Fractionated Doses Mylotarg in First Relapsed AML Patients Is Effective and Safe: Results of a Prospective French Alfa Group Study.

Gemtuzumab ozogamicin (GO, Mylotarg®) is a humanized calicheamicin-conjugated monoclonal antibody targeted to the CD33 antigen (ag) present on myeloblasts. Phase 2 single agent trials have demonstrated a 28% response rate in acute myeloblastic leukaemia (AML) in first relapse. In these trials GO was used at a dose of 9 mg/m2 for two doses at 14 days interval. GO reaches its site of action by binding to CD33 ag with subsequent internalisation of the drug-receptor complex. A continuous renewed expression of CD33 ag on the cellular surface after a first exposure to GO was described. This observation led to hypothesize that treatment with frequent pulses exposures could facilitate intracellular drug loading and augment its cytotoxicity. Accordingly, we conducted a pilot study testing fractionated doses of GO, 3 mg/m2 administered at days 1, 4 and 7 for the induction course in 11 patients (pts) with poor prognosis AML. Analysis showed 4/11 (36%) responses and good haematological and liver tolerance. We decided to confirm these results in a phase 2 sequential non controlled trial, planned to detect a 15% benefit in response rate with Mylotarg® from the control rate of 15% (with type I and II error rate at 5% and 15%, respectively). Inclusion criteria were: de novo AML CD33+ adult pts in first early relapse with a duration of first remission < than 18 months. 56 pts were included between 11/12/03 and 24/07/05. At the second interim analysis, based on the first 26 pts the trial was stopped because of the crossing of the upper boundary, with the conclusion of benefit of Mylotarg®. Results of this second interim analysis: Median age was 65 years (22 to 75 years). Median duration of first remission was 9 months (4 to 18 mo). Karyotype was of intermediate risk in 17/24 (71%) pts and poor in 7. All pts received GO intravenously at a dose of 3 mg/m2 on days 1, 4 and 7. Response criteria were those of the revised criteria of NCI. Eight pts reached CR and two CRp (platelets below 100 10.⁹/ L) for a total response rate of 10/26 (38.5%). Median time to ANC recovery to 0.5 10.⁹ /L for CR and Crp pts was 24 days (10 to 38 days); median time to platelets recovery to 50 10.⁹/L for CR pts was 18 days (4 to 38 days); median number of red blood cells transfusions was 4 (0 to 18), median number of platelets transfusions was 4 (0 to 28). A transient rise in ASAT or ALAT or bilirubin (17 grade 1, 2 grade 2 and only 1 grade 4) was observed in 20/26 (77%) pts. No VOD was observed. Infections grade 3 were observed in 11 patients. All responders pts received consolidation chemotherapy with HD AraC. In summary: This study confirms that fractionated doses of GO allows to achieve response rate close to that observed in previous phase 2 studies, permits an early recovery of platelets and ANC, and has a low toxicity profile, leading to the possibility for all responsive patients to receive additional consolidation therapy. Final analysis with data of all patients and follow-up will be presented.