The pressures imposed by a rapidly expanding knowledge base and escalating time constraints create a dilemma for all physicians and, particularly, for diagnostic radiologists. How can diagnostic radiologists consistently provide quality diagnostic interpretations? This dilemma becomes particularly acute for low volume studies such as Tc-99m MAG3 renography where many radiologists have had limited training and experience. Our long-term objectives are to assist diagnostic radiologists, enhance the care of nephrourology patients, reduce interpretative errors and provide a template for the development and implementation of heuristic and predictive Decision Support Systems (DSS) in radiologic diagnosis. In particular, we propose to complete the development of a comprehensive and fully informed DSS (iDSS) that will analyze MAG3 renal scans, automatically identify and correct quality control problems, and incorporate clinical data to interpret MAG3 renograms in regard to renal function, obstruction and renovascular hypertension as well as experts with full access to clinical information. We expect to confirm our hypothesis the iDSS will improve the performance of general radiologists compared to their perfomance without iDSS assistance.
These dual heuristic (RENEX) and statistical (STATEX-2) approaches complement and inform each other to maximize performance; the resulting iDSS will define the essential interpretative criteria, foster standardized interpretation and, importantly, provide the rules and supporting data justifying a specific iDSS interpretation. This last feature will allow radiologists and trainees to query the iDSS to learn the basis for a particular interpretation; this process will educate trainees and radiologists and provide them with the knowledge and skills required to interpret renal scans as competently as experts.
Although our specific application is development of a decision support system to interpret renal scans, the knowledge based and predictive methodology we are developing can be broadly applied to other areas of diagnostic imaging and medical care. Computer assisted diagnosis and computer assisted patient management will have to play a critical role in the future if the country is to control future health care costs and simultaneously maintain a high standard of care.
This web page summarizes and demonstrates our methodology for acquiring and analyzing quantitative nuclear renography studies, and for using RENEX to interpret diuresis renography studies. Our work also includes innovative statistical approaches; these are described in the publications and a more complete description will be added to these pages in the near future.
The following sections are accessed by the links in the table at the upper left.