Previous series including our earlier study of survivors of childhood acute lymphoblastic leukemia have documented the presence of nonclonal chromosome abnormalities in peripheral blood lymphocytes. To investigate whether or not these abnormalities persist, we restudied nine patients, all of whom had received radiotherapy and a minimum of 3 years of systemic chemotherapy. Therapy had been discontinued a median of 3.9 years prior to study in our first report and a median of 7.2 years prior to study in this follow-up report. Thirty metaphases from short-term cultures of phytohemagglutinin-stimulated peripheral blood from each of nine patients and nine matched controls were analyzed using G-banding. Nonclonal chromosome abnormalities were present in cells from six of the nine patients and 3.0% of patients' cells overall, as compared to seven of ten patients and 3.5% of patients' cells in the earlier study. All patients studied twice had abnormalities on one or both occasions. No evidence for the development of cytogenetically abnormal clones was found. Combined data from the two evaluations indicate that all of the patients have a small population of long-lived lymphocytes with nonclonal chromosome abnormalities, the frequency of these cells appears to remain remarkably constant over time, and the proportion of cells with chromosome abnormalities is approximately threefold greater in patients than in controls. It has become apparent that survivors of acute lymphoblastic leukemia are at increased risk for development of second malignant neoplasms. The relationship between therapy-induced chromosome damage and the development of second malignant neoplasms is currently unknown. We speculate that some cells, possibly in tissues other than peripheral blood, are preneoplastic by virtue of therapy-induced chromosome rearrangements involving critical genes and, following a long latent period, may undergo subsequent genetic or cellular changes that result in malignant transformation. Unfortunately, however, estimation of genetic damage through cytogenetic studies of peripheral blood T-lymphocytes does not appear to be a useful method for prediction of second malignant neoplasms following acute lymphoblastic leukemia, since most or all of the patients have a low level of abnormal cells.