Patients confronting advanced cancer are encouraged to participate in clinical trials.
Clinical trials are employed to develop new drugs and refine treatment strategies.
Each type of trial serves a different purpose.
Purposes of Each Type of Trial
Phase I trials determine whether a new drug is safe for administration. Phase II trials determine whether a new drug is active in a given disease, that is, colon cancer versus lung or breast cancer. Phase III trials then determine whether a new drug or combination is superior to an existing treatment.
Phase I trials are usually relatively small and focused. They determine whether a new drug is safe and the best schedule for its administration.
Phase II trials are larger, as they determine whether the drug has a role in the treatment of advanced disease and might include dozens of patients, with multiple administration schedules.
The truly large trials are the phase III trials. These are designed to prove that the new drug or combination is superior (or in some instances non-inferior) to the standard of care.
The size of these trials is dictated by the need to achieve statistical significance.
A biostatistician confirms the findings to a given degree of certainty. A 95% degree of certainty is known as P = 0.05. A 99% degree of certainty is known as P = 0.01. And a 99.9% degree of certainty is known as P = 0.001.
Larger trials provide the opportunity to establish the highest level statistical proof which is why some trials literally have thousands of patients.
The Big Question
The question is why clinical trials aren’t more successful and why drug development is so slow?
Let's start by examining how clinical trials are designed and conducted.
First, the investigator poses a question: Is treatment A better than treatment B?
They then design the trial with the expectation that A will be better than B by some amount i.e. 25%, 50%, 75%, etc.
Patients receiving the new drug, called A, should get, for example, 50% more response. They might also design the trial to determine whether patients receiving drug A have a longer duration of benefit than those on treatment B. This is known as "time-to-progression".
The most rigorous of all determinations is whether drug A provides superior survival over drug B. That is the highest hurdle of all.
Let's Look at an Example
Here we can conduct a thought experiment to develop a theoretical clinical trial.
Our theoretical trial will determine whether a hormone therapy (treatment A) prevents breast cancer.
One half of our patients will receive the hormone therapy (treatment A) and the other half will receive a sugar pill known as a "placebo" (treatment B).
The trial is designed, to the best of the investigator’s knowledge, so that there are no other factors (confounding variables) that might alter the outcome. Thus, patients are of similar age, similar disease stage, have similar prior treatments and are of good general health.
Once these criteria are met then patients are randomly assigned to each arm.
Randomization is the process, not unlike flipping a coin, that determines who goes on treatment A and who goes on treatment B.
The intent of randomization is to say that of all the known variables have been measured but that some cannot be determined, and those determinants of outcome are (randomly) distributed to avoid bias of one side over the other.
In our theoretical study of treatment A (hormone intervention) versus treatment B (placebo) the results are found to be inconclusive.
That is, the hormonal intervention cannot be proven to prevent breast cancer.
Could We Have Avoided This Unsuccessful Trial?
At this point, one of the smarter investigators realizes that they failed to ask one additional question: Were the participants who went on the clinical trial men or women?
In retrospect, they realize that this profoundly influenced the outcome.
As men rarely suffer breast cancer (only 1% as often as women) a treatment to prevent breast cancer would be hard pressed to prove its worth where the impact of the therapy has been so severely diluted by the inclusion of men in the study.
If half of the subjects in the two arms are men, then the impact of hormonal therapy would be very difficult to prove. One might argue that only half of the people in both arms are men, so the effect should ultimately be seen.
And in fact, if you put enough people on the trial, the intended treatment benefit might have been seen. But it would take a gargantuan trial, likely requiring thousands of patients, to prove a point that could easily have been proven had the right population been studied in the first place i.e. women.
Continuing along the same line of reasoning, if the randomization process had accidentally put more men than women in the treatment arm (as sometimes happens entirely by chance) the results, instead of being inconclusive, might have been frankly negative, incorrectly concluding that hormonal prevention is worse than placebo.
How could such a mistake be made? While the example may seem extreme and academicians might cry “straw man argument” this is actually not out of the realm of clinical experience.
Although it is obvious that men and women differ and can be distinguished one from the other relatively simply, similarly profound distinctions are not always so obvious, although they can be equally impactful on the results of a trial.
