Vulnerability of the ventricular conduction axis during transcatheter aortic valvar implantation: A translational pathologic study

The ventricular components of the conduction axis remain vulnerable following transcatheter aortic valvar replacement. We aimed to describe features which may be used accurately by interventionalists to predict the precise location of the conduction axis, hoping better to avoid conduction disturbances. We scanned eight normal adult heart specimens by 3T magnetic resonance, using the images to simulate histological sections in order accurately to place the conduction axis back within the heart. We then used histology, tested in two pediatric hearts, to prepare sections, validated by the magnetic resonance images, to reveal the key relationships between the conduction axis and the aortic root. The axis was shown to have a close relationship to the nadir of the right coronary leaflet, in particular when the aortic root was rotated in counterclockwise fashion. The axis was more vulnerable in the setting of a narrow inferoseptal recess, when the inferior margin of the membranous septum was above the plane of the virtual basal ring, and when minimal myocardium was supporting the right coronary sinus. The features identified in our study are in keeping with the original description provided by Tawara, but at variance with more recent accounts. They suggest that the vulnerability of the axis during transcatheter valvar replacement can potentially be inferred on the basis of knowledge of the position of the aortic root within the ventricular base. If validated by clinical studies, our findings may better permit avoidance of new‐onset left bundle branch block following transcatheter aortic valvar replacement.

the nadir of the right coronary leaflet, in particular when the aortic root was rotated in counterclockwise fashion. The axis was more vulnerable in the setting of a narrow inferoseptal recess, when the inferior margin of the membranous septum was above the plane of the virtual basal ring, and when minimal myocardium was supporting the right coronary sinus. The features identified in our study are in keeping with the original description provided by Tawara, but at variance with more recent accounts. They suggest that the vulnerability of the axis during transcatheter valvar replacement can potentially be inferred on the basis of knowledge of the position of the aortic root within the ventricular base. If validated by clinical studies, our findings may better permit avoidance of new-onset left bundle branch block following transcatheter aortic valvar replacement.

K E Y W O R D S
aortic root, conduction system, heart block, left bundle branch block, three-dimensional imaging, transcatheter aortic valvar replacement 1 | INTRODUCTION Injury to the ventricular components of the atrioventricular conduction axis remains to be a problem subsequent to transcatheter aortic valvar replacement (Bis et al., 2021;Sammour et al., 2021), with new left bundle branch block relating to an increased risk of cardiac death (Faroux et al., 2020). Widely different outcomes, furthermore, are reported from center to center, with significantly lower rates of conduction damage at established centers of excellence (Desai et al., 2021). It is axiomatic that avoidance of damage to the conduction axis will be enhanced by accurate knowledge of the precise locations of the conducting tissues adjacent to the aortic root. Previous recommendations in this regard have largely focused on determining the depth of implantation of devices relative to the plane of the virtual basal ring of the aortic root, and to the inferior margin of the membranous septum (Hamdan et al., 2015;Jilaihawi et al., 2019;Maeno et al., 2017;Sánchez-Quintana et al., 2022;. In this regard, however, many drawings used to show the location of the left bundle branch, (Aslan et al., 2022;Jørgensen et al., 2022;Vijayaraman et al., 2018;Zhang et al., 2022) and commentaries on this feature, (Liang & Bogun, 2022) are at variance with the initial account provided by Tawara. (Tawara, 1906) The particular difference relates to adjacency of the superior fascicle of the left bundle branch to the nadir of the right coronary aortic leaflet. This, in turn, may reflect the challenges that remain in precisely relating the conduction axis to the complex three-dimensional (3D) anatomy of the aortic root . As has been emphasized in a recent vignette commemorating the sesquicentennial anniversary of the birth of Tawara (Mori et al., 2023), fully to appreciate his findings it is necessary to put the atrioventricular conduction axis back into the heart. Our aim in this study, therefore, was to expand our previous study, in which we confirmed the findings of Tawara (Tawara, 1906), but using a twodimensional (2D) approach (Macías et al., 2022). In our current study, we obtained 3D representations of the aortic root from autopsied adult heart specimens, subsequently using histological sections so as to place the ventricular conduction axis back into the structurally normal heart. In this way, we aimed to identify variable features of the aortic root within the ventricular base that might be identified during life, thus providing inferences which might permit interventionalists better to avoid the conduction axis during transcatheter aortic valvar replacement.

