Purpose To report on the long-term results of a randomized trial comparing a standard dose (400 mg/d) versus a higher dose (800 mg/d) of imatinib in patients with metastatic or locally advanced GI stromal tumors (GISTs). Patients and Methods Eligible patients with advanced CD117-positive GIST from 56 institutions in 13 countries were randomly assigned to receive either imatinib 400 mg or 800 mg daily. Patients on the 400-mg arm were allowed to cross over to 800 mg upon progression. Results Between February 2001 and February 2002, 946 patients were accrued. Median age was 60 years (range, 18 to 91 years). Median follow-up time was 10.9 years. Median progression-free survival times were 1.7 and 2.0 years in the 400- and 800-mg arms, respectively (hazard ratio, 0.91; P = .18), and median overall survival time was 3.9 years in both treatment arms. The estimated 10-year progression-free survival rates were 9.5% and 9.2% for the 400- and 800-mg arms, respectively, and the estimated 10-year overall survival rates were 19.4% and 21.5%, respectively. At multivariable analysis, age (< 60 years), performance status (0 v ≥ 1), size of the largest lesion (smaller), and KIT mutation (exon 11) were significant prognostic factors for the probability of surviving beyond 10 years. Conclusion This trial was carried out on a worldwide intergroup basis, at the beginning of the learning curve of the use of imatinib, in a large population of patients with advanced GIST. With a long follow-up, 6% of patients are long-term progression free and 13% are survivors. Among clinical prognostic factors, only performance status, KIT mutation, and size of largest lesion predicted long-term outcome, likely pointing to a lower burden of disease. Genomic and/or immune profiling could help understand long-term survivorship. Addressing secondary resistance remains a therapeutic challenge.