A prognostic model that makes quantitative estimates of probability of relapse for breast cancer patients

Citation
M. De Laurentiis et al., A prognostic model that makes quantitative estimates of probability of relapse for breast cancer patients, CLIN CANC R, 5(12), 1999, pp. 4133-4139
Citations number
37
Categorie Soggetti
Oncology
Journal title
CLINICAL CANCER RESEARCH
ISSN journal
10780432 → ACNP
Volume
5
Issue
12
Year of publication
1999
Pages
4133 - 4139
Database
ISI
SICI code
1078-0432(199912)5:12<4133:APMTMQ>2.0.ZU;2-Z
Abstract
Tumor-node-metastasis (TNM) staging is the standard system for the estimati on of prognosis of breast cancer patients. However, this system does not ex ploit information yielded by markers of the biological aggressiveness of br east cancer and is clearly unsatisfactory for optimal-treatment decision-ma king and for patient counseling. We have developed a prognostic model, base d on a few routinely evaluated prognostic variables, that produces quantita tive estimates for risk of relapse of individual breast cancer patients. We used data concerning 2441 of 2990 consecutive breast cancer patients to de velop an artificial neural network (ANN) for the prediction of the probabil ity of relapse over 5 years. The prognostic variables used were: patient ag e, tumor size, number of axillary metastases, estrogen and progesterone rec eptor levels, S-phase fraction, and tumor ploidy, Performances of the model were evaluated in terms of discrimination ability and quantitative precisi on. Predictions were validated on an independent series of 310 patients fro m an institution in another country. The ANN discriminated patients accordi ng to their risk of relapse better than the TNM classification (P = 0.0015) , The quantitative precision of the model's estimates was accurate and was confirmed on the series from the second institution. The 5-year relapse ris k yielded by the model varied greatly within the same TNM class, particular ly for patients with four or more nodal metastases, The model discriminates prognosis better than the TNM classification and is able to identify patie nts with strikingly different risks of relapse within each TNM class.