Imatinib (Gleevec) for Unresectable or Metastatic GIST

Imatinib (Gleevec) was approved by the United States FDA for unresectable or metastatic GIST based on the results of clinical trials discussed below. This page answers the following questions (click the question to skip to that section).

Does imatinib help all GIST patients?
Does imatinib cure GIST?
How long does it take to show a response?
Does imatinib keep working indefinitely?
Does mutation status affect response to imatinib?
What is the appropriate dose of imatinib?

Does imatinib help all GIST patients?

About 85% of patients experience benefit from imatinib (Gleevec). Oncologists use a description system called RECIST (Response Criteria in Solid Tumors) to categorize tumor responses to treatments (Therasse et al, 2000). The categories are:

  • Complete response (CR): tumor is no longer detectable.
  • Partial response (PR): a decrease of 30% or more in the longest dimension of a single tumor or in the sum of the longest dimensions of a group of target lesions.
  • Stable disease (SD): a decrease of less than 30% in the longest dimension of a single tumor (or summed dimensions of a group of tumors) or an increase of less than 20% in the same dimension (or sum).
  • Progressive disease (PD): an increase of at least 20% in the monitored longest dimension (or sum).

The tumor “best response” results from all the available trials of imatinib are summarized in the following table. The results may vary partly as a function of the duration of follow-up, since some tumors may take a long time to show enough shrinkage to qualify for “partial response.” For each trial the most recently reported results are shown, and an earlier analysis may also be included for comparison.









I EORTC Van Oosterom et al (2002) 35 0% 54% 37% 9%
II US and Finland
Demetri et al (2002)
Blanke et al (2006)
147 0%
II EORTC Verweij et al (2003) 27 4% 69% 18%
III North American
Intergroup S0033
Benjamin et al (2003)
Rankin et al (2004)
746 0%
III EORTC / ISG / AGITG Verweij et al (2004) 946 5% 47% 32% 11%

The patients with an objective response (complete or partial), plus the patients with stable disease, all receive benefit from the treatment. Between about 50-70% of patients showed enough tumor shrinkage to qualify as an objective response, but only a handful achieved a complete response. Between about 18-32% of patients attained stable disease. Most patients with stable disease show a small amount of shrinkage. Importantly for quality of life, however, the symptoms from large tumors may be controlled or reversed even without tumor shrinkage. In summary, about 85% of patients do benefit from imatinib.

Studies cited in the table above (click PMID or ASCO abstract link to see abstract):

Demetri GD et al (2002), PMID: 12181401 (free access paper)
van Oosterom AT et al (2002) PMID: 12528778
Verweij J et al (2003), PMID: 12957454
Verweij J et al (2004), PMID: 15451219

The response of tumors to imatinib also depends partially on the mutation type shown. The table below summarizes the response rates on initial administration of imatinib (Gleevec) by tumor mutation type in adults. Note that this does not apply to secondary mutations that may develop after the patient has been taking imatinib for some period of time. Also, the row for wild-type tumors (showing no KIT or PDGFRA mutations) does not apply to pediatric/adolescent GISTs.


Predicted imatinib response and GIST characteristics based on the mutation of the primary tumor in the KIT or PDGFRA genes.



to Imatinib

cell type


KIT gene 80-85% mostly spindle
(long tapered

any site</st rong>

exon 9



small bowela

exon 11




exon 13


~100% b


exon 17




PDGFRA gene 5-10% dummy epithelioid
(round shape)
or mixed
(spindle and
usually stomach
exon 12


some b

exon 14



exon 18



Wild Type adult GIST c
(no KIT or
PDGFRA mutation)

more with stable disease than with
partial response

any type any site
* reported as “clinical benefit” including response or stable disease based on combined data from 3 clinical trials: B2222, EORTC phase I-II, EORTC Phase III.
a Rarely gastric GIST can be exon 9. About a third of small intestinal GISTs are exon 9 mutant (remainder are exon 11).
b Few data available.
c Pediatric/adolescent GISTs are usually wild-type but differ from adult wild-type GISTs in terms of histology and drug responsiveness.


The response predictions shown in the table above may change over time as more data become available.

back to top

Does imatinib cure GIST?

