Study The Risk Assessment Parameters Of Biological Potential Grading Criteria For Localized Gastrointestinal Stromal Tumors (GISTs)

Objective: To evaluate the implication of Fletcher and Miettinen biological potential grading criteria in native localized GISTs. Meanwhile, from histologic features, mutational types and immunophenotypes, to study the risk assessment parameters and provide new clue and theory for complement of the biological potential grading criteria.Methods: (1) Two hundred and twenty localized GISTs with complete clinicopathologic and follow-up date were evaluated their biologic potential by Fletcher and Miettinen grading criteria. The implications of the two grading criteria were compared by survival analysis. (2) The correlation between histologic features and clinical prognosis are studied by survival analysis in two hundred and twenty localized GISTs. (3) One hundred and sixty-eight localized GISTs were evaluated for the presence of c-kit and PDGFRA mutations by PCR amplication and direct sequencing. The correlation between mutational types and clinical prognosis are studied by survival analysis. (4) One hundred and sixty formalin-fixed paraffin-embedded localized GISTs were selected and Tissue microarry (TMA) block and section slides were prepared. Sections were immunohistochemically stained with antibodies against Ki-67, p16, p53 and CD44. The correlation between the protein expression and clinical prognosis are studied by survival analysis. (5) Cox's multivariate regression analysis were performed to evaluate the effect of all significant markers about histology, mutational types and immunophenotypes on prognosis in one hundred and fifty-four localized GISTs. (6) Cox's multivariate regression analysis were performed to evaluate the effect of all significant markers about histology, mutational types and immunophenotypes on prognosis in one hundred and three localized GISTs with c-kit exon 11 mutations.Results: (1) Evaluated by Fletcher grading criteria, the overall and disease-free survival rate of high risk GISTs were lower than that of very-low, low and intermediate GIST; while the overall and disease-free survival rate of very-low, low and intermediate risk GISTs had no statistical difference. In the high risk GISTs, the overall and disease-free survival rate of small intestinal and rectal GISTs was lower than that of gastric GISTs; while in the intermediate risk GISTs, the disease-free survival rate of small intestinal GISTs was lower than that of gastric GISTs. Evaluated by Miettinen grading criteria, the overall and disease-free survival rate of high risk GISTs were lower than that of very-low, low and intermediate GIST; while the overall and disease-free survival rate of very-low, low and intermediate risk GISTs had no statistical difference. In the risk subgroup of GISTs, the overall and disease-free survival rate of gastric, small intestinal and rectal GISTs had no statistical difference. (2) Univariate analysis demonstrated the following as poor prognostic histologic factors for overall and disease-free survival: high cellularity, nuclear pleomorphism, tumor necrosis, tumor invasion and perivascular collar-like pattern. (3) the overall and disease-free survival rate of the GISTs with c-kit exon 9 mutations was lower than that the GISTs with c-kit exon 11 mutations and wide-type GISTs. the overall and disease-free survival rate of the GISTs with deletions in c-kit exon 11 was lower than that the GISTs with duplications or missence in c-kit exon 11 mutations, the overall and disease-free survival rate of the GISTs with deletions affecting codons 557 to 558 in c-kit exon 11 was lower than that the GISTs without deletions affecting codons 557 to 558 in c-kit exon 11. (4) Univariate analysis demonstrated the following as poor prognostic immunophenotypes for overall and disease-free survival in the protein expression of Ki-67, p16, p53 and CD44: Ki-67 labeling index>5% and pl6 loss. (5) Multivariate analysis indicated that Miettinen high risk index, perivascular collar-like histologic pattern and p16 loss were the independent poor prognostic factor of localized GISTs. Multivariate analysis indicated that Miettinen high risk index, perivascular collar-like histologic pattern, pl6 loss and deletions affecting codons 557 to 558 in c-kit exon 11 were the independent poor prognostic factor of localized GISTs with c-kit exon 11 mutations.Conclusions: (1) Fletcher grading criteria is simple and easy to use, while Miettinen grading criteria for evaluating biological potential by anatomic site is more critical. (2) the histologic features of high cellularity, nuclear pleomorphism, tumor necrosis, tumor invasion and perivascular collar-like pattern are poor prognostic factors of localized GISTs. (3) The mutational types of c-kit gene in localized GISTs have effect on patient's prognosis. (4) Ki-67 labeling index>5% and pl6 loss are poor prognostic factors of localized GISTs. (5) perivascular collar-like histologic pattern, pl6 loss and mutational types of c-kit gene may be the risk parameters for the completement of Miettinen grading criteria and have important reference implication for the selection of high risk patients for targeted adjuvant treatment.