Skip Navigation U.S. Department of Health and Human Services
Agency for Healthcare Research Quality
Archive print banner
Report on the Relative Efficacy of Oral Cancer Therapy for Medicare Beneficiaries Versus Currently Covered Therapy

This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: Let us know the nature of the problem, the Web address of what you want, and your contact information.

Please go to for current information.

Part 2. Imatinib for Gastrointestinal Stromal Tumors (GISTs) (continued)


In this section we summarize the findings of the review in terms of answering the key questions initially posed, and then discuss the clinical and research implications of these data.

Prior to the advent of imatinib, unresectable or metastatic GIST had an exceptionally poor prognosis.5,16  Single-agent doxorubicin, single-agent ifosfamide or combination chemotherapy including these agents were the standard of care since GIST was treated in the same manner as any other sarcoma of soft tissues, although response rates were exceedingly low and short-lived, and generally indicative of the ineffectiveness of any conventional chemotherapy approach.31,32  Importantly, these chemotherapeutic studies were done in an era when it was hard to differentiate GIST from other soft tissue sarcomas (STS), so the reported response rates are for the entire group of tumors rather than GIST specifically.16  With the discovery of CD117 and the KIT or PDGFRA tyrosine kinase proteins on the surface of most GISTs, it became possible to designate any individual STS as GIST or as some other histopathologic subtype of STS.  The use of imatinib for CD 117 positive advanced GISTs quickly followed.  Efficacy and tolerability have only been compared with historical norms from studies of the more general advanced STS therapy of single-agent doxorubicin, single-agent ifosfamide, or combination chemotherapy.  Head-to-head comparisons are not available.  Given the great improvements in efficacy witnessed in the phase II and III studies, it is unlikely that head-to-head studies would be conducted or would be ethical.  More recent analyses are starting to evaluate the role of imatinib in the adjuvant and neo-adjuvant settings, but this work is early and the role of imatinib in these peri-surgical settings remains unclear.  

1. In patients with GIST, what is the effect of imatinib compared to doxorubicin and ifosfamide on overall survival, disease free survival, time to progression, CR, PR, and quality of life?

There is consistent convincing evidence from high quality phase II and III studies that imatinib for unresectable or metastatic GIST yields complete response (CR) rates of 0-6 percent, partial response (PR) rates of 45-67 percent, and stable disease (SD) rates of 19-47 percent, with an overall response rate (CR + PR) of 49-71 percent.  This is substantially better than historical response rates of 15-34 percent (median 26 percent) for single agent doxorubicin and 7-38 percent (median 26 percent) for single agent ifosfamide (studies of general STS).31  Experts argue that these comparative efficacy rates are high, noting that the GIST-subgoup within the studies of all STS had response rates in the 0-5 percent range.68,5  Exact estimates are difficult to determine due to the historical difficulty with distinguishing GIST from other STS prior to the advent of CD117.

Data from the Verweij, et al. (2004) study49 provides the most complete estimates of survival.  One-year overall survival (OS) with imatinib can be estimated at 85 percent and 2-year OS at 72 percent; 2-year progression-free survival (PPS) with imatinib can be estimated at 44-50 percent depending upon the imatinib dose.   The most widely reported survival estimates from earlier STS studies are from a phase III trial of combination chemotherapy with doxorubicin, dacarbazine, ifosfamide, and mesna (MAID) vs. doxorubicin and decarbazine for advanced STS.37  In that study, 1-year OS is approximately 45-50 percent for both groups (estimated from Kaplan-Meier graphs) and 2-year OS is approximately 25-30 percent %; 2-year PFS was approximately < 5 percent.  In the NICE systematic review of non-imatinib treatments for advanced GIST, OS was 72 percent (18-100 percent) at 1 year, 40 percent (30-66 percent) at 2 years, and 16 percent (0-40 percent) at 3 years.16  The interventions reviewed were heterogeneous including novel chemotherapeutics and/or standard sarcoma chemotherapy regimens; studies included patients with GISTs and a broad range of other histopathologic subtypes of STS.  In the Bramwell, et al. meta-analysis of randomized trials of doxorubicin-containing regimens for advanced STS overall, median survival ranged from 7.3-12.7 months and OS was not reported.32

