Risks and benefits of phase I oncology trials
After having spent much of last week blogging about the 60th anniversary of V-E Day (and here), Pat Buchanan's idiotic comments about World War II, the intelligent design creationism craziness going on in Kansas (and here), and slapping down ID apologists, the start of a new week would seem to be a good time to get back to more medical topics, specifically the science of clinical trials.
A couple of months ago, I read an article in the New England Journal of Medicine about the risks and benefits of phase I oncology trials, and I thought it might be a good opportunity to clear up some misconceptions that many people have about phase I clinical trials. Normally, drug development follows a rather rigidly defined pathway. Compounds that serve the basis of drugs are usually discovered either through screening natural compounds or synthesizing new compounds based upon either compounds that have known activity or based on molecular knowledge of the intended drug target, known as "intelligent" or "rational" drug design (not the same thing as "intelligent design"). Promising new compounds, however they are found or synthesized, are then generally tested in cell culture systems for in vitro activity. Those that show promise there can then go on to animal testing. Finally, those that show an ability to shrink experimental tumors in animals without unacceptable toxicity to those animals are chosen for human studies.
The very first study a new drug will go through in humans is known as a phase I trial. Phase I trials are different from phase II and Phase III trials in that the primary purpose of a phase I trial is not to determine if the drug has any efficacy. The primary purposes of a phase I trial are to determine the optimal dose and identify toxicities. They are not the only purposes of such trials, but the determination of whether there is an objective tumor response to a new drug is definitely a secondary consideration in phase I trials. Drugs that are shown not to have excessive toxicity are then evaluated in phase II and phase III trials to compare them against existing therapies for efficacy. This is one reason that the ethics of phase I trials have been questioned, because, in the mind of some ethicists, phase I trials imply the promise of a benefit that is unlikely to be there. Although patients generally take part in clinical trials out of altruism, they also do hope for some benefit for themselves. That's why great care needs to be taken to make sure that potential patients understand that the chances of their personally benefitting from the trial are very small. Patients are hoping for a complete response (complete disappearance of their tumor) or a partial response (defined as a greater than 50% shrinkage of their tumor), but the odds of that are small.
Given this background, this article is of interest because it examines critically what exactly the risks and potential benefits of phase I oncology trials are. The authors evaluated all nonpediatric phase I trials sponsored by the National Cancer Institute between 1991 and 2002 for rate of response to treatment, toxicity, and treatment-related deaths. These trials included single agents, combinations of investigational agents, combinations of FDA-approved agents with investigational agents, antiangiogenesis agents, signal transduction blockers, and vaccines. Complete response rates ranged from 0% to 8%; partial response rates ranged from 0% to 20.4%. The overall response rate (partial and complete responses) was 10.6%. In addition, in 34.1% of participants, either stable disease (no growth) or less-than-partial responses were observed, for an estimate of 44.7% of patients who have a chance of benefiting. In contrast, deaths from toxic events ranged from 0% to 2.02%, with an overall treatment-related death rate of 0.49%. Finally, 14.3% of patients had at least one grade 4 toxic event.
So what does this all mean? First, the response rates observed are higher than previously reported response rates of 4% to 6%, although the mortality rate is roughly the same. Second, it implies that the risk-benefit ratio of such trials is suprisingly favorable, given that these are usually trials representing the first use of a new experimental agent or combination of experimental and/or previously approved agents. The risk of death is at most 0.5%, which in my mind is surprisingly low, given the patient population, which consists of patients with advanced cancer who are likely to be physiologically fragile to begin with. Given that the chance of a partial or complete response is approximately 10.6%, that makes the odds of potential benefit (I say "potential" because complete and partial responses may not actually improve survival) versus the risk of death from the experimental agent approximately 22:1. True, the risk-benefit ratio becomes less favorable when you factor in severe complications, such as grade 4 toxicities, falling to 3:4, but it becomes more favorable if we add in stable disease and lesser responses, with a ratio of stable disease/response to severe complication/death of 3:1. Also, the observation that as many as 44.7% of patients may have stable disease or better in phase I trials becomes more important in light of the increasing appreciation that stable disease or minimal tumor shrinkage can be of clinical benefit, with the recent introduction of less toxic agents that do not kill tumors but rather prevent their growth, such as antiangiogenic agents. Moreover, another important result of this study is that phase I studies are more heterogeneous than most clinicians assume, with a wide variety of classes of agents, all with different chances of efficacy and toxicities. For classes of drugs for which the toxicities are lower, the risk benefit ratio of phase I trials may therefore be even more favorable.
