Forty years ago, President Richard Nixon announced a “war on cancer.” Some of that war’s first battles were won in the field of acute myeloid leukemia (AML) with two agents, cytarabine and daunorubicin, receiving U.S. Food and Drug Administration (FDA) approval based on their ability to produce typically only transient remissions. Since then, little has happened except for FDA approval of all-trans retinoic acid (aka tretinoin, ATRA) and arsenic trioxide (Trisenox) for acute promyelocytic leukemia (APL), a rare subset of AML. The war on AML appears to be at a standstill.
Emblematic of this standstill are the recent FDA decisions to not grant approval for decitabine (Dacogen) for the treatment of older patients with AML, and to withdraw its approval for gemtuzumab ozogamicin (Mylotarg) for the treatment of older patients with relapsed AML. We question those decisions and suggest that the policies they exemplify are, to at least some extent, responsible for the standstill.
Decitabine Decision
In a recent Comments and Controversies article in the Journal of Clinical Oncology, Sekeres and Steensma comment on the FDA Oncologic Drug Advisory Committee (ODAC)’s negative decision regarding decitabine. They describe the decision as rooted in “rigorous statistical principles” and cite decitabine as another example of the “boulevard of broken dreams” characteristic of drug development in AML.1 They cite failure of the difference in survival (the primary endpoint) between decitabine recipients (median survival, 7.7 months) and the control group, wherein physician-advised patients received their treatment choice (median survival, 5.0 months) to reach statistical significance as a rigorous justification for nonapproval.2 The authors, however, raise two additional issues that influenced the vote and that both diverge from the rigorous scientific principles they espouse.
First, they note the response rate in the low-dose cytarabine group (treatment choice for 88% of control arm) was lower than historical rates. Comparing response rates between trials is highly problematic and can introduce bias due to differences in patient populations and treatment that cannot be accounted for by multivariate analyses; hence, the need for randomized trials.2
Second, they cite improved survival of the treatment choice arm in Western Europe compared to other geographic regions. However, they fail to mention that the hazard ratio (HR) favored decitabine (HR range = 0.75–0.93) in all other geographic regions and that the P value for the Western European comparison was greater than the 0.10 that led to rejection of decitabine in the primary analysis. Furthermore, patients from Western Europe only comprised 17.5% of patients; their randomized distribution was not 1:1, due to lack of geographic stratification; patients on the decitabine arm had worse baseline characteristics compared to the treatment choice arm; and patients on the treatment choice arm more often received azacitidine (Vidaza) or decitabine poststudy.2,3 In the overall study population, forest plots of subset analyses of response and survival also favored the decitabine arm in virtually all subgroups.
Conflicting Analyses
Unfortunately, ODAC appears to have regarded a more mature but unplanned survival analysis (95% of patients dead), which continued to show an advantage for decitabine (median survivals of 7.7 vs 5.0 months—exactly the same as in the primary analysis but now with HR = 0.82, 95% confidence interval = 0.68–0.99, P = .037), as less satisfactory than the primary less mature analysis because it was “unplanned.” This begs the question of how ODAC might have reacted had the less mature but planned analysis shown P = .037 and the more mature but unplanned analysis P = .10, and prompts us to question the consequences of our seeming infatuation with P = .05.5,6
ODAC also did not adequately consider the effect of subsequent therapy that could have reduced the benefit from decitabine. Sensitivity analyses—in which patients who received subsequent disease-modifying therapy (eg, intensive chemotherapy, hypomethylating agents) were censored at the date of such therapy or excluded from analyses—showed improved survival with decitabine (HR = 0.8, P = .044 for censoring: HR = 0.74, P = .006 for exclusion).
Based on the totality of data, including the significant prospectively defined secondary study endpoints, the European Medicines Agency (EMA) Committee for Medicinal Products for Human Use (CHMP) came to a different conclusion and granted approval of decitabine in the European Union. That committee concluded that “the benefits with Dacogen are its ability to show an increase in overall survival.”4 The committee obviously took into account the later unplanned but more mature survival data and other sensitivity analyses.
