was not statistically significant compared to 22.7 months from BMs treated with HSRT alone (P=0.094). Our study suggested that there was no significant difference in median OS between patients with 4-10 BMs and 1-3 BMs treated with HSRT (28.5 months vs 28.3 months, P=0.785).
The multivariate analysis showed that the independent adverse prognostic factors  Table 1).

The intracranial local progression-free survival time
The median iLPFS of patients in our study was 19.3 months, ranging from 1 to 32 months. The median iLPFS of patients who received HSRT combined with other antitumor therapies reached 22.0 months, and the difference was statistically significant compared to 12.0 months in patients whose BMs were treated with HSRT alone (P=0.033). Our study suggested that there was no significant difference in 6-month iLPFS rates between patients with 4-10 BMs and 1-3 BMs treated with HSRT (81.3% vs 72.9%, P=0.998).
The multivariate analysis showed that the independent adverse prognostic factors of iLPFS  Table 2).

The intracranial regional progression-free survival time
The median iRPFS of patients in our study was 20.5 months, ranging from 1 to 32 months.

Secondary analysis of other antitumor therapies
The median OS of HSRT combined chemotherapy for the same period was 30.1 months, the combined target was 31.1 months, and both combined chemotherapy and target was 27.9 months, which better than the 22.7 months of BMs treated with HSRT alone. The median iLPFS of HSRT combined chemotherapy for the same period was 15.6 months, the combined target was 22.1 months, and both combined chemotherapy and target was 20.8 months, which better than the 12.0 months of BMs treated with HSRT alone. The median iRPFS of HSRT combined chemotherapy for the same period was 17.4 months, the combined target was 22.1 months, and both combined chemotherapy and target was 19.9 months, which better than the 16.5 months of BMs treated with HSRT alone.
Above all, our study suggested that compared to patients who did not receive other antitumor treatments during the same period of HSRT, whether combined with platinumcontaining chemotherapy, targeted therapy, or combined chemotherapy and targeted therapy could improve OS, iLPFS, and iRPFS.
The specific antitumor therapy received in the same period of HSRT treatment for BMs was included in the multifactor analysis of OS, iLPFS and iRPFS. The multifactor analysis showed that compared to HSRT of BMs without receiving other antitumor treatments, whether combined with platinum-containing chemotherapy, targeted therapy, or combined chemotherapy and targeted therapy could improve OS (P=0.027), iLPFS (P=0.050) and iRPFS (P=0.124) (Figure 4). The differences in OS and iLPFS were statistically significant, but there was no significant difference in iRPFS (Supplementary Table 4 Discussion SRS is becoming a major force in the treatment of BMs. It can be used for single or multiple BMs, and the median OS can be up to 20 months (14-16). It can be used as adjuvant therapy after surgery for BMs, with a median OS up to 12.2 months (17). SRS can be applied to re-SRS that has previously received SRS treatment for more than 6 months and confirming tumor recurrence, and the second SRS median OS could be up to 11 months (18, 19). It can be used for rescue treatment after whole brain radiotherapy (WBRT)failure, delay or avoid WBRT (20-22). In our study, HSRT was used to treat 80 patients with NSCLC BMs, with a median OS of 29.2 months, a median iLPFS of 19.3 months, a median iRPFS of 20.5 months, and a lower incidence of toxic reactions, which was better than the results of most previous studies.
In a retrospective trial, analyzed 2,966 patients who received a single SRS treatment for BMs (not limited to primary tumors), and statistical analysis suggested that patients with KPS ≥ 80, primary cancer under control, and no extracranial metastasis may have longer OS and a higher risk of radioactive complications (23). In our study, a multifactor analysis suggested that KPS < 70 is an independent adverse prognosis factor affecting OS (P = 0.024). It did not suggest that no extracranial metastasis and the control of extracranial and primary cancer were the independent prognostic factors affecting OS, iLPFS and iRPFS. Rosell et al. (24) showed that adenocarcinoma had a longer survival period than squamous cell carcinoma (8.8 months vs 5.4 months, P = 0.01). Our study suggested that non-adenocarcinoma is an independent adverse prognostic factor affecting OS, iLPFS, and iRPFS (P < 0.05), consistent with previous studies. GPA and RPA can predict the survival of patients with BMs (25-27). In our study, the single-factor analysis suggested that GPA had some effect on OS and iRPFS (P < 0.1), and RPA had some effect on OS, iLPFS and iRPFS (P < 0.05). However, the multivariate analysis did not suggest that GPA and RPA were the independent prognostic factors affecting OS, iLPFS and iRPFS.
At present, the main treatment for BMs of NSCLC include the combination or nonunion of local therapy and systemic therapy. Several studies have shown that intracranial tumor remission rates can be improved by combined/noncombined platinum-based chemotherapy (whether sequential or synchronous) in patients with BMs of NSCLC, but there was no significant difference between PFS and OS, and radiotherapy combined chemotherapy increased the incidence of adverse reactions (28, 29). BM is more likely in NSCLC patients with positive driver genes, with a rate of 60% BM in patients with EGFR mutation and an incidence of BM of approximately 45%-70% in ALK-positive patients (30,31). The discovery of driver genes such as EGFR, ALK, ROS1 and the emergence of related targets provide a more simple and effective treatment for patients with BMs, further prolonging survival. The combination of EGFR-TKI