Phase I trials are critical for the development of new cancer therapies. Also, even though the primary purpose of such trials is not to determine efficacy against specific tumors, but rather to determine dosage and toxicities of investigational agents, indications of a potential response observed in phase I trials often serve as both a major impetus for phase II studies and as a guide for which tumors should be studied. Of course, overall risks and benefits cannot be generalized to individual trials from the results of this study, and it's impossible to predict them for any single trial. However, this study suggests that the risks of these trials are not as high as some have suggested, nor are the chances of benefit as low as feared. The bottom line is that it would appear that more patients probably benefit from phase I trials than are harmed by them. Consequently, previous objections on ethical grounds appear less compelling than they perhaps once were, and, given the new generation of targeted therapies, enrollment of patients in phase I trials has the potential to benefit more patients at lower risk than ever.
A couple of months ago, I read an article in the New England Journal of Medicine about the risks and benefits of phase I oncology trials, and I thought it might be a good opportunity to clear up some misconceptions that many people have about phase I clinical trials. Normally, drug development follows a rather rigidly defined pathway. Compounds that serve the basis of drugs are usually discovered either through screening natural compounds or synthesizing new compounds based upon either compounds that have known activity or based on molecular knowledge of the intended drug target, known as "intelligent" or "rational" drug design (not the same thing as "intelligent design"). Promising new compounds, however they are found or synthesized, are then generally tested in cell culture systems for in vitro activity. Those that show promise there can then go on to animal testing. Finally, those that show an ability to shrink experimental tumors in animals without unacceptable toxicity to those animals are chosen for human studies.
The very first study a new drug will go through in humans is known as a phase I trial. Phase I trials are different from phase II and Phase III trials in that the primary purpose of a phase I trial is not to determine if the drug has any efficacy. The primary purposes of a phase I trial are to determine the optimal dose and identify toxicities. They are not the only purposes of such trials, but the determination of whether there is an objective tumor response to a new drug is definitely a secondary consideration in phase I trials. Drugs that are shown not to have excessive toxicity are then evaluated in phase II and phase III trials to compare them against existing therapies for efficacy. This is one reason that the ethics of phase I trials have been questioned, because, in the mind of some ethicists, phase I trials imply the promise of a benefit that is unlikely to be there. Although patients generally take part in clinical trials out of altruism, they also do hope for some benefit for themselves. That's why great care needs to be taken to make sure that potential patients understand that the chances of their personally benefitting from the trial are very small. Patients are hoping for a complete response (complete disappearance of their tumor) or a partial response (defined as a greater than 50% shrinkage of their tumor), but the odds of that are small.
Given this background, this article is of interest because it examines critically what exactly the risks and potential benefits of phase I oncology trials are. The authors evaluated all nonpediatric phase I trials sponsored by the National Cancer Institute between 1991 and 2002 for rate of response to treatment, toxicity, and treatment-related deaths. These trials included single agents, combinations of investigational agents, combinations of FDA-approved agents with investigational agents, antiangiogenesis agents, signal transduction blockers, and vaccines. Complete response rates ranged from 0% to 8%; partial response rates ranged from 0% to 20.4%. The overall response rate (partial and complete responses) was 10.6%. In addition, in 34.1% of participants, either stable disease (no growth) or less-than-partial responses were observed, for an estimate of 44.7% of patients who have a chance of benefiting. In contrast, deaths from toxic events ranged from 0% to 2.02%, with an overall treatment-related death rate of 0.49%. Finally, 14.3% of patients had at least one grade 4 toxic event.
