Background: Children with acute lymphoblastic leukemia with multiple poor prognostic factors and who have a lymphomatous mass at diagnosis, whether of T- or non-T-immunophenotype, are at increased risk of short term remission and extramedullary recurrence, and are in need of better therapies.

Methods: Six hundred and ninety-four eligible patients ranging in age from 1-20 years were entered on the study. Sixty-five percent of the patients had T-cell immunophenotype. Of these, 678 were randomized to one of four regimens: Regimen A: Berlin-Frankfurt-Munster (BFM) 76/79; Regimen B: LSA2-L2 with cranial irradiation; Regimen C: LSA2-L2 without cranial irradiation; and Regimen D: the New York (NY) regimen.

Results: Complete remission was induced in 97% of patients. The overall event free survival (EFS) +/- the standard deviation was 60 +/- 4% 6 years after diagnosis, in contrast to 36 +/- 6% in a comparable historic group. The EFS of the 371 T-cell patients was 62 +/- 7%. EFS was best on the NY (67 +/- 7%) and the BFM (67 +/- 6%) arms. These were significantly better than the EFS on the 2 LSA-L2 regimens, with an EFS of 53 +/- 8% (Regimen B) and 42 +/- 11% (Regimen C) (P = 0.03 and 0.0003 for NY vs. Regimen B and NY vs. Regimen C; P = 0.01 and 0.0001 for BFM vs. Regimen B and BFM vs. Regimen C). Regimen C had a 3-fold greater central nervous system (CNS) recurrence rate than the identical chemotherapy Regimen B (16 +/- 5% vs. 6 +/- 4%; P = 0.02), although the difference in overall EFS did not reach the required level for significance. Testicular recurrence varied from 2-8% in comparison with 20% in the historic group. EFS was not influenced by age, gender, CNS disease at diagnosis, morphology, or immunophenotype. In addition to treatment regimen and early response rate, initial leukocyte count, hemoglobin level, liver, spleen, and lymph node enlargement, and the presence of a mediastinal mass had univariate prognostic influence on EFS. In multivariate analysis, only the kinetics of response, leukocyte count (unfavorably, P < 0.0001), and mediastinal mass status (favorably, P = 0.01) were prognostic.

Conclusions: The adverse prognostic implications of lymphomatous ALL can be minimized by the NY and BFM regimens. Cranial irradiation resulted in better CNS disease control when added to the LSA2-L2 regimen, but did not improve the overall disease free survival. With improved systemic chemotherapy, there was no excess of lymph node, testicular, or other local recurrence without prophylactic irradiation to sites of initial bulk disease or to the testes.