Incidence of Extramedullary Relapse of Acute Myeloid Leukemia Following Transplantation with Busulfan-Based Conditioning Regimens.

Rates of extramedullary relapse are reported to be more than twice as high following transplantation with busulfan-based conditioning regimens (Simpson et al, Bone Marrow Transplantation 1998, Lee et al, Bone Marrow Transplantation 2000) compared to total body irradiation (TBI)-based regimens (Mortimer et al, J. Clin Oncol 1989) in acute myeloid leukemia (AML). Up to 50% of relapses following transplantation with busulfan-based therapies are reported to be extramedullary, with more than 25% as isolated extramedullary relapses. We analyzed 228 consecutive AML patients who underwent allogeneic (n=192) or autologous (n=36) transplantation following busulfan-based conditioning regimens between 2/84 and 12/03. One hundred twenty-three patients were in remission at the time of transplantation, 67 had relapsed disease, and 38 had never achieved remission. One hundred twenty-seven patients received BuCy, 77 received BuCyVP16, and 24 (all autologous) received BuVP16. All patients had a lumbar puncture demonstrating no evidence of leukemia, and all received intrathecal methotrexate prophylaxis prior to receiving busulfan. Seventy-nine patients (35%) relapsed (any site) at a median of 6 months (range, 1 to 44 months) after transplantation. Fifteen patients (7%) relapsed at extramedullary sites, including 7 in the CNS. Ten patients (4%) had isolated extramedullary relapses at a median of 13 months (range, 1 to 44 months) after transplantation. Four had isolated CNS relapses, and 4 had isolated cutaneous relapses. Of patients transplanted in remission, only 2 relapsed at extramedullary sites, including one patient with an isolated CNS relapse.

This retrospective study includes a substantially larger cohort of patients with longer follow-up than prior studies examining extramedullary relapse in patients with AML following transplantation with busulfan-based conditioning. In contrast to other reports, these results demonstrate that the incidence of extramedullary relapse is comparable to that reported with TBI-based conditioning regimens for patients with AML.