ASCO Chief Medical Officer Richard L. Schilsky, MD, and other oncology drug experts presented a panel on drug dosing at a recent meeting, cosponsored by the Friends of Cancer Research and the Brookings Institution, in Washington, DC.1 The presentations made it clear that issues surrounding drug dosing are among the thorniest problems in oncology.
“We want patients to live longer and enjoy a better quality of life,” said Dr. Schilsky, “but we need to do a better job of balancing the benefits and risks of therapy, that is, identifying the drug dose at which efficacy is maximized and toxicity minimized. A dose that is too high can render an otherwise effective drug intolerable, and one that is too low can negate the expected benefit.”
Because cancer is life-threatening, a high degree of toxicity is accepted, as is the need to develop new drugs quickly. These two factors often take precedence over finding an optimal dose.
“We have accepted too high a degree of toxicity in cancer drugs, and it has become a serious problem,” said Richard Pazdur, MD, Director of the U.S. Food and Drug Administration (FDA) Office of Hematology and Oncology Products. “We do a terrible job of exploring dosing prior to a phase III trial. The emphasis in clinical trials is primarily on efficacy, and drug companies don’t want to do phase II dosing studies to determine whether the maximum tolerated dose is the optimal dose,” he said.
“We simply don’t have the license to cause patients so much toxicity—even in the face of the threat to life. We have got to find a better balance between doing good and doing harm,” Dr. Pazdur added.
Easier Said Than Done
The “right” dose is often elusive, said Dr. Schilsky, and even if it could be pinpointed, it may be only an estimate. “For any drug dose, a range of beneficial and toxic effects will vary based on the unique characteristics of each patient.”
Dr. Schilsky noted that many recently approved oncology drugs are labeled for use at doses that are either too high or too low for at least some patients. “We need to devise a strategy for drug development that includes dose optimization but doesn’t delay market entry.”
Phase I trials determine the highest tolerable dose; small numbers of patients are given increasing doses until dose-limiting toxicity is reached. The amount of drug immediately below that is the maximum tolerated dose, which is then used for phase II and III trials, which rarely evaluate doses other than maximum tolerated dose. “The [maximum tolerated dose] may not be appropriate for targeted therapy,” said Atiqur Rahman, PhD, Director, FDA Division of Clinical Pharmacology V. “Moreover, dose selection based on [maximum tolerated dose] does not mean there will not be serious toxicity.”
This process has other limitations as well: For example, it does not evaluate patient variability in treatment response and toxicity. “Doses used in phase II and III often achieve concentrations that may substantially surpass those needed to inhibit or stimulate the intended target(s),” said Dr. Rahman. This can lead to a high rate of dose reductions in clinical trials, as well as failure to identify patients who might benefit from a higher dose.
Moreover, pharmacodynamic endpoints that assess target inhibition might be more relevant for agents such as hormonal and targeted therapies. In addition, oncology drug development does not evaluate long-term cumulative toxicity or changes in tolerability over time. This is significant because patients tend to stay on therapy longer now than in the past. Finally, decisions about dosing and tolerability are different among the heavily pretreated, usually sicker, patients in phase I trials as compared to those in phase II and III trials.
Approach to Dose Determination
The way to approach dose determination is to understand the relationship between drug exposure and clinical outcomes. “Defining ranges of toxic and therapeutic concentrations may let us monitor patient drug levels, which could be used to guide decisions, particularly for chronic treatment,” said Dr. Schilsky.
A key component is randomized dose-comparison studies, which are not typically done in oncology. Phase I trials should include sufficient pharmacokinetic sampling to clearly determine the pharmacokinetic properties of the drug. Phase II should go beyond assessment of drug activity to include adaptive designs and/or randomized exploration of doses. Phase III should incorporate population pharmacokinetic sampling to further evaluate the relationship between drug exposure and clinical outcomes.
Regarding toxicity, when possible, patient-reported outcomes should be assessed and measured in all phases of clinical trials as one guide to optimal dosing. Moreover, pharmacokinetic and patient-reported outcomes data could be used in the postmarketing setting with an eye toward individual dose adjustment. But to do all this, exposure data are required.
The label of an approved drug represents the average response of trial patients, but many factors affect the amount of drug that patients are exposed to. Some oral drugs are affected by taking them with food, and drug metabolism is affected by genetic polymorphisms. Then there are concomitant medications, age, body weight, hepatic and renal function, and comorbidities. Pharmacokinetic data can reveal drug concentrations and lead to estimation of a therapeutic index for individual patients or a defined patient population.
