Central nervous system relapse in children with acute lymphoblastic leukemia

Abstract

Relapses in children treated for acute lymphoblastic leukemia (ALL) have become significantly less common over the past 25 years as primary therapy has improved. In spite of our improved ability to identify children at higher risk of relapse and the use of therapy based on the relative risk of treatment failure, however, relapses do continue to occur. In the past, efforts to produce long-term disease-free survival in patients who have experienced a relapse have yielded a rather discouraging salvage rate ranging from about 10% for patients with bone marrow relapses to as high as 70% for patients with an initial testicular relapse (7). Patients with central nervous system (CNS) relapses have had an outcome somewhere between these values as summarized in a recent review (6). The report by Kumar et al. suggests that about half of the patients with an initial CNS relapse may be responsive to salvage therapy without major sequelae in terms of neurotoxicity (4). ity to detect and quantitate residual disease will be to clarify the progression of disease that leads up to and then follows an “isolated” CNS relapse (2). Whichever mechanism is operative, the successful management of CNS relapse requires therapy that addresses the CNS and the high probability of subsequent bone marrow relapse. Current therapy for newly diagnosed patients with ALL generally includes chemotherapy systemically and intrathec~ly to prevent CNS relapse. If i~adiation is used, it is generally reserved for patients with a high risk of relapse, perhaps 30% of the total population, and for those few patients presenting with CNS disease. In the series reported here, irradiation was deferred to enable moderate dose methotrexate to be given intravenously without a high risk of leukoencephalopathy. Patients who now develop a CNS relapse are quite likely to have not received prior irradiation, as in this series in which 17 of the 18 relapses occurred in unirradiated patients. By common convention we use the term isolated CNS The results of the study by Kumar and associates are relapse to refer to the presence of lymphoblasts present similar to prior trials of cr~iospinal radiotherapy for isoin the cerebra-spinal fluid with an increased cerebro-spilated meningeal retapse in children with acute lymphonal fluid cell count (or much less commonly a mass lesion blastic leukemia. Nearly 10 years ago, the Childrens Canproducing a focal neurologic deficit) in the absence of cer Group reported 60% 3-year survival among 95 childemonstrable leukemic cells elsewhere. The key word is dren with isolated meningeal relapse (l), which may be “demonstrable.” The observation that CNS relapse is compared with the 47-97% 95% confidence interval (our frequently followed by relapse in the bone marrow within estimate) for the corresponding rate in the study of 18 a relatively short period of time forces us to consider that patients by Kumar et al. This was achieved with intratheat the time of CNS relapse, many patients already have cal methotrexate alone during induction, rather than with lymphoblasts proliferating in the bone marrow, although the triple drug therapy used by Kumar and his colleagues, at a level too low to be detected by commonly available with a lower dose of spinal radiotherapy, 12 Gy, rather techniques. An alternative explanation for the pathogenethan with the 15 Gy used in the study reported herein, sis of the subsequent bone marrow relapse is that the and with maintenance intrathecal methotrexate at 3-month marrow is repopulated from lymphoblasts that have prointervals for 2 years. The use of craniospinal irradiation liferated in the meninges or in the cortical perivascular for treatment of CNS relapse dates back almost 20 years spaces and that because of their residence in these pharto a small but convincing study conducted by the British macologically protected sites, they may emerge from the Medical Research Council (8). In that trial, craniospinal CNS with a significant degree of drug resistance. A goal irradiation offered significant protection against subsefor future studies that use techniques that extend our abilquent CNS relapse as compared to cranial irradiation.