The conventional tumor response criteria of CR and PR represent the conventional goal in oncology to eliminate the tumor to the greatest extent possible in an effort to ultimately improve patient outcomes and survival.  Recently there is an evolving change to this convention.  Current studies suggest that if targeted therapy stabilizes disease, it may prolong survival despite failure of the tumor to shrink sufficiently to show a response according to conventional criteria.69  In essence the tumor is changed into a more chronic disease, quiescent until resistance occurs.

The role of imatinib in other clinical settings is still unclear.  Clearly, complete surgical resection is still the therapy of choice in patients with primary presentation of limited GIST in whom the disease can be completed resected without unacceptable functional morbidity.3,16  It is unknown whether pre-operative (so-called "neo-adjuvant") imatinib can make complete resection more feasible and recent studies are just starting to address this question (Table 5).  The role of adjuvant imatinib for resected GISTs at high risk of recurrence is also unknown, but is being addressed by a large, prospective, placebo-controlled clinical trial sponsored by the U.S. NCI that is ongoing (ACOSOG trial Z9001).  Meta-analysis of adjuvant doxorubicin-containing regimens for general STS suggest that they improve recurrence rates but not overall survival;28 subsequent studies of adjuvant ifosfamide-containing regimens suggest the same.29

Quality of life outcomes have been poorly studied.  The only reported relevant findings were from one study;46 patients receiving imatinib had demonstrable improvement in performance status with only 42 percent of patients fully functional at baseline and 64 percent fully functional 4 months after initiation of imatinib.

2. In patients with GIST, what is the effect of imatinib compared to doxorubicin (Adriamycin) and ifosfamide on adverse effects, tolerability, and compliance with treatment?

Imatinib has far fewer adverse effects (any grade and grade 3/4) compared with single-agent doxorubicin or ifosfamide.  Imatinib's most common side effects are edema, nausea and diarrhea, which are rarely grade 3 or 4.  Any grade 3 or 4 side effects occur in ≤36 percent of patients at the 600 mg daily dose or lower and hemorrhagic or hematologic effects occur in ≤8 percent.  At 800 mg daily, 30-50 percent of patients will have grade 3 or 4 toxicities, which are primarily hemorrhagic or hematologic effects.  Compliance with treatment was not reported.

Doxorubicin’s most concerning toxicities are its cardiotoxicity, nausea/vomiting, mucositis, and myelosuppression.  In the Bramwell, et al. meta-analysis of randomized trials of doxorubicin-containing regimens for advanced GIST, toxicities were variably reported and included severe hematologic 28-53 percent and moderate/severe nausea/vomiting 2-42 percent.32  At total doxorubicin doses  < 400 mg/m2 the incidence of congestive heart failure (CHF) is 0.14 percent.70  The incidence of CHF increases as the cumulative dose increases; at total doses of 550 mg/m2 the incidence is 7 percent and at 700 mg/m2 the incidence is 18 percent.  Ifosfamide’s most frequently reported toxicities include bladder toxicity, nephrotoxicity, nausea/vomiting myelosuppression, and neurotoxicity.  Summary toxicity data were not identified in the literature; grade 3-4 toxicity from individual studies varied widely with increasing rates as doses increased.  For example, Antman, et al. reported neurotoxicity in 19 percent of patients who received 2g/m2 x 4 days,34 and van Oosterom, et al. reported up to 63 percent with grade 3/4 leukopenia at 3 g/m2 x 4 days.35  In the MAID combination, the incidence of severe life-threatening toxicities includes the following:  leucopenia (89 percent), granulocytopenia (79 percent), thrombocytopenia (26 percent), anemia (22 percent), nausea/vomiting (19 percent), mucositis (9 percent), neurotoxicity (6 percent), and diarrhea (4 percent).37  Life-threatening cardiotoxicity due to doxorubicin was not seen in the MAID phase III trial.  Further, these chemotherapeutic agents also have the inconvenience and increased cost of requiring parenteral (intravenous) administration directly under the supervision of an oncologist and treating nurse.

