Prospective Evaluation of Radiation Dose Escalation in Patients With High-Risk Neuroblastoma and Gross Residual Disease After Surgery: A Report From the Children's Oncology Group ANBL0532 Study.

PURPOSE A primary objective of the Children's Oncology Group (COG) ANBL0532 phase III study was to assess the effect of increasing local dose of radiation to a residual primary tumor on the cumulative incidence of local progression (CILP) in patients with high-risk neuroblastoma. PATIENTS AND METHODS Newly diagnosed patients with high-risk neuroblastoma were randomly assigned or assigned to receive single or tandem autologous stem-cell transplantation (SCT) after induction chemotherapy. Local control consisted of surgical resection during induction chemotherapy and radiotherapy after last SCT. Patients received 21.6 Gy to the preoperative primary tumor volume. For patients with incomplete surgical resection, an additional boost of 14.4 Gy was delivered to the gross residual tumor, for a total dose of 36 Gy. CILP (primary end point) and event-free (EFS) and overall survival (OS; secondary end points) were compared with the COG A3973 historical cohort, in which all patients received single SCT and 21.6 Gy without a boost. RESULTS For all patients in ANBL0532 receiving radiotherapy (n = 323), 5-year CILP, EFS, and OS rates were 11.2% ± 1.8%, 56.2% ± 3.4%, and 68.4% ± 3.2% compared with 7.1% ± 1.4% (P = .0590), 47.0% ± 3.5% (P = .0090), and 57.4% ± 3.5% (P = .0088) for all patients in A3973 receiving radiotherapy (n = 328), respectively. Five-year CILP, EFS, and OS rates for patients in A3973 with incomplete resection and radiotherapy (n = 47) were 10.6% ± 4.6%, 48.9% ± 10.1%, and 56.9% ± 10.0%, respectively. In comparison, 5-year CILP, EFS, and OS rates for patients in ANBL0532 who were randomly assigned or assigned to single SCT and received boost radiotherapy (n = 74) were 16.3% ± 4.3% (P = .4126), 50.9% ± 7.0% (P = .5084), and 68.1% ± 6.7% (P = .2835), respectively. CONCLUSION Boost radiotherapy to gross residual tumor present at the end of induction did not significantly improve 5-year CILP. These results highlight the need for new strategies to decrease the risk of locoregional failure.