Gastrointestinal stromal tumors (GIST): a prospective evaluation of risk factors and prognostic scores

Many risk scores for malignancy in gastrointestinal stromal tumors (GISTs) are based on the combination of size, mitotic index, and location in order to predict recurrence, appearance of metastases, or survival.

This is a prospective analysis of prognosis factors (size, mitotic index, Ki-67, and others) and malignancy risk scores (Fletcher's, modified NIH, Miettinen, NCCN, and A or B Goh's scores). This is a study of the sensitivity and specificity of the different malignancy risk scores in the prognosis of recurrence and survival.

We operated on 52 patients from 1 January 2002 to 10 January 2008. Mean follow-up was 35 months. We used SPSS 13.0 for Windows for the statistical analysis. The differences between frequencies and means were calculated using the chi-square test, Student's t test, and analysis of variance. Kaplan-Meier and the Cox multiple regression methods were used to calculate overall and disease-free survival. Sensitivity and specificity were depicted graphically as receiver operating characteristic (ROC) curves.

Thirteen patients suffered recurrences (27.7%). Eleven died from disease-related causes. Two-year and 5-year actuarial survival was 80% and 76%, respectively, and the mean survival was 60 months (CI 95% 51.9-69.5). Two-year and 5-year actuarial disease-free survival was 76% and 72% respectively, and the mean disease-free survival was 52 months (CI 95% 43.5-61.1). The score proposed by Fletcher has the greatest sensitivity and specificity in our series to predict overall survival (AUROC = 0.761; SE 0.084, p = 0.009; CI 95% 0.597-0.926) and recurrence (AUROC RECURRENCE = 0.693; SE 0.087, p = 0.042; CI 95% 0.523-0.864) compared with other scores.