Statistical and Clinical Contexts
The ODAC recommendation for rejection and subsequent FDA actions should have taken into account the totality of evidence that we believe argues that decitabine is likely to produce clinical benefit in older patients with AML. Large randomized clinical trials represent significant investments by patients, clinicians, and sponsors. Results should be interpreted in both statistical and clinical contexts. We are concerned that, in this case, rigorous statistical principles were applied inconsistently and, more importantly, that balance between clinical medicine and statistical analysis was not achieved.
Fundamental to the former is the concept of benefit-to-risk ratio. Decitabine is relatively well tolerated and offers low-intensity therapy that improved survival by 2.7 months, a 50% benefit. Although the benefits of decitabine are perhaps modest, they are real and the drug’s risks are small.
Moreover the application before ODAC was not for initial drug approval but for a second indication. With the loss of orphan drug exclusivity in 2013, generic formulations of decitabine will be inexpensive. Hence we believe that the AML community would have been better served with a positive ODAC vote. The negative vote impedes clinical care and research and leads to the perception of a “boulevard of broken dreams.”
Gemtuzumab Reconsidered
Gemtuzumab raises different drug development and regulatory issues.7,8 Based on three open-label studies, gemtuzumab was approved by the FDA in May 2000 under accelerated approval authority for the treatment of patients with CD33-positive AML in first relapse who are at least 60 years old and are not considered candidates for cytotoxic chemotherapy.
However, on June 21, 2010, the FDA and Pfizer announced that gemtuzumab would be voluntarily withdrawn, after considering the results of a confirmatory trial mandated by the FDA and conducted by the Southwest Oncology Group (SWOG) in adults under age 60 with de novo newly diagnosed AML. The trial failed to show that patients randomly assigned to standard therapy (“3+7” induction regimen of cytarabine and an anthracycline) plus gemtuzumab had superior survival to those randomly assigned to standard therapy alone, and that the 30-day death rate was higher in the former.9
The decision to withdraw did not account for several facts. First, four randomized trials (two by the Medical Research Council/National Cancer Research Institute [MRC/NCRI] in the United Kingdom, one in older and the other in younger patients; and two in France including patients up to age 70 years) found that certain patients consistently lived longer if given gemtuzumab.10,11 These were patients with core binding factor AML, who also had superior survival with gemtuzumab in the SWOG study, and many patients with a normal karyotype. Indeed, the MRC developed a prospectively validated index that can identify patients who will have superior survival with gemtuzumab.
Second, the requirement for an aggregate benefit ignores the salient clinical and biologic feature of AML: that it is several, and probably many, diseases, each of which will almost certainly require distinctive treatments. Certainly a randomized trial comparing 3+7 with or without all-trans retinoic acid in AML would have missed the important activity of all-trans retinoic acid in APL.
Finally, none of the other randomized trials found a difference in 30-day mortality. Gemtuzumab appears to have been a casualty of a drug approval process that has unrealistic expectations of uniform benefit in an era of “personalized medicine” and that—as in the case of decitabine—did not consider the totality of the evidence.
Lack of Advocacy
We are struck by the sense of resignation, if not apathy, with which the AML community has greeted the decitabine and gemtuzumab decisions. This is in sharp contrast with the solid tumor community’s strong advocacy of several drugs that have received FDA approval, although in most cases those agents failed to produce better survival (relative to standard therapy) than decitabine or gemtuzumab.12-15
Since 2005, the FDA approved seven agents for renal cell cancer, based on progression-free survival improvements of 2 to 6 months, with only one study (temsirolimus [Torisel]) showing a survival benefit. In lung and colorectal cancer, cetuximab (Erbitux) showed survival improvements of 1.2 to 1.7 months.12-14 In pancreatic cancer, erlotinib (Tarceva) was approved for improving survival by 10 days.15
AML is a relatively uncommon disease, and appears to have no strong advocacy groups. In contrast to the market share of billions of dollars for drugs approved in lung, breast, and other common cancers, the market share of gemtuzumab was about $20 million. Could this have played a role in the lukewarm advocacy or the manufacturer’s decisions? After all, a multitude of articles were written and strong protests were voiced when questions about the efficacy of bevacizumab (Avastin) were raised in particular tumors.
