Abstract
Background/Aim: Stereotactic radiosurgery (SRS)—used for brain metastases (BMs) with a tumor diameter of ≤2 cm—has a high local control rate, however, it can cause symptomatic radiation-induced brain necrosis. Hypofractionated stereotactic radiation therapy (HFSRT) is not commonly used for such lesions and its effectiveness remains unknown. Herein, the efficacy of 30 Gy 5-fraction HFSRT for treating BMs of <2 cm was retrospectively evaluated. Patients and Methods: Patients who received HFSRT and had a gross tumor volume (GTV) of ≤2 cm in maximum diameter were included in the study (49 patients; 179 BMs; median follow-up period, 11.9 months). Results: The mean GTV Peripheral Dose (D95) was 36.2 Gy. The local control (LC) rates at 1 and 2 years were 93.0% and 81.5%, respectively, for all lesions. The 1-year LC rates were 93.6% and 92.0% for ≤1.0-cm and 1.0-2.0-cm lesions, respectively. Multivariate analysis revealed that the only significant difference was in GTV maximal tumor diameter (HR=1.961, p=0.0002). Notably, only one patient had asymptomatic radiation necrosis. Conclusion: Owing to the high toxicity of SRS, 5-fraction HFSRT can be an effective treatment strategy for BMs of <2 cm.
- Hypofractionated stereotactic radiation therapy
- brain metastases
- local control
- radiation oncology
- radiation necrosis
Brain metastases (BMs) are the most frequent cancer of the central nervous system. Nearly 30% of cancer patients will develop BM during their illness (1). A crucial component of the care of BM is radiotherapy (RT). Stereotactic radiosurgery (SRS) and Hypofractionated stereotactic radiotherapy (HFSRT) have emerged as the most popular RT treatments for single and multiple BMs, respectively (2-5). HFSRT, currently used for relatively small lesions in addition to large tumor lesions, allows for great local control (LC) and minimizes the risk of symptomatic radiation brain necrosis. In this context, some results have been observed (6-9). With high-dose RT, radiation necrosis (RN) is the most frequent long-term toxicity occurring between 6 months and several years after SRS/fSRS, with a median time to onset of approximately a year (2). The Hypofractionated Treatment Effects in the Clinic (HyTEC) group advises keeping tissue volumes (including target volume) receiving 12 Gy (single fraction) below 5 cm3 to limit the probability of RN risk below 10% in order to prevent RN (2). For multi-fraction treatments, the group recommends a dose of 20 Gy (3 fractions) or 24 Gy (5 fractions) to be kept below 20 cm3 for the symptomatic RN risk to be below 4% (2). With a single dose of 18-24 Gy, the HyTEC group demonstrated 85-95% LC at 1 year in patients with a maximal tumor size of ≤2 cm in diameter [defined as small tumors based on Radiation Therapy Oncology Group (RTOG) classification] (10). Some studies have also reported cases of symptomatic radiation-induced brain necrosis (2, 11, 12). Neurotoxicity from SRS is high, and attempts are being made to score the risk of neurological death (13).
Although the HyTEC dose-response model speculates on fractionated regimens of 18-24 Gy in a single fraction, little research on actual outcomes has been published. Our institute has been performing 5-fraction HFSRT for tumors smaller than 2 cm in diameter. In our institution, we performed a retrospective analysis of the effectiveness and safety of 5-fraction HFSRT for the treatment of BMs under 2 cm diameter.
Patients and Methods
Patient characteristics. Among106 patients who received stereotactic brain irradiation at Fujieda Municipal General Hospital between November 2016 and November 2021 and were diagnosed with BM, we found 347 lesions. Those who had a gross tumor volume (GTV) of 2 cm or below in maximum diameter and at least one post-treatment contrast-enhanced MRI follow-up were included in this research. Individuals who had received whole-brain radiation therapy (WBRT) in the past or at the same time were excluded. Re-irradiated lesions were also excluded. Patients who underwent numerous sessions at various time points for SRT lesions at other sites were included. The results excluded 42 lesions in 33 patients with tumors >2 cm in diameter, 12 lesions in 6 patients with the previously irradiated lesions, 113 lesions in 42 patients for whom no follow-up information was provided, and 1 lesion in 1 patient with a non-standard irradiation schedule. The operated-on lesions had never been subjected to surgery before.