Examples include studies that were conducted before the routine use of the estrogen receptor (ER) or HER-2 in breast cancer, the use of BRCA 1 and 2 in ovarian cancer or K-RAS in colon cancer.
Only after the completion of numerous large clinical trials in breast, colon and lung cancers did scientists go back and re-examine the results using newly recognized yet highly impactful patient features like ER, HER-2 or K-RAS.
A Look at Equipoise (Balance)
The term applied to the errors that lead to these unsuccessful clinical trials is known as a failure of "equipoise".
Equipoise, or balance, stipulates that when a patient is randomized to one arm or another of a clinical trial that every effort has been made to determine whether there is a confounding variable that makes them more likely to respond to one of the arms over the other before they go on the treatment.
In my theoretical example this was male versus female, but ER (+) vs. ER (-); HER2 (+) vs. HER2 (-) represent very real experiences, where trials were conducted in patients who were not equally balanced in their likelihood of response.
Until you know to look for a given patient feature, and until you know that it confounds (influences) your clinical trial, you will continue to accrue patients to a treatment study that is likely to fail.
This is the dilemma that cancer patients face.
Has their physician determined at every level available that the treatment offered on the trial meets equipoise?
More often than not, the answer is no.
Whether it is ignorance of the feature, or an unwillingness to conduct additional studies, patients regularly enter clinical trials where one of the arms for them is likely to be better, yet they are randomized to the opposite arm.
In my practice, I do all that in my power to avoid this mistake.
A Look at Advanced Pancreatic Cancer Treatment
The current management of pancreatic cancer exemplifies this very issue.
Today, patients who present with advanced pancreatic cancer will be given either FOLFIRINOX or Abraxane plus Gemcitabine. There are other treatments, but in the U.S. today, these are the two principal drug combinations.
As the drugs in these two regimens are quite different, it is unlikely that every patient will be equally sensitive to both combinations.
Clearly, each patient’s biology dictates that one drug or combination may be uniquely better for them.
Some patients are very sensitive to Irinotecan, others would absolutely benefit from abraxane, but when drug selections are made, the physicians do not take the time, energy or effort to make that final crucial determination.
The Consequences of Random Drug Selection
Failing to address a patient's likelihood response can result in patients receiving treatments that don’t work while other treatments might have worked better.
More concerning is that while patients suffer through the side effects of one treatment, their clinical condition may deteriorate to the point where they may never receive the right drug for them.
Failure of equipoise may be the greatest problem confronting medical oncology clinical trials.
No one would conduct a clinical trial in lung cancer today without measuring EGFR and ALK mutations.
No one would conduct a breast cancer trial without measuring ER and HER2, and no one would conduct an ovarian cancer trial without conducting BRCA1 and 2 mutation analysis.
Yet, oncologists continue to give patients treatments for many diseases where they have not taken the trouble of determining the true likelihoods of response to the drugs they administer (such as cytotoxic drugs, signal transduction inhibitors or other targeted agents) even though these can be evaluated prior to treatment.
Patients Should Get More Information Before Joining Trials
This is the reason we encourage patients to find out if they are sensitive to drugs and to then get the treatments that work best for them.
Until clinical trialists are willing to incorporate into their trial designs the tools that can provide equipoise, patients should look carefully at the criteria under which they are being "randomized."
Participation in a clinical trial is a noble and often selfless undertaking on the part of patients. Physicians should be equally noble in their respect for the patient's well-being.
To randomize a patient to a clinical trial when an investigator has failed to assess whether one arm is better than the other for that individual, runs the risk of squandering the good will and well-being of those patients.
This remains a patient’s clinical trial dilemma.
As always, I appreciate your thoughts and comments.
Dr. Robert Nagourney, has been internationally recognized as a pioneer in cancer research and personalized cancer treatment for over 20 years. He is a TEDX SPEAKER, author of the book OUTLIVING CANCER, a practicing oncologist and triple board certified in Internal Medicine, Medical Oncology and Hematology helping cancer patients from around the world at his Nagourney Cancer Institute in Long Beach, California. For more info go to NAGOURNEYCANCERINSTITUTE.COM