Results and Significance of Six Randomized Trials in Four Consecutive ALL-BFM Studies

Abstract

Beyond all biological properties and clinical features of individual patients with acute leukemias, therapy by itself and its quality has emerged as the most important predicting factor for success or failure. The acute leukemias still need adequate management, no matter how the geno- or phenotype of the leukemic clone is characterized. Enforced by the antileukemic efficacy of improved therapy, some risk factors of the past have lost importance, whereas others are still meaningful. Dependent on quality of chemotherapy, the proportion of survivors in acute lymphoblastic leukemia (ALL) varied between 30% and 40% with the standard therapy in the early 1970s, and between 50% and 70% with intensified therapy of the 1980s. All ądvanced therapy concepts are fraught with danger. They have to be balanced by knowledge, specific competence, commitment and equipment. Thus, incompetence and insufficient infrastructures in individual institutions are adverse prognostic factors. The majority of treatment failures still occur unpredictably and are rather unexpected. Only early relapses are more easily suspected as a consequence of an initial inappropriate therapy response. In addition, the prognostic significance of two chromosomal translocations (4/11, 9/22) is recognized. In ALL, allogenic bone marrow transplantation is only an option in front-line therapy for a small minority of patients, whereas after (early) relapse this method offers the only potential rescue chance. In the acute nonlymphoblastic leukemias, only a selected few patients need this method, according to risk factors that have been characterized over the last years. Only therapy distinguishes readily between success and failure. Risk factor therapy can be assessed only in adequate randomized trials. In four stratified BFM trials performed between 1976 and 1986, answers to six randomized therapy questions have been summarized as follows: In high-risk patients in studies with BFM 76 and 79 (36%, risk index >2), either early or late intensification (protocol II) improved event-free interval (EFI) dramatically if compared with the preceding study, BFM 70 (0.64 vs. 0.38, p = 0.0001). Non-risk patients (62%, risk index <2), three vincristine/prednisone reinductions (BFM 79 study) had no major influence on prognosis (EFI 0.76 vs. 0.71, p = 0.44).

Methotrexate, as applied in the BFM 81 study, could not adequately replace standard CNS preventive therapy (brain irradiation 18 Gy) if applied to patients (57%) with a risk factor of 0–1.2 (EFI 0.68 vs. 0.78, p = 0.08). In patients in the BFM 83 study with a risk factor of <0.8 (27%), no protocol III impaired EFI significantly (EFI 0.62 vs. 0.84, p = 0.01). 12 Gy brain irradiation proved to be as effective as 18 Gy irradiation in patients (25%) with a risk factor of 0.8-1.2 (BFM 83 study, EFI 0.74 vs. 0.70). Patients in BFM 81 and 83 studies (85%, n=746), randomized for 24 months total duration of therapy (vs. 18 months) had an EFI of 0.83 (vs. 0.77, p = 0.04). For the vast majority of patients in these trials, only tumor burden, age, and sex are important prognostic factors beyond the quality of therapy and initial response.