Welcome, everybody, to day three of HRS 47. I feel like I have the distinct responsibility as a San Diegan to apologize for the sprinkles that occurred this morning. That's my bad, personally, and I hope that the rest of the day is a little sunnier. And welcome to all those that came directly from closing time of the bars and are waddling in here. And I have a distinct pleasure of presenting a few speakers that are going to truly have a potpourri of different novel concepts that I'm excited to share with you all. If you have not downloaded the app and done the QR code thing to answer any questions, if it's all not working, just flag me down and ask me the question, and I'll give it to the speaker. With that, you keep wearing them, so you can actually hear yourself speak. Oh, no. All right, Ivor. That's all right. So I have Ivor Othalos from CHOP presenting h-ventricular accessory pathway localization from rusting EKG in children with WPW, and he has some nice AI memes embedded in the PowerPoint, so I'm excited to hear from you. Cool. All right. Hi, guys. So we'll get started. If I can figure out how to... There we go. All right. I have no disclosures. So as you all know, especially in this room, Wolff-Parkinson-White syndrome is characterized by the presence of an h-ventricular accessory pathway with antigrade conduction. Patients with persistent pre-excitation are referred for ablation. But knowing where the pathway is before arriving to the cath lab can still be helpful in multiple ways. Knowing whether or not you will need to perform transeptal puncture can be helpful when talking to the family and for pre-procedural planning. Knowing whether the pathway is close to the conduction system may inform timing of the procedure based on patient size and whether or not you plan on using cryoablation. In an error of cryoablation, the location of the pathway can be exactly determined by the site of acutely successful elimination of extranodal conduction. This error generated many of the most popular location algorithms such as the ones from Chiang and Aruta. Many of you probably carry a copy of one of these algorithms around with you either in your phone or in your head. And according to their own internal validation metrics, they work like gangbusters with reported accuracies all above 90% for exact location of the pathway. However, when performance is measured by an external group on an external cohort, accuracy tends to drop by about 20 percentage points. The next problem is that all these original algorithms were derived from and validated in adults. Predictably, the performance drops even further when these algorithms are used in children. When Ren and colleagues looked at this and validated this externally in children, they found that accuracy ranged from as little as only from 35 to 40%. Identifying the exact location is a tall order, and in most cases, just being able to distinguish between septal, right, and left is usually sufficient. Right versus left versus septal accuracy has been assessed for these adult algorithms in children by three different studies as you can see here, and accuracies range from as little as 5% to a max of 85%. To address this gap, two algorithms have been specifically derived and validated in children, but again, accuracy and external validation is no greater than 55% for both, and laterality achieved in actually over 80% in neither algorithm. So we hypothesized that using artificial intelligence may localize accessory pathways with at least as good accuracy as published manual algorithms. So we set out to develop and internally validate a convolutional neural network to predict accessory pathway location from surface ECG in children. One criteria of the study was limited to patients 0 to 21 years of age who had an acutely successful ablation for a single atrial ventricular accessory pathway with integrated conduction in an otherwise structurally normal heart. Applying these exclusions resulted in a cohort of about 905 ablations, which we merged with our ECG database of about a million ECGs. This left us with about 6,400 ECGs. After we excluded ECGs that were performed on or after the date of ablation, or ECGs didn't have pre-excitation on them, this left us with our final cohort of 2,400 ECGs from 788 unique patients. The data split was 85% to a training dumbbell inset, and 15% to a holdout test set at the patient level to prevent data leakage. The median age and time of cath was just over 13 years. The median age and time of ECG was just a few months younger. A slight majority of the cohort was male, about two-thirds of patients were white, with about six of patients being black or other. The deep neural network we developed was a convolutional ResNet. This model takes as input the entire 15-lead ECG. This is then passed through 16 convolutional residual blocks. Then the demographics that we input are just sex and race. They get passed into it before a terminal dense network, and then this ultimately goes to a multiclass classifier that outputs one of 14 pathway locations. The primary outcome was accuracy, defined as the sum of true positives and true negatives as a percentage of all predictions. We calculated accuracy in four ways. Was the predicted pathway exactly correct? Was it within one contiguous neighbor? Just comparing right versus left, and whether or not it was right versus left versus septal accurate. We classified