Prognostic significance of factors affecting disease free interval and overall survival for Stage II breast cancer in Greece - A multivariate cohort study

Background: Univariate analysis evaluates the impact of a prognostic factor on survival rates, either disease free: (DFI) or overall (OS). Since many of the factors are interrelated, it is difficult to predict the prognosis o f an individual patient. Multivariate analysis is therefore required in ord er to allow factors act together thus ending in the best possible combined predicting result. Methods: A step-up procedure (Cox's Proportional Hazards Regression model) was used to include various prognostic parameters, relat ing to patients themselves, to the pathology of their tumours and to the tr eatment schedule followed. Two hundred and sixty-nine Stage II breast cance r Greek patients, treated from 1981 until 1991 and with a median 12-year of follow-up are studied. Results: five factors were found to be significant for patients DFI. in order of relevant importance, these were the number of infiltrated nodes, tumour size, postoperative radiotherapy, adjuvant chemo therapy and patients age. Regarding patients OS, tumour size, number of pos itive nodes, patients' age at entry and ER/PR status were the most importan t ones. Conclusion: Our long-term (12-year), single institution, single are a results, suggest that, the prognostic factors for patients DFI and OS are the same with those of series from Europe and USA. Additionally, they rema in unchanged after long-term follow-up, compared to a previously reported s hort-term national-wide study from this country. (C) 2001 Elsevier Science ireland Ltd. All rights reserved.