Multifactorial analysis of survival in thyroid cancer. Pitfalls of applying the results of published studies to another population.

Multifactorial analyses (MFAs) of survival data of thyroid carcinoma patients have already been published by the European Organization for Research and Treatment of Cancer EORTC, the Institut Curie (IC), and the Institut Gustave Roussy (IGR). The aim of this study was to find out if the results obtained in these MFAs could be applied to a new population of 480 patients. To determine this, three MFAs were carried out on the new population. Each of these MFAs is based on exactly the same methodology as one of the three reference studies. Using the EORTC methodology, five significant factors were retained: age, principal cell type, anaplastic, T-category, and metastatic sites (The significant factors in the EORTC study also included the sex factor, which was not significant for the current study). Using their own prognostic index instead of the EORTC index, gave the authors a better classification of the patients and better survival prediction. Using the IC methodology, four significant factors were retained: age, histology, T-category, and metastatic sites (the significant factors in the IC study were age, histology and sex). Using the IGR methodology, two significant factors were retained: age and N-category (the significant factors in the IGR study also included sex and histology). The authors' results differ from those published by the previous groups. Therefore, one should hesitate before applying the results obtained by other groups for thyroid carcinoma to another population. It would seem necessary to carry out one's own MFA before determining a prognostic index.