pathways into one of 14 locations. The data set is heavily skewed, with right postural septal and left lateral being the most common, and together those combined account for more than half of all pathways. When dichotomizing the results to just right versus left, a slight majority, 54%, were right-sided. For right versus left versus septal classification, we characterized right and left postural septal and right and mid-anteroseptal as septal. Using those three buckets, about 40% are left-sided and septal each, and the remaining 16% are right-sided. So how well did the algorithm do? This is a confusion matrix where the vertical axis is the site of successful ablation, and then on the horizontal axis is where the algorithm predicted the pathway was going to be based on the surface ECG. If we limit correct predictions to just those that are exactly on the money, these green boxes in diagonal, then the accuracy is 55%. So for example, if you look at that first row for right postural septal, of the 85 right postural septal pathways, 78 of them are correctly predicted to be right postural septal. However, when the ACG predicts the pathway and it's wrong, the correct location is usually next door. So for example, if you look at the 35 pathways that were predicted to be left lateral but were not, two-thirds of those are either left postural lateral or left anterolateral, immediately adjacent to the prediction. So if we extend the definition and be a little bit looser, and say the correct prediction has to be within one contiguous neighbor of where the successful ablation site was, the accuracy improves to 82%. If all you want to know is whether or not this is right or left-sided, or whether or not you need to do transeptal, the accuracy is even higher, it's 94%. The sensitivity, the specificity, and both the negative and positive predictive values are also all above 90%. And then lastly, if we look at right versus left versus septal, 88% of right-sided pathways are correctly identified, 93% of left-sided pathways, and 81% of septal pathways. So if we take these results and we put them back on the leaderboard of pathway localization accuracy exclusively in children in child cohorts, we see that the exact accuracy is right in the middle of the mix, at about 55%. This is an unfair fight though. The AIECG predicts 14 different locations, whereas the algorithms shown here predict only six to nine locations, and the fewer the locations, the easier it is to be correct. So if we keep the number of locations the same, and look at right versus left versus septal accuracy, then you'll see that the AIECG outperforms all other algorithms on all three locations. One of the inherent limitations of currently published algorithms is that main decision points are subjective. Even veteran-like physiologists looking at the same ECG using the same algorithms will sometimes come up with a different location. The concordance rates, the percentage of times that two readers predict the same location, range from as low as 40% to 80%, disagreeing on average 34% of the time. The AIECG, however, always produces the same prediction for the same ECG. There's just no human element involved. The concordance is thus 100%, and the kappa is one by necessity. Despite these promising results, this study has a number of limitations. First off, it was a single-center, retrospective study with all its attendant limitations, including the fact that validation was done only internally. Accessory pathways at our institutions were and continue to be described by the relation to the septum and not their location within the chest, although I expect this is probably the case at most of the institutions here as well. Concordance is also likely worsened by the inherent subjectivity of describing where the pathway is. One operator's left posterolateral is another operator's left lateral. Despite overweighting minority classes, this algorithm doesn't work as well for the most uncommon locations, for example, left anterior pathways. And then lastly, this algorithm just did not work for multiple manifest pathways, although we tried it. In conclusion, a deep neural network utilizing the entire ECG and the demographics may predict accessory pathway better in children, or may predict accuracy with better accuracy in children than currently published adult and pediatric algorithms, although admittedly the bar is currently just not that high. It achieves this without any human input, and so the concordance is perfect and no expertise is required. So catchy title aside, is this the one algorithm to rule them all? I think a deep neural network will be the one to obviate all current algorithms, but this particular algorithm is still being forged. The learning curve suggests that we can still get a lot of performance by increasing training volumes, and we're looking to collaborating with other people to acquire additional ECGs from successfully ablated patients, both for additional training volume and for external validation. For this work, I'm most indebted to Tammy Sweeten, our EP specialist, who has been assiduously maintaining our database for successful ablations since before I could read. The Tree Lab, which was where I did my post-doc, and so all the operators on whose successful ablations this study relies, primarily to Molly Shaw, who I think did a disproportionate number of these successful ablations over the last two