COMPARATIVE RESULTS OF 2 INTENSIVE TREATMENT PROGRAMS FOR CHILDHOOD ACUTE LYMPHOBLASTIC-LEUKEMIA - THE BERLIN-FRANKFURT-MUNSTER AND DANA-FARBER-CANCER-INSTITUTE PROTOCOLS

We analyzed two of the most successful trials for childhood acute lymphoblastic leukemia, from the Berlin-Frankfurt-Munster group (BFM) and the Dana-Farber Cancer Institute (DFCI), to determine their similarities and differences and suggest ways to improve future treatments. Protocol ALL-BFM 81 (n = 611) was conducted between 1981-1983 in 37 centers in West Germany and Austria; Protocol DFCI 81-01 (n = 286) was conducted between 1981-1985 in seven centers within the United States. The BFM study used a risk score, based on peripheral lymphoblast count, liver size, and spleen size at the time of diagnosis to assign patients into one of the three treatment arms. The DFCI study classified patients into two groups: standard risk patients (age 2-9 years, with white blood count < 20,000/mm3, and with no T-cell markers, mediastinal mass, nor central nervous system disease), and high risk patients (all others). Both studies involved intensive chemotherapy, although treatment strategies, cumulative drug doses and cranial radiation protocols differed. Study populations in the two trials were comparable with respect to age, sex, white blood count and proportion of patients with T-cell markers at the time of diagnosis. Using DFCI risk group criteria, 37% of BFM and 38% of DFCI patients were classified as standard risk. Median follow-up times were 6.8 years and 6.1 years, respectively. Six-year event-free survival percentages (+/- standard error) were nearly identical: BFM 69% (+/- 2%) and DFCI 70% (+/- 3%) overall; BFM 79% (+/- 3%) and DFCI 80% (+/- 5%) for standard risk patients; and BFM 63% (+/- 2%) and DFCI 63% (+/- 4%) for high risk patients. In both studies, outcomes for boys and girls were similar. The BFM study had a higher incidence of testicular relapse (p = 0.004) and CNS relapse (p = 0.004; especially among the 177 non-irradiated patients), than the DFCI study, but a lower incidence of death in remission (p = 0.20). While none of the five infants less than one year old in the DFCI protocol were successfully treated, seven of 10 infants in the BFM trial remained disease-free (p = 0.0004). Because the patterns of relapse differed, we recommend the integration of major components of each program into a single, new protocol in order to improve outcome.