Artificial Neural Networks in Forecasting Cancer Therapy Methods and Costs of Cancer Patient Treatment. Case Study for Breast Cancer

Forecasting of cancer therapy and its cost is a key factor in the process of managing the oncological patient's treatment. This research is based on two predictive models of artificial neural networks (ANNs). The paper scope also includes a statistical analysis of correlation of the ANN variables. Research shows that ANNs are a viable alternative to traditional method of estimating the cost of a cancer therapy, especially in situations of poor recognition of the nature of medical relations between costs and their cost drivers, or in situations of non-linear, multidimensional relations between variables. Studies have also shown that ANN is an adequate method for predicting cancer therapy. The first contribution of this study is application of an innovative, complex approach based on two ANN models in the management of oncological patient treatment. The second contribution is the use of medical specifications required by the oncological patient directly to estimate the costs of cancer therapy.