decades. Thank you so much. All right, thanks, Ivor. I don't see any questions from the app, so I remind you to use it for the next sessions too, but I've got a question for you, I guess. I'm sure I'll see this data in 2026, but how is this already affecting practice in your lab, in terms of, hey, maybe I don't need to reach for that, or don't even open the needle yet. I'm going to see what's going on here first. Currently right now, we do not use this algorithm to decide where we're going to go in the lab. One of the limitations of it is, you may have heard about this if you were at some of the AI talks the other day, is that the input for this is digital ECGs, and so if somebody comes from a referral center, all you have is a picture or a PDF of their ECG, you're not going to be able to input into this model. And even if they've done it locally, which is the majority of our patients who came for a pre-visit, the amount of work it takes to just get that ECG into Python and run this inference is usually more, it's a longer slide, a longer run than the slide is worth. Perfect. I had a question. How do you, what was considered true in regards to the location? Was it the reports? Were all the cases re-analyzed? And was there a single person choosing, because you said that there was perfect concordance, but ultimately a human being had to tell the computer where was the successful relation, so how did you adjudicate that? Yeah, so the report in our database is considered the ground truth, so if the operator said this is left posterior, then that's where we said it was. I think a corollary to your question might be, how do you know that you were exactly correct? If we sort of carpet bombed, you know, the back of the left atrium or the mitral annulus, are we exactly sure where it was? And so one way to do this is to say, hey, you needed to have gotten it relatively quickly, and you needed to have gotten it on a relatively early ablation. It couldn't have been like your 14th ablation at, you know, second 15 is when you finally got loss of pre-excitation. As a follow-up, would there be any way to train the computer to look at the image either on 3D electroanatomical mapping or FLORA to create its own analysis of where the successful lesion was, which might then really take out the concordance issues? I love that question. One of the issues with the way the algorithm currently works is the output is one of these 14 classes, and the algorithm gets penalized exactly the same whether or not it is one away or seven are on the exact wrong annulus. And so the best case scenario for as far as I'm concerned is that you would basically take the annulus, put it into two-dimensional space, say this is where it was, this is where it predicted, and the loss function now becomes the distance between them instead of this class is totally different than that class. So I love that idea. All right. Thank you very much, Ivor. And up next, we have Chelsea Boyd from Texas, who's going to tell us all that we're waiting too much in the lab after an ablation. Yeah. Thank you so much, everyone. I really appreciate the opportunity to speak today, and that's a perfect segue. We just talked about, you know, the 10, 30 minutes before the procedure, and now you've done the procedure, you've got that pathway, and we're going to talk about the 10 to 30 minutes after the procedure. So let's get on started. All right. So pediatric-specific data on waiting period recurrence are limited, and in the context of 3D mapping and ablation techniques, the exact utility of the waiting period is not well-defined. Nevertheless, most or many operators observe for at least 30 minutes post-ablation to assess for return of pathway conduction. This raised the central question that led to our investigation. How often do we actually detect recurrences during the conventional post-ablation waiting period? From there, a key follow-up emerged. Could a shorter 10-minute waiting period reliably identify recurrences that would otherwise appear at 30 minutes? Although shortening the conventional 30-minute waiting period may seem insignificant, and you might say, why not wait if I could prevent a repeat procedure? We should consider the potential benefits of a shortened waiting period. A reduced observation time could minimize patient exposure to anesthesia, shorten overall procedure duration, improve lab efficiency and case throughput, and help prevent staff fatigue, particularly during prolonged or after-hours cases. It may also offer cost benefits by decreasing anesthesia time and reducing overtime staffing. The question of a 10-minute waiting period was recently studied in an adult cohort. The early study in 2023 assessed adults undergoing RF accessory pathway ablation and reported a 7.6% 30-minute recurrence rate. This adult population differs from our pediatric cohort, both in age and procedural factors, as ablations were performed under conscious sedation with non-irrigated catheters. The authors found that a 10-minute post-ablation waiting period had poor sensitivity for detecting conduction recurrence at 30 minutes, detecting only 37.5% of 30-minute recurrences. They concluded that a 10-minute waiting period may not be sufficient to universally replace the 30-minute waiting period in this population. Turning our attention to our pediatric cohort, we had three objectives. First, we aimed to determine the incidence of recurrence