The FDA can use these data to assess the need for postmarketing studies. The panel agreed that randomized dose-comparison studies should be included in phase II, and exposure-response analyses should be part of phase III.
“Physicians are trained to recognize serious side effects and usually capture accurate adverse events, but they tend to underreport effects that they consider merely ‘bothersome,’” said Lori Minasian, MD, Deputy Director, National Cancer Institute (NCI) Division of Cancer Prevention. Nevertheless, “bothersome,” if it lasts long enough, can make a patient stop taking a drug.
Toxicities are currently reported using the NCI Common Terminology Criteria for Adverse Events (CTCAE). This is a valuable tool, but clinicians often underreport patient symptoms. In come cases, early patient reports of mild-to-moderate events have presaged poor long-term treatment tolerability or an increased risk of severe toxicity.
Dr. Minasian noted the often-significant discrepancies between clinician and patient reports of adverse events. She pointed out that in a study looking at reporting of symptoms by patients and their physicians, about a third of the time there is discrepancy, with physicians underreporting the symptom.
“Patients’ report of side effects correlates with function and overall health status and may better reflect tolerability over time,” she said.
A new tool potentially could be used to capture both physician-observed adverse events and patient-reported outcomes. Dr. Minasian described the PRO-CTCAE instrument as follows:
In phase I, it can gauge side effects relative to dose escalation, as well as refine measurement of adverse events. In phase II, it can define toxicity in depth by assessing the tolerability of the recommended dose, and it can identify chronic toxicity. In phase III, it can assess overall risk and benefit for a particular regimen and more deeply evaluate efficacy and tolerability. In phase IV (postmarketing), it can optimize tolerability and tailor regimens for vulnerable populations (older patients and those with comorbidities).
Improving Clinical Outcomes
Dose-comparison studies in phase II could improve the design of phase III trials. They also could identify drugs amenable to dose escalation or reduction. For instance, a comparison trial might indicate that while antitumor responses can be achieved at multiple doses, some patients do not respond at lower doses and some do not tolerate higher ones. If the trial design had adaptations built in, individual dosing could lead to a better, more realistic response rate.
For most oncology drugs, the label provides only a starting dose that may need to be modified, but if exposure and tolerability data existed for a range of doses, it would be easier to define a threshold dose as well as peak exposure that correlates with excess toxicity. This also would make it easier to monitor and adjust doses in clinical practice.
Targeted therapies are usually taken for a longer time than classical cytotoxics, and their toxicities are different—ie, later onset and more cumulative. For example, a patient has been on a targeted drug for several months with low-grade toxicity. Eventually he or she may reach a point where treatment is no longer tolerable, but if the dose can be lowered, unnecessary discontinuation might be forestalled.
Randomized dose-comparison studies prior to registration trials could improve the chance of approval by minimizing the likelihood of excessive toxicity. Moreover, collection of drug exposure and clinical outcomes data in the postmarketing setting could improve understanding of real-world patient experience, as well as identify vulnerable populations.
Nevertheless, there are downsides to this design. If approval is in doubt or the drug is on a fast track, dose comparison could be burdensome and take too long. In the postmarketing arena, patients may be reluctant to be in a trial with an approved drug, so sponsors would have to look for participants outside the United States, where the FDA has limited enforcement powers once a drug has been approved.
One solution is for companies to conduct such trials after registration studies are complete but before approval is granted. This window of opportunity may appeal to patients who want access once it becomes known that the drug has met the goals of the registration trials, especially if there is no current standard of care.
Drug companies can benefit as well. If they are confident of approval, it would be to their advantage to determine whether a lower dose would decrease efficacy at an early endpoint such as response rate. But if a lower dose works as well as a higher one, patients could stay on the drug longer, and it’s possible that the drug would increase the company’s competitiveness.
In short, said Dr. Schilsky, knowing more about tolerable dosing regimens would benefit everyone: The FDA would have more information to guide the approval process, clinicians would know more about optimal dosing and have more choices, and patients would be more likely to tolerate their treatment. ■
Disclosure: Dr. Schilsky is Chief Medical Officer, ASCO. Dr. Pazdur is FDA’s Director of the Office of Hematology and Oncology Products. Drs. Rahman and Minasian reported no potential conflicts of interest.