Abstract
Gemtuzumab Ozogamycin (GO, Mylotarg), a humanized CD33-Chalechiamycin conjugate has been incorporated as an integral part of therapy of relapsed and de novo AML. Early data generated in relapsed AML using single agent GO demonstrated higher degree of response to GO in patients with high CD33 expression that correlated with lower multi drug resistance (MDR). There is little data on the efficacy of GO in those with low CD33 expression in de novo setting using GO in combination with conventional chemotherapy. Current therapies that utilize GO include patients with low or no CD33 expression. Children’s Oncology Group (COG) AML pilot protocol COG AAML03P1 studied the feasibility of combining GO with MRC based chemotherapy in selected courses in patients with de novo AML regardless of CD33 expression level. We hypothesized that patients with low or no CD33 expression may have a different outcome (more inferior) from those with higher CD33 expression in this GO-based therapy. We prospectively evaluated the CD33 expression level of diagnostic blast cells and correlated the CD33 expression level with disease characteristics and clinical outcome. Of the 341 patients treated in COG AAML03P1, 240 diagnostic specimens were available for evaluation. CD33 expression level of the diagnostic leukemic blasts was determined by measuring the CD33 Mean Fluorescence Intensity (MFI). MFI of the blast population varied over 2 log fold from a baseline of 6.33 to 1551 with a median of 117.7. Study population was divided into quartiles based on CD33 expression: 1st Q: (MFI 6.33 – 57.054), 2nd Q: (MFI 57.1 – 117.7), 3rd Q: (MFI 117.7 – 209.3) and 4th Q (MFI 209.3 and higher). Remission induction rate (CR) at the end of induction and the clinical outcome was compared between the four groups. Patients with the lowest CD33 expression had a similar CR rate as those with higher CD33 expression (86% vs. 85%, = 0.980). In patients with the lowest CD33 expression (1st quartile), relapse risk and disease-free survival was not different than those in 2nd quartile (p=0.5 and 0.1, respectively), suggesting that paucity of CD33 expression does not portend poor outcome. Surprisingly, patients with the highest CD33 expression (4th quartile) had a worse outcome with disease-free survival of 45%, compared to that of 62% in those with lower CD33 expression (hazard ratio=1.69; p=0.02). Disease characteristics and molecular and cytogenetic make up of the patients were evaluated within the four groups. CD33 expression level was not associated with age, gender, diagnostic WBC or blast percentage. However, there was an inverse association between CD33 expression and the prevalence of CBF AML where in those with the lowest CD33 expression (1st quartile), 45% had CBF AML (inv(16) or t(8;21)). There was a step-wise decrease in the prevalence of CBF AML of 28%, 20% and 13% in the 2nd, 3rd and 4th quartiles, respectively. There was also an increased prevalence of FLT3/ITD with increasing CD33 expression with a prevalence of 3%, 12%, 15% and 18% in quartiles 1–4, respectively. Conversely, median CD33 MFI in favorable risk patients (CBF AML, NPMc or CEBPA) was 75 compared to those with FLT3/ITD or high risk cytogenetics who had a CD33 MFI of 170. This study demonstrates significant heterogeneity of CD33 expression and inverse association of CD33 expression level with clinical outcome, where those with low CD33 expression have a favorable outcome. This association with clinical outcome is directly linked to the underlying molecular/cytogenetic make up of the disease.
Disclosures: No relevant conflicts of interest to declare.
Author notes
Corresponding author