during the conventional 30-minute waiting period. Second, we assessed the sensitivity of an abbreviated 10-minute waiting period in detecting recurrence at 30 minutes. Third, we assessed 24-hour recurrence with two goals. We aimed to determine whether reablation following a waiting period recurrence effectively reduced recurrence rate at 24 hours. We also examined timing of recurrence for manifest pathways, focusing on whether an extended waiting period could feasibly capture any of these recurrences. We conducted a retrospective review of all pediatric accessory pathway RF ablations performed at our institution between September 2020 and February 2025, all performed under general anesthesia with 3D mapping and irrigated catheters. Patients were excluded for ages greater than 21 years, prior ablations, multiple accessory pathways, coexisting arrhythmia mechanisms, use of prior ablation during the primary ablation, and for pathways in which we were unable to adequately affect the accessory pathway. A total of 237 patients met inclusion criteria. All 237 were included for the assessment of overall waiting period recurrence. In 2022, a 10-minute waiting period was introduced into clinical practice. Therefore, patients were divided into two cohorts. 131 patients underwent a 30-minute waiting period alone. 106 patients underwent both 10- and 30-minute waiting periods. For comparison of the 10- and 30-minute waiting periods, we included only those patients who underwent both waiting periods. Waiting period assessment was performed using adenosine and or pacing. A positive waiting period test was defined by any of the following, return of pre-excitation, SVT or re-entry beats suggestive of pathway conduction, or response to adenosine that was consistent with return of pathway conduction. Demographics of the total cohort are shown in the red column. The subgroup that underwent both 10- and 30-minute waiting periods is shown in blue, included to demonstrate that there are no differences between the specific subset and the total cohort. Most patients were young adolescents with a male predominance. The majority had structurally normal hearts. Diagnoses were fairly evenly distributed among pre-excitation only, SVT, and WPW with SVT. Focusing on pathway characteristics, there was a predominance of left-sided pathways, and most had bidirectional conduction. We started with the full cohort of 237 patients to evaluate total recurrence within the conventional 30-minute observation window. Six out of the 237 patients, or 2.5%, demonstrated recurrence during the full waiting period. Of the six pathways highlighted with red stars, there were four right-free wall pathways, one septal, and one left-sided pathway that recurred during the total waiting period. All six then underwent immediate reablation, followed by a second 30-minute waiting period, at which time point none of these six pathways demonstrated recurrence. We then focused only on the 106 patients who underwent both 10- and 30-minute waiting periods. Our central question was whether recurrences that appear at 30 minutes could be detected as early as 10 minutes. Among the 106 patients who underwent both waiting periods, three accessory pathways recurred. Of these, one was a left-sided pathway, and one septal pathway. Both of those, marked with red stars, were successfully detected at the 10-minute waiting period. The third recurrence was a right-free wall pathway, marked with a blue star. Recurrence here was detected at 30 minutes, but was not detected at 10 minutes. Thus, our analysis of the abbreviated 10-minute waiting period found a sensitivity of 66.7% for detecting recurrences at 30 minutes. It's important to note that this sensitivity is based on a small number of recurrences, with the 10-minute waiting period detecting two out of a total of three, which limits the clinical applicability of this figure. Acknowledging this limitation, we observed a trend consistent with our clinical experience. Specifically, a shorter waiting period may be adequate for left and septal pathways, while a longer observation time could be warranted for higher-risk locations, like the right-free wall. We then examined 24-hour recurrence. We evaluated whether waiting period recurrence and reablation effectively reduced 24-hour recurrence. We additionally identified the timing of recurrence for pathways that recurred during the first 24 hours to determine whether there might be a possibility of identification during an extended waiting period. For this analysis, we focused only on manifest pathways and defined 24-hour recurrence as the return of pre-excitation, or SVT, on the 24-hour post-ablation holter. Of manifest pathways, nine of the 168 recurred during the first 24 hours, or 5.4%. None of these pathways had recurrence during the waiting period. Of the pathways that did have recurrence during the waiting period and were reablated, none of the six had recurrence at 24 hours. By pathway location, pathways that did not recur during the waiting period but recurred within 24 hours included three right-free wall, four septal, and two left-sided pathways. With respect to their time at initial 24-hour recurrence, patients fell into three categories. First, in some cases, recurrence was documented in the recovery unit prior to the start of the holter monitoring, representing the earliest identifiable recurrences outside of the waiting period. All