3. What patient or tumor characteristics distinguish treatment responders from non-responders and have potential to be used to target therapy?

There was little consistency in studies seeking to identify possible prognostic factors.  No factor was evaluated in more than two studies.  The most relevant predictors of response relate to the mechanism of action of imatinib.  Patients with GIST (presumably expressing CD117) are more likely to respond to imatinib than patients with other STS.48  Patients with identified c-kit mutations, especially those on exons 11 and 9, are less likely to progress on imatinib and have longer overall survival.47,59  These findings are striking, since c-kit mutations including those on exons 11 and 9 lead to the more malignant GIST phenotype when imatinib is not used.12,17-21

Radiological predictors suggestive of response may be useful for prognostication and tailoring therapy.  In particular, early response on PET scan at day 8 predicts clinical response to imatinib.61  In the highest quality study of two similar studies, PET was more sensitive than CT at determining tumor response and combined PET-CT was most sensitive.62  A second study of lower quality contradicted the finding of PET vs. CT.  Other clinical predictors were varied, unrelated to the mechanism of action of imatinib, and uncorroborated.

Taken together, these data suggest that the patients most likely to get benefit from imatinib are those with c-kit mutations, especially those on exons 11 or 9.  Early response PET may be a good indicator of overall treatment effect.

Current State of Clinical Use

Imatinib is quickly becoming the standard of care for unresectable and/or metastatic GIST worldwide.  A recent National Comprehensive Cancer Network (NCCN) 2004 Task Force Report and Guideline advocated its use3, as does the NICE systematic review.16  The NCI clinical guide at does not recommend imatinib or other specific therapy for recurrent/metastatic GIST currently; 30 experts argue that this guideline is out of date and does not represent current clinical care.  According to the NCCN, baseline CT with or without PET is advocated, with three monthly reassessments to evaluate response to therapy.  Imatinib is started at a minimum dose of 400 mg daily; dose escalation can be considered for tumor progression or disease recurrence.  The NCCN recommends imatinib as the only chemotherapy for primary, metastatic, postsurgical adjuvant, or progressive recurrent treatment. The use of imatinib in the post-operative adjuvant setting remains the focus of ongoing clinical research studies.

The optimal dose continues to be unclear.  The current FDA indication is for 400–600 mg daily.43  The randomized phase II trial of 400 mg vs. 600 mg daily does not show clear benefit of one dose over another.43,46  In the randomized phase III trial of 400 mg vs. 800 mg, there was a trend (in the U.S./Canadian trial) or a statistically significant increase (in the European/Australasian trial) in progression-free survival but not overall survival with the higher imatinib dose.  However, the higher dose had significantly more adverse effects.71,72,73,54 Patients on the lower dose were able to cross over to the higher dose when disease progressed and some anti-tumor activity was seen in these crossover patients.49,50 

Current phase III studies are investigating the question of 400 mg vs. 800 mg daily further and the option for intermittent dosing.  Early analyses of the two dosing levels suggest that the two doses have equal efficacy, and that increasing to the higher dose in the setting of tumor progression at the lower dose is possible.53,54  In the trial of continuous vs. intermittent dosing, patients are provided continuous imatinib for the first year (dose unknown) and then randomized to continue the imatinib or stop the imatinib and resume it in the setting of tumor progression.52  It is too early to interpret the results from this study.  Based on current information, the current dosing plan of 400 mg daily on an uninterrupted schedule and then increasing to 800 mg daily in the setting of progressive tumor appears to be a rational strategy.   

