Leukemic relapse remains a major problem after both autologous and allogeneic transplantation. In the single-arm ALL autograft study our results were very encouraging and suggest that Melph/TBI can produce disease-free survival results at least as good as with other conditioning regimens. The role of postautograft maintenance remains unclear, but we feel our results are sufficiently encouraging to justify a randomized study, particularly as we studied a group of patients with relatively poor prognoses. In our study comparing Cy and TBI with Melph and TBI in AML, we have shown a significant increase in antileukemic activity after transplantation following the latter conditioning regimen. The retrospective study of Melph/TBI in autologous versus allogenic transplantation suggested that in AML this antileukemic effect may derive from increased GvHD and is not present in the autologous setting. We hope that by increasing the intensity of our GvHD prophylaxis we can reduce the toxicity of Melph/TBI and preserve its antileukemia effect. Our experience with GM-CSF has been a little disappointing: despite facilitating neutrophil recovery, we were unable to demonstrate a clinical benefit in the treatment arm. We hope to further investigate the use of cytokine combinations in the transplant setting.