I write from an unusual perspective. I am a patient who has had to cope with a challenging clinical problem, and I am an academic whose research aims to improve the evidence and methods used to inform patient care.
I have had to cope with a succession of recurrences of melanoma. Following initial diagnosis in 2012 of a tumor on my back, I experienced recurrences in 2014, 2016, and 2022. Each was surgically removed. I have been under continual surveillance.
In 2022, I was offered the opportunity to complement surveillance with adjuvant immunotherapy. The protocol called for administration of nivolumab every 4 weeks for up to 1 year or until disease recurrence or occurrence of unacceptable immune-related adverse events (irAEs). Concerned with treatment efficacy and toxicity, I inquired how this regimen (dose, schedule, and duration) was determined. I learned that the regimen was based on one evaluated in the CheckMate 238 trial,1 revised to double the duration between doses used in the trial. I asked whether this regimen is optimal. The oncologist responded that this is unknown because phase 3 trials with other regimens have not been performed.
Having online access to medical literature, I searched to learn about the efficacy and toxicity of nivolumab. I found reason for concern when I read statements such as this: “There is still large room for improvement in the treatment of metastatic melanoma by addressing two of its major problems: resistance and treatment-related adverse events.”2 I eventually agreed to initiate immunotherapy. The first three treatments were without incident. After the fourth, I felt no symptoms, but lab tests revealed a diminution in thyroid function, a common but treatable irAE. Levothyroxine stabilized the condition. After the sixth treatment, I experienced a less common irAE, a combined condition of dry mouth, lips, and eyes. A 1-week course of prednisone mitigated the issue to a limited extent but did not cure it. After considerable deliberation, I made the difficult decision to terminate treatment after 6 of 13 doses, out of concern that I would experience more severe irAEs if I were to continue.
The uncertainty with which the oncologist and I had to cope was enormous, there being no randomized controlled trial evidence on the effectiveness and toxicity I would face if I were to continue or stop. Dealing with the uncertainty as a patient was difficult. I took solace in my recognition as a researcher that medical decision-making is replete with uncertainties. Indeed, I recently published a book on the subject.3
Pros and Cons of Different Regimens
My experience illustrates the clinical problem of determining optimal regimens of immunotherapies. Clinicians seeking to achieve optimal adjuvant care for cancers must compare the consequences of immunotherapy and surveillance. Immunotherapy aims to lower the risk of disease recurrence, but it may have toxicities that generate irAEs, lowering quality of life or even resulting in death. There is good reason to expect that, as the intensity of the regimen (dose, schedule, and duration) increases, the risk of recurrence falls and the risk of irAEs rises. The optimal regimen appropriately weighs these opposing risks, which may be patient specific and, of course can be complicated if some toxicities are idiosyncratic and not dose-related.
The prevalent practice in phase 3 studies of immunotherapies has been to perform a two-group trial that compares a specified regimen of an innovative therapy with an existing therapy or a placebo. These trials have informed clinicians about patient outcomes with the specified regimens, but they have provided no information about outcomes with alternative regimens that might yield better combinations of efficacy and toxicity.
A new draft Guidance for Industry from the U.S. Food and Drug Administration (FDA)4 recognizes the dearth of knowledge in the context of dosing for treatment of oncological diseases. Unfortunately, it does not clearly explain what it means to optimize dosage. The FDA writes: “Optimal dosage is the dosage that can maximize the benefit/risk profile or provide the desired therapeutic effect while minimizing toxicity.” These concepts are relevant beyond the field of oncology because efficacy and toxicity commonly increase with dosage in treatment of other diseases. The optimal dosage appropriately weighs the opposing risks. The tension between these forces makes it impossible to find a regimen that simultaneously maximizes therapeutic effect and minimizes toxicity. The optimal regimen must appropriately weigh the risks as they manifest for individual patients.
What We Need
Part of what we need is careful design of multigroup trials that compare alternative regimens. The FDA Guidance appropriately recommends the performance of such trials, but it focuses on preliminary “dose-finding trials” undertaken in support of phase 2 studies. It only briefly mentions the possibility of adding dosage groups to phase 3 registration trials. Comparison of alternative dosages in phase 3 trials, which have much larger sample sizes than do phase 2 trials, is essential for conducting adequately powered subgroup analyses to support personalization of regimens.
There is also a need for increased sophistication in the statistical analysis of existing trials. One aspect of increased sophistication is to cease the prevalent practice of studying treatment efficacy and toxicity in isolation from one another, viewing the former as the primary outcome and the latter as a secondary outcome. It would be better to give balanced attention to the risk of disease recurrence and the risk of irAEs, defining a composite measure of patient health that depends on both efficacy and toxicity. See, for example, the discussion of composite measures in Evans et al.5
Another aspect of increased sophistication is to permit analysis of dosage trials to move away from the practice of viewing each trial group as qualitatively distinct. Consider, for example, a three-group trial comparing treatments (A, B, and C). In standard analysis, the findings with group A are assumed to reveal nothing about groups B and C, and vice versa. This assumption is well grounded when treatment groups really are qualitatively different from one another. However, this is not so when comparing dosages. If A, B, and C are increasing dosages of a medication, then it will typically be reasonable to assume that both effectiveness and toxicities weakly increase as dosage rises.
If these needs are met, I hope that medical research will be able to provide more informative guidance about regimens. Improving guidance cannot eliminate the difficulty of treating cancer. Still, my experience has increased my appreciation that it can do much to help oncologists and patients cope with their shared decision problem.
Disclosure forms provided by the authors are available with the full text of this article.
Disclosure Forms (evidpp2300028_disclosures.pdf)
Weber J, Mandala M, Del Vecchio M, et al. Adjuvant nivolumab versus ipilimumab in resected stage III or IV melanoma. N Engl J Med 2017;377:1824-1835.
Trojaniello C, Luke JJ, Ascierto PA. Therapeutic advancements across clinical stages in melanoma, with a focus on targeted immunotherapy. Front Oncol 2021;11:670726.
Manski CF. Patient care under uncertainty. Princeton, NJ: Princeton University Press, 2019.
U.S. Food and Drug Administration. Optimizing the dosage of human prescription drugs and biological products for the treatment of oncologic diseases. January 2023 (https://www.fda.gov/media/164555/download).
Evans SR, Zeng L, Dai W. The data and safety monitoring board: the toughest job in clinical trials. NEJM Evid 2023;2(2).
Copyright © 2023 Massachusetts Medical Society.
Published online: March 28, 2023
Published in issue: March 28, 2023
Select the format you want to export the citation of this publication.