Concurrent optimization of timing delays and electrode positioning in biventricular pacing based on a computer heart...

The efficacy of cardiac resynchronization therapy through biventricular pacing (BVP) has been demonstrated by numerous studies in patients suffering from congestive heart failure. In order to achieve a guideline for optimal treatment with BVP devices, an automated non-invasive strategy based on a computer model of the heart is presented.

The presented research investigates an off-line optimization algorithm regarding electrode positioning and timing delays. The efficacy of the algorithm is demonstrated in four patients suffering from left bundle branch block (LBBB) and myocardial infarction (MI). The computer model of the heart was used to simulate the LBBB in addition to several MI allocations according to the different left ventricular subdivisions introduced by the American Heart Association. Furthermore, simulations with reduced interventricular conduction velocity were performed in order to model interventricular excitation conduction delay. More than 800,000 simulations were carried out by adjusting a variety of 121 pairs of atrioventricular and interventricular delays and 36 different electrode positioning set-ups. Additionally, three different conduction velocities were examined. The optimization measures included the minimum root mean square error (E(RMS)) between physiological, pathological and therapeutic excitation, and also the difference of QRS-complex duration. Both of these measures were computed automatically.

Depending on the patient's pathology and conduction velocity, a reduction of E(RMS) between physiological and therapeutic excitation could be reached. For each patient and pathology, an optimal pacing electrode pair was determined. The results demonstrated the importance of an individual adjustment of BVP parameters to the patient's anatomy and pathology.

This work proposes a novel non-invasive optimization algorithm to find the best electrode positioning sites and timing delays for BVP in patients with LBBB and MI. This algorithm can be used to plan an optimal therapy for an individual patient.