Back to Search Start Over

Immunohistochemical Analysis and Prognostic Value of Cathepsin D Determination in Laryngeal Squamous Cell Carcinoma

Authors :
Seiwerth, S.
Stambuk, N.
Konjevoda, P.
Masic, N.
Vasilj, A.
Bura, M.
Klapan, I.
Manojlovic, S.
Danic, D.
Source :
Journal of Chemical Information and Computer Sciences (now called Journal of Chemical Information and Modeling); May 22, 2000, Vol. 40 Issue: 3 p545-549, 5p
Publication Year :
2000

Abstract

Cathepsin D, a protease with the capability of degrading matrix proteins, is implicated in the process of breast and colorectal cancer invasion and metastasis. Biochemical studies in laryngeal cancer have shown a potential prognostic significance of cathepsin D content determination. We studied immunohistochemical positivity of cathepsin D in tumor epithelium and stroma of 61 surgical specimens of squamous cell laryngeal cancer. Immunohistochemical reaction was quantitatively assessed using a PC-based image analysis system SFORM-VAMS. The results were correlated to clinical and morphological parameters and survival. Immunohistochemical positivity was noted in neoplastic cells and tumor stroma. Significant prognostic value for cathepsin D was established separately for epithelial tumor component and tumor stroma using log-rank test, the Cox proportional hazards regression model, and C4.5 machine learning system. In all groups, patients above the median cathepsin D staining showed significantly shorter survival time. C4.5 machine learning system extracted cutoff values for the decision tree that defines the probabilities of patients survival and death with high sensitivity (92.8% alive, 73.6% dead), 100% specificity, and 86.9% accuracy. This makes immunohistochemical cathepsin D estimation an independent prognostic parameter in laryngeal carcinomas within a 5-year period from the time of tumor surgery.

Details

Language :
English
ISSN :
00952338 and 15205142
Volume :
40
Issue :
3
Database :
Supplemental Index
Journal :
Journal of Chemical Information and Computer Sciences (now called Journal of Chemical Information and Modeling)
Publication Type :
Periodical
Accession number :
ejs1108082