David A. Tong

Validation of the EINTHOVEN model-based computerized electrocardiogram rhythm analysis system with three classes of clinical arrhythmias.

Rhythm analysis by commercial systems does not meet clinical needs well, because (1) differential diagnosis of complex rhythms is not performed, (2) common rhythms are often misdiagnosed, and (3) transitions between rhythms are not described. We have developed a model-based diagnostic software system named EINTHOVEN that is designed to address the above limitations. A demonstration is available on the World Wide Web at http:@einthoven.uokhsc.edu. The system has been validated using simple rhythms from introductory electrocardiogram (ECG) textbooks. We present here the results of evaluation with more complex rhythm strips taken from clinical records and intermediate-level ECG textbooks. Rhythm strips were described by the onset and offset of each electrical event (P wave, QRS complex, and T wave) and by a morphology classification for each event. The rhythms included a variety of supraventricular and ventricular rhythms. The analysis was considered correct if it named all correct diagnoses in a rhythm strip, incorrect if it completed the analysis and failed to name the correct diagnoses, and indeterminate if it failed to complete the analysis. The system was designed not to complete an analysis if it could not explain an entire rhythm by at least 1 pathophysiological model. The test rhythms were not used to develop the system. Forty-six of 56 test rhythms were diagnosed correctly, and 8 were not analyzed completely. The 2 incorrect diagnoses were atrial tachycardia with variable conduction (diagnosed as intermittent complete heart block) and atrial fibrillation (diagnosed as irregular junctional tachycardia). All 56 rhythms were diagnosed correctly after minor technical improvements to the system. The processing time of the system was 7.6-fold (range 1.5-to 16.9-fold) faster than the elapsed time of the individual records. These preliminary results suggest (1) that computer-based interpretation of complex rhythms is possible, (2) that further software development is necessary to reach a clinical level of accuracy, and (3) that there are no theoretical obstacles to achieving this goal.

An algorithm for complete enumeration of the mechanisms of supraventricular tachycardias that use multiple atrioventricular, AV nodal, and/or Mahaim pathways

The EINTHOVEN system is a model-based expert system that interprets the cardiac rhythm from the electrocardiogram. It simulates the expected behavior of realistic semi-quantitative cardiac models constructed by heuristic rules to generate interpretations that include both text descriptions and event-by-event causal explanations in the form of ladder diagrams. The simulation has been limited by an inability to predict all possible behaviors of hearts with more than one reentrant circuit. We now describe an algorithm that overcomes this limitation. Its output has been validated by an independent possibility-tree analysis. Timing and storage measurements are presented for models with up to three slow atrioventricular nodal pathways, four atrioventricular pathways, and a single atriofascicular (Mahaim) pathway. This is the first report in the literature of an algorithm that enumerates all possible mechanisms for reentrant supraventricular tachycardias that use atrioventricular, atrioventricular nodal, and/or atriofascicular pathways in humans.

A semi-quantitative, three-dimensional model of cardiac electrophysiology

Catheter-mediated radiofrequency ablation is a medical procedure performed by highly trained and experienced cardiology sub-specialists. Even though these physicians are highly trained, the massive amount of data produced during these procedures creates a data overload problem that can impede the performance of even the best practitioners. Performance may be effected if the physician overlooks important signal features, misinterprets the signals, and/or misinterprets catheter locations in the heart, which may lead to increased procedure duration and/or applications of radiofrequency energy to the wrong part of the heart. To assist physicians performing catheter-mediated radiofrequency ablation procedures cope with the massive amount of data generated by the procedure, the authors began developing a computer-based system for analysing the signals generated by the procedure. As part of this effort, they have developed a semi-quantitative three-dimensional model of cardiac electrophysiology. This model has been implemented as part of the EINTHOVEN system, a computer-based system aimed interpreting the electrical signals from the heart. The authors incorporated the model into the EINTHOVEN system and developed and implemented new model-based algorithms for the system that together enable the system to interpret intracardiac electrograms in near real-time. The semi-quantitative three-dimensional model of cardiac electrophysiology and the model-based reasoning algorithms implemented to enable the EINTHOVEN system for analyzing intracardiac electrograms are presented.

Automated analysis of intracardiac electrograms obtained during extrastimulus tests using a three-dimensional electrophysiology model

Catheter ablation procedures are performed by highly trained and experienced cardiology subspecialists. Yet the massive amount of data produced during these procedures creates a data overload problem that can impede the performance of even the best practitioners. This may be evidenced by (1) overlooking important signal features, (2) misinterpreting the signals, and (3) misinterpreting catheter locations in the heart, all of which can lead to increased procedure duration, applications of radiofrequency energy to the wrong part of the heart, or both. This article presents the first results from a project aimed at developing a model-based system for interpreting intracardiac electrograms in near real time. The system is intended to assist physicians in interpreting the enormous amounts of data recorded during catheter ablation studies. It is an extension of the Einthoven system that has been extended to account for the three-dimensional relationships in the cardiac conduction system as recorded in the various intracardiac electrograms. The new three-dimensional cardiac conduction model and the enhancements to Einthovens reasoning algorithms are presented. The locus of this study is on interpreting the results of ventricular extrastimulus tests. Data collected for this study and the output generated by the system are presented.

