More study needed to determine optimal Gleevec dose

Several large clinical trials have compared different doses of Gleevec in an attempt to identify the optimal dose. Differences in the initial results of the large phase III trials conducted by the U.S./Canadian group and the European/Asian/Australian group have been reported. The European study found a longer time to progression at 800 mg. vs. 400 mg. The U.S. study did not find any such correlation between dosage level and time to progression. A recent Life Raft Group study of patient-reported data showed fewer relapses at higher doses, especially when the analysis was based upon actual dose delivered instead of the “intent to treat” (starting) dose.

A phase II study of Japanese patients was also reported at the 2004 meeting of the American Society of Clinical Oncology. This study compared 400 mg. vs. 600 mg. This study found a higher partial response rate at 600 mg. vs. 400 mg. (60.9 percent vs. 46.4 percent), but found that this difference was not statistically significant. It is interesting that the original phase II trials found a similar partial response rate when comparing 600 mg. to 400 mg. (58 percent vs. 49.3 percent as of 8/15/02).

Clinical trials use the dose that the patient was randomized to receive as the reported dose. That means the dose reported in a trial may not reflect the dose the patient actually receives. For example, a patient with unacceptable side effects at 800 mg. may have his or her dose reduced to 400 mg. (or in a few cases, less than 400 mg.). This patient could receive 400 mg. for most of the time they were in the trial -- but they would still be counted in the 800 mg. arm. Using the starting dosage as the basis for statistical analysis is called “intent to treat.”

Drug treatment produces both desired effects and undesired effects. The acceptable level of side effects with oncology drugs is much higher than for most other diseases. One of the primary reasons to use “intent to treat” analysis is to try to measure the “real world results” of benefit plus toxicity. A patient on 400 mg. will probably have some level of benefit and some level of toxicity. A patient receiving 800 mg. may have some level of benefit greater than at 400, but the patient will probably have a greater level of toxicity. Intent-to-treat analysis attempts to answer which dose has the most clinical benefit. In theory, it should take into account beneficial effects, toxicity, a patient’s ability to tolerate the drug, and even dose reductions.

The intent-to-treat method of analysis can be a useful tool and has had a valuable place in clinical trials. However, it falls short in the Gleevec-for- GIST trials when used as the sole reporting method. The problem: It fails to adequately account for the fact that side effects, for most patients, improve over time. For many patients, side effects get MUCH better. Dr. Allan Van Oosterom, Dr. Ian Judson, and their European colleagues have provided a possible explanation for this effect. They have found that the body’s ability to get rid of Gleevec increases over time, resulting in patients having lower drug concentrations.

When I look at the results of all of the GIST/Gleevec trials, it is apparent that a starting dose of 800 mg. causes significant side effects for most patients, and is probably too high a starting dose. My impression is that 600 mg. may still be too high of a starting dose for many patients. In a Japanese study, 70 percent of the patients started at 600 mg. required dose reductions. Clearly there is a concern about the greater side effects that occur at higher doses.

Where do we go from here? After the remarkable early results in the original phase II Gleevec-GIST trials, the phase III studies were organized and started in a mere five weeks! This was an amazing accomplishment -- and desperately needed by GIST patients with no other effective therapy. At that time however, no one really knew that levels of Gleevec in the body would tend to fall over time. As a result, the trials were not really optimized to consider this phenomenon.

At the Life Raft Group membership meeting in April and at the ASCO meeting in June, Dr. Van Oosterom suggested that phasing-in higher doses of Gleevec might be possible. Interestingly, this is the method that was used by Dana-Farber Cancer Institute prior to the early phase III results. This is also the method that I favor.

While I believe that investigators will still learn more from the current phase III trials, I have my doubts that these trials can establish an optimal dose. Deficiencies also exist in the Life Raft Group study on dose. These include: patients not randomized (may introduce bias) and lack of central review (non-standardized response criteria).

What I would like to see is a doseoptimization trial of Gleevec where the high-dose arm has the higher dose phased in. This trial would have one arm at 400mg. vs. another arm where the dose was “optimized” for each patient. This would require the key pharmacokinetics researchers and clinicians reaching a consensus on what parameters need to be included to determine an optimum dose.

In my vision of this trial, it would start the “optimized” patients at 400 mg., measure drug concentrations and variables at one month, then, as tolerated, raise dose to “optimal” levels after another month, again at the third month, then every three months (scheduled to match CT scans). The timing of the phase in would have to be determined by the experts.

This type of trial is in line with current medical practice of proving what we think we know. It also makes better use of the intent-to-treat principal. If we think that phasing-in a dose is the way to go, then the intent-to-treat principal will test whether this method truly is better. I do believe, however, that the actual dose should generally be reported in addition to the intent-totreat dose in most clinical trials. I think this will help identify potential concerns in some cases. I believe that analyzing actual dose was one factor in focusing attention on the question of optimal dose in GIST. The other key factor was the finding by the Europeans that Gleevec concentrations tend to decrease over time.

Until we can either have the current phase III studies analyzed by both intent- to-treat AND actual dose, or we can do a trial similar to the one proposed above, patient and their doctors will have to educate themselves about the pluses and minuses of each study and make the best decision about what dose to administer. The questions that I urge patients to ask themselves are: Do I fit the profile described in the LRG study? That is, patients with metastatic GIST who had initial shrinkage? If the answer to that question is yes, then ask: Have my side effects gotten better over time? If the answer to both these questions is yes, then consider that a dose increase may be needed to maintain the drug concentration you started with.

Of course, no matter what you decide, you will still have to convince your doctor to change your dose.