BACKGROUND: Approval of imatinib for adjuvant treatment of gastrointestinal stromal tumours (GIST) raised discussion about accuracy of prognostic factors in GIST and the clinical significance of the available risk stratification criteria. METHODS: We studied the influence of a new modification of the NIH Consensus Criteria (the Joensuu risk criteria), NCCN-AFIP criteria, and several clinicopathological factors, including tumour rupture, on relapse-free survival (RFS) in a prospectively collected tumour registry series consisting of 640 consecutive patients with primary, resectable, CD117-immunopositive GIST. The median follow-up time after tumour resection was 39 months. None of the patients received adjuvant imatinib. RESULTS: The median RFS time after surgery was 50 months. In univariable analyses, high Joensuu risk group, tumour mitotic count >5/50 HPF, size >5 cm, non-gastric location, tumour rupture (7% of cases; P = 0.0014) and male gender had adverse influence on RFS. In a multivariable analysis mitotic count >5/50HPF, tumour size >5 cm and non-gastric location were independent adverse prognostic factors. Forty, 151, 86 and 348 patients were assigned according to the Joensuu criteria to very low, low, intermediate and high risk groups and had 5-year RFS of 94%, 94%, 86% and 29%, respectively. CONCLUSION: The Joensuu criteria, which include 4 prognostic factors (tumour size, site, mitotic count and rupture) and 3 categories for the mitotic count, were found to be a reliable tool for assessing prognosis of operable GIST. The Joensuu criteria identified particularly well high risk patients, who are likely the proper candidates for adjuvant therapy.
BACKGROUND: Approval of imatinib for adjuvant treatment of gastrointestinal stromal tumours (GIST) raised discussion about accuracy of prognostic factors in GIST and the clinical significance of the available risk stratification criteria. METHODS: We studied the influence of a new modification of the NIH Consensus Criteria (the Joensuu risk criteria), NCCN-AFIP criteria, and several clinicopathological factors, including tumour rupture, on relapse-free survival (RFS) in a prospectively collected tumour registry series consisting of 640 consecutive patients with primary, resectable, CD117-immunopositive GIST. The median follow-up time after tumour resection was 39 months. None of the patients received adjuvant imatinib. RESULTS: The median RFS time after surgery was 50 months. In univariable analyses, high Joensuu risk group, tumour mitotic count >5/50 HPF, size >5 cm, non-gastric location, tumour rupture (7% of cases; P = 0.0014) and male gender had adverse influence on RFS. In a multivariable analysis mitotic count >5/50HPF, tumour size >5 cm and non-gastric location were independent adverse prognostic factors. Forty, 151, 86 and 348 patients were assigned according to the Joensuu criteria to very low, low, intermediate and high risk groups and had 5-year RFS of 94%, 94%, 86% and 29%, respectively. CONCLUSION: The Joensuu criteria, which include 4 prognostic factors (tumour size, site, mitotic count and rupture) and 3 categories for the mitotic count, were found to be a reliable tool for assessing prognosis of operable GIST. The Joensuu criteria identified particularly well high risk patients, who are likely the proper candidates for adjuvant therapy.