Purpose: An analysis of survival outcome following isolated central nervous system (CNS) relapse treated with craniospinal irradiation (CSI) and additional chemotherapy in children with acute lymphoblastic leukemia (ALL) was conducted.
Methods and materials: Eighteen of 344 pediatric patients with ALL who attained initial complete remission on the St. Jude Children's Research Hospital "Study XI" prospective protocol (1984-1988) developed a CNS relapse as first adverse event. Median interval to isolated CNS relapse was 7.5 months (range = 2-40 months) after achieving initial complete remission. At diagnosis, 14 of the 18 children were categorized as "high risk" for subsequent leukemic relapse. Preventive cranial irradiation [PCI (18 Gy)] was delivered as planned to one of the 14 "high-risk" children. The other 13 "high-risk" patients experienced a CNS relapse during the first year of continuation therapy prior to week 52 of planned PCI. All four "low-risk" patients experienced a CNS relapse beyond the first year of continuation therapy; none were scheduled to receive PCI. Following isolated CNS relapse, all 18 patients were treated on a prospective contingency of "Study XI" trial consisting of intensified reinduction chemotherapy, weekly intrathecal methotrexate/hydrocortisone/Ara-C x 4-6 injections, craniospinal irradiation (cranium to 24.0 Gy and spine to 15.0 Gy at 1.5 Gy/fraction) and maintenance systemic therapy for a minimum of 1 year.
Results: Ten of 18 patients remain in continuous complete secondary remission at 17 to 50 months post-CNS relapse. Second sites of relapse in the remaining eight children were as follows: CNS in four, bone marrow in three, and bilateral testicular in one patient. Each of these eight patients died of progressive leukemia. At a median followup of 40 months post-initial CNS relapse, the 3-year secondary Kaplan-Meier survival and event-free survival are 72% and 56%, respectively. Minimal long-term neurotoxicity was associated with the treatment regimen. The most important prognostic factors predicting continuous secondary remission included white blood cell count at diagnosis (p = 0.05), and duration of initial remission (p = 0.04).
Conclusion: This trial demonstrates that more than one-half of patients may be successfully salvaged with intensified chemotherapy and craniospinal irradiation without significant morbidity following an isolated CNS relapse, despite previous multiagent chemotherapy though virtually no prior PCI in childhood ALL.