| METHODS
We procured eight structurally normal adult donor hearts, fixed in Kaiserling III solution, from individuals without known cardiovascular disease from the National Disease Research Interchange, along with two hearts obtained from children undergoing autopsy. Available details of the age, size, and weight of the donors are shown in Table S1. We initially used the relatively smaller pediatric hearts to assess the feasibility of making serial histological sections. For the adult donor hearts, once received, we first acquired high-resolution anatomic images using a 3T research magnetic resonance scanner (Achieva; Philips Healthcare; Best, the Netherlands). 3D volumetric acquisition was performed using an 8-element phased array coil for signal reception, using a repetition time of 2500 milliseconds, echo time of 65 milliseconds, an acquired voxel size of 0.4 x 0.4 x 0.4 millimeters cube, and five signal averages. On this basis, we provided 3D representations of the aortic root in each heart, showing the variable relationships of the roots within the ventricular base. We then used the magnetic resonance images to simulate the histological sections obtained by subsequent serial sectioning ( Figure S1). Having placed the histological sections within the reconstructed root, we were then able accurately to measure, using the histological sections, the distance of the components of the conduction axis from the components of the roots themselves. This information was then correlated with the identified location of the roots within the ventricular base. This allowed us to make inferences regarding the vulnerability of the conduction axis during valvar replacement. Subsequent to scanning, the hearts had been dissected to reveal the details of the triangle of Koch on the right side, and the left ventricular outflow tract and aortic root on the left side. The rotation of the root within the base of the left ventricle was determined according to the relationship of the interleaflet triangle between the left and non-coronary leaflets to the midline of the aortic leaflet of the mitral valve. (Amofa et al., 2019;Tretter et al., 2018) Clockwise rotation, as opposed to a central position of the aortic root, or counterclockwise rotation, was then determined as viewed from the apex of the left ventricle. We also used the 3D datasets to measure the angle of the crest of the muscular ventricular septum relative to the plane of the virtual basal ring. Prior to removing the block to be used for histology, we transilluminated from the right side to permit identification of the membranous septum. The magnetic resonance dataset, along with the histologic findings, then permitted the location of the conduction axis to be determined relative to the aortic root.
For the histological sectioning, we had removed a block of tissue extending from the base of the triangle of Koch to the aortic root, including the hinges of the right and non-coronary leaflets, as shown in Figure S1. We had bisected the block perpendicular to the virtual basal ring of the root along a line between the right and non-coronary aortic valvar sinuses. Each block was sectioned perpendicular to the plane of the virtual basal ring at 5 microns thickness, using Masson's trichrome technique to stain at intervals of 0.2 mm. Multiplanar reformatting of the magnetic resonance datasets permitted us accurately to simulate the location of the histological sections. Using the 3D reconstructions, we then segmented the extent of the inferoseptal recess of the left ventricle, the membranous septum, the plane of the virtual basal ring, the nadirs of the right and non-coronary aortic valvar sinuses, and the apex of the inferior pyramidal space. In this way, we placed the locations of the ventricular components of the conduction axis, as assessed from the histological sections, back into the 3D heart, showing their precise relationships to the components of the aortic root. Comparison with the variation noted in the fibrous components supporting the aortic root, specifically those parts identifiable by clinical imaging, permitted us to draw inferences regarding the vulnerability of the ventricular conduction axis. We have previously described variations of the atrioventricular node within the triangle of Koch. (Cabrera et al., 2020;Macías et al., 2022) The study, in accord with the Helsinki Declaration, was approved by the Institutional Review Board at Cincinnati Children's Hospital Medical Center.

