In this study, patients who have been diagnosed with gastrointestinal stromal tumor (GIST) and have been treated with adjuvant imatinib for 3 years after surgery will be randomly allocated in a 1:1 ratio to receive imatinib (Gleevec) for 2 more years (Arm A) or to stop imatinib (Arm B). The study participants are required to have histologically verified GIST with a high risk of GIST recurrence despite removal of all macroscopic GIST tissue at surgery and 3 years of adjuvant imatinib. The high risk of GIST recurrence is defined as one of the following: gastric GIST with mitotic count >10/50 high power fields (HPFs) of the microscope, non-gastric GIST with mitotic count >5/50 HPFs, or tumor rupture. Study participants allocated to Arm A will receive imatinib 400 mg/day for 24 months after the date of randomization. All study participants will be followed up using blood tests and computerized tomography (or MRI) of the abdomen. The computerized tomography examinations will be performed at 6 month intervals. A total of 300 patients will be entered to the study. The study hypothesis is that adjuvant imatinib given for a total of 5 years may prevent some of the GISTs to recur as compared to patients who receive adjuvant imatinib for 3 years, and there may be a difference in the rate of GIST recurrence between the two groups.

The study will accrue patients in several countries in the Europe.

Inclusion Criteria:

• Age ≥ 18 years.
• Morphological and immunohistological documentation of GIST (immunostaining for KIT and/or DOG-1 positive, or mutation of KIT or PDGFRA present in tumor tissue).
• Macroscopically complete surgical resection of GIST (either R0 or R1 resection).
• Mutation analysis of KIT and PDGFR genes has been carried out.
• A high risk of GIST recurrence; either gastric GIST with mitotic count >10/50 HPFs, or non-gastric GIST with mitotic count >5/50 HPFs, or tumor rupture.
• Eastern Cooperative Oncology Group performance status ≤ 2.
• Adequate organ function.
• Female patients of childbearing potential must have a negative pregnancy test within 14 days before initiation of study drug dosing. Postmenopausal women must have amenorrhea for at least 12 months to be considered of non-childbearing potential. Male and female patients of reproductive potential must agree to employ an effective barrier method of birth control throughout the study and for up to 3 months following discontinuation of study drug.
• Patient willing to be followed up at the study site regardless of the result of randomization.
• Patient has provided a written, voluntary informed consent prior to study-specific screening procedures.

Exclusion Criteria:

• Presence of distant metastases or local recurrence of GIST.
• Not willing to donate tumor tissue and/or blood samples for the study molecular studies.
• Presence of a substitution mutation at PDGFRA codon D842 (usually D842V).
• Administration of adjuvant imatinib longer than for 3 years is planned regardless of the result of randomization, or "life long" imatinib administration is planned.
• Prior adjuvant (+ neoadjuvant) therapy with imatinib mesylate for at least 35 months has not been completed, or the total duration of prior adjuvant (+ neoadjuvant) imatinib administration exceeds the total duration of 37 months.
• Neoadjuvant imatinib for a duration that exceeds 9 months.
• Longer than 4-week break during adjuvant imatinib administration.
• The dose of imatinib at completion of 3 years of adjuvant imatinib was 200 mg per day or less or greater than 800 mg per day.
• Patient has received any investigational anti-cancer agents during adjuvant imatinib or between completion of adjuvant imatinib and the date of randomization.
• Patient has been free of another malignancy for less than 5 years except if the other malignancy is not currently clinically significant nor requiring active intervention, or if the other malignancy is a basal cell skin cancer or a cervical carcinoma in situ. Recent existence of any other malignant disease is not allowed.
• Patient with Grade III/IV cardiac disease as defined by the New York Heart Association Criteria (i.e., congestive heart failure, myocardial infarction within 6 months of study entry).
• Female patients who are pregnant or breast-feeding.
• Severe and/or uncontrolled medical disease (i.e., uncontrolled diabetes, severe chronic renal disease, or active uncontrolled infection).
• Known diagnosis of human immunodeficiency virus (HIV) infection.
• Patient with a significant history of non-compliance to medical regimens or with inability to grant reliable informed consent.