Reduced-dose WBRT combined with SRS for 1–4 brain metastases aiming at minimizing neurocognitive function deterioration without compromising brain tumor control

Highlights • The addition of whole-brain radiotherapy (WBRT) to stereotactic radiosurgery (SRS) reduces the risk of brain tumor recurrence but standard-dose WBRT (SD-WBRT) accompanies the risk of neurocognitive decline.• Reduced-dose WBRT (RD-WBRT) combined with SRS provides intracranial tumor control rate comparable to that after SD-WBRT + SRS.• RD-WBRT could reduce the risk of neurocognitive decline compared to that after SD-WBRT.


Background
Brain metastasis is the most common brain tumor, as it develops in 20 % -40 % of patients with systemic cancers. The number of patients diagnosed with brain metastases has been constantly increasing, due in part to the improvement of systemic therapy and the increasing quality and prevalence of MRI in recent years. [1] The standard treatment for brain metastases has been whole-brain radiation therapy (WBRT), but there has also been concern about the cognitive deterioration of patients as a result of the toxicity of WBRT. Treatment with stereotactic radiosurgery (SRS) without WBRT is thus becoming more widely used for patients whose number of brain metastases is limited to 3-4. [2,3] However, the avoidance of WBRT results in a higher risk of brain tumor recurrence (BTR) at distant sites in the brain (BTR-distant) with or without BTR at local sites that received SRS (BTR-all). More importantly, the higher risk of BTR could translate to an impairment of overall survival in some subsets of patients. [4] In addition, there remains large regions of the world where frequent monitoring by enhanced-MRI after SRS-alone or advanced forms of radiation therapy such as intensity modulated radiation therapy including hippocampal-avoidance [HA]-WBRT are not commonly available. The optimal use of WBRT thus remains to be determined.
The risk of developing one or more radiation-induced late adverse effects, including cognitive decline, is closely related to the total radiation dose and the dose-per-fraction. The current standard dosefractionation regimens of WBRT, such as 30 Gy in 10 fractions, were established in the 1970s -80s when MRI and stereotactic radiosurgery (SRS) were not commonly available, and the primary treatment goal of WBRT was its therapeutic effect on already visualized metastases rather than non-visualized micro-metastases. Today, both WBRT and SRS are readily available worldwide, and thus we considered that the role of WBRT could be limited to merely avoiding the progression of nonvisualized micro-metastases or slightly enhancing the treatment effect for already visualized metastases (especially large metastases).
Reduced dose (RD)-WBRT, such as 25 Gy in 10 fractions, was used in the clinical trials examining the role of prophylactic cranial irradiation (PCI) for limited-stage small cell lung cancer (SCLC) [5] or locally advanced non-small-cell lung cancer, [6] and the incidence of cognitive decline in these scenarios is reported to be significantly lower than that by standard-dose WBRT (SD-WBRT). [5] In addition, in a single-arm study conducted using 25 Gy/10 fractions in patients with SCLC, transient decreases were observed in executive function and language after PCI, but the decreases improved to the pretreatment levels in the long term. [7] We conducted the present non-randomized, single-arm study to determine whether the combination of RD-WBRT and SRS could be used to minimize the risk of cognitive decline without compromising the brain tumor control for patients with 1-4 brain metastases.

Study design and patients
This was a multi-institutional phase II study by the Japanese Radiation Oncology Study Group (JROSG 13-1). Adult patients (20-80 years old) with 1-4 brain metastases, all of which were ≤ 3 cm in diameter, were eligible for the trial. The eligibility criteria also included a Karnofsky Performance Status (KPS) ≥ 70 and pathological confirmation of an extracranial tumor site. The exclusion criteria included a past history of surgery or radiation to the brain, the presence of metastasis to the brainstem or leptomeningeal dissemination, and inability to take cognitive function tests or quality of life (QOL) surveys. Brain metastases from small cell cancers, germ cell tumors, or lymphoma were also excluded. Each participating institution provided institutional review board approval, and each patient provided written informed consent. This trial is registered with the UMIN Clinical Trial Registry (UMIN000009055).

