Back to the start: Evaluation of prognostic markers in gastrointestinal stromal tumors

Abstract

The aim of this study was to provide a standardized risk stratification model for gastrointestinal stromal tumors (GISTs) based on tumor localization, tumor size, involved lymph nodes and metastases, as well as mitotic activity and other morphological and molecular markers, in order to improve the risk evaluation scheme for recurrence, metastatic spread and survival for patients with GIST. A total of 201 cases of patients with GIST were investigated according to standardized morphological markers, including nuclear pleomorphism, tumor cell necrosis, mucosal infiltration, ulceration, skeinoid fibers and growth pattern. In addition, all cases were immunohistochemically analyzed using a tissue microarray platform for various markers of differentiation (CD34, CD44, CD117, desmin, discovered on GIST 1, platelet-derived growth factor receptor alpha, S-100 and smooth muscle actin) and proliferation (B-cell lymphoma 2, P16, P53, phosphohistone H3 and Ki-67). These findings were correlated by uni-and multivariable analyses with clinicopathological characteristics, including recurrence, metastasis and survival. The general clinicopathological parameters of this GIST specimen cohort were comparable to previous studies. While several parameters exhibited clear associations to each other and to the defined clinical endpoints, the multivariate analysis reduced the number of relevant prognostic variables to localization, margin status, growth pattern and hematoxylin and eosin-based mitosis/Ki-67-based proliferation of GISTs. With the exception of CD34, none of the applied markers of differentiation and proliferation were found to be independent prognostic markers in GIST and the classical risk factors of GIST remain important prognostic factors. Additionally, growth pattern may predict the risk of recurrence and metastasis in GIST patients. Additional independent molecular prognostic markers remain to be identified and validated.Copyright © 2016, Spandidos Publications. All rights reserved.