Stereotactic Radiotherapy vs. Whole Brain Radiation Therapy for Patients with 1-10 Brain Metastases from Small Cell Lung Cancer: Results of the Randomized ENCEPHALON (ARO 2018-9) Trial

Purpose/Objective(s) Although SRT is preferred for limited brain metastases from most histologies, WBRT has remained the standard of care for patients with SCLC. It remains unclear whether the benefit of WBRT to SRT for the intracerebral tumor control outweighs the potential neurocognitive risks in SCLC patients. Materials/Methods This pilot-trial is a single-center prospective, randomized, two-arm Phase II study. The primary endpoint is neurocognition after cerebral irradiation in SCLC patients treated with WBRT or SRT (radiosurgery (SRS) with 20 Gy or 18 Gy or hypofractionated SRT with 30 Gy in 5 Gy fractions for lesions >3 cm) defined as a drop of at least 5 points from baseline in Hopkins Verbal Learning Test–Revised (HVLT-R) total recall subscale at 3 months after baseline. Eligible patients had histologically confirmed SCLC, MRI-confirmed cerebral metastasis (not resected, maximum number of 10), Karnofsky performance score >50 and no prior irradiation to the brain. Patients were randomly assigned (1:1) to either SRT or WBRT. Secondary endpoints included survival parameters, quality of life, toxicity and neurocognitive assessments. Results 56 patients were randomized to either WBRT or SRT. The modified intention-to-treat (mITT) set included all randomized patients, who started study treatment with WBRT (n=25) or SRT (n=26). Prior to imputation, the primary endpoint was reached in 7.7% (n=2) of patients in the SRT group and 24.0% (n=6) of patients in the WBRT group (mITT set). After multiple imputation via predictive mean matching, the primary endpoint was analyzed using the Cochrane-Mantel-Haenszel test stratified for time of appearance (p=0.0723). For preliminary OS analysis in the mITT set, data up to 6 months were considered. Patients not having reached the endpoint were censored at 181 days. There was no significant difference in survival probability between treatment groups (p=0.36). Median time to death (at 6 months) was 124.0 (Q1 43.0- Q3 139.5) days in the SRT group and 131.0 (Q1 107.0-Q3 150.0) in the WBRT group. Conclusion SCLC patients in the WBRT group were at a greater risk of a significant decline in neurocognitive function 3 months after baseline compared with the SRT group. SRT should be considered one of the standards of care for patients with brain metastases from SCLC. Although SRT is preferred for limited brain metastases from most histologies, WBRT has remained the standard of care for patients with SCLC. It remains unclear whether the benefit of WBRT to SRT for the intracerebral tumor control outweighs the potential neurocognitive risks in SCLC patients. This pilot-trial is a single-center prospective, randomized, two-arm Phase II study. The primary endpoint is neurocognition after cerebral irradiation in SCLC patients treated with WBRT or SRT (radiosurgery (SRS) with 20 Gy or 18 Gy or hypofractionated SRT with 30 Gy in 5 Gy fractions for lesions >3 cm) defined as a drop of at least 5 points from baseline in Hopkins Verbal Learning Test–Revised (HVLT-R) total recall subscale at 3 months after baseline. Eligible patients had histologically confirmed SCLC, MRI-confirmed cerebral metastasis (not resected, maximum number of 10), Karnofsky performance score >50 and no prior irradiation to the brain. Patients were randomly assigned (1:1) to either SRT or WBRT. Secondary endpoints included survival parameters, quality of life, toxicity and neurocognitive assessments. 56 patients were randomized to either WBRT or SRT. The modified intention-to-treat (mITT) set included all randomized patients, who started study treatment with WBRT (n=25) or SRT (n=26). Prior to imputation, the primary endpoint was reached in 7.7% (n=2) of patients in the SRT group and 24.0% (n=6) of patients in the WBRT group (mITT set). After multiple imputation via predictive mean matching, the primary endpoint was analyzed using the Cochrane-Mantel-Haenszel test stratified for time of appearance (p=0.0723). For preliminary OS analysis in the mITT set, data up to 6 months were considered. Patients not having reached the endpoint were censored at 181 days. There was no significant difference in survival probability between treatment groups (p=0.36). Median time to death (at 6 months) was 124.0 (Q1 43.0- Q3 139.5) days in the SRT group and 131.0 (Q1 107.0-Q3 150.0) in the WBRT group. SCLC patients in the WBRT group were at a greater risk of a significant decline in neurocognitive function 3 months after baseline compared with the SRT group. SRT should be considered one of the standards of care for patients with brain metastases from SCLC.