Procedure
The radiation dose of single-fraction SRS (SF-SRS) prescribed to the 95 % of the gross tumor volume defined as the enhanced area on MRI was 22-24 Gy for lesions ≤ 2 cm and 18-22 Gy for lesions >2 cm. The use of hypo-fractionated SRS (HF-SRS), which has an effect that is biologically identical to that of various protocols of SF-SRS, such as 28-35 Gy in 4 fractions or 26-30 Gy in 3 fractions, was allowed. RD-WBRT (25 Gy in 10 fractions) was started within 1 week after the final date of the patient's SRS.

End points
This study's primary endpoint was the BTR-distant-free survival of patients at 6 months after the completion of radiation therapy. The secondary endpoints included overall survival (OS), local tumor control, cognitive functional change, radiation-related adverse effects, and the cause of death. The cumulative incidences of BTR-distant and BTR-all were estimated by the competing risk method to account for the competing risk of death. Gray's test was used to test for significant differences in the cumulative incidence of BTR-distant and BTR-all.
Contrast-enhanced MRI was taken at baseline and 4, 6, 9, and 12 months and every 6 months thereafter. The standardized neuropsychological test battery [8] was used, which included the Hopkins Verbal Learning Test Revised (HVLT-R) for memory (both immediate and delayed recall and recognition), the Controlled Oral Word Association Test (COWA) for language/verbal fluency, the Trail Making Test Part A (TMT-A) for visual and spatial scanning, attention, sequencing, and speed, and the Trail Making Test Part B (TMT-B) for executive/frontal lobe skills at baseline and 4, 8, and 12 months and every 6 months thereafter. The quality-of-life measures included the European Organization for Research and Treatment of Cancer (EORTC) Quality-of-Life Questionnaire (QLQ-C30) and the Brain Cancer Module 20 (BN20). [5] All treatment-related toxicities and adverse events were recorded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 4.0.

Sample size and statistical analyses
The target accrual was 40 patients with 33 as the number of occurred events. This target was calculated using the 95 % confidence interval (95 %CI) for an exponential model, based on having 80 % power to detect the non-inferiority of RD-WBRT combined with stereotactic radiosurgery (SRS) of JROSG13-1 compared to that of SD-WBRT (30 Gy in 10 fractions) in the JROSG99-1 study [2] (a randomized clinical trial [RCT] between SRS + SD-WBRT and SRS alone) with a one-sided significance level of 0.05 assuming that the 6-month BTR-distant-free survival of the JROSG99-1 study was 81 %, and defining noninferiority as the same 6-month disease-free survival of >71 %. [9] We performed a one-sided test for the non-inferiority of the JROSG13-1 Abbreviations: RD-WBRT, reduced-dose whole brain radiation therapy; SD-WBRT, standard-dose whole brain radiation therapy; SRS, stereotactic radiosurgery Table 3 Cumulative incidence of brain tumor recurrence (BTR) in the JROSG13-1, JROSG99-1 and N0574 trials Cumulative incidence Treatment 6 months 12 months Abbreviations: RD-WBRT, reduced-dose whole-brain radiation therapy; SD-WBRT, standard-dose whole-brain radiation therapy; SRS, stereotactic radiosurgery study by comparing the observed 6-month BTR-distant-free survival rate with a margin of 10 % (i.e., a null hypothesis that the hazard ratio of the JROSG13-1 study was ≥ 1.625). In addition, a comparison of BTRdistant-free survivals accounting for the entire follow-up period of patients in each study was also conducted as a reference. The statistical analyses were performed with SAS software, ver. 9.4 (SAS Institute, Cary, NC) and EZR (a modified version of R commander). [10] The cumulative incidences of BTR-distant and BTR-all were estimated by the competing risk method to account for the competing risk of death, and the statistical difference derived from the results of the JROSG99-1 study was compared by Gray's test. The OS and BTR results of the JROSG99-1 trial [2] were calculated from the date of the last day of treatment for the comparison with the results of the present study.

