The definitive assessment of a drug’s activity within a given population cannot be derived from extrapolation or retrospective data mining but rather requires a randomized controlled clinical study, which has always been the gold standard.— Mansoor Raza Mirza, MD
Tweet this quote
An article that appeared in the April 10 issue of The ASCO Post (“Olaparib Maintenance Prolongs Progression-Free Survival in Ovarian Cancer”) suggests unwarranted conclusions from the phase III SOLO2 trial of olaparib (Lynparza) in patients with platinum-sensitive relapsed ovarian cancer and mutations in BRCA1/2, presented at the 2017 Society of Gynecologic Oncology (SGO) Annual Meeting.
The article correctly states that for patients randomized to receive maintenance olaparib, median progression-free survival was 19.1 months, vs 5.5 months with placebo, which was the primary endpoint of this study. Three times, however, the article emphasizes and elevates data from an analysis by blinded independent central radiologic review (BICR) without clarifying the following important points: (1) the BICR results represented a sensitivity analysis, (2) the sensitivity analysis excluded nearly 25% of the progression-free survival events in the olaparib group captured in the primary analysis (26 of 107 progression-free survival events), without explanation, and (3) the sensitivity analysis excluded the impact of clinical progression.
In the article, co-investigator Richard Penson, MD, suggests the seemingly large difference in the findings may reflect some “statistical artifact.” A scientifically rigorous and balanced discussion would not elevate a single sensitivity analysis seemingly because it is favorable to the drug being tested.
Important Details Omitted
Importantly, neither the SOLO2 presentation at SGO nor this article included mention of important details related to the progression-free survival analysis and their impact on the median progression-free survival. In the case of SOLO2, the scan schedule used for the determination of radiographic progression was highly favorable to the drug treatment arm since after week 72, patients were evaluated for progression only once every 6 months, and nonscheduled scans were imputed forward to the next scheduled scan date.
Furthermore, the statement attributed to Dr. Penson, suggesting that the Study 19 population will be retrospectively evaluated for its benefit in patients with a “BRCA-ness” phenotype, is misleading. The fact is that a retrospective analysis of the Study 19 population has already been conducted and presented.1 The conclusion of that presentation was clearly stated: Olaparib did not achieve a statistically significant difference in progression-free survival compared to controls, in populations defined by two different measures of “BRCA-ness”—homologous recombination deficiency–high (P = .14254) or homologous recombination repair wild-type (P = .30).
Neither SOLO2, which only studied patients with a BRCA mutation, nor the retrospective analyses of Study 19 have provided evidence for the activity of olaparib in patients without a BRCA mutation. The definitive assessment of a drug’s activity within a given population cannot be derived from extrapolation or retrospective data mining, but rather requires a randomized controlled clinical study, which has always been the gold standard. ■
—Mansoor Raza Mirza, MD
Department of Oncology
Copenhagen University Hospital
Disclosure: Dr. Mirza serves on the advisory boards of Advisory Board of AstraZeneca, Tesaro, and Clovis Oncology.