Premium
Cox multivariate regression models for estimating prognosis of patients with endometrioid adenocarcinoma of the uterine corpus who underwent thorough surgical staging
Author(s) -
Nishiya Masashi,
Sakuragi Noriaki,
Hareyama Hitoshi,
Ebina Yasuhiko,
Furuya Mitsuko,
Oikawa Mamoru,
Yamamoto Ritsu,
Fujino Takafumi,
Fujimoto Seiichiro
Publication year - 1998
Publication title -
international journal of cancer
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.475
H-Index - 234
eISSN - 1097-0215
pISSN - 0020-7136
DOI - 10.1002/(sici)1097-0215(19981023)79:5<521::aid-ijc13>3.0.co;2-4
Subject(s) - medicine , proportional hazards model , multivariate statistics , adenocarcinoma , multivariate analysis , oncology , gynecology , general surgery , cancer , computer science , machine learning
The International Federation of Gynecology and Obstetrics (FIGO) adopted surgical staging criteria in 1988. Many studies have shown that histologic grade, nuclear grade, lymph‐vascular space invasion and cell type are also important predictors of survival. It has not been clarified, however, how to integrate these histopathologic variables into the process of estimating individual prognosis. We performed Cox multivariate regression analysis to create models that incorporate various histopathologic factors for estimating the prognoses of patients with endometrioid adenocarcinoma of the uterine corpus. Our study was based on data from 206 patients who underwent complete surgical staging, including systematic pelvic and para‐aortic lymph node dissection. Two models resulted: one included depth of myometrial invasion, para‐aortic node metastasis and the number of sites involved by the tumor among the cervix, ovary and pelvic lymph nodes (which we designated as extracorporeal spread score, ECS) and the other incorporated nuclear grade and lymph‐vascular space invasion as variables. These 2 models enabled the prognosis for patients with endometrioid adenocarcinoma to be stratified into several levels according to hazard ratio. Comprehensive integration of the histopathologic prognostic factors, categorized into those relating to tumor extent and those relating to tumor virulence, should facilitate the estimation of individual prognosis more accurately than FIGO staging alone. Int. J. Cancer (Pred. Oncol.) 79:521–525, 1998.© 1998 Wiley‐Liss, Inc.