In a recent publication,2 Dr Kimmelman looked at how often patients in phase 1 oncology trials receive therapeutic doses of a drug that will eventually be approved by the FDA for their cancer type. He and his colleagues found that only 1 in 83 patients received a drug at a dosage that eventually received FDA approval for their cancer type. Combining that proportion with the number needed to treat (NNT) for various cancer drugs, he calculated that between 350 and 1350 patients need to participate in a phase 1 trial in order for 1 person to benefit.

“One of the ethical challenges of these studies is that you want to minimize the number of people who are receiving the drug at toxic levels or at levels that aren’t high enough to trigger a tumor response,” he said. “If people are getting low doses, you can feel a little bit better if that dose still has the prospect of having some activity.”

American Society of Clinical Oncology’s (ASCO’s) official position is that phase 1 trials have the potential to provide patients with clinical benefit,3 but Dr Kimmelman questions how this is communicated to patients. “If we want to make progress in cancer and develop new drugs, we have to do phase 1 studies,” he said. “My view is that these studies ought to be presented not as therapeutic opportunities, but rather as opportunities for patients to contribute to the care of future patients.”

Dr Chen said he hopes the report will get people thinking about new ways of constructing phase 1 trials to better suit the new drugs being tested. “Different people will have different interpretations of what this shows,” he said. One possible interpretation is that for some drugs, it may be appropriate to take a lower dose forward into phase 2 and 3 trials, he said. Increased testing of biomarkers could also make phase 1 trials more informative, he said.

Continue Reading

“With these new agents, once you figure out a ballpark where the dose should be, you don’t need to fine-tune the dose,” he said. “It’s more about evaluating, ‘What is it doing under the microscope? Are there biomarkers in the blood and in the biopsy that can tell us more information?’”

“Really honing down the mechanism should be part of the phase 1 study,” he added. “That could help us understand why some people respond and why some don’t.”

Related Articles

Dr Siu added that evaluating pharmacokinetic biomarkers could help determine appropriate dosage. “If you see similar clinical activity and similar molecular activity in both low and high doses, you have to think twice about whether you want to expose the patient to the higher dose, if the toxicity is a lot higher,” she said. “Having some biomarker correlates to show if you’re seeing biological activity is very reassuring.”

One limitation the OHSU team encountered while reviewing the phase 1 literature, however, was the high number of trials that did not report complete response data for all doses. Of 175 trials that met their criteria, only 93 published enough data to be included in the meta-analysis.

“When we report phase 1 studies … dose level-response data are going to be relevant for us to make more sense of [them] in the future,” said Dr Siu. “When a study is finished, [those] kind of data should be in the public domain.”


  1. Hazim A, Mills G, Prasad V, Haslam A, & Chen EY. Relationship between response and dose in published, contemporary phase 1 oncology trials. J Natl Compr Canc Netw. 2020; 18(4):428-433.
  2. Zhang SX, Fergusson D, Kimmelman J. Proportion of patients in phase 1 oncology trials receiving treatments that are ultimately approved [published online April 1, 2020]. J Natl Cancer Inst. doi: 10.1093/jnci/djaa044
  3. Burris HA. Correcting the ASCO position on phase I clinical trials in cancer. Nat Rev Clin Oncol. 2019;17:125.