three right-free wall recurrences fell into this category, adding further evidence that an abbreviated waiting period may not be optimal for this population. Second, the two left-sided pathways and one of the four septal pathways, recurrence was present at the beginning of the holter monitor. In these cases, we were unable to determine the precise timing of recurrence that would occur during the first few hours post-ablation. Finally, three septal pathways recurred later during the holter recording, allowing us to identify that, on average, recurrence occurred 9.5 hours after the final ablation lesion. No reasonable extension of the waiting period would be able to capture these recurrences. Taken together, right-free wall pathways had the highest waiting period recurrence, and 10-minute waiting period failed to capture a recurrence at 30 minutes. All 24-hour recurrences at this location occurred early in the recovery unit, together indicating that a prolonged or conventional waiting period may be more appropriate than an abbreviated waiting period in this location. In septal pathways, one recurrence occurred in the waiting period, and it was detected at 10 minutes. Although this location was also the most likely to recur at 24 hours, the mean identifiable time to recurrence was 9.5 hours after the last ablation lesion, well beyond a practical waiting period. Thus, extending the observation window beyond 10 minutes is unlikely to improve detection. Left-sided pathways had the lowest overall recurrence, and waiting period recurrence was detected at the 10-minute window. Precise timing of 24-hour recurrence was the least likely, sorry, we were unable to determine the exact timing, but it was the least likely to recur in total. Taken together, the 10-minute waiting period may be sufficient to identify recurrences in this group. There are a few important limitations to consider. First, the overall recurrence rate in our cohort was low, limiting our ability to draw definitive conclusions and to establish non-inferiority at a clinically reasonable effect size margin. Additionally, this represents our single center experiences, and these findings may not be fully generalizable to other institutions. In conclusion, with the use of modern 3D mapping and ablation techniques, recurrence during the waiting period is very low, following pediatric accessory pathway RF ablation. Because of this, extended waiting periods may not always be necessary. For left and septal pathways, a shorter waiting period may be a reasonable approach. However, in higher risk locations, like the right free wall, a conventional waiting period may still be preferable. Again, I'd like to thank you all for your time, and extend a very special thank you to all the folks here from Texas Children's, especially Dr. Mike Bruno, who started this project and really brought it to life. That's a question. I have a question. Congratulations. Wonderful presentation. Thank you. Were patients ablated with standard temperature control RF or with irrigated tip, or was it a mix of it? And can you comment on whether you think the type of RF used would have an impact on this? It's a great question, and we did review that for our population. All used irrigated catheters in our population, and I think that's one of the differences between us and the adult study. Potentially, they use non-irrigated catheters, and their recurrence rates in the waiting period were very different than ours. Time to pathway elimination. We haven't quite looked at it yet, but it's on our radar. Are you able to look at the pathway and see how it's going to recur? Yes. If you have that really quick effect. That being said, those that did recur in the waiting period did have quick effect, but they returned, so they had transient effect. It was a theme we noticed in our recurrences. So it's a great point and something we'll look at moving into manuscript, but haven't quite collected that yet. This is great data. You presented it really well, and you pointed out the limitations of the small sample size. You're clearly underpowered to make practice-changing recommendations. Absolutely. That's absolutely a great question. That's what our 24-hour data is raising. The question it's raising, we don't have the granularity in this data to know, is it 45 minutes? Is it 47 minutes? Is it 60 minutes? And it's also your risk tolerance. Some people not waiting at all. Some people would prefer to wait an hour and a half if they could not bring a patient back. So I think we take this in an individualized way, and that's certainly something that we could assess for in the future. Thank you. Okay. So, why do we have a waiting period is essentially the question, and let me know if I don't answer your question correctly. So I think this is our opportunity to modify our recurrence. It's the only opportunity we have in the lab to modify how much recurrence we're going to have. I think some people would say, I would wait longer if I didn't have to bring a patient back to the lab. Some people might say, 2.5 percent, I'm not going to wait on every patient. And so I think that's really provider-dependent by why we do a waiting period. And I was talking to some folks today who don't have to do a waiting period at all, it seems. So I think it's very provider-dependent. Does that sort of address what you're getting at, or do you want to ask your question in another way