This study was approved by the Ethical Review Committee of Fujieda Municipal General Hospital (R-FY22-6) and was directed in accordance with the ethical principles of the Declaration of Helsinki. Informed consent was obtained from all individuals included in this study.
Treatment. Patients were fixed with a stereotactic thermoplastic mask, and CT for simulation was performed. Within 3 days of the CT scan, a contrast-enhanced MRI was conducted with a slice thickness of 1.0 mm. The planning target volume (PTV) margin was set at 1-3 mm. The total dose of 30 Gy was administered in 5 fractions, and the prescribed dose covered 95% of the PTV. An iplan was used to carry out each treatment plan (Brain LAB AG, Heimstetten, Germany). Image registration for MRI and CT images was conducted using the treatment planning system, Monaco version 5.1.1 (Elekta AB, Stockholm, Sweden). The X-ray Voxel Monte Carlo method was the calculation algorithm employed. Irradiation was performed on an Infinity system (Elekta AB) equipped with MLC Agility and 6 MV X-ray Flattening Filter Free was administered. A single isocenter was used with the volumetric modulated arc therapy irradiation technique. Cone-beam CT XVI (Elekta AB) and stereo 2-way radiography ExacTrac (Brain LAB AG) were used to check daily patient setup and ensure precise placement right prior treatment. The time between CT imaging for simulation and the start of irradiation was within one week.
Follow-up. Following therapy, contrast-enhanced MRI was performed every 2- to 3-months. The LC rate for all treated lesions commencing on the first day of irradiation was calculated by the Kaplan–Meier method. For the analysis of LC, response criteria were defined as follows based on RECIST (14). No tumor apparent to the naked eye or no contrast enhancement was considered a complete response (CR). Partial response (PR) was defined as 30% or greater tumor reduction, and progressive disease (PD) was defined as 20% or greater increase. The term stable illness was applied to all other reactions (SD). The three types of LC were CR, PR, and SD. Incidents of grade 3 or higher toxicity were noted, according to the RTOG 90-05 central nervous system toxicity criteria (15-17). Lesions with heterogeneously enhanced contrast-enhanced MRI images that did not continue to advance over time were considered to have RN (12). Toxicity was assessed using CTCAE ver. 5.0.
Statistical analysis. Statistical analysis was performed using EZR software program, version 1.5 (Saitama Medical Center, Jichi Medical University, Saitama, Japan) (18). The Kaplan–Meier (KM) method was employed to calculate LC rates; The Cox proportional hazards model was used in univariate (UVA) and multivariate (MVA) studies to find potential LC factors. Statistical significance was set at p<0.05.
Results
Patient and treatment characteristics. There were 49 patients. The total number of lesions was 179, and the median follow-up period was 11.9 months (2.0-42.4). The average age of the patients was 69 years. The most common presenting illness was non-small-cell lung cancer (39 patients, 80%), followed by breast cancer (5 patients, 10%), SCLC (2 patients, 4%), and others (3 patients, 6%), with no renal cell carcinoma or melanoma.
The Karnofsky Performance Status (KPS) was 100 in 37 patients (76%), 90 in 4 patients (8%), 80 or less in 6 patients (12%), and unidentified in 2 patients (4%). The PTV margin was 2 mm in 92 cases (51%), 3 mm in 81 cases (45%), and in the remaining cases it was 1 mm or unknown. GTV Peripheral Dose (D95) was 36.2 (30.9-46.5) Gy. Patient data are shown in Table I, and lesion details and several dosimetric details are displayed in Table II.
Patient characteristics.
Tumors and dosimetric characteristics.
LC outcomes and toxicity. Based on Kaplan–Meier calculation, the 1- and 2-years for BMs of <2 cm were 93.0% and 81.5%, respectively (Figure 1). Subgroup analysis by tumor size showed that the 1-year LC rate was 93.6% for lesions ≤1.0 cm (n=137), compared to 92.0% for lesions 1.0-2.0 cm (n=42), (p=0.006) according to subgroup analysis of tumor size (Figure 2).
Local control outcomes for all tumors.