| RESULTS
The major finding of our study was our confirmation of the original account of Tawara (Tawara, 1906), which described a close relationship between the superior fascicle of the left bundle branch and the nadir of the right coronary aortic leaflet. This relationship itself, however, was variable. In three of the eight adult hearts, the overall conduction axis was relatively remote to the virtual basal ring. The virtual basal ring is the plane that can be constructed by joining together the nadirs of the semilunar hinges of the valvar leaflets ( Figure 1). In these hearts, on average, the axis was 3 mm inferior to the virtual plane at its closest point, with the superior fascicle of the left bundle branch positioned from 7.3 to 9.6 mm inferior to the nadir of the right coronary aortic leaflet. In two of these hearts, the aortic root was centrally positioned, but showed significant clockwise rotation in the third ( Figures 1C and 2A). The entirety of the branching atrioventricular bundle was positioned inferior to the non-coronary aortic sinus ( Figure 2A). In all three hearts, the inferior margin of the membranous septum was itself inferior to the plane of the virtual basal ring. All three hearts had extensive myocardial support of the right coronary aortic sinus ( Figure 3A).
In the remaining five adult hearts, in contrast to the first three, the conduction axis was significantly closer to the plane of the virtual basal ring throughout its ventricular course ( Figure 4). On average, at its closest point, the superior fascicle of the left bundle branch was located from 1.0 to 3.6 mm inferior to the nadir of the right coronary aortic sinus. In three of these hearts, the root was rotated counterclockwise, being centrally positioned in the other two. In four of the five hearts, the superior fascicle of the left bundle was the component closest to the virtual basal ring ( Figure 4C-F). In heart #4, in which the root was centrally positioned, the origin of the left bundle was closest ( Figure 4B). In all five hearts, the roof of the inferoseptal recess, defined as the area of fibrous tissue interposed between the aortic leaflet of the mitral valve and the septum, was narrow. In heart #6, with a counterclockwise-rotated root, it was almost non-existent. In this heart, the right fibrous trigone, defined as the thickened rightward end of the area of aortic-to-mitral fibrous continuity, was almost directly attached to the steeply angled crest of the muscular ventricular septum. The conduction axis itself in this heart was located more anteriorly within the root, with the superior fascicle of the left bundle branch positioned well above the plane of the virtual basal ring ( Figure 4D). In four of the five hearts, the anterior aspect of the inferior margin of the membranous septum was also superior to the virtual basal ring. In heart #5, despite the origin of the left bundle branch being 5.1 mm inferior to the plane of the virtual basal ring ( Figure 4C), the superior fascicle reached within 1 mm of the nadir of the right coronary aortic leaflet. In four of these hearts, no myocardium was incorporated at the base of the right coronary aortic sinus ( Figure 3B).
Taken overall, therefore, and in keeping with our previous investigation using a 2D approach (Macías et al., 2022), significant variability was found in the location of the conduction axis relative to the aortic root (Tables S2 and S3). When assessing these findings, we considered only the rightward extent of fibrous continuity between the leaflets of the aortic and mitral valves as the right fibrous trigone. This was because the 3D datasets showed this thickening of the rightward end of the area of mitral-to-aortic continuity to be distinct from the area of fibrous continuity between the leaflets of the mitral and tricuspid valves. This latter fibrous area formed the roof of the inferoseptal recess of the left ventricular outflow tract ( Figure 5). It is F I G U R E 1 The panels show virtual dissection reconstructions of the three-dimensional (3D) magnetic resonance datasets in three-chamber cuts demonstrating the conduction axis in the three specimens with a greater distance between the aortic root and atrioventricular conduction axis. For each heart, virtual reconstructions of the 3D magnetic resonance datasets are cut in a long axis plane that allows visualization of the circumference of the virtual basal ring. Comparisons between the images provided by magnetic resonance interrogation and the histological sections permitted the atrioventricular conduction axis accurately to be placed back into the heart. The translucent black oval shows the atrioventricular node, the translucent black line the penetrating bundle, the yellow line the non-branching bundle, and the white line the left branching bundle. The red arrow then shows the proximal course of superior fascicle of left bundle. The membranous septum is colored blue, with the inferoseptal recess colored orange. noteworthy that it was this area of fibrous continuity between the leaflets of the mitral and tricuspid valves that was emphasized by Keith and Flack as forming the larger part of the central fibrous body in their study published in 1906 (Keith & Flack, 1906) to endorse the original account of Tawara. When traced rightward, it was continuous with the membranous septum, which was divided into interventricular and atrioventricular components by the septal leaflet of the tricuspid valve ( Figure 5B-D).
