Bayesian Network Based Classification of Mammography Structured Reports

In modern medical domain, documents are created directly in electronic form and stored on huge databases containing documents, text in integral form and images. Retrieving right informations from these servers is challenging and, sometimes, this is very time consuming. Current medical technology do not provide a smart methodology classification of such documents based on their content. In this work the radiological structured reports are analysed classified and assigning an appropriate label. The text classifier is used to label a mammographic structured report. The experimental data are real clinical report coming from a hospital server. Analysing the structured report content, the classifier labels the patient structured report as healthy or pathological. The present work uses Information Retrieval techniques to improve the classification process. These technique provide a light semantic analysis to remove negative terms, a removing stop-word step and, finally, a thesaurus is used to uniform used words. The structured reports are classified using a Bayes Naive Classifier. The experimental results provide interesting performance in terms of specificity and sensibility. Others two indexes are computed in order to assess system's robustness: these are the A(z) (Area under Curve ROC) and sigma(Az) (A(z) standard error).