Local control outcomes based on tumor size [gross tumor volume (GTV) ≤1 cm and GTV=1-2 cm].
The mean time to retreatment was 12.4 months for the 17 lesions that underwent re-irradiation. A total of 19 (11%) lesions exhibited local failure (5.6-30.1 months). Thirteen lesions were treated with repeat SRT, one lesion with extended local irradiation, and whole-brain irradiation was used to treat three lesions (Table III).
Failure or radiation necrosis and re-irradiation.
Results of the Cox proportional hazards model for potential predictors of local failure are displayed in Table IV. Univariate analysis revealed significant differences in GTV maximal tumor diameter and PTV D5 with HR=1.938 (95%CI=1.367-2.748), p=0.0002 and HR=1.133 (95%CI=1.011-1.269), p=0.031, respectively. However, multivariate analysis revealed that the only significant difference was detected in GTV maximum tumor diameter (HR=1.961, 95%CI=1.369-2.807, p=0.0002).
Cox Proportional Hazard Model analysis of covariates that contribute to local failure.
Based on a toxicity study, one patient was found to have asymptomatic RN, which got better throughout the follow-up. There were no patients who developed symptoms or underwent surgical resection, and there were no cases of obvious acute toxicity.
Discussion
Herein, we report the clinical outcomes of HFSRT, including those of 30 Gy dose administered in 5 fractions, for BMs of <2 cm in diameter. To the best of our knowledge, this is the first study to report results of HFSRT at doses similar to those used in traditional SRS and limited to small BMs. A few cases of relatively small-size metastases were reported in previous investigations of clinical outcomes of treatment of BMs in general (6-9). Most of the cases in those studies included inconsistent prescriptions as well as postoperative and post-WBRT cases. This research included only non-postoperative and non-post-WBRT BMs. Additionally, we confirmed that a uniform prescription from a single institution is included to provide a highly reproducible procedure. In addition to the dose prescription, dosimetric parameters were clearly reported.
Redmond KJ et al. reported a model of tumor control probability following SRS and SRT for BMs as part of the American Association of Physicists in Medicine Working Group on Biological Effects of Hypofractionated Radiotherapy/SBRT (WGSBRT), based on dose and clinical data from an earlier study (10). Although some of these reports describe prospective findings for fractionated regimens for small tumors, 99% of the results were calculated from SRS data; thus, the results exhibited great uncertainty.
This research revealed that the 1- and 2-year local control rates for small tumors less than 2 cm treated with 30 Gy in 5 fractions were 93.0% and 81.5%, respectively, with no RN needed therapeutic intervention. Contemplating the toxicity of high doses of SRS, HFSRT for small BM might be an appropriate treatment strategy. In a subgroup evaluation of patients with lesions ≤1 cm versus 1-2 cm, the 1-year local control rate was significantly lesser in the 1-2 cm group than in the 1 cm group, identifying that even among small lesions ≤2 cm, which are categorized as small by RTOG classification, there are size-specific differences in results.
There are no previously published reports limiting lesion size to 2 cm or less, but there are several reports that pooled high numbers of small-to-medium lesions (3, 19-25).
Leonie Johannwerner et al. reported that outcomes for BMs at biological effective dose (BED) 49.6-66.7Gy12 [α/β=12 (linear-quadratic model)] in 69 patients with 1-4 BMs; LC at 1 year was 81% and prescriptions above 63 Gy12 were not associated with LC. However, BMs larger than 2.1 cm were included in 21 of all subjects. Furthermore, the prescriptions were not uniform (26).
Samanci et al. compared SRS with FSRT in 208 gamma knife-treated lesions with BMs <2.9 cm (median 1.1 cm) (6). The authors reported an excellent 6-month LC rate of 99% for both techniques. However, only 74 lesions were less than 2 cm in size and only 29 non-postoperative and non-post-WBRT BMs were observed. Although the authors specified a BED10 >50 Gy10, the number of fractions was not specified.
Yan M et al. reported on HFSRT of 133 metastases (21 lesions ≤1 cm and 112 lesions >1 cm) and showed that at 12 months, local failure was 17.8% for all lesions, 22.1% for lesions >1 cm, and 13.7% for lesions <1 cm (7). The prescriptions were not standardized and the results were based on groups that included postoperative lesions.
