MATERIALS AND METHODS

Detailed study inclusion criteria, data management issues, and statistical considerations are available in an accompanying article by Karrison et al.(2002). Briefly, a detailed data collection record was completed electronically as fully as possible by the participant (s) representing each center or collaborative group (Appendix A of Karrison et al.(2002)). Data were acquired on patients from individual centers anonymously by assigning a unique patient number to each patient so that no information on the patient's identity was ever transmitted to the central database. All treatment of the patients was conducted in centers with the usual standards of informed consent at the time of their treatment, and no patient was contacted to obtain additional information for this study. Data for some specific items were therefore incomplete for some patients, and it is indicated in each report when data are missing. Patients were divided into six subgroups (A through F) based on the balanced rearrangements present in the karyotype that was obtained at diagnosis of the therapy-related hematologic disease before any cytotoxic treatment was undertaken. The first four, 11q23, 21q22, inv (16)/t (16; 16), and t (15; 17), are frequently recurring breakpoints or aberrations seen in de novo hematologic disease. The fifth subgroup, Rare aberrations, seen at a low frequency, involved balanced rearrangements identified at least twice in published reports of hematologic disease or at the Workshop. The exception is the t (9; 22)(q34; q11. 2), which occurs at a frequency of> 1%, but was present in only 10 cases in the current series and thus was too small to be treated separately; it was therefore assigned to the Rare aberrations subgroup. The final subgroup, Unique aberrations, included all other cases with balanced aberrations, which were present only once in this series and were not reported previously as recurrent findings.

RESULTS