No. When surgery is done to remove GIST (primary tumors or metastases) in patients who have been pre-treated with imatinib (Gleevec), some viable tumor cells remain (Scaife et al, 2003; Bauer et al, 2005). If the drug is discontinued after a period of good response in patients with remaining visible disease, GIST reactivates. These findings indicate that imatinib does not kill all cells in an unresected tumor and should be continued as long as disease is present (unless the patient switches to a different drug).

back to top

How long does it take to show a response?

Patients may start to notice subjective relief from symptoms within a few days of beginning to take imatinib (Gleevec). PET scans can also confirm tumor response within a few days to two weeks. PET response is based on tumor metabolic changes (reduced uptake of tracer), not tumor size changes. However, PET is seldom used outside of clinical trials at the start of imatinib therapy because

  • Some patients GISTs do not show PET activity on baseline
  • PET is more expensive and less widely available than CT
  • If there is no alternate treatment, it is less important to assess response so rapidly. Since a second anti-GIST drug was approved in early 2006 (sunitinib, or Sutent), this reason is less applicable than previously.

However, for patients whose characteristics make them less likely to respond to imatinib and for whom alternate treatments should be considered, then early PET monitoring is useful to determine as soon as possible whether imatinib is working.

Objective imaging results to demonstrate measurable tumor size changes take longer. Blanke et al (2006) reported a median time-to-response on CT scan of 12 weeks, with 75% of “best responses” occurring by 23 weeks (about 6 months). In addition to the ASCO webcast ilnked above, you can download a pdf of Dr. Blanke’s poster on this topic. Notably, the maximum time to achieve a partial response was 171 weeks, which is over 3 years. Dr. Blanke stated “Late responses are often seen in patients with initial stable disease.”

Verweij et al (2004), reporting on the EORTC/ISG/AGITG Phase III trial, found a median time to “best response” of 107 days (about 3.5 months). However, 25% of patients showed best response before 58 days of treatment (about 2 months), and 25% needed over 172 days (about 6 months). At the time of this report, only 58% of subjects in this trial had been followed for 2 years or more. Therefore, additional late responses may be captured in subsequent analyses.

back to top

Does imatinib keep working indefinitely?

Many patients develop resistance to imatinib (Gleevec) after a period of time. The research study providing the longest duration of follow-up is the US-Finland Phase II trial most recently updated in February 2006 (Blanke et al, 2006). You can view Dr. Blanke’s webcast and see his entire poster at the preceding link. Take a look at the figure reproduced below.

As illustrated in the above figure, patients who achieved partial response and those with stable disease show very similar durations of survival in the US-Finland study. Dr. Blanke stated in his oral presentation (available via ASCO webcast) that 33% of patients were still progression-free on imatinib at 4.4 years of treatment. Some individuals have been stable on treatment for over 5 years.

The two large Phase III imatinib trials also provide data regarding progression-free survival and overall survival, although these studies have not been ongoing as long as the Phase II study described by Dr. Blanke. The following table compares the progression-free and overall survival rates reported so far. The results of both Phase III studies are very similar, showing that about half of patients are stable on imatinib therapy at 2 years, and about a third are still stable at 3 years.

Percent of subjects with progression-free versus overall survival at 1, 2, and 3 years by imatinib dose group in the two Phase III trials. (NR = not reported.)

Trial Dose Progression-Free Survival Overall Survival
1-year 2-year 3-year 1-year 2-year 3-year
US S0033 400 mg 71% 50% 29% 86% 76% NR
800 mg 70% 53% 33% 85% 72% NR
EORTC/ISG/AGITG 400 mg 65% 50% 31% 85% 69% NR
800 mg 73% 56% 35% 86% 74% NR
Combined “MetaGIST”results 400 mg NR NR 30% NR NR 60%
800 mg NR NR 34% NR NR 61%

When resistance to imatinib does occur, it may take two forms:

  • General progression of all tumors
  • focal progression of only one or two tumors, or one portion of one tumor (a phenomenon known as “a nodule within a mass”) (Shankar et al, 2005)

The initial strategy for overall progression is to increase the dose of imatinib (for patients taking less than 800 mg) or to consider a different drug. For focal progression, the current strategy is to consider surgical intervention for the progressing tumor(s) plus a potential dose increase. This is discussed on our page Surgery for Metastatic GIST. Also see the NCCN Sarcoma Guidelines and the ESMO GIST Consensus conference report (summarized on our site in a slide show).

back to top

Does mutation status affect response to imatinib?