Putz F et al. reported on 98 patients treated with 10-fraction SRT (median tumor size of 1.8 cm) and 92 patients treated with SRS (median tumor size of 1.0 cm) (8). The 12-month LC rate was not statistically different between FSRT and SRS for lesions <1.0 cm in diameter, however it was considerably higher for FSRT versus SRS for lesions >1.0 cm in diameter, according to the authors.
Moraes FY et al. compared the prescribed doses of 15 Gy and 21 Gy for lesions smaller than 1 cm that were treated with SRS. The authors discovered that the risk of local failure and RN was similar for the two doses. Nevertheless, for 1-2 cm lesions, 15 Gy was insufficient to give LC and 21 Gy was needed to obtain this (27). Then, numerous groups reported that the 1-year LC rate for SRS was lower for lesions larger than 1 cm compared to lesions smaller than 1 cm (28-32). These outcomes may imply that 1 cm signifies the clinical cut-off for LC.
Contrarily to these studies, Matsuyama T et al. reported no significant difference in 1-year LC rates for SRT comprising relatively small lesions (median 0.8 cm) between the 1 cm or smaller and 1-2 cm groups. Their prescription was BED10 (α/β ratio=10) at approximately 80 Gy10 with minimal fractionation. The most frequent prescription was 32 Gy/2 Fr (peripheral) (9). Contrarily, our prescription was 48 Gy10 (BED10). Considering the research of Matsuyama T et al., we believe that 65 Gy10 (BED10) or more is required to control a 1-2 cm tumor.
In our research, toxicity was considered acceptable with only one case diagnosed with asymptomatic RN. RTOG recommends 18-24 Gy of SRS, but this has been demonstrated to result in a small number of symptomatic cases (2, 15). The crude overall risk of permanent morbidity or necrosis necessitating craniotomy after SRS for BMs is reported to be 4.7% to 7% (11, 33, 34). In prior research containing small lesions treated with SRT, the risk of symptomatic RN was very small, with some reports demonstrating no evidence of symptomatic RN (6-8, 19). This is in line with our findings, implying that a prescription of 30 Gy/5 Fr is associated with acceptable toxicity.
In contrast, Matsuyama T et al. reported progressive RN in 6 patients despite the prophylactic use of corticosteroids. Their prescription had a median peripheral BED2 of 288 Gy2, whereas for our prescription, it was 120 Gy2. This indicates that there is a capacity for increasing the dose to enhance the control rate for the 1-2 cm group.
This study has some limitations. First, with a single-center retrospective study, there is a possibility of patient selection bias. Additionally, differentiation between LC and RN is still a clinical question, and although a method to examine adverse radiation effects has been reported, it is not discussed in this investigation (2, 19, 33, 35). Although The Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) has recently been recommended for identifying treatment efficacy in brain tumors, our criteria for determining LC were based on RECIST (36). Many studies are still transitioning to RANO-BM and thus we estimate that it will take some time before consistent reporting of LC is achieved. Finally, there is no mention of interactions with systemic therapy in our research. The effects of the administration of targeted agents and immune checkpoint inhibitors, which have been used often in recent years, is thus unknown (37-39).
Conclusion
We report on HFSRT with 30 Gy in 5 fractions for metastatic BM <2 cm in diameter carried out at our institution. Considering the high toxicity of SRS, 5-fraction HFSRT may be a suitable treatment strategy for BM <2 cm. The outcomes also indicate the need for dose escalation in the 1-2 cm group. Future clinical trials for the treatment of BMs should integrate appropriate HFSRT prescriptions according to tumor size.
Acknowledgements
The Authors would like to thank Enago (www.enago.jp) for the manuscript review and editing support.
Footnotes
Authors’ Contributions
All Authors have contributed to the study conception, design and collaborated in the acquisition of clinical data. Data analysis and statistical analysis were performed by YK, and TO was responsible for statistical analysis. The manuscript was written by YK. All Authors have read and approved the manuscript.
Conflicts of Interest
The Authors declare no competing interests in relation to this study.
- Received July 18, 2023.
- Revision received August 29, 2023.
- Accepted August 30, 2023.
- Copyright © 2023 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).