Key Points

QuestionÌý Can contrast-enhanced cardiac computed tomography accurately localize the atrioventricular conduction system by identification of the fat that insulates the conductive myocardium?

FindingsÌý In a cohort of 20 patients referred for atrial fibrillation ablation, myocardial low attenuation on contrast-enhanced computed tomography (CECT) indicating fat deposition at the top of the triangle of Koch was within the registration error range of the intracardiac His signal in all cases.

MeaningÌý The atrioventricular conduction system is insulated from surrounding atrial and ventricular myocardium by fat, and results suggest that the fat can be accurately localized using CECT.

ImportanceÌý Noninvasive localization of the compact atrioventricular node and the proximal specialized conduction system (AVCS) would enhance planning for transcatheter aortic valve and complex or congenital heart disease surgical procedures.

ObjectiveÌý To test the hypothesis that preprocedure contrast-enhanced cardiac computed tomography (CECT) can accurately localize the AVCS by identification of the fat that insulates the conductive myocardium.

Design, Setting, and ParticipantsÌý This was a prospective cohort study that took place at an academic tertiary care center. Included in the study were patients with CECT acquired less than 1 month before atrial fibrillation ablation and electroanatomic localization of the His electrogram signal on electroanatomic mapping (EAM) between January 2022 and January 2023.

ExposuresÌý Preprocedure CECT.

Main Outcomes and MeasuresÌý The distance from the His electrogram signal to the fat segmentation encompassing the AVCS on CECT, after registration of the images to EAM.

ResultsÌý Among 20 patients (mean [SD] age, 66 [10] years; 15 male [75%]) in the cohort, the mean (SD) attenuation of the AVCS fat segmentation was 2.9 (21.5) Hounsfield units. The mean (SD) distance from the His electrogram to the closest AVCS fat voxel was 3.3 (1.6) mm.

Conclusions and RelevanceÌý Results of this cohort study suggest that CECT could accurately localize the fatty tissue that insulates the AVCS from surrounding atrial and ventricular myocardium and may enhance the efficacy and safety of procedures targeting the conduction system and structures in its proximity.

The atrioventricular cardiac conduction system (AVCS) contains myocytes responsible for transmitting excitatory impulses from the atria to the ventricles.¹ The AVCS is histologically recognized as myocytes insulated from adjacent myocardium by fibrofatty tissue,² with fat deposition demonstrated along the compact atrioventricular node and His bundle,³ once it becomes insulated from the atrial myocardium, within the septum.⁴ The superior extension of the inferior atrioventricular groove is sandwiched between the right atrial wall of the triangle of Koch and the crest of the muscular ventricular septum and can be visualized on contrast enhanced cardiac computed tomography (CECT),⁵ a technique that also identifies myocardial fat with high specificity.⁶ The AVCS penetrates the atrioventricular myocardial discontinuity either at the apex of the inferior pyramidal space or lower within the inferoseptal recess of the left ventricular outflow tract.⁷ We sought to test the hypothesis that CECT can accurately localize the penetrating portion of the AVCS, ie, the His bundle, in a cohort of patients undergoing atrial fibrillation (AF) ablation.

Between January 2022 and January 2023, we prospectively identified patients with CECT acquired less than 1 month before AF ablation at the Hospital of the University of Pennsylvania. To enhance clarity regarding data generalizability, patient race (African American, White, or undeclared) was included in study data. The Hospital of the University of Pennsylvania atrial fibrillation (HUP-AF) ablation registry is a prospectively collected and abstracted database that enrolls all patients undergoing AF ablation at HUP. Collection of consent-exempt registry data was approved by the HUP institutional review board. The research adhered to Helsinki Declaration guidelines.

CECT was performed with a 2 × 192-detector scanner (Somatom Force [Siemens]; collimation = 512 × 0.6 mm, gantry rotation time = 610 milliseconds, in-plane resolution, 0.45 × 0.45 × 0.45 mm, prospectively or retrospectively triggered at 70% of the RR interval [ie, the time elapsed between 2 successive R waves of the QRS signal], 120-kV tube voltage). Heart rate was maintained at less than 70 beats per minute by intravenous metoprolol. Iodinated intravenous contrast (100 mL, rate of 4-5 mL per second, Isovue-370 [Bracco Diagnostics]) was injected and timed to optimize myocardial perfusion. Image reconstruction was performed at 70% of the RR using soft-tissue convolution kernel.

Fat density, ie, very low attenuation (<30 Hounsfield units [HU]), in the inferior pyramidal space on axial images was identified (Figure 1). Adjacent musculature of the right atrium, left atrium, right ventricle, and left ventricle were identified and contoured. The top of the triangle of Koch, bounded posteriorly by the tendon of Todaro (ie, line connecting the anterior inferior vena cava ostium to the superior septal aspect of the tricuspid valve), anteriorly by the septal leaflet of the tricuspid valve, and inferiorly by the coronary sinus, was interrogated for low attenuation representing AVCS fibrofatty deposition connecting the atrial and myocardial contours. The search area within the triangle of Koch was focused on the tissue near the apex of the inferior pyramidal space and/or the region beneath the right coronary cusp of the aorta. Image analyses were performed by ADAS 3D (ADAS 3D Medical SL).

Our mapping and AF ablation techniques have been previously described.⁸ Electroanatomic maps were created with the Carto 3 system (Biosense Webster) using multipolar PentaRay (20 electrodes with 2-6-2–mm spacing) catheters during distal CS pacing. The location of the His electrogram was marked on the EAM using Thermocool Smarttouch SF (Biosense Webster) catheter with 5g or greater force to confirm a contact signal. Pulmonary vein (PV) isolation was performed with confirmation of bidirectional PV entrance and exit block. Non-PV triggers were targeted when identified.