So what does this all mean? First, the response rates observed are higher than previously reported response rates of 4% to 6%, although the mortality rate is roughly the same. Second, it implies that the risk-benefit ratio of such trials is suprisingly favorable, given that these are usually trials representing the first use of a new experimental agent or combination of experimental and/or previously approved agents. The risk of death is at most 0.5%, which in my mind is surprisingly low, given the patient population, which consists of patients with advanced cancer who are likely to be physiologically fragile to begin with. Given that the chance of a partial or complete response is approximately 10.6%, that makes the odds of potential benefit (I say "potential" because complete and partial responses may not actually improve survival) versus the risk of death from the experimental agent approximately 22:1. True, the risk-benefit ratio becomes less favorable when you factor in severe complications, such as grade 4 toxicities, falling to 3:4, but it becomes more favorable if we add in stable disease and lesser responses, with a ratio of stable disease/response to severe complication/death of 3:1. Also, the observation that as many as 44.7% of patients may have stable disease or better in phase I trials becomes more important in light of the increasing appreciation that stable disease or minimal tumor shrinkage can be of clinical benefit, with the recent introduction of less toxic agents that do not kill tumors but rather prevent their growth, such as antiangiogenic agents. Moreover, another important result of this study is that phase I studies are more heterogeneous than most clinicians assume, with a wide variety of classes of agents, all with different chances of efficacy and toxicities. For classes of drugs for which the toxicities are lower, the risk benefit ratio of phase I trials may therefore be even more favorable.
Phase I trials are critical for the development of new cancer therapies. Also, even though the primary purpose of such trials is not to determine efficacy against specific tumors, but rather to determine dosage and toxicities of investigational agents, indications of a potential response observed in phase I trials often serve as both a major impetus for phase II studies and as a guide for which tumors should be studied. Of course, overall risks and benefits cannot be generalized to individual trials from the results of this study, and it's impossible to predict them for any single trial. However, this study suggests that the risks of these trials are not as high as some have suggested, nor are the chances of benefit as low as feared. The bottom line is that it would appear that more patients probably benefit from phase I trials than are harmed by them. Consequently, previous objections on ethical grounds appear less compelling than they perhaps once were, and, given the new generation of targeted therapies, enrollment of patients in phase I trials has the potential to benefit more patients at lower risk than ever.
So what do you think of the new drug Revlimid, that treats melodysplastic syndrome, and supposedly puts 60-70% of patients into "deep remission"?
ReplyDeleteWhat happens after the trial? If a patient in a phase I shows a response, are they given the option to continue the drug when the trial is concluded? This could be an ethical problem, because on one hand you have no objective data supporting that it works; on the other hand, if the drug is helping the patient then taking it away would be withholding care. I'm not arguing against phase I trials, I just want to know how this particular situation should be handled.
ReplyDeleteUnfortunately, I don't know much about Revlimid.
ReplyDeleteAs for what happens after the trial, the results of phase I trials do have a lot of impact on what drugs get to move on to phase II and phase III trials. There's only so much money to go around to fund these trials. Consequently, if a Drug A shows signs of activity in a phase I trial and Drug B does not, that doesn't mean Drug B might not be worth studying further. (Phase I trials, for example, tend to study patients with all sorts of advanced malignancies, rather than just one and to have relatively small numbers; real responses can be missed.) However, clearly Drug A would be the one that the NIH or the drug company would be more interested studying further and of providing the funding for.
By reciting the Hippocratic Oath, physicians commit to benefit the health of their patients with minimal harm. Interesting to note that IRB's ensure the rights and welfare of study subjects (and not patients) participating in clinical trials, thus assuming that harm may and probably will be inflicted on the individual.
ReplyDeleteThank you for the insightful post.