and radiotherapy for BMs can improve Os and PFS, but the combination mode of EGFR-TKI and radiotherapy has not been determined. EGFR status was differentiated in only a few clinical studies of EGFR-TKI concurrent radiotherapy for BMs, and the results varied. A Small Phase II study suggested that in EGFR mutated patients with BMs of NSCLC, concurrent EGFR-TKI and radiotherapy have a slight OS and PFS benefit compared to radiotherapy alone (22.0 months vs 7.5 months) (32). Chang W et al. (33) retrospectively analyzed 351 newly diagnosed EGFR mutant lung adenocarcinoma BMs patients who were not subjected to targeted therapy and studied the sequence of EGFR-TKI and radiation therapy for BMs. The results suggested that patients with the SRS sequence EGFR-TKI had longer OS (45 months vs 25 months, P < 0.001) and that delayed radiotherapy reduced OS. The BRAIN study suggested that for patients with EGFR mutations, iPFS improved when EGFR-TKI used prior to BMs radiotherapy (10.0 months vs 4.8 months, P = 0.014) (34). For patients with NSCLC BMs in ALK-positive or ROS1-positive, crizotinib is widely used in clinical practice (35)(36)(37). Vascular-targeted therapy (e.g., bevacizumab, etc.) is also widely studied in patients with NSCLC BMs (38, 39).
In our study, the multivariate analysis suggested that other mutations of non-EGFR mutations was independent prognostic factor affecting OS (P = 0.045) and iLPFS (P = 0.030). We considered whether this was because there were fewer patients (7.5%) with non-EGFR mutations in our study: three patients (3.8%) had ALK-positive mutations, two patients (2.5%) had ROS1-positive mutations, and one patient (1.3%) had KRAS mutations.
Patients that were ALK-positive and ROS1-positive were treated with crizotinib, which was unevenly distributed across the groups, resulting in a large difference in prognosis. Of the 28 patients (35.0%) in our study, 14 patients (17.5%) received EGFR-TKI targeted therapy, 5 patients (6.3%) received ALK-TKI targeted therapy, and 9 patients (11.3%) received vascular targeted therapy; there were no significant differences in OS (P = 0.597), iLPFS (P = 0.711) and iRPFS (P = 0.670) in the univariate analysis. The multifactorial analysis was not included because of the small number of cases and uneven stratification.
For patients with NSCLC BMs without a driver gene mutation, systemic chemotherapy is still the main method of therapy. The study on the combination of brain radiation and targeted therapy in driver gene-positive NSCLC patients with BMs is mostly retrospective, with small sample sizes and inconsistent results. WBRT was the most common choice in the current study of combined chemotherapy/targeted therapy and radiotherapy for patients with BMs from NSCLC. There is a lack of research on HSRT combined with chemotherapy and targeted therapy for BMs. Our study suggests that, compared to patients who did not receive other antitumor treatments during the same period of HSRT, a combination with platinum-containing chemotherapy, targeted therapy, or both chemotherapy and targeted therapy could improve OS (P = 0.027), iLPFS (P = 0.050) and iRPFS (P = 0.124), but the best combination should be determined by large-scale prospective studies.
In a prospective trial SRS-JLGK 0901, 1,194 patients with newly diagnosed BMs were recruited, and all patients were treated with SRS for BMs. The median OS of patients with 1 BM (N = 455) was 13.9 months. The median OS of patients with 5-10 BMs (N = 531) was not significantly different compared with the OS of patients with 2-4 BMs (N = 208), at 10.8 months (P < 0.0001). There was no significant difference in the incidence of adverse events between the two groups (P = 0.89) (40). This experiment shows that SRS of 1-3 BMs or multiple BMs results in similar OS and similar toxicity and side effects.
In our study, when HSRT was used to treat 4-10 BMs compared with 1-3 BMs, and there was no significant difference in median OS (28.5 months vs 28.3 months, P = 0.785), 6month iLPFS rates (81.3% vs 72.9%, P = 0.998) and 6-month iRPFS rates (92.9% vs 74.8%, P = 0.974), consistent with other studies. The median OS of patients with 4-10 BMs was slightly higher than that of patients with 1-3 BMs in our study, and the 6-month iLPFS rates and 6-month iRPFS rates were higher than those of patients with 1-3 BMs, which may be due to the large baseline difference between the two groups, with the majority (65.0%) having 1-3 BMs and only a small minority (15.0%) having 4-10 BMs.

Limitations
There are many limitations to our study. These include: the limiting factors inherent in retrospective studies, the small sample size, the imbalance of the baseline of patients after stratification in the group, the short follow-up period, the difficulty in assessing the associated nervous system side effects after brain radiotherapy and the complexity of HSRT doses. Therefore, the results of our study need to be carefully interpreted, and large-scale prospective studies are required to verify the credibility of the conclusions.

Conclusion
The use of HSRT to treat BMs of NSCLC based on linear accelerators is safe and effective.
Use of HSRT in combination with other antitumor therapies to treat BMs is a favorable prognostic factor that can affect OS and iLPFS, whether combined with platinumcontaining chemotherapy, targeted therapy, or both chemotherapy and targeted therapy.
HSRT treatment of 4-10 BMs or 1-3 BMs resulted in similar OS, and there were no significant differences in 6-month iLPFS rates and 6-month iRPFS rates. KPS<70 is an independent adverse prognostic factor for OS. The non-adenocarcinoma subtype is an independent adverse prognostic factor for OS, iLPFS and iRPFS.   VMAT, volumetric modulated arc therapy; PTV, planning target volume.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.