I can understand? Yeah, there's definitely variability here. It makes it a little challenging to study, but I think it's still worthwhile, and we're having a great conversation about it. It makes me really happy that we looked into this. Yeah, and we're just running a little short on time, but one overall question, a show of hands question. So Jim Perry in our lab calls this obligatory farting around time. We're waiting around. And as a show of hands, who is waiting 30 minutes after an ablation? Yeah, the majority. Hands down. Who's waiting 10 minutes? All right. And who's waiting longer? Who's waiting 60 minutes? All right, cool. Perfect. Thanks so much, Chelsea. Yeah, thank you all. I really appreciate it. Thank you. All right. All right. Next up is Molly Shaw, who's going to give us another update on some of her device work. Okay. Well, we're waiting. These were two excellent presentations. So I'm actually a little nervous. Two hard acts to follow. Well, thank you for the opportunity to present some very preliminary results of the safety and electrical performance of a novel small diameter lumenless ICD lead. That's my disclosure. All right. So I don't need to remind this audience that ICD therapy in children and congenital heart disease patients is a life-saving treatment, but is often hampered by poor long-term lead survival. Lead failures have been attributed to younger age at implant, diameter, lead design, and lead coating materials. Since the failure of two small caliber leads, and those of you who are as old as I am will remember the disastrous time we had with the Riata and the Sprint Fidelis, there have been no new ICD lead designs. In 2021, Medtronic introduced the new Omnia Secure ICD lead for clinical trial. This is a 4.7 French simplified lumenless coaxial lead design with a flexible cable conductor. The lead has a six centimeter defibrillation coil. It has bipolar sensing. The lead is catheter delivered and compatible with multiple available catheters. Seen here is the lead implanted as part of a CRTD system, and you can see that the Omnia Secure lead size is even thinner than the seven French atrial lead. The Omnia Secure lead was then evaluated in the worldwide pivotal leader trial. But just to let folks know who are familiar with implanting the Select Secure lead, that this ICD lead essentially leverages the 3830 pacing lead design, and as most implanters know, has an excellent track record for long-term durability and electrical performance, not just in adults, but also in pediatrics. Now, in contrast to the Select Secure pacing lead and the Omnia Secure ICD lead, other commercially available leads have a much more complex lead design. It's essentially a multi-lumen assembly. Many factors of this lead to lead failures and decreased longevity. So additionally, learning from the Sprint Fidelis experience, the Omnia Secure ICD lead actually underwent an advanced cardiac lead reliability model, and I want to just explain this model in a little bit of detail. This model has two main inputs, an in vivo clinical imaging data analysis with 3D reconstruction of the center lines of where the leads bend, and also an in vitro testing of the lead for analysis of fatigue strength. These tests are conducted to replicate the bending cycles due to cardiac and patient motion experienced over a projected 10-year period. So this would include testing for 400 million cycles to represent 10 years of cardiac motion at 78 beats a minute, and 5 million cycles to represent 10 years of arm motions for a population of highly active patients. In fact, they actually tested this lead with the pediatric and adolescent population in mind when they bench tested this lead and augmented their testing for lead bending and heart rate by 20 percent. So after the failure of two thin leads over a decade ago, one might wonder, why do you need another ICD lead and another thin one? Well, the goal of this lead design is actually to improve lead durability, and for older adults it would be a lead for life, reduce lead complications, and also enhance targeted ICD lead placement. And then for the smallest patients with small vasculature, this lead would be extremely suitable. So I'm going to very briefly go over the leader study. Our institution was part of the pivotal clinical trial for which the inclusion criteria were patients over the age of 18 years who met de novo ICD implantation criteria. And briefly, over an 18-month period, there were 643 leads implanted worldwide. The trial was completed in November 2022, and the clinical results of the trial actually exceeded the study defibrillation efficacy and safety goals with 97 percent defibrillation efficacy, and there were no lead fractures at 12 months. So coming back to the present study, the aim of our study was to evaluate implantation success, safety, and electrical performance in patients under the age of 18 years. So we conducted a retrospective review of pediatric patients implanted with the Omnia Secure ICD lead, and this was done after obtaining local ethics approval, manufacturer agreement, and compassionate use approval from FDA under their expanded access medical device pathway. All patients were implanted at participating leader clinical trial sites only. All patients underwent the study protocol. All patients had procedural imaging, pre-procedural imaging with an echo, and intra-procedural venograms. The lead delivery