Study patients
Between April 2012 and November 2018, 40 patients were enrolled at seven participating institutions in Japan. The data were fixed in November 2019. All analyses were undertaken after all patients had been potentially followed for ≥ 12 months. All patients completed the pretreatment cognitive and patient-reported QOL assessment. The patients' baseline characteristics are summarized in Table 1: there were 22 males and 18 females, and the median age was 69 years (range 43-80 yrs). The primary tumor site was a lung in 28 patients (70 %). All 9 patients with EGFR-or ALK-positive lung adenocarcinomas received a treatment with tyrosine kinase inhibitors (TKIs) before and/or after brain radiation therapy.

Discussion
The cumulative incidences of BTR-distant and BTR-all after RD-WBRT in the present study were comparable to the corresponding values after SD-WBRT in the JROSG 99-1 trial and were significantly lower than those after SRS-alone. Therefore, the dose-fractionation of WBRT when combined with SRS could be safely reduced to 25 Gy in 10 fractions without increasing the risks of BTR-distant and BTR-all.
The incidences of cognitive decline in this study according to the different definitions of cognitive decline are summarized in Table 4, and we compared these values to the results of previous brain metastasis trials in which cognitive function after radiation was assessed by the same standardized cognitive battery. [8] The definitions of cognitive decline are not standardized and they differ among clinical trials; results should be carefully interpreted when they are compared with those of other trials, taking into account the definition of cognitive decline used. [11] When we defined decline as [>2.0 SD in ≥ 1 test] in the present study, the rate of decline was 48.6 % (Fig. 2). In the N0574 trial comparing SD-WBRT + SRS and SRS-alone, [3] the rates of decline were 72.9 % and 42.9 %, respectively. The decline rate in the present study was thus much lower than that in the SD-WBRT + SRS group in the N0574 trial, and it was close to that in the SRS-alone group.
When we used [>1.0 SD in ≥ 1 test] as the definition of decline, the cognitive decline rate was 75.0 %. Similar definitions were used in the N107C/CEC3 trial (Surgery + SD-WBRT vs. Surgery + SRS), [12] the N0574 trial, [3] and the SAKK 15/12 trial (Hippocampal-avoidance [HA]-RD-WBRT). [13] The rates of cognitive decline in these studies were as follows: 85 % (N107C/CEC3) and 91.7 % (N0574) in the SD-WBRT group, 52 % (N107C/CEC3) and 63.5 % (N0574) in the SRSalone group, and 65.8 % (SAKK15/12) in the HA-RD-WBRT group. Compared to those numbers, the decline rate in our present investigation is somewhere between SD-WBRT and the others. In addition, it is noteworthy that Vees et al. concluded in the SAKK15/12 trial that the rate of neurocognitive decline after HA-RD-WBRT was not significantly different from that of RD-WBRT. [13] However, it should be noted that the [>1.0 SD in ≥ 1 test] criterion has high sensitivity but low specificity. [11] When we used [>2.0 SD or > RCI (reliable change index) in ≥ 1 test] as the definition of cognitive decline, the decline rate was 59.7 % (Fig. A5). This definition was also used in the RTOG0614 trial (SD-WBRT + memantine vs. SD-WBRT + placebo). [14] The rate of cognitive decline in the present study was similar to the 64.9 % after SD-WBRT with placebo and 53.5 % after SD-WBRT with memantine in the RTOG0614 trial. In the NRG CC001 trial, the definition [>RCI in ≥ 1 test] was used as the definition of cognitive decline, and the decline rate was 68.5 % after SD-WBRT + memantine and 59.5 % after HA-SD-WBRT + memantine. [15] The decline rate of 56.7 % in the present study In addition to the dose-fractionation schedule of WBRT, age is another factor that could strongly affect the rate of decline in patients' cognitive function. [5] The median age of the present patients was rather high at 69 years, and it was 6-9 years higher than the ages of the patients in the other studies cited herein (Table 4). It is quite likely that the ages of our patient population negatively influenced the cognitive preservation rate. In other words, it may be that the rate of cognitive decline after RD-WBRT would be lower than that after SD-WBRT, even though more elderly patients were registered in this study; this indicates the possibility of RD-WBRT as a new standard WBRT schedule for patients who are indicated for SRS from the viewpoint of the preservation of cognitive function.
As SRS becomes more and more widely adopted, and as new systemic therapies with some efficacy against brain metastases emerge, what are the indications for adding WBRT to SRS? As a patient's prognosis improves, the significance of controlling brain tumors becomes more important in regard to OS as well as the maintenance of QOL. [4] Apart from the reduction of BTR-distant, another important role of WBRT combined with SRS is to enhance the local tumor control compared to that provided by SRS alone. The results of all four of the randomized trials comparing SRS alone and SRS + WBRT demonstrated that not only the BTR-distant-free survival but also the local tumor control rate by SRS + WBRT was significantly higher than that by SRS alone. [2,3,16,17] The significant benefit on local tumor control is especially prominent in medium-to-large tumors (≥15-20 mm). [18,19] In the above-mentioned JROSG99-1 trial, the local tumor control rates of lesions ≥ 20 mm at 12 months were 81.5 % after SD-WBRT + SRS and 59.8 % after SRS alone, and the discrepancy in the rates expanded further at 24 months to 81.5 % versus 19.9 % ( Table 2). The use of hypofractionation might improve local tumor control to some extent compared to single-fraction SRS, [18,19] but we observed that the combination of HF-SRS and RD-WBRT in the present study provided tumor control comparable to that of SD-WBRT + SRS for tumors < 20 mm and ≥ 20 mm at 12 months (94.2 % vs. 67.9 %), and more importantly, this effect was maintained for the next 12 months. The long-term tumor control observed in this study is especially relevant for patients who can expect a good prognosis (i.e., a median OS of 17 months as achieved in this study), in order to achieve long-term maintenance of QOL while avoiding neurologic death, despite the modest risk of neurocognitive decline. Therefore, RD-WBRT combined with HF-SRS might be particularly indicated for patients with favorable prognosis and harboring medium-to-large brain metastases, since medium-to-large brain metastases would be difficult to control by SRS with or without systemic therapies, or could be refractory to systemic therapies. RD-WBRT plus HF-SRS might also be appropriate for patients in whom frequent monitoring by enhanced-MRI after treatment with SRS alone would be difficult for financial or geometrical reasons.

