Author:
Di Vizio Dolores,Demichelis Francesca,Simonetti Sara,Pettinato Guido,Terracciano Luigi,Tornillo Luigi,Freeman Michael R,Insabato Luigi
Abstract
Abstract
Background
Gastrointestinal stromal tumors (GIST) exhibit an unpredictable clinical course and can rapidly progress to lethality. Predictions about the biological behavior of GIST are based on a number of canonical clinical and pathologic parameters whose validity in distinguishing between a benign and a malignant tumour is still imperfect. The aim of our study was to investigate the role of morphologic parameters and expression of cells cycle regulators as prognosticators in GIST.
Methods
We performed an immunohistochemical analysis for Ki67, p27Kip1, Jab1, and Skp2, on a Tissue Microarray (TMA) containing 94 GIST. Expression of the above proteins was correlated to classically used prognosticators, as well as to risk groups. Clinical significance of histologic and immunohistochemical features were evaluated in 59 patients for whom follow-up information was available.
Results
Overexpression of Ki67 and Skp2, and p27Kip1 loss directly correlated with the high risk group (p = 0.03 for Ki67 and Skp2, p = 0.05 for p27Kip1). Jab1 expression did not exhibit correlation with risk. In 59 cases provided with clinical follow-up, high cellularity, presence of necrosis, and Ki67 overexpression were predictive of a reduced overall survival in a univariate model. The same parameters, as well as mitotic rate, tumour size, and p27Kip1 loss were indicative of a shortened relapse free survival interval. High cellularity, and high mitotic rate retained their prognostic significance by multivariate analysis.
Conclusion
Our data suggest that a number of histologic parameters in combination with immunohistochemical expression of cell cycle regulators can facilitate risk categorization and predict biologic behavior in GIST. Importantly this study demonstrates, for the first time, that Skp2 expression correlates with Ki67 expression and high risk in GIST.
Publisher
Springer Science and Business Media LLC
Subject
Cancer Research,Genetics,Oncology
Cited by
21 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献