Today's cardiac imaging field requires diagnosticians to master an ever-expanding knowledge base (KB) while the time to master this knowledge, apply it to specific tasks and reimbursement are steadily shrinking. These constraints pose a serious healthcare problem that inevitably leads to physician's errors. Thus new tools are required to assist physicians to timely apply comprehensive, up-to-date objective knowledge and the available patient data to specific clinical problems. The long-term objective of our Fast-Track proposal is to improve the care of cardiac patients and reduce the cost of cardiac image interpretation by developing new tools for a WEB-accessible cardiac toolbox that provides decision support to increase the accuracy of detecting coronary heart disease (CHD). Specifically, we propose a WEB-based system where acquired ECG-gated myocardial perfusion SPECT (MPS) raw images are uploaded to be automatically reconstructed and analyzed to extract regional quantitative parameters of myocardial perfusion and function. These parameters are converted to certainty factors of abnormality and submitted to an imaging decision support system (DSS) that is continuously updated with the latest scientific/clinical knowledge to reach an impression of the patient's heart status. These conclusions reached by the DSS and justifications for each conclusion are used to automatically generate a web-based structured report for the diagnostician to easily review, learn from the justifications, and either modify and/or approve for optimal accuracy of the diagnosis and prognosis of CHD. Specifically we propose to: 1) develop a novel left ventricular (LV) quantitative algorithm that automatically extracts parameters of left LV regional perfusion and function used to diagnose CHD;2) develop the LV expert (LVX) DSS and 3) design and implement the LV quantification and DSS algorithms using the .net platform so that they can be integrated into our Syntermed infrastructure and deployed over the web and/or used as conventional stand- alone work-stations. In Phase I we will develop a proof-of-principle system where LV perfusion information from MPS studies is analyzed and DSS interpreted for automatic report generation and physician review. In Phase II, the system will be: a) extended to include quantification and DSS of myocardial function, ischemia, viability and clinical risk factors, b) extended to include a methodology to continuously update LVX's KB, c) automated to link all the reconstruction, processing, quantification, interpretation, and reporting applications, and d) deployed in .net on the web with database and eCommerce accounting capability. Using this process we expect to confirm our primary hypothesis that diagnosticians using our decision support will provide a faster, more accurate diagnosis and prognosis of CHD than those provided by the same diagnosticians without the aid of this system. The system will be commercialized using Syntermed's successful strategy of other Emory software through: 1) licensing to major instrumentation manufacturers, 2) direct sales to clients that use PC workstations and 3) per WEB-access fee using the existing Syntermed Live network.
Physicians are required to master an ever-expanding knowledge base and take into account an ever increasing amount of patient-specific clinical information while time available to master this knowledge base and apply it to specific tasks is steadily shrinking. There is also an increasing shortage of cardiac diagnosticians [Fye04] who primarily interpret nuclear cardiology studies and an ever increasing number of aging """"""""Baby Boomers"""""""" who are becoming patients [Kni02]. This project is to develop software tools that will use the latest pertinent clinical and imaging knowledge from the medical literature and domain experts and make it WEB-available to physicians to support their medical decisions so they can make faster and more accurate diagnosis and avoid misdiagnosis and patient mismanagement.
|Garcia, Ernest V; Klein, J Larry; Moncayo, Valeria et al. (2018) Diagnostic performance of an artificial intelligence-driven cardiac-structured reporting system for myocardial perfusion SPECT imaging. J Nucl Cardiol :|
|Garcia, Ernest V; Klein, J Larry; Taylor, Andrew T (2014) Clinical decision support systems in myocardial perfusion imaging. J Nucl Cardiol 21:427-39; quiz 440|