Perhaps, in AML, the “boulevard of broken dreams” has become a self-fulfilling prophecy. ■
Dr. Kantarjian is Professor of Medicine and Chair of the Department of Leukemia, The University of Texas MD Anderson Cancer Center, Houston, and Dr. Estey is Professor of Medicine, Department of Hematology, University of Washington School of Medicine, and Member of Fred Hutchinson Cancer Research Center, Seattle.
Disclosure: Dr. Kantarjian has received research grants from ChemGenex. Dr. Estey reported no potential conflicts of interest.
References
1. Sekeres MA, Steensma DP: Boulevard of broken dreams: drug approval for older adults with acute myeloid leukemia. J Clin Oncol 33:4061-4063, 2012.
2. Kantarjian HM, Thomas XG, Dmoszynska A, et al: Multicenter randomized open-label, phase III trial of decitabine versus patient choice, with physician advice, of either supportive care or low dose cytarabine for treatment of older patients with newly diagnosed acute myeloid leukemia. J Clin Oncol 30:2670-2677, 2012.
3. Eisai Inc: FDA ODAC Briefing Document NDA 021790/S-010: Dacogen (decitabine). Indication: Acute Myeloid Leukemia, February 9, 2012. Available at http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/OncologicDrugsAdvisoryCommittee/UCM290512.pdf. Accessed May 1, 2013.
4. European Medicines Agency: Dacogen (decitabine) Summary of opinion (initial authorization), July 19, 2012. Available at http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/002221/smops/Positive/human_smop_000404.jsp&mid=WC0b01ac058001d127. Accessed May 1, 2013.
5. Berger JO, Berry DA: Statistical analysis and the illusion of objectivity. Am Sci 76:159-165,1988.
6. Goodman SN: Toward evidence-based medical statistics. 1: The p value fallacy. Ann Intern Med 130:995-1004, 1999.
7. Ravandi F, Estey EH, Appelbaum FR, et al: Gemtuzumab ozogamicin. J Clin Oncol 30:3921-3923, 2012.
8. Estey E: Treatment of AML: Resurrection for gemtuzumab ozogamicin? Lancet 379:1468-1469, 2012.
9. Petersdorf S, Kopecky K, Stuart RK, et al: Preliminary results of Southwest Oncology Group Study S0106. Blood (ASH Annual Meeting Abstracts), 114:Abstract 790, 2009.
10. Burnett AK, Hills RK, Milligan D, et al: Identification of patients with acute myeloblastic leukemia who benefit from the addition of gemtuzumab ozogamicin: Results of the MRC AML15 trial. J Clin Oncol 29:369-377, 2010.
11. Castaigne S, Pautas C, Terre C, et al: Effect of gemtuzumab ozogamicin on survival of adult patients with de-novo acute myeloid leukaemia (ALFA-0701): A randomised, open-label, phase 3 study. Lancet 379:1508-1516, 2012.
12. Cunningham D, Humblet Y, Siena S, et al: Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med 351:337-345, 2004.
13. Pirker R, Pereira JR, Szczesna A, et al: Cetuximab plus chemotherapy in patients with advanced non-small-cell lung cancer (FLEX). Lancet 373:1525-1531, 2009.
14. Cunningham D, Humblet Y, Siena S, et al: Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med 351:337-345, 2004.
15. Moore MJ, Goldstein D, Hamm J, et al: National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 25:1960-1966, 2007.