Additional findings had emerged from the study of the two pediatric hearts. Both had significant counterclockwise rotation of the root. The inferoseptal recess was small in both, with the anterior aspect of the inferior margin of the membranous septum coursing F I G U R E 2 Legend on next page.
superior to the plane of the virtual basal ring. The superior fascicle of the left bundle branch was visible grossly in the first pediatric heart.
Its close relationship with the nadir of the right coronary aortic leaflet ( Figure 6A) was confirmed by histology ( Figure 6B,C). No myocardium was incorporated under the base of the right coronary aortic sinus in either heart.

| DISCUSSION
Our study has revealed marked variability in the location of the ventricular components of the atrioventricular conduction axis relative to the aortic root, with potential inferences regarding its vulnerability during transcatheter valvar replacement. The main findings pre-F I G U R E 2 Panels A and B show short axis virtual dissections of the three-dimensional magnetic resonance datasets cut at the plane of the virtual basal ring viewed from above (arterial view), demonstrating how the rotational position of the aortic root impacts the longitudinal relationship to the atrioventricular conduction axis. Panel A shows Heart 3 with a clockwise positioned aortic root. The non-coronary sinus is related to a significant proportion of the atrioventricular conduction axis. Panel B shows the relationship in Heart 6 with a counterclockwise positioned aortic root. It is now the right coronary sinus that is relates to a larger proportion of the conduction axis. In this heart the conduction axis coursed superior to the plane shown, depicted where the yellow line transitions from being translucent to not being translucent, at the black arrow, and remains superior to this plane throughout the remainder of its course. The inferoseptal recess as seen in Panel A is much broader than in Panel B. A greater proportion of the proximal atrioventricular conduction axis is positioned within the circumference below the plane of the virtual basal ring in the heart with a shallow inferoseptal recess (Panel B). The components of the conduction axis, and the central fibrous body, are shown as for Figure 1. Panels C and D demonstrate clinical examples corresponding with the rotational position demonstrated in Heart 3 (Panel A) and Heart 6 (Panel B), respectively. However, while Panels A and B demonstrate the plane of the virtual basal ring as viewed from above (surgical or arterial view), Panels C and D depict the aortic root viewed from the apex of the left ventricle (ventricular view). Historically, it is the latter view which determines designating clockwise (more anterior position of the right coronary leaflet) versus counterclockwise position (more posterior postion of the right coronary leaflet). The rotational position is judged by comparing the nadir of the non-coronary leaflet (N) with the atrial buttress (green triangle), and quantitatively can be measured as demonstrated in Figure 9E1. Panels E and F depict corresponding drawings of clockwise (Panel E) and counterclockwise rotation as shown using the arterial view (Panel F). The corresponding features dictating the position of the ventricular components of the conduction axis are demonstrated, as discussed for Panels A and B. L, left coronary leaflet; R, right coronary leaflet.  (Figure 7), a F I G U R E 4 The panels show virtual dissection reconstructions of the three-dimensional magnetic resonance dataset in three-chamber cuts in the specimens with a shorter distance between the conduction axis and the aortic root. Panel A depicts this plane in Heart 5, zoomed out for reference, with the myocardium translucent. The red box represents the zoomed up field of view presented in the remaining panels. The components of the conduction system and the aortic root are shown as for Figures 1 and 2 F I G U R E 5 The images show the features of the so-called central fibrous body. Panel A, a histological section, shows the roof of the inferoseptal recess (orange star) formed by fibrous continuity between the septal leaflet of the tricuspid valve and the aortic leaflet of the mitral valve. The yellow asterisk shows the branching component of the conduction axis. Panels B-D are virtual dissections from a computed tomographic dataset of a living patient. The four-chamber cut (Panel B) replicates the histological cut, with the red box representing the field of view. The floor of the inferoseptal recess is colored orange. The short axis cut, in Panel C, demonstrates the central fibrous body to be formed by the membranous septum, colored blue, the floor of the inferoseptal recess in orange, and the right fibrous trigone, shown by the white circle. The left fibrous trigone is marked with the gray circle. Note the apex of the inferior pyramidal space overlaps on the rightward and posterior aspect of the inferiorly directed apex of the inferoseptal recess. Panel D demonstrates the components in a three-chamber cut, with the plane of the virtual basal ring represented by the green line technique now commonly used prior to transcatheter valvar insertion.
