Combining Artificial Neural Networks and Transrectal Ultrasound in the Diagnosis of Prostate Cancer
October 01, 2003
Arguably the most important step in the prognosis of prostate canceris early diagnosis. More than 1 million transrectal ultrasound (TRUS)-guided prostate needle biopsies are performed annually in the UnitedStates, resulting in the detection of 200,000 new cases per year. Unfortunately,the urologist's ability to diagnose prostate cancer has not keptpace with therapeutic advances; currently, many men are facing theneed for prostate biopsy with the likelihood that the result will beinconclusive. This paper will focus on the tools available to assist theclinician in predicting the outcome of the prostate needle biopsy. We willexamine the use of "machine learning" models (artificial intelligence),in the form of artificial neural networks (ANNs), to predict prostatebiopsy outcomes using prebiopsy variables. Currently, six validatedpredictive models are available. Of these, five are machine learningmodels, and one is based on logistic regression. The role of ANNs inproviding valuable predictive models to be used in conjunction withTRUS appears promising. In the few studies that have comparedmachine learning to traditional statistical methods, ANN and logisticregression appear to function equivalently when predicting biopsyoutcome. With the introduction of more complex prebiopsy variables,ANNs are in a commanding position for use in predictive models. Easyand immediate physician access to these models will be imperative iftheir full potential is to be realized.