LAUSR

1627 such as progression-free survival or overall survival and are limited in assessing safety. Nevertheless, Dr Agrawal believes that oncologists can have confidence in the approvals based on single-arm studies. “Many approvals based on single-arm trials were for drugs previously approved in other disease settings and with well-understood safety profiles, which is also reassuring,” he says. He emphasizes, however, that oncologists should be aware that accelerated approvals, some of which are based on data obtained from single-arm trials, require confirmation of clinical benefit or may be removed from the market. (For a list of accelerated-approval drugs withdrawn from the market to date, visit https://www.fda. gov/drugs/resources-information-approved-drugs/withdrawncancer-accelerated-approvals.) Additional perspectives on issues for oncologists to consider when they are weighing the evidence of single-arm studies/ approvals can be found in commentaries published in Nature Reviews Clinical Oncology and Annals of Oncology.2,3 Thomas G. DeLoughery, MD, of Oregon Health and Science University in Portland, and Vinay Prasad, MD, MPH, of the University of California, San Francisco, discuss the unintended consequences of approving drugs on the basis of single-arm trials through the regular FDA approval process, and they highlight that such drugs are not required to undergo mandatory postsurveillance efficacy trials as is normally required for accelerated approval.3 Robert Glassman, MD; Grace Kim, DO; and Marc J. Kahn, MD, MBA, describe four criteria (instinct, comparative analysis, statistically sound approaches, and like-for-like comparisons) for assessing the outcomes of single-arm studies that will help with clinical decision-making.2 Gini Fleming, MD, the medical director of gynecologic oncology at UChicago Medicine in Illinois, expresses no concern with using a drug on the basis of single-arm data per se, but there is a caveat. “I personally have no concerns for approval of a single agent in the setting of pretreated metastatic disease, if the drug activity is high enough and the toxicity low enough,” she says, adding that she would have more qualms about using combination drug regimens based on single-arm data because it can be unclear how much each agent contributes to the efficacy of the combination. However, for a single agent with a demonstrated response rate in patients with pretreated metastatic disease and acceptable toxicity, she sees the clinical value. “As a clinical oncologist, if your patient has symptoms of cancer and their cancer shrinks, it usually causes the symptoms to get better,” she says. “So, in a setting of metastatic disease, particularly if the patient is symptomatic, I feel that response rate is often predictive of benefit to the patient as long as the drug is not too toxic.”