The findings, achieved using 3D reconstructions of the root, expand our previous findings obtained using more traditional histological techniques (Macías et al., 2022). These showed that, in the variations placing the conduction axis at greatest risk, the component of the axis closest to the aortic root is the superior fascicle of the left bundle branch. This had been described by Tawara (Tawara, 1906), who showed that, having taken origin from the nonbranching atrioventricular bundle, this fascicle continued to encircle the left ventricular outflow tract, passing beneath the nadir of the right coronary aortic leaflet ( Figure 8A). This adjacency to the nadir of the right coronary leaflet, however, has not been shown in several drawings provided to date to show the potential vulnerability of the conduction axis (Aslan et al., 2022;Jørgensen et al., 2022;Vijayaraman et al., 2018;Zhang et al., 2022). In these drawings, as emphasized in a recent commentary (Liang & Bogun, 2022), the left bundle branch is shown as descending the septum directly at the superior margin of the membranous septum ( Figure 8B). This finding could also be pertinent to a recent suggestion that the component of the conduction axis directly related to the nadir of the right coronary aortic leaflet of the aortic valve is the so-called "dead end tract" (Zhang et al., 2022). In reality, as we have now shown, the component in question is the superior fascicle of the left bundle branch. As we showed in our previous 2D study, the dead-end tract can be identified in only a small proportion of hearts. When present it continues beyond the take-off of the right bundle branch. It is quite distinct from the superior fascicle of the left bundle. As we have also shown, nonetheless, there is marked variability in the proximity of the superior edge of the left bundle branch to the nadir of the right coronary leaflet. This variability had also been found in our previous study based solely on 2D methods (Mori et al., 2023). The added value of access to our 3D datasets now permits us to identify features, potentially available to clinicians using modern imaging techniques, which might permit better prediction of the vulnerability of the conduction axis. The observed variation in the position of the aortic root, marked even in these normal hearts, includes not only its rotation, but also the angle between its long axis and that of the left ventricle, and the extent of its wedging between the mitral valve and the septal surface of the left ventricle. (Amofa et al., 2019;Mori et al., 2017;Tretter et al., 2018. To date, those seeking to predict the features that might reflect the vulnerability of the conduction axis during transcatheter valvar replacement have focused on the location of the membranous septum. It is now accepted that clinical measurements need to acknowledge these variations, including the angle of its inferior margin. (Amofa et al., 2019;Faroux et al., 2020;Jørgensen et al., 2022;Tretter et al., 2018) Failure also to account for the variation on the location of the aortic root within the base of the left ventricle (Figure 9), which to the best of our knowledge have not previously received attention, may serve to underestimate the risk of iatrogenic injury.  Our findings suggest that this variation, now measureable using clinical techniques (Figure 7), dictates the relationships of the aortic valvar sinuses to the ventricular components of the conduction F I G U R E 6 The panels show the relationship between the aortic root and and the conduction axis in Pediatric Heart 1. Panel A shows the gross anatomy, with counterclockwise rotation of the aortic root relative to the base of the left ventricle, revealed by the relationship of the purple arrow, which shows the midline of the aortic leaflet of the mitral valve, relative to the interleaflet triangle between the non-coronary leaflet (NCL) and left coronary leaflet (LCL). The right and left fibrous trigones are shown by the white and gray circles, respectively. The membranous septum, outlined in blue, has been transilluminated, along with the small inferoseptal recess, outlined in orange. Note that the majority of the membranous septum, including its anterior aspect, courses above the plane of the virtual basal ring (green line). The superior fascicle of the left bundle branch (outlined with yellow dashed lines) is grossly visible in close proximity to the nadir of the right coronary leaflet (RCL). The approximate section planes for Panels B and C are shown by the black lines, with Panel B demonstrating the superior fascicle of the left bundle to be within 1 mm of the nadir of the right coronary leaflet, and Panel C demonstrating the very proximal branching bundle to be 1.4 mm inferior to the attachment line of the non-coronary leaflet. The majority of the branching bundle relates to the right coronary leaflet.
