THE POTENTIAL APPLICATION OF REGION MARKING SOFTWARE TOOL IN MEDICAL DIAGNOSTIC EDUCATION

Abstract

Subjective quality evaluation of medical images is mostly achieved using some version of ROC (Receiver Operating Characteristic) evaluation. The evaluation is conducted in two phases: data gathering and data analysis. In the data gathering phase observers perform diagnostic tasks by marking regions of interest. A region marking software tool is used for data gathering. Statistical data extracted from these markings are used for medical image evaluation in the data analysis phase of ROC evaluation. Medical image influence on diagnostic accuracy is evaluated by measuring whether the same results are achieved compared to some gold rule. In this paper we will demonstrate that a region marking software tool can be used as a software platform for education in medical image diagnostics. Specifically, diagnostic tasks are performed the same way in both ROC evaluation and medical image diagnostic education. The difference occurs in later stages. When the observer reaches the end of the ROC evaluation, hers/his answers are converted to ROC statistical data. The results are not reported to the observer. In medical image diagnostic education the observer’s results are compared to the gold standard, a report is generated, and the results are presented to the observer. It is possible to view each of the images evaluated with the observer’s marks compared to the gold standard. Therefore, a region marking software tool is a potential tool for education in medical diagnostic. We implemented a region marking software tool for ROC evaluation of medical images. We had several goals in mind: (1) it should resemble software tools used in everyday clinical tasks: (2) it should be accessible over distance; (3) it should be usable in education; and (4) it should be free. Therefore, we based our solution on web technologies, such as HTML, PHP, JavaScript, and MySQL. The region marking software tool for ROC evaluation of medical images has been used during our research on quality evaluation system of medical images in PACS systems (Picture Archiving and Communication System). We will present a prototype software tool built upon the region marking software tool to support educational use. Its foundations in web technologies make it an ideal tool for presence and distance education and training in medical image diagnostics.