catheter was selected by the operator. Electrical parameters were checked at implant, and then at each follow-up visit, and as well as through remote monitoring, and defibrillation testing and ICD programming was per operator's discretion. So this is just an example of the fact that this is an integrated bipolar lead. So once the lead is implanted, you have to pull back the catheter and expose the whole coil for electrical parameters. So the results of this case series was that there were six pediatric patients from three centers who had the Omnia secure lead implanted. The median age at implant was 14 years, the youngest patient was 8 years, the median weight was 52 kilos, the smallest patient was 25 kilos. There were three primary prevention ICD indications and out of the devices implanted, two were CRT defibrillators, one dual chamber and three single chamber ICDs. At median follow-up of 15 months, the longest being 27 months, there were no inappropriate shocks. Two patients had appropriate shocks. One patient had T-wave over-sensing, there was no P-wave over-sensing and no complications. Electrical parameters were stable at implant as well as at follow-up. So because this is a case series, I thought I would just go through each case and discuss some important points. The first case was a young patient, an 11-year-old small patient with Timothy syndrome, status post cardiac arrest, who had an epicardial dual chamber ICD with coil failure. In this patient, the patient's only 28 kilos and in our institution this would be sort of a borderline weight and size for a transvenous lead implant, but we felt that with the Omnia secure lead, the lead was implanted successfully in the mid-septal area. Even though the patient was small, we were able to get the coil completely in the RV. Most notably, this patient had very long QT intervals, but there was no T-wave over-sensing on this lead. Second case is a patient with dilated cardiomyopathy with syncope and AV node dysfunction. This patient had an epicardial CRT system from the age of seven, now needing upgrade because of a lead fracture. And again, this patient, we felt that the Omnia secure lead would be a very good option because we would be putting three leads in a transvenous space of a small patient. And in this patient, also the RV anatomy was challenging. The RV was extremely vertical because of the dilated LV and because this is a catheter driven lead, you can actually implant it wherever you want. And so we were able to get a nice low septal position. Third case is a more complex patient. This was a CCTGA double switch patient with post-operative heart block, not congenital, who had an epicardial dual chamber pacemaker with atrial lead failure and left ventricular ejection fraction of 20%. The patient was 100% ventricular paced. So we decided to upgrade the system to a CRT, but decided to do a hybrid. And we maintained the left ventricular, so the anatomic left ventricular lead. And again, we were able to pass this thin lead through a sending baffle and actually target the best site, which gave us the greatest distance from the LV lead. So again, an excellent choice for targeted lead placement. The next case is just a dilated cardiomyopathy with a very big heart. The patient was young, but her heart was huge. So there's absolutely no problem getting the whole coil in the RV. The fifth patient is a 15-year-old with multiple epicardial and transvenous lead fractures. And therefore, this lead was really selected for the lead durability. But at the time of implant, it was found that the right ventricle cavity was completely obliterated by hypertrophy. And with a traditional lead, it would have been very difficult to even pass the lead in the RV cavity. But with a catheter deliver system, it made the procedure much more easier. This patient then went on to have an appropriate shock in the follow-up period. And last is a young patient, small patient with Brugada syndrome. This is a secondary prevention ICD case. And you can see that this patient, the only way to get the whole coil in the RV was to do an RVOT implant. So in summary, the Omnia Secure ICD lead was implanted successfully in all patients. The catheter-directed lead positioning allowed for targeted site implantation. There were no procedural complications. Due to the larger coil, incomplete placement in the RV cavity is a concern, but none of our patients had P-wave over-sensing. And this lead may be a good option for select pediatric patients. Limitations are obvious. This is just a case series. Pediatric patients were not actually included in the leader trial, and we don't have enough follow-up time to make any more comments. But exciting. The Omnia Secure lead is now being trialed for a left bundle branch area position so that we can get conduction pacing and defibrillation with the same lead. And this is currently undergoing a global trial. And hot off the press, this lead just got FDA-approved, so might be worthwhile checking it out at the Medtronic booth to see if you like it. Thank you for your attention. Thank you, Molly. You stole the thunder of a couple questions asking about conduction system pacing. One question we have from the audience is, were any of the Omnia Secure leads that initially had the entire coil in the RV at implant have a change in lead configuration such that the coil moved back into the RA on follow-up? No, no. Do you mostly