axis. With significant clockwise rotation, much more of the axis is directly related to the non-coronary aortic sinus, often associated with a broader and deeper inferoseptal recess, making it less vulnerable during valvar implantation. And, as we have now shown, the greater extent of myocardium incorporated into the base of the clockwiserotated right coronary aortic sinus increases the distance between left bundle branch and the base of the sinus. In their study of 1906, Keith and Flack also noted the significance of this "subaortic musculature" (Keith & Flack, 1906). In their report, they commented that its functional significance was "unknown to us." Its importance is that, when present, it makes the axis less vulnerable during transcatheter valvar implantation. In the setting of counterclockwise rotation of the root, F I G U R E 7 Three-dimensional reconstructions from a computed tomographic dataset of a normal patient, as shown in panels A through D, demonstrate the normal aortic root anatomy, emphasizing its relationship to the membranous septum (colored blue), the roof of the inferoseptal recess (colored orange), and the right and left fibrous trigones (white and gray circles, respectively). In panels E1 and E2, we then show comparative two-dimensional images of the short axis of the aortic root (Panel E1) and plane of the virtual basal ring (Panel E2; the circumference of the virtual basal ring is colored green). These structures are viewed from the ventricular aspect, in contrast to the surgical views demonstrated in Panels A and B. The images show how the rotational angle of the aortic root can be measured as the angle between a line extending from the triangular attachment of the atrial septum across the aortic root (red solid line) compared to a line extending from the center of the non-coronary sinus (N) extending across the zone of apposition between the right (R) and left coronary sinuses (L) (red hashed line). Normative data of the rotational position of the aortic root in an adult population utilizing this same method has previously been described (20). The width of the roof of the inferoseptal recess at its base can be measured at the plane of the virtual basal ring (Panel E2). These sections are continued inferiorly (Panel E3), ending at the inferiorly directed apex of the inferoseptal recess (Panel E4). The superimposed circumference of the virtual basal ring in these inferior panels allow measurement of the depth of the inferoseptal recess at its inferior apex. The recess extends inferior to the circumference extending below the plane of the virtual basal ring. This is also demonstrated in the orthogonal plane depicted in Panel F2, with the black solid lines representing this circumference. A center point placed within the plane of the virtual basal ring can then be established, and rotated around to produce the orthogonal long axis Panels F1-F4 to assess the height and depth of the roof of the inferoseptal recess (Panel F2), and the distance from the plane of the virtual basal ring to inferior margin of the membranous septum (Panel F3). This measurement is referred to as the inferior distance, and should be assessed in at least three points from its posterior to anterior aspect as demonstrated in Panel C. The black hashed lines in Panel F2 depict the planes of the orthogonal short axis images produced in Panels E1-E4. Diastolic phase measurements are depicted. Further clinical investigations are required better to understand dynamic changes which may occur throughout the cardiac cycle and conformational changes which may occur following valve deployment. The only feature from this study which currently cannot be assessed by standard clinical imaging is the transition from the crest of the interventricular septum to the outflow tract myocardium underneath the right coronary sinus. LMCA, left main coronary artery; RCA, right coronary artery F I G U R E 8 The drawing in Panel A, based on the image initially provided by Tawara,[16] shows the correct relationship of the ventricular components of the atrioventricular conduction axis relative to the aortic root. As was shown initially by Tawara, and confirmed by our study, the superior fascicle of the left