Yes, the particular mutation shown by a patient’s GIST is clearly related to imatinib (Gleevec) response. For a list of labs that perform mutation testing, see our page Mutation Testing. The differential response to imatinib (Gleevec) by mutation status in the US-Finland Phase II trial is illustrated in Figure 6 from Blanke et al (2006), reproduced below.

The group of patients with exon 11 mutations showed a partial response rate of 87%, while the exon-9-mutation group showed a partial response rate of 48%. Including stable disease plus partial response, over 90% of exon-11-mutant patients benefited, compared to about 74% of exon-9-mutant patients. The durations of survival were also longer for the exon-11-mutation group, as shown in the following figure from Dr. Blanke’s poster.

In 2007 Dr. van Glabbeke et al presented responses by mutation type among patients in the MetaGIST analysis of 1640 patients (combined subjects from the US 0033 and the EORTC/ISG/AGITG Phase III trials.  Patients with exon 11 KIT mutations had a longer progression-free survival and longer overall survival when treated with imatinib than did those with other mutation types, as shown in the following table.

Imatinib benefit in Meta-GIST analysis of 1640 patients as reported by van Glabbeke et al in webcast of ASCO 2007 abstract 10004.

Median Progression-
Free Survival
3-year PFS
Median Overall
3-year OS
KIT exon 11 26 months 38% 60 months 69%
KIT exon 9 13 months 11% 31 months 44%
wildtype 16 months 27% 43 months 57%
other 11 months 9% 34 months 46%

You can read more about mutations in a paper entitled Biology of Gastrointestinal Stromal Tumors by Corless, Fletcher, and Heinrich (2004), available via the link above.

back to top

What is the appropriate dose of imatinib?

The evidence on this question indicates that 400 mg of imatinib pe day is the appropriate dose for most GIST patients, but that patients with an exon 9 mutation of the KIT gene benefit more at a higher dose.

All mutations types combined

As shown in a table above, the results for the two Phase III trials have so far yielded very similar responses for imatinib (Gleevec) doses of 400 mg/day versus 800 mg/day when mutation status is not considered. However, in both trials dose reductions due to side effects were much more common in the 800-mg groups than in the 400-mg groups.

Percent of patients with dose reductions

400 mg 800 mg
US S0033 10% 44%

When Dr. Charles Blanke reviewed this question in 2005, he concluded that until there is clear evidence “it remains prudent to start patients with advanced GIST at 400 mg daily” (Blanke, 2005). In a later analysis of the Phase II B2222 trial, Blanke et al (2008) stated “No difference in outcome was seen between patients taking imatinib 400 or 600 mg/day. This result parallels the findings from two Phase III studies that assessed imatinib 400 vs 800 mg/day. Neither of those trials showed a progression-free survival advantage for the initial higher dose.”

Exon 9 mutant patients

Dr. Heinrich’s plenary lecture (referenced in the preceding question) discusses a trend for better response at 600 mg/day for exon-9-mutant patients than at a lower dose. This important question was addressed in the van Glabbeke et al 2007 ASCO presentation of meta-GIST data (ASCO 2007 abstract 10004) analyzing the combined 1640 subjects from the ongoing Phase III trials. As shown in the following table, exon-9-mutation patients assigned to the 800 mg dose had a longer progession-free survival and overall survival.

Imatinib benefit in Meta-GIST analysis of 91 patients with exon 9 mutations of the KIT gene as reported by van Glabbeke et al in webcast of ASCO 2007 abstract 10004.

Median Progression-
Free Survival
3-year PFS
Median Overall
3-year OS
400 mg 6 months 5% 28 months 37%
800 mg 19 months 17% 35 months 49%

Dose increase upon progression

Until recently, all available data agreed that between 25 and 40% of patients who try a dosage increase of imatinib due to tumor progression will regain tumor control at the higher dose at least temporarily. However, Debiec-Rychter et al (2006), analyzing data on mutation status from part of the subjects in the EORTC trial, found that exon-9-mutant GIST patients had a much higher percentage of responses to dose increase (57%) than did exon-11-mutant patients (only 7%). In this analysis patients with exon-9 mutation had superior progression-free survival at a dose of 800 mg/day of imatinib, especially if given within the first 6 months of treatment.

Therefore, it now appears that 400 mg/day is the appropriate dose for all except exon-9-mutant patients, who benefit from higher doses beginning early in treatment. Hopefully additional data analyses of other trials will be forthcoming to confirm these findings.

back to top