Presence of Mid-Myocardial Fibrosis in Ventricular Endocardial Electrograms is Related to Power of Patient-Specific Frequency Bands.

Ventricular tachycardia (VT) is an important cause of morbidity and mortality in patients with structural heart disease. During catheter ablation procedures to control VT, electrograms (EGMs) are acquired from the endocardial surface of the heart. Standard-of-care EGM amplitudes can fail to identify important fibrotic substrate responsible for VT. Frequencydomain information has been shown to improve prediction of presence of fibrosis within patients. Our objective was to explore what specific frequency characteristics enable improved prediction of fibrotic regions and whether these characteristics were generalizable to larger patient populations. In nonischemic cardiomyopathy patients undergoing VT ablation, we used registered preprocedural cardiac computed tomography (CCT) and late iodine enhancement (LIE) imaging for indication of presence of fibrosis and multitaper spectral analysis for estimation of the frequency spectra. We find that total unipolar and bipolar spectral power contains comparable information about presence of LIE-CCT-defined fibrosis to that of standard-of-care voltage metrics. Additionally, we find that the shape of the frequency domain contains useful information for tissue characterization within patients, but, thus far, does not have consistent characteristics across patients. This work elucidates characteristics of the frequency domain that have contributed to the improved tissue characterization resultant from its use.