A Dose Escalation and Phase II Study of Gemtuzumab Ozogamicin (GO) with High-Dose Cytarabine (HiDAC) for Patients (pts) with Refractory or Relapsed Acute Myeloid Leukemia (AML): CALGB 19902.

GO is a calicheamycin toxin-conjugated humanized monoclonal antibody with specificity for the CD33 antigen expressed on blasts from 90% of AML pts and is approved for single agent use in CD33-positive older adults with relapsed AML not deemed chemotherapy candidates. We sought to define the safety and efficacy of combining GO (kindly supplied by Wyeth) with HiDAC for the treatment of adults with relapsed or primary refractory (after 2 cycles of induction therapy) AML. Eligibility included CD33 expression on >20% of blasts, no BMT within 6 mo or HiDAC within 3 mo, no prior hepatic disease or bilirubin >2 mg/dL. The initial cohort of 9 pts were treated with 2 doses of GO alone at 9 mg/m² on d1 and 4.5 mg/m² on d8; however, dose-limiting toxicity (DLT), defined as prolonged nadir or gr 4 or irreversble gr 3 toxicity, was difficult to determine given poor blast clearance and advanced disease. The protocol was amended to base DLT on treatment-related deaths (TRD) and to add eligibility restrictions of PS=0-2, no serious infection, and no cytotoxic therapy within 2 mo. No TRD were observed in the subsequent 9 pt cohort (IA) treated with HiDAC (ara-C 3.0 g/m² over 3h daily x 5d) followed by GO 9 mg/m² on d7, but 3/7 pts (cohort II) treated with HiDAC plus GO 9 mg/m² d1 and 4.5 mg/m² on d 8 died, indicating that this regimen was intolerable. Thereafter, all pts (n=44) received therapy as in cohort IA. The phase II portion of the study included 37 pts with relapsed AML and 7 with primary refractory disease (1/7 achieved complete remission [CR]). The accrual goal was met and the study was closed. The median age of the 37 pts with relapsed AML was 64 years (interquartile range: 55–69); 16 (43%) were female, 2 (5%) were African-American and 1 (3%) was Asian. The baseline performance status was 0 in 23 (62%), 1 in 10 (27%) and 2 in 4 pts (11%). 7 (19%) achieved CR, 2 (5%) CRp (no peripheral blasts, ≤ 5% marrow blasts, neutrophil recovery, platelet transfusion-independence, but platelets less than 100 K/ul), 15 failed to achieve CR (41%) and data are outstanding in 13 (35%). 12 patient deaths have been reported (5 treatment-related [14%], 4 disease-related, 1 unrelated to treatment or disease and 2 died of unknown causes). The following gr ≥ 3 toxicities occurred with a frequency of at least 10% based on data currently available for 25/37 patients with relapsed AML on the phase II component of the study: ileus (12%), hyperbilirubinemia (gr 3 in 20%, gr 4 in 4%), elevated SGOT (gr 3 in 16%, gr 4 in 8%), hypoalbuminemia in 12% (all grade 3), infection (gr 3 in 20%, gr 4 in 4%, and fatal in 8%) and hypoxia in 12% (all gr 3). No gr ≥ 4 hepatic veno-occlusive disease was reported. In summary, HiDAC (3.0 g/m2 over 3h qd on d 1–5) plus GO (9 mg/m2 on d 7) is reasonably well tolerated and associated with at least a 19% CR rate in pts with relapsed AML. This regimen merits further study, both in this setting and as a consolidation therapy for adult pts with AML in first remission.