Following the ACOSOG Z9001trial, imatinib received market authorization in Europe for patients with GIST at significant risk of relapse in the adjuvant setting, according to the classifications of Miettinen and Joensuu. Thereafter, the SSG XVIII / AI trial proved the need to revise the recommendations of the European Society for Medical Oncology regarding the optimal duration of treatment, which is currently three years. Patients at low risk of recurrence should not receive adjuvant treatment with imatinib and recommendations cannot be made from the literature data as to the indication of adjuvant treatment for patients with an intermediate risk of relapse. The provision of prognostic molecular markers in this group of so-called intermediate-risk subjects would facilitate the identification of responders to imatinib and avoid overtreating some patients and undertreating others who would benefit from Imatinib. Recently, Lagarde et al. have shown that the Genomic Index (GI = A ² / C, where A is the total number of alterations gains or losses and C is the number of chromosomes involved in these alterations in Comparative Genomic Hybridization array(CGH array)) could have prognostic value in GIST, particularly in intermediate risk GISTs. More recent work by the same author in 100 cases of GISTs with intermediate prognosis according to the classification of Miettinem identified two prognostic groups based on GI. The rate of metastatic relapse at 2 years was 30.6% in the group with GI greater than 10 versus 5.4% in the group with GI less than 10 (manuscript under preparation). Thus, it is legitimate to set up a randomized trial to study the effectiveness of adjuvant treatment with imatinib in the GIST population at intermediate risk of relapse and with a high GI. This study is a prospective randomized clinical trial: a phase III, open-label, 2 parallel groups, multicenter study. The primary objective of this study is to assess the efficacy of adjuvant Imatinib on rate of metastatic relapse at 2 years in patients with intermediate-risk gastrointestinal stromal tumor presenting a high Genomic Grade Index. The second objectives of this study are to compare the two therapeutic approaches in terms of metastasis-free survival at 1 year, 2 years and 3 years, overall survival, clinical and biological tolerance, safety and Quality of life of patients and caregivers. The eligible subjects must meet all of the following criteria : subject with a gastrointestinal stromal tumor, intermediary risk from the Armed Forces Institute of Pathology classification [Miettenen 2006], subject with Genomic Grade Index higher than 10 determined by CGH array, subject with surgery for primary tumor performed from 2 weeks to 2 months before starting adjuvant Imatinib mesylate, subject with no evidence of residual macroscopic disease after surgery and with a medical decision to prescribe imatinib. Subjects meeting any of the following criteria must not be enrolled : subject who have experienced spontaneous tumor rupture before surgery, subject whose tumor has a PDGFRA D842V mutation evidenced by sequencing from tumor Block and subject whose mutational status meets the wild phenotype definition as evidenced by sequencing from tumor Block.

The Standard Group will receive adjuvant imatinib at a dose of 400 mg per day for a period of 3 years. Patients will be assessed for metastases every three months for three years with thoraco-abdominal and pelvic CT scan. The Experimental Group will receive the same thoraco-abdominal and pelvic CT scan. The estimated proportion of subjects relapsing at 2 years will be 30% in the experimental group and 2.5% in the standard group: alpha risk, 5%, power 80%. A total of 80 subjects (40 in each arm) will be included. This is a trial combining two learned societies that already are taking part in many clinical trials in France (French Sarcoma Group and French Digestive Cancer Federation). The expected benefits for patients are : not treat subjects for whom this treatment would offer too little benefit weighed against the disadvantages and treat subjects in whom this treatment would provide a real benefit and reduce the cost of treatment in patients who would not benefit from being treated by imatinib. The originality of this study is that it will include molecular data in the therapeutic decision and demonstrate the concept of individualized treatment in this patient population. This could ultimately change the current recommendations.

Inclusion Criteria:

• Man or woman 18 years old or over and PS:0-2
• No prior radiation therapy, no prior chemotherapy, no molecular targeted or biological therapy
• Subject with a gastrointestinal stromal tumor, intermediary risk from the Armed Forces Institute of Pathology classification [Miettenen 2006]
• Subject with Genomic Grade Index higher than 10 determined by CGH array;
• Subject with surgery for primary tumor performed from 2 weeks to 2 months before starting adjuvant Imatinib mesylate;
• Subject with no evidence of residual macroscopic disease after surgery (RO). Microscopically infiltrated margins, or supposed to be are allowed (R1)
• Subjects with absence of distant metastases
• Subject with a medical decision of treatment with its in accordance with imatinib marketed authorization.

Exclusion Criteria:

• Minors or pregnant or breast-feeding women.
• Subject with a contraindication to Imatinib, a known hypersensitivity to the active substance or to any of the excipients (ambivalence clause);
• Subject treated with medicinal products that induce CYP3A4;
• Subject who have experienced spontaneous tumor rupture before surgery (risk of spread);
• Subject whose tumor has a PDGFRA D842V mutation evidenced by sequencing from tumor block;
• Subject whose mutational status meets the wild phenotype definition as evidenced by sequencing from tumor block