Additional clinical trials are underway to determine whether adjuvant imatinib for one, two or three years post-resection will impact progression-free survival.74

Forthcoming Evidence and Implications for Future Research

Understanding of the role of imatinib in GIST is quickly evolving.  At the American Society of Clinical Oncology (ASCO) meeting in June 2005, 19 abstracts included information specifically about imatinib for the treatment of GIST.  Published ASCO abstracts were reviewed to develop a horizon view of emerging data and upcoming clinical trials.  Of the 19 abstracts, 3 were case reports,75-77 3 were retrospective reviews78-80, and 2 focused on radiographic issues81,82 and therefore excluded from this discussion.  Eleven abstracts were fully reviewed.

Three abstracts present information from the S0033 Phase III trial of 400 mg vs. 800 mg of imatinib for the management of advanced GIST.83-85  A total of 746 participants were randomized in this study.  On the 400mg arm (N=350), 62 percent had at least one dose delay and 10 percent had at least one dose reduction unusually due to side effects such as rash, edema, or gastrointestinal hemorrhage.83  On the 800 mg arm, 56 percent had at least one dose delay and 44 percent had at least one dose reduction due to edema, nausea, and fatigue.  Crossover from 400 mg to 800 mg was allowed for non-responders at the lower dose.  This occurred in 112 patients with 23 percent having at least one dose delay and 16 percent one dose reduction (data available on 77 of 112 crossover patients).  Pathological data and tumor response information were available for 414 participants.84  Patients with KIT positive GIST (N=377) and KIT negative GIST (N=14) had similar response rates and PFS at 2 years (KIT+ 49 percent, KIT- 43 percent).  Patients with non-GIST histology (N=16) had substantially poorer outcomes with 13 percent PFS at 2 years.  Median survival for GIST patients had not been reached at the time of the abstract analysis; median survival for non-GIST patients was 8 months.  Unblinded results of this RCT are pending.  Among KIT+ GIST patients, 86 percent had KIT mutations and 1 percent had PDGFRA mutations, with an overall mutation frequency of 87 percent.85  Patients with the exon 11 mutant KIT isoform were more likely to have an objective response (OR) to imatinib (OR=67 percent) than those with the exon 9 mutation (OR=40 percent) or no mutation (OR=39 percent, p=0.0022).  There was also a trend toward better overall survival for patients with the exon 11 mutation.  Current analyses of S0033 suggest that there are differential rates of toxicity for 400 mg vs. 800 mg, that crossover to higher dose imatinib is feasible when patients progress on the 400 mg dose, that imatinib is efficacious for both KIT+ and KIT- GIST but not tumors other than GIST, and that patients with KIT exon 11 mutations have the best prognosis.

The BFR14 study is a phase III trial of the French Sarcoma Group evaluating continuous vs. interrupted imatinib for the management of advanced GIST; two abstracts regarding this trial were presented.86,87  Patients who were progression-free after one year were randomized to either discontinue imatinib until evidence of further progression or to continue imatinib until evidence of progression.  Thus far, 198 patients have been enrolled in the study and 58 patients were free from progression at one year and therefore randomized.  This study is ongoing and final results are pending, however at the time of the ASCO abstract, 66 percent of patients in the interrupted treatment plan arm had progressed vs. 15 percent of patients who received uninterrupted imatinib (median follow up 21 months, p < 0.0004 for difference in PFS).87  Imatinib reintroduction allowed tumor control in 79 percent of patients.  One year OS rates were similar (89 percent vs 87 percent, p = 0.46).  Evaluation for predictors of OS and PFS among enrolled participants demonstrated that elevated lymphocyte count (HR 1.25, 95 percent CI 1.04-1.49), CD34 negative phenotype (HR 5.18, 95 percent CI 1.98-13.6), and performance status > 1 (HR 5.32, 95 percent CI 1.75-16.2) independently predicted OS while liver metastases ((HR 0.40, 95 percent CI 0.20-0.83) and CD34 positive phenotype (HR 0.45, 95 percent CI 0.22-0.93) independently predicted higher PFS.86  Current analyses of BFR14 suggest that PFS is poorer when imatinib is interrupted, but that reintroduction of the drug allows further tumor control in the majority of patients, thus far without a difference in overall survival.