Study limitations
Limitations of this study include the single-arm design and the small number of patients (n = 40); therefore, these data cannot be used to conclude the superiority of RD-WBRT + SRS over SRS with or without SD-WBRT in terms of avoiding BTR and the preservation of neurocognitive function. In addition, the effect of the use of TKIs was not negligible, though no significant difference was observed between patients who received TKI-therapy for EGFR/ALK-positive lung-adenocarcinoma and the other patients in terms of either BTR-distant or BTRall in the present study. Nonetheless, our findings are encouraging and  merit further investigation in one of the arms of a prospective randomized study designed to identify the optimal treatment for patients with a limited number of brain metastases.

Conclusions
By achieving durable brain tumor control, the combination of SRS and RD-WBRT may be an optimal treatment method for patients who are expected to have a good life expectancy and to maintain their QOL. The information obtained in this study will be important for physicians in regions of the world where the routine use of HA-WBRT and/or the administration of memantine for patients with brain metastases is not possible.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data Availability Statement for this Work
Research data are not available at this time.

Acknowledgements
The effort of Yukiko Morita and Mika Hasegawa, research assistants at the Department of Radiology and Radiation Oncology, Niigata University Graduate School of Medical and Dental Sciences; and of Miki Nagai and Kumiko Shirai of the Clinical and Translational Research Center, Niigata University Medical and Dental Hospital, Niigata, Japan should be acknowledged.

Funding
This work was partly supported by a grant from the Japan Society for the Promotion of Science KAKENHI program (no. 15H04903, 22H03008), a Health and Labor Sciences Research Grant (no. 19-EA1-010), and funds from the Niigata University Clinical Trial Support Project 2011.

Previous presentation
The 2020 Annual Meeting of the American Society for Radiation Oncology, Web, 2020

Table A1
Characteristics of the patients in the present study (JROSG13-1) and the historical data (JROSG99-1)      * Scores range from 0 to 100, with a higher score representing a higher level of functioning or health status. ** Scores range from 0 to 100, with a higher score representing a higher degree of symptoms.

Table A6
Number