Since we have gotten a number of questions about Gleevec (imatinib) plasma levels and testing, I think it’s time to revisit the subject, especially for newer members. Since this is a very complex subject, this may take a few posts and a few days.
In this first post, I am just going to discuss what plasma level testing is and state some of the questions surrounding plasma testing. In a future post(s), I go into detail about some of the outstanding questions.
Gleevec and most other drugs currently used in GIST are dosed based on fixed dosing (same dose given to everyone, with the exception of higher doses for metastatic KIT exon 9 patients). For patients receiving the same 400 mg dose of Gleevec, some will have high concentrations of Gleevec in the blood (actually in a part of the blood, the plasma) and some will have low concentrations. In fact, the levels of Gleevec in the plasma can vary as much as 10 fold or more. In part, this helps to explain why some patients experience very mild side effects, while others experience very toxic side effects for the same drug given at the same dose.
Gleevec is approved for metastatic/advanced GIST based on the results of a phase II trial for this group of patients, the B2222 trial. 
This trial started in July, 2000 and compared 400 mg of Gleevec to 600 mg of Gleevec. Although there was a small advantage in progression-free survival for the 600 mg arm, there was no overall survival difference when both groups were compared based on the dose group that they were assigned to at the beginning of the trial (dose changes were ignored). This type of analysis is called intent-to-treat.
Two larger phase III trials comparing 400 mg to 800 mg found essentially the same results and form the basis of most thinking about dosage today. Current guidelines for metastatic GIST, such as the FDA prescribing information, NCCN guidelines and ESMO guidelines are primarily based on the trial results from these three trials. All adjuvant trials have only tested 400 mg, so they don’t really add anything regarding the question of dosing.
The Life Raft Group began raising questions about the optimal Gleevec dose as early as 2004, based on the actual dose that patients were taking. This was reported initially at CTOS in 2004 and formally published in 2009. There were still many questions since intent-to-treat is the gold standard, not the dose that patients actually took (it’s a long discussion to get into the biases and pros and cons of each method, so I am not going to do that in this post).
In January, 2008, Dr. George Demetri reported new data from the original phase II trial (B2222) at the 2008 GI ASCO meeting. When the data that had originally been analyzed by dose was re-analyzed by the concentration of imatinib in the blood, there was a correlation between low imatinib plasma levels and premature failure of imatinib.
Plasma levels TAKEN AFTER 29 DAYS OF GLEEVEC, were available for 73 of the 147 patients enrolled in the trial. These plasma levels were grouped into four equal groups (quartiles) according to imatinib trough plasma concentrations (the level of drug in the blood at its lowest point during the day, just before taking the daily Gleevec capsule). Plasma levels were not available for any time point after 29 days. This presents some problems that I will talk about in another post.
The quartile with the lowest imatinib plasma levels (below 1110 ng/mL, which is generally rounded to 1100 ng/mL) had a median response of 11.3 months to imatinib with a 44% objective response rate (a measure of tumor shrinkage). In contrast, those in the 2nd and 3rd quartiles had a median response of 30.6 months with a 67% response rate and those in the highest quartile had a median response of 33.1 months with a 74% (objective) response rate. For exon 11 patients (very sensitive to imatinib), the response rate was 55.6% for the lowest quartile, 94.1% for quartiles 2 and 3 and 92.3% for quartile 4.
The authors concluded that, “Exposure to adequate drug levels of imatinib appears to correlate with clinical benefit; patients with the lowest imatinib levels show lowest objective response and shortest time to progression. These results suggest that monitoring pharmacokinetic/ pharmacodynamic relationships may provide novel predictive markers and that exposure to adequate IM trough plasma concentrations (>1,110 ng/mL) is important for optimal clinical response.”
In 2009, Demetri and his colleagues reported their results in the Journal of Clinical Oncology. The full text link is here: http://jco.ascopubs.org/content/27/19/3141.long and the PDF is here: http://jco.ascopubs.org/content/27/19/3141.full.pdf+html
A video of Dr. Demetri giving an interview at the 2008 GI ASCO meeting can be found at this link: http://www.medpagetoday.com/MeetingCoverage/ASCOGI/8107
Similar results (slightly lower 1st quartile threshold of 1002 ng/mL) were obtained for CML (leukemia) patients, by Stephane Picard and colleagues. There were reported in 2007: http://bloodjournal.hematologylibrary.org/content/109/8/3496.full
Based on these results, several things happened:
- Novartis set-up/funded imatinib plasma level testing sites at a number of locations around the world. They also supported a website with information about plasma testing and some other info such as neo-adjuvant Gleevec. While I believe that Novartis had good intentions, the FDA saw this as unauthorized marketing for an indication that was not approved and issued a warning letter to Novartis. This immediately ended Novartis-sponsored plasma testing and severely curtailed the ability to get plasma testing in the United States.
- A clinical trial was designed to rigorously test the question of benefit for GIST patients based on imatinib plasma levels. This trial was sponsored by Novartis with SARC as a partner. This trial recruited 5 patients in 15 months and was then terminated for “failure to recruit”. This effectively ended the quest for “rigorous scientific data” into the subject of plasma level testing for GIST.
A little additional data has come out about plasma levels in GIST patients, but for the most part, they raise more questions than they answer. Some of these are:
- Several series have reported a lower “1st quartile” threshold than the 1100 ng/mL level found in the B2222 trial.
- Several groups have reported that imatinib plasma levels drop about 30% over time.
- Is 1100 ng/mL the right threshold? Should it be something else? If so, what?
- Confusion exists over the minimum threshold. For example:
Which patients does it apply to?
- Gleevec-sensitive?
- Gleevec-resistant?
- Less Gleevec sensitive?
- Exon 9?
- Wildtype?
- D842V?
- If 1100 ng/mL is the right level at 29 day, what is the right level at 3 months, at 6 months, at 1 year?
- Can patients that are tolerating Gleevec well and on 400 mg but have high plasma levels reduce their dose? (before anyone does this, I am going to give my opinion; the answer is no. I will talk about why in another post).
- How is the test done? Where can patients get it done?
- Logistic hurdles to testing.
- Can testing be integrated with what we already know about dose?
In future posts, I will explore these and other questions further. I welcome your comments and questions.