1051-1 A Telemedicine Arrhythmia Analysis Tool for Rural Physicians

Introduction Rural health care practitioners are often faced with the task of interpreting complex heart rhythms. Usually these practitioners do not have specialized training in cardiology or ECG interpretation and available commercial systems for interpreting 12-lead ECG have been well-documented to perform rhythm analysis poorly. A computer-based tool is being developed in our laboratory that will provide rural health care practitioners with an automated system for interpreting complex arrhythmias. Methods A prototype system was developed using resources readily available in on the Internet: NCSA Mosaic, a hypermedia document browser; NCSA Collage, an interactive distributed white-board system; and GNU Ghostscript/Ghostview, a PostScript language interpretation system. The prototype was created by modifying and integrating the functionality of the individual systems into a single, easy to use system. The prototype was developed in Con UNIX workstations using the X display protocol and is based on the client-server model. The system may be accessed by any computer on the Internet with X display capability. On the client side, the system assists the user in scanning an ECG, identifying the waves in the scanned ECG, and transmitting the scanned ECG and annotations to the server. On the server side, the system analyzes the annotations using an knowledge-based rhythm analysis system being developed in our laboratory. The system produces a PostScript file containing the interpretation(s) and accompanying ladder diagrams and a case-specific help file containing detailed descriptions and therapeutic indications. These files are transmitted back to the users computer and displayed to the user by the client-side of the system. The design of the system provides for an “on-line,” interactive consultation with a cardiologist. Results Development of the prototype system required 2.5 man-months of effort to complete. The prototype system is undergoing alpha testing. For the prototype stage, all interpretations of clinical records are overread by a clinical cardiologist prior to transmission to the user. Conclusions This system may be beneficial in increasing the level of care of patients in the rural setting by providing the ECG interpretation expertise of an experienced cardiologist to rural practitioners on demand, which may also lead to decreasing the cost of rural health care. In addition, the system may be modified for applications in other health care domains.

901-36 Computer-based Interpretation of Intracardiac Electrograms: Identifying Rhythm Mechanisms from Ventricular Extrastimulus Tests

Introduction A computer-based system for analyzing complex cardiac rhythms is being developed in our laboratory. The system analyzes rhythms by adapting and simulating one or more semi-quantitative electrophysiologic models based on a sequence of events (P waves, QRS complexes, His bundle potential, etc.) on an event-by-event basis. The electrophysiologic model developed for the system can represent anatomies consisting of up to 4 accessory pathways, up to 3 AV nodal slow pathways, and 1 Mahaim pathway. We hypothesized that this system can (1) generate the correct electrophysiologic model(s) based on the observed events, (2) correctly differentiate between AV Nodal Reentrant Tachycardia (AVNRT), AV Reentrant Tachycardia (AVRT), and Atrial Tachycardia (ATACH) based on intracardiac electrogram data obtained during ventricular extrastimulus pacing protocols, and (3) generate ladder diagrams describing the identified rhythm mechanisms. Methods Typical cases were identified for each of the desired rhythms (AVNRT, AVRT, and ATACH). Using software written in our lab, the cases were annotated to extract wave timing and morphology from the body-surface electrocardiogram and wave timing of the high right atrial. His bundle, and right ventricular intracardiac electrograms. Due to the anatomic complexity underlying reentrant tachyarrhythmias, we developed novel algorithms for enumerating the complete set of conduction paths through complex electrophysiologic models. We also developed the logic necessary for reasoning with multiple model paths and simultaneous activation of anatomically distinct model regions. The system was developed in C on a UNIX workstation and runs on UNIX workstations or IBM PC-compatible computers. Results In all cases, the system (1) generated the correct set of electrophysiologic models, (2) eliminated or supported the appropriate models (i.e. rhythm mechanisms) based on the ventricular extrastimulus, and (3) generated ladder diagrams describing the rhythm mechanisms. Conclusions This is the first report of a computer-based system capable of deriving differential diagnoses with supporting ladder diagrams from intracardiac electrogram recordings. This demonstrates the potential for automatic analysis of complex rhythms by computer-assisted analysis of intracardiac electrograms. This system may be useful in the clinical electrophysiology laboratory by assisting physicians with the enormous amount of data generated during the typical electrophysiology study.