Up-Front Window Trial of Gemtuzumab Ozogamicin (GO) in Previously Untreated Elderly Patients with AML: An EORTC Leukemia Group Study.

Gemtuzumab ozogamicin, a conjugate of a humanized anti-CD33 monoclonal antibody linked to the cytotoxic antibiotic calicheamicin, has shown significant antileukemic activity in relapsed AML with a favorable toxicity profile. The EORTC-LG performed a phase II study to investigate the activity and toxicity of GO used alone and in combination with conventional chemotherapy for remission induction in medically fit elderly pts (age between 61 and 75 years; WHO PS 0-1) with untreated AML. GO was administered iv over 2 hrs at the FDA-approved dose of 9 mg/m2. A second dose was given on day 15 in the absence of progressive disease or excessive toxicity and response was assessed four weeks later. This was to be followed by a course of conventional chemotherapy with mitoxantrone, cytarabine and etoposide (MICE regimen) and no further treatment in complete responders. Between 09/00 and 10/01 64 pts were enrolled , 57 of whom were evaluable for response and toxicity. The median age was 68 yrs (range 61–73); 43 pts had primary AML; CD33 positivity (≥20% marrow blasts) was documented in 44 pts (85%). GO therapy was completed by 47 of the 57 pts (82%): primary reasons 10 pts did not receive the scheduled second dose were disease progression (4), early death (2), toxicity (2) and protocol violation (2). The rate of initial response to GO was 35.1% (95% CI, 22.9%–48.9%) with 13 pts achieving complete remission (CR) and 7 CRp (complete remission with incomplete platelet recovery); 6 additional pts entered partial remission (PR). There were 3 (5.3%) toxic deaths (1 ARDS, 1 VOD, 1 infection) and 28 pts had resistant disease. Severe myelosuppression was the primary toxicity to GO. Grade 3 or 4 non-hematologic toxicities in more than 5% of pts during GO therapy included febrile neutropenia (39%), infection (28%), elevations of bilirubin, SGPT and creatinine of 8%, 6% and 6%, respectively. A clinical picture compatible with hepatic VOD developed in 3 pts (5%) resulting in 2 deaths (1 early, 1 in CRp). No baseline characteristics predicted for response to GO, but pts expressing CD33 in >80% marrow blasts tended to respond more favourably to the immunoconjugate (CR+CRp 7/14, 50%). Altogether, 51 pts survived GO therapy and 38 of them (75%) went on to receive a course of MICE after a median interval of 49 days (range 12–77) from the first dose of GO. Excessive toxicity (11 pts), protocol violation (1 pt) and treatment refusal (1 pt) were the main reasons for not starting chemotherapy. The ultimate best response rate was 54.4% (31/57; CR 35.1% and CRp 19.3%) increasing up to 65.8% among the 38 pts who were able to complete the whole induction sequence. Five pts died of toxicity during the MICE segment (2 VOD, 1 infection, 1 heart failure, 1 multi-organ failure) for an overall treatment related mortality of 14.1% (8/57). With a median follow-up of 3 yrs, 2-year overall survival (OS) and disease-free survival (DFS) estimates were 11.0% (SE=4.1%) and 12.1% (SE=3.5%), respectively. We conclude that: 1) single agent, standard dose GO has significant activity in untreated AML of the elderly with an acceptable safety profile; 2) sequential combination with standard chemotherapy is feasible and may improve treatment outcome in this poor risk disease, but myelotoxicity and liver dysfunction are of concern. An ongoing randomized trial will determine the contribution of GO given at reduced dose (6 mg/m2) to the observed antileukemic effect and toxicity.