Peripheral blast counts at diagnosis of late isolated bone marrow relapse of childhood acute lymphoblastic leukemia predict response to salvage chemotherapy and outcome. Berlin-Frankfurt-Münster Relapse Study Group.

PURPOSE: In newly diagnosed childhood acute lymphoblastic leukemia (ALL), a high tumor burden indicates a poor prognosis, while no such link has been established yet after relapse. The impact of the absolute peripheral blast count (PBC) at the time of relapse on the response to salvage chemotherapy after a late isolated bone marrow (BM) relapse is the subject of this prospective analysis. PATIENTS AND METHODS: Since 1983, 260 children with a first isolated BM relapse of ALL that occurred 6 months or later after elective cessation of front-line therapy were enrolled onto four consecutive multicenter trials of the Berlin-Frankfurt-Münster (BFM) Relapse Study Group. All patients received intensive multiagent induction and consolidation chemotherapy for 6 months, followed by maintenance therapy with methotrexate (MTX) and thioguanine for 2 years. Treatment of subclinical meningeal leukemia consisted of high-dose intravenous MTX and intrathecally administered cytostatic drugs, which was augmented by cranial irradiation since 1988. RESULTS: At the time relapse was diagnosed, PBC varied considerably among patients (median, 1,060/microL; range, 0 to 106,800/microL). Achievement of a second complete remission (CR) was not significantly different in children without detectable circulating blasts at relapse (37 of 38) and those with moderate (1 to 9,999/microL) PBC (165 of 171). In contrast, only 42 of 51 children with high PBC (> or = 10,000/microL) achieved a second CR (P = .0015). At a median follow-up time of 40 months, the 10-year event-free survival (EFS) probability was significantly (P = .0001) higher in children without circulating blasts (.64) than in children with moderate PBC (.32) or high PBC (.10). There was a preponderance of boys in the group without detectable circulating blasts, while the three PBC-defined groups did not differ with respect to frontline treatment, age at initial diagnosis, age at relapse, time off therapy, or salvage treatment protocol. On sequential univariate and multivariate analysis, only duration of first remission > or = 48 months was an additional independent indicator of adverse prognosis, while preventive cranial irradiation improved outcome independently of PBC. CONCLUSION: The absence of blasts on peripheral-blood smears at the time of a first late isolated BM relapse of childhood ALL is associated with a favorable response and prognosis in chemotherapy-treated children, who should be regarded as ineligible for bone marrow transplantation (BMT) unless a second round of chemotherapy has failed to produce a response.