Drug Development for Leukemia : Comprehensive Analysis of Clinical Trial Success Rates Presented at the American Society Hematology (ASH) Annual Meetings

Introduction: Slow progress in drug development is an ongoing issue even more critical for oncologic drugs, which have lower likelihood for approval than other therapeutic areas. Early clinical trials are frequently conducted and few eventually lead to drug approvals. We analyzed success rates of anti-leukemic drugs presented at the annual ASH meetings to identify rates of progression to advanced development and factors associated with clinical trial (CT) success.

Methods: Our eligibility criteria included ASH meetings abstracts reporting CT assessing investigational anti-leukemic drugs conducted in the United States from 2011 to 2015. We searched ClinicalTrials.gov, PubMed, and Food and Drug Administration (FDA) websites to track the transition at each clinical phase (from Phase 1 to Phase 3), regulatory history, and publication as full-length manuscripts. We assessed the success rate (number of CT approved by FDA divided by the total number of CT) by study, drug and outcomes features using chi-square or Fisher's exact test. To verify the correlation between the impact factor (IF) of the journal in which the CT was published with the success of the study, the FDA-approval status (No vs Yes) was considered ordinal variable and the Pearson correlation coefficient was calculated.

Results: Among the 3,929 abstracts for leukemia reviewed, 488 were main CT reports. Based on our eligibility criteria, 235 abstracts were identified including Phase 1 (121), Phase 2 (100), and Phase 3 (14) CT. Of these CT abstracts, only 45 (19.1%) eventually led to FDA approval. Among the unsuccessful studies, 37 (15.7%) progressed to the next one or two phases, and 153 (65.1%) did not transition any phase forward. Abstracts in acute myeloid leukemia (AML, 47%) and chronic lymphocytic leukemia (CLL, 30.2%) were the most common, followed by acute lymphocytic leukemia (ALL, 15.6%) and chronic myeloid leukemia (CML, 6.5%). Investigational New Drug (IND) combined with FDA-approved regimens (41.7%) or as monotherapy (40.9%) were the most commonly investigated therapy modalities. Combinations of multiple IND (9.8%) and new uses for FDA-approved drugs or regimens (7.7%) were evaluated less frequently.

The overall success rate of CT was 6%. Although CML and ALL were the least studied types of leukemia, they had the highest approval success rates, 47% and 24.4%, respectively (p<0.001). Monotherapy INDs were more frequently approved by FDA (32.3%) than when combined with FDA-approved drugs (11.2%, p<0.001). FDA approval was also more frequent for advanced CT phase (p<0.001) and more recent (by year) presentation of the abstract (p=0.008). Tumor status (untreated vs relapsed and/or refractory [R/R] vs untreated and R/R) and the age of the patients included in the CT (adults including older adults vs older adults only vs pediatric vs adults and pediatric) were not significantly associated with approval rates (p>0.05). CT including drugs and regimens that led to FDA approval were published in journals with higher impact factors (R=0.368, p=0.0002), such as Lancet Oncology (24.1%), Blood (17.2%), Lancet Haematology (13.8%), and New England Journal of Medicine (13.8%). Blood was the journal that published the highest number of unsuccessful CT (15.3%).

During the study period, CT assessing Enasidenib as monotherapy for R/R AML with IDH2 mutation, Ponatinib for CML and Philadelphia chromosome (Ph+) ALL, and Venetoclax for R/R CLL transitioned directly from Phase 2 to FDA approval.

Conclusion: CT success rates for hematologic malignancies improved modestly over the last years. The high investment in CT for AML and CLL management has not resulted in increased approval rates proportional to reported CT.