bundle usually continues to ascend the crest of the muscular interventricular septum, anterior to the membranous septum, as it wraps around the left ventricular outflow tract. Failure to note such variations may serve to underestimate the risk of iatrogenic injury. However, it has been common place to incorrectly depict the superior fascicle as descending rapidly when the axis reaches the superior margin of the membranous septum, with a more distant relationship to the right coronary leaflet (Panel B) F I G U R E 9 The panels demonstrate the vulnerability of the conduction axis following simulation of prosthesis implantation in a low risk (Panels A and C) versus high risk (Panels B and D) arrangement. In these hearts, this difference in risk is despite both hearts having a comparable distance between the plane of the virtual basal ring and inferior margin of the membranous septum. The components of the conduction system and the aortic root are shown as for Figures 1 and 4. LMCA, left main coronary artery; RCA, right coronary artery in contrast, the inferoseptal recess becomes extremely narrow, or almost nonexistent. The axis itself is then positioned more anteriorly relative to the aortic root. Its more proximal components are then brought within the circumference of the left ventricular outflow tract extending below the plane of the virtual basal ring. The superior fascicle of the left bundle branch is appreciably closer to the nadir of the right coronary leaflet in this setting. These features in the counterclockwise positioned aortic root suggest greater vulnerability of disturbance to the axis following transcatheter valvar replacement.
We recognize that our current findings are based on study of a small number of normal hearts. They extend, nonetheless, our previous histological evaluation of 20 normal adult hearts (Macías et al., 2022). In the majority of hearts in both studies, the most vulnerable part of the conduction axis was that which was below the right coronary sinus (Macías et al., 2022). As far as we are aware, our study is the first to make inferences linking this variability to the 3D arrangement of the aortic root. By use of 3D imaging datasets, and multiplanar reformatting, the aortic root can now accurately be related to its surrounding structures (Tretter et al., 2020;Tretter et al., 2021;Tretter et al., 2022). These features, namely the rotational position of the aortic root, the width, height and depth of the roof of the inferoseptal recess, and the site of the inferior margin of the membranous septum relative to the plane of the virtual basal ring, can all now be assessed clinically by computed tomographic interrogation ( Figure 7). We suggest that, if validated by additional studies in patients with valvar stenosis, these findings can set the stage toward more accurate clinical prediction and avoidance of iatrogenic conduction damage.

| LIMITATIONS
We recognize that our study is limited by the small number of hearts used. The received donor hearts randomly obtained from the donor source were from Caucasian female donors. We are not aware of any correlation, nonetheless, between gender or race regarding variations in cardiac anatomy. Furthermore, these hearts were fixed in a nonhemodynamic state. The process from acquisition of the magnetic resonance image acquisition of the intact autopsied heart specimen, toward dissection to allow gross internal visualization, and finally to the preparation of the blocks for histology, leads to additional incremental changes in deformation. Although this potentially creates limitations for comparison between the imaging and histological datasets, by simulating the histological sections from the 3D images, we showed that such caveats were unfounded. The findings from the small number of hearts in our current study endorsed the comparable variability noted in our study of the much larger group of hearts achieved using 2D techniques (Macías et al., 2022). We also recognize, nonetheless, that both of these studies were conducted using anatomically normal hearts. It will now be necessary to undertake comparable investigation of hearts obtained from individuals known to have suffered from aortic valvar stenosis.