The cardiac anatomy CECT image segmentation was registered to the EAM generated from the clinical procedure using the PV ostia as landmarks. After the procedure, the Euclidean distance, ie, the length of the line segment between the His electrogram signal and the closest AVCS fat segmentation voxel in 3-dimensional space, was measured using ADAS 3D software. If more than 1 His electrogram signal location was tagged, the mean distance for all His locations in that patient was used. Continuous variables are expressed as mean and SD and categorical variables as count and percentage. The independent-samples t test, Fisher exact test, or χ² test was used to compare differences across groups, as appropriate. All P values were 2-sided, and a P value <.05 was considered statistically significant. Statistical analyses were performed using Stata, version 18 SE (StataCorp).

The cohort included 20 patients (mean [SD] age, 66 [10] years; 5 female [25%]; 15 male [75%]; 3 African American [15%]; 15 White [75%]; 2 undeclared [10%]). Of all patients, 17 (85%) had hypertension, and 6 (30%) had persistent AF. At baseline, the mean (SD) PR interval during sinus rhythm was 160 (24) milliseconds. Coronary artery disease was prevalent in 8 patients (40%), and history of first- or second-degree AV block was noted among 5 patients (25%). Patient characteristics are summarized in the Table.

On preprocedural CECT, the mean (SD) left atrial volume index was 32.7 (8.0) mL/m², and the mean (SD) echocardiographic left ventricular ejection fraction was 54.0% (9.9%).

Low attenuation was noted at the apex of the triangle of Koch in all patients (Figure 2 and eFigure 1 in Supplement 1). The mean (SD) attenuation of the AVCS fat segmentation was 2.9 (21.5) HU. On EAM, the mean (SD) distance from the His electrogram to the closest AVCS fat voxel was 3.3 (1.6) mm (eFigure 2 in Supplement 1). There was no association between the HU intensity or volume of the AVCS-fat on CT and the PR interval.

This cohort study demonstrates the ability of widely available conventional CECT to accurately localize the AVCS validated by invasive electroanatomic mapping.

Shinohara et al⁹ previously demonstrated the feasibility of AVCS (including the compact node, His bundle, and both left and right bundle branches) visualization as a low-density structure on phase-contrast CT in ex vivo whole-heart human specimens, validated against histological analysis. Micro-CECT has also been used to visualize the AVCS in nonhuman mammalian hearts.¹⁰ Our study builds on these findings by demonstrating the ability of CECT to accurately delineate the AVCS, with simple image processing techniques that are easy to replicate, and does not necessitate specialized image acquisition protocols or contrast agents.

This study has some limitations. This was a single-center study with a small sample size and a homogeneous population of patients with AF without congenital heart disease, which reduces the generalizability of our results. Although attenuation less than 30 HU is specific for fat on CECT, the AVCS segmentation, intended to identify the fat surrounding the conduction system, is expected to contain not only the specialized conducting myocardium but also some fibrous tissue due to spatiotemporal resolution limitations and volume averaging. Future studies are needed to validate our findings, particularly when the expected AVCS anatomy is congenitally altered. A prior study has implicated excessive fatty infiltration in the compact AV node transitional area in conduction disturbances.¹¹ In this study with a small sample size, we did not find an association between HU intensity or fat volume and PR interval or first- or second-degree AV block.

In conclusion, results of this cohort study suggest that CECT accurately localized the AVCS compared with invasive electroanatomic mapping. Noninvasive visualization of the AVCS may provide utility in planning for electrophysiologic, interventional, and surgical procedures by providing an anatomic roadmap. This may be potentially useful in ablation of para-Hisian atrial and ventricular arrhythmias or anterior and midseptal accessory pathways and transcatheter valve replacements. If the findings can be validated in patients with congenital heart disease, CECT may lower the incidence of iatrogenic conduction system injury during congenital heart surgery.

Accepted for Publication: May 24, 2024.

Published Online: July 24, 2024. doi:10.1001/jamacardio.2024.2012

Corresponding Author: Saman Nazarian, MD, PhD, Division of Medicine, Section of Cardiac Electrophysiology, Hospital of the University of Pennsylvania Pavilion, Second Floor City Side, Office 6, One Convention Ave, Philadelphia, PA 19104 (saman.nazarian@pennmedicine.upenn.edu).

Author Contributions: Dr Nazarian had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Pourmousavi Khoshknab, Markman, Mavroudis, Desjardins, Nazarian.

Acquisition, analysis, or interpretation of data: Pourmousavi Khoshknab, Zghaib, Xu, Markman, Desjardins.

Drafting of the manuscript: Pourmousavi Khoshknab, Desjardins.

Critical review of the manuscript for important intellectual content: All authors.

Statistical analysis: Pourmousavi Khoshknab, Xu, Desjardins.

Obtained funding: Pourmousavi Khoshknab.

Administrative, technical, or material support: Pourmousavi Khoshknab, Mavroudis, Nazarian.

Supervision: Pourmousavi Khoshknab, Markman, Desjardins, Nazarian.

Conflict of Interest Disclosures: Dr Xu reported receiving grants from the American Heart Association during the conduct of the study. Dr Markman reported receiving personal fees from Biosense Webster, Abbott, Medtronic, and Boston Scientific outside the submitted work. Dr Nazarian reported receiving nonfinancial support from ADAS Software during the conduct of the study; grants from Biosense Webster, ADAS Software, and the National Institutes of Health/National Heart, Lung, and Blood Institute; and personal fees from Biosense Webster, and Dyne Pharmaceuticals outside the submitted work. No other disclosures were reported.

Data Sharing Statement: See Supplement 2.