Validation of neural network analysis to predict prognosis in breast cancer patients

In order to analyze the possibility to utilize alternative techniques for statistical evaluation of biomedical data, the authors have applied neural network analysis for prediction of probability of relapse in a series of 101 node negative breast cancer patients treated by radical surgery without any further systemic therapy. On the basis of the authors' preliminary results, the neural model, based on Kohonen map, seems able to individualize specific clusters of patients with different clinical pathological characteristics and probability of relapse.<
>