An Intelligent "Virtual Signout" System for Pathology Training: An Efficient and Objective Approach Using Whole Slide Images
J Zhang, B Yang, C Johns, A Kulkarni, C Handorf. The University of Tennessee Health Science Center, Memphis, TN
Background: There are several challenges in pathology residency training faced by both the programs and the trainees. These include: i) Programs and the board requirements focusing on total time of exposure and not necessarily the quality or duration of exposure to different entities and diagnoses. ii) The case exposure is not systematically structured to achieve efficient learning in all the desired areas. iii) Lack of an objective method for competency evaluation.
Design: To address these issues, we are developing a web based software solution that, backed by a large database of digitized images from archived cases, will simulate resident learning process. Case exposure is individualized and designed to avoid redundancy based on each user's diagnostic skills for multiple disease conditions. User performance is recorded and analyzed to provide objective, real-world evaluation of competency using different measures such as accuracy for diagnosis and ancillary studies.
Results: 134,000 cases from the recent archives at the University of Tennessee Health Science Center Pathology Department are scrutinized. pathology cases representing a wide variety of diagnosis are selected. Relevant case information, including gender, age, clinical history, specimen type and gross descriptions are extracted using scripting language PERL. Slides are digitized using Scascope®XT at 20X magnification. Website and database backend are constructed using Microsoft Visual Studio.
Conclusions: Initial user feedbacks suggest that the efficiency of pathology training can be improved by using the power of information technology.
Tuesday, March 23, 2010 9:30 AM
Poster Session III # 104, Tuesday Morning