ABSTRACT
This chapter demonstrates the ability of artificial neural networks (ANNs) to classify patients and to identify significant input vectors. ANNs have been applied to the difficult task of preoperative diagnosis of Renal cell carcinoma (RCC). In addition, ANNs have been utilised to examine outcomes data and quality of life issues which are becoming increasingly important in the choice of treatment. In most cases of RCC, metastases are related to primary tumour size; in a small portion, however, metastases are identified regardless of primary tumour size leading many surgeons to be more aggressive, resecting tumours as small as 1.5 cm routinely. Urologic malignancies, which include renal, testis, bladder, and prostate cancers, are a challenging heterogeneous group of malignancies. Research assessing the usefulness of neural networks in urologic oncology, although in its infancy, has demonstrated high accuracy in assessing characteristics of nonlinear datasets, identifying patterns, and generalising what it has “learned” to new datasets.
