Evidence of a Treatment Dose Response in Acute Nonlymphocytic Leukemias Which Occur After Therapy of Non-Hodgkin's Lymphoma

We evaluated the occurrence of second cancers among 517 patients with non-Hodgkin's lymphoma (NHL) treated at the National Cancer Institute. Nine cases of acute nonlymphocytic leukemia (ANL) were observed compared to 0.08 cases expected (ratio of observed to expected cases, 105; 95% confidence limits, 48; 199). The excess risk of ANL was 4.1 cases per 1000 patients per year; the cumulative risk of ANL at 10 years was 7.9 +/- 3.2% (S.E.). A case-control study within the NHL cohort revealed that patients treated with both radiation and chemotherapy were at greater risk of ANL than were patients who received single-modality therapy (relative risk, 6.0; p less than 0.05), especially if the therapy included total-body or hemibody radiation. A positive correlation between cumulative radiation dose to the bone marrow and risk of ANL was demonstrated, independent of chemotherapy duration. A similar correlation between chemotherapy dose and risk of ANL was suggested but could not be proven with the available data. An apparent association between ANL risk and indolent NHL histological subtypes was due to the significantly larger amounts of potentially leukemogenic therapy to which these patients were repeatedly exposed. Only one case of ANL occurred among NHL patients whose initial therapy produced a durable complete remission. Our data are compatible with a multistep model of leukemogenesis and also underscore the need for curative NHL treatment regimens which minimize the duration and quantity of therapy required for optimum patient management.