A multifactorial prognostic model for adult soft tissue sarcoma considering clinical, histopathological and molecular data.

Soft tissue sarcomas (STS) are malignant mesenchymal lesions with a high degree of prognostic variability. Different prognostic markers such as grading, staging, tumour type and localisation are known. The establishment of these markers was based on the evaluation at results of extensive cohorts of patients. Therefore, only the established markers provide us with information about probabilities in relation to other qualities. Considering as many different markers as possible in one prognostic statement should increase the value of the resultant information. Therefore, we developed a model involving known prognostic markers to formulate an individual prognostic index. In a retrospective analysis, different prognostic factors of 198 adult STS patients with histological tumour free resection margins were evaluated using a multifactorial analysis. On the basis of a Cox-Regression-Model with proportional hazards, the prognostic factors (tumour type, staging, localisation and type of surgical resection) were selected using previous knowledge and a statistical step backward selection procedure adjusting the immunohistochemical status of p53/Mdm2 expression. On the basis of the baseline survival function of our cohort (S 0 (t)), the cumulative probability of survival for two S (2) and five S (5) years was estimated. As a result of our analysis the equations S (2) = (e -00393 ) P and S (5) = (e -00869 ) P can be used to estimate the individual two and five-year probability of survival in our cohort. Here p is the result of the amount of the estimated regression- coefficients of the exact variables of the respective individual patient. This model makes it possible to include all the evaluated prognostic factors which, in turn, increases the accuracy of the prognostic information for individual patients underlining the proportional hazards assumption.