ACASOG Z9000 is a phase II evaluation of adjuvant imatinib at 400 mg for 1 year in patients with primary high risk GIST following complete resection.88  Data on 106 evaluable patients who had been in the study for at least one year were presented in one ASCO abstract (median age 58, 67 percent male).  Toxicity profiles were similar to other studies of imatinib; nineteen (18 percent) patients did not complete therapy either due to toxicity (N=6) or withdrawal of consent (N=12).  This study is ongoing with survival outcomes pending.  The current assessment of Z9000 is that adjuvant imatinib is well tolerated.

A phase II efficacy study including 7 patients with unresectable advanced GIST treated with imatinib 400 mg daily at Oncology Hospital “Siglo XXI” IMSS (Mexico) demonstrated an overall response rate of 72 percent.89  One patient achieved a complete response.  Average follow up was 9 months (range 5-10).

In addition to the predictors of response and survival described in the S0033 and BFR14 trials, there were three other abstracts addressing predictors of response.  In an analysis of 15 advanced GIST patients treated with imatinib at 400 mg daily, PDGFR overexpression (N=10 of 15) predicted shorter time to progression (p=0.02).90  In an analysis of 68 cases of advanced GIST, p53 mutations were identified in 12 cases (18 percent).  Two-year OS was better in p53 mutation negative patients than p53 positive patients (89 percent vs. 75 percent, p=0.0156).91  In an analysis of 55 resected GIST patients, survival was predicted by size of the tumor mass (tumor <5cm with 3-year OS 86 percent and >5cm 66 percent, p=0.023) and mitotic activity (<10 mitoses per high power field 3-year OS=90 percent, >10 OS=64 percent, p=0.0368).92  In this study, patients with a deletion or insertion of KIT exon 11 had poorer 3-year OS of 35 percent vs. 64 percent for all other patients (p=0.0383).  These findings contradict those of the S0033 study where exon 11 mutations predicted better survival.85  Overall, molecular predictors analyses presented at ASCO suggest that KIT status does not affect GIST response, that the effect of KIT exon 11 mutations on outcomes with imatinib is unclear, that we have a lot to learn about the CD34 phenotype, and that p53 may portend poorer prognosis in GIST.

Only one abstract focused on side effects of imatinib for GIST.  Mean corpuscular volume (MCV) was monitored for 33 advanced GIST patients treated with 400 mg daily.93  A total of 42 percent developed an elevated MCV over the upper limit of normal, without coincident anemia or explanation.  This side effect was asymptomatic.

Overall, preliminary review of this forthcoming evidence suggests that new data will soon be available to inform ideal dosing, ideal dosing schedules, timing with surgery, and likelihood of response.  In addition to ongoing phase II and III studies, future clinical studies will likely focus on refining the molecular predictors of response and developing related tests for routine clinical use.  The other active area of GIST clinical research is to develop radiological tests that are less invasive and are more predictive of response to imatinib and outcome.  Given the growing number of recent articles about neoadjuvant and adjuvant imatinib, further prospective and coordinated studies of the use of imatinib in this clinical context need to be undertaken.  Further, many patients with GIST still have progressive or recurrent disease despite imatinib.  Imatinib resistance is a main topic in CML research.  Research on imatinib resistance in GIST is likely to be forthcoming, including an understanding of the molecular basis of this resistance and new methods to overcome it.  And, like in CML, combinations of imatinib and other chemotherapies will likely be studied in the future.

Return to Contents
Proceed to Next Section

The information on this page is archived and provided for reference purposes only.


AHRQ Advancing Excellence in Health Care