NCOG-01. PRESERVATION OF NEUROCOGNITIVE FUNCTION (NCF) WITH HIPPOCAMPAL AVOIDANCE DURING WHOLE-BRAIN RADIOTHERAPY (WBRT) FOR BRAIN METASTASES: PRELIMINARY RESULTS OF PHASE III TRIAL NRG ONCOLOGY CC001

Abstract

Purpose/Objective(s): Based on preliminary evidence that radiation to the neuroregenerative hippocampal stem cells plays a role in NCF decline, the phase II NRG/RTOG 0933 trial demonstrated memory-preservation following HA-WBRT. The phase III NRG-CC001 trial of WBRT plus memantine without (WBRT+M) or with hippocampal avoidance (HAWBRT+M) sought to validate these findings. Materials/Methods: Adult patients with brain metastases were stratified by RPA and receipt of prior radiosurgery/surgery and randomized to WBRT+M versus HA-WBRT+M (30 Gy in 10 fractions). Standardized NCF tests were performed at baseline, 2, 4, 6, and 12 months (mos). The primary endpoint was time to NCF failure defined as decline using the reliable change index on at least one of the following tests: Hopkins Verbal Learning Test-Revised, Trail Making Test, or Controlled Oral Word Association. Cumulative incidence was used to estimate time to NCF failure (death without NCF failure was treated as a competing risk). Betweenarms differences were tested using Gray’s test. To detect an 11% absolute reduction in 6-month NCF failure, 382 analyzable patients were required for 90% power with two-sided aZ0.05. Due to possible non-compliance, the sample size was increased by 25% (510 patients). Results: 518 patients were randomized from July 2016 to March 2018. Median age was 61.5 years. Treatment arms did not differ in baseline characteristics. Grade3 toxicity did not differ (pZ0.88). Median followup for alive patients was 6.1mos. NCF testing compliance was 69% at 6mos and 61% at 12mos. Treatment arms did not differ in baseline NCF, overall survival (hazard ratio (HR)Z1.13, 95% confidence interval (CI): 0.89-1.44, pZ0.31) or intracranial progression-free survival (HR 1.12, 95% CI 0.90-1.39, pZ0.33). Time to NCF failure was significantly longer in favor of HA-WBRT+M. The NCF failure rates following WBRT+M vs. HA-WBRT+M were 12.8% (95% CI 8.5-18.0%) vs. 11.2% (7.1-16.3%) at 2mos, 63.0% (55.6-69.5%) vs. 53.7% (46.1-60.8%) at 4mos, and 69.1% (61.8-75.3%) vs. 58.0% (50.2-64.9%) at 6mos (pZ0.012). In analyses adjusted for stratification factors, HA-WBRT+M (HRZ0.72; 95% CI: 0.56-0.94, pZ0.016) and age
61 years (HRZ0.61, 95% CI: 0.46-0.81, pZ0.0006) predicted for longer time to NCF failure. Test for interaction between treatment arm and age was non-significant (pZ0.67). Conclusion: Preliminary analysis confirms our hypothesis that conformal avoidance of the hippocampal neuro-regenerative stem cell niche during WBRT preserves NCF while achieving similar intracranial control and survival. While age independently predicts for NCF, the NCF benefit of hippocampal avoidance does not differ by age. Supported by grants UG1CA189867 (NCORP), U10CA180868 (NRG Oncology Operations), DCP from the National Cancer Institute