use the C315 sheath, or did you have to use a C304 lead to get to the apical septum? I would say that five out of the six is a 315 catheter. The other one is the patient with a very dilated heart is the only patient where a C304 was used. And then I have more and more questions popping in, and I think you're gonna get mobbed after this talk, but we're running out of time. So one more question. Do you worry about coil adherence to the tricuspid valve and future extraction as we attempted to implant in small kids? Yeah, that is an excellent question. So for the Select Secure lead, there are very well described extraction strategies. In fact, for people who ultimately are going to use this lead, I highly recommend just reading this paper, which really describes the various mechanisms by which you can extract a Select Secure lead. And I think if we're going to implant this lead, we should all learn a good way to extract it. Now, this lead has a defibrillation coil, so I anticipate that there is going to be lead adherence, but I think it, just like any transvenous lead, which is 9 French with a greater coil area, I would think that this would be safer for tricuspid valve adherence, because the surface area is so much lower. There is an abstract at APHRS on four leads that were extracted from the initial leader trial, but the dwell time is only two years, so I don't know what really you can determine from that. Yeah, you can extract away from that. Okay, all right. Thank you very much, Molly. Thank you. Lastly, we have Joey Needleman from Emory joining us with a talk on leadless pacing and transvenous pacing in the young. Come on up, Joey. Great. Thank you all for joining me. My name is, as mentioned, Joey Needleman, and I'm here to talk about our research, Ready for the Big Time, comparing early outcomes of leadless and transvenous pacing in the young. As noted, I have no conflicts of interest to share. So, leadless devices have been around since the early 1970s, but were only first approved by the FDA back in April of 2016. Since that time, the pediatric population receiving these devices has steadily grown, or, in a sense, as we've become more comfortable placing these in smaller and smaller patients, shrunk as well. Back in 2021, PACE's expert consensus did recognize that there are knowledge gaps about these devices. However, since then, there have been limited publications in the pediatric and young adult population. Dr. Shaw et al., back in 2023, published a series of 62 micropacemakers in a multi-center pediatric and young adult population, showing that these devices are both safe and effective. Overall, compared to a transvenous system, these devices have numerous potential benefits, including reduced long-term vascular and baffle obstruction, possibly reduced AV valve injury and risk of endocarditis, and, given the lack of a pocket, lack of leads, improvements to quality of life and activity. There are, of course, some potential risks and barriers to these devices, including vascular injury, given the size of the catheters required, limitation to technology, particularly in the pediatric population, and limits to battery life, and, given our lack of experience, extraction and possible risk of intracardiac clutter if we place multiple devices in the pediatric patient. So, what we intended to do was look at the intra- and post-operative outcomes of lupus and transvenous pacemakers in this pediatric and young adult population. We did this study at the Children's Healthcare of Atlanta, a 330-bed hospital in Atlanta, Georgia, while the hospital predominantly focuses on the Atlanta population, given the fact that we're part of one of the largest pediatric health systems in the country, we do largely cover Georgia as a whole. All patients who were aged 8 to 25 who received a Lelus pacemaker were included, and we matched them to controls who had received a transvenous system. We looked at patients who received these devices between January 1st of 2016 and March 30th of 2024. We'd also have an additional arm not reviewed here, looking at the ACHE population at Emory, also located in Atlanta. Our Lelus patients were age-matched to transvenous patients at the time of placement. We matched them in a one-to-two ratio, so for every one Lelus patient, we matched them to two transvenous patients and grouped them into 8 to 11, 12 to 17, and 18 to 25-year-old age groups. We did try also, I should note, to match them to calendar year placement. If we weren't able to, we focused on devices based after and then before to better align with provider experience. Data collected included the BSA, BMI, sex, race, and ethnicity of patients, medical history including arrhythmia and history of structural heart disease, as well as prior device history. We looked at access, personnel present at time of placement, site of implantation, and procedural timelines, and then we looked at procedural and post-procedural device settings and readings throughout one year post-placement. Our statistical analysis included a chi-square test and Fischer-Xas test for categorical variables and two sample t-tests for continuous variables. We used a standard p-value and this analysis was supported by the Pediatric Biostatistics Corps. Both the Emory and Children's Healthcare of Atlanta IRBs reviewed and approved this study. So, diving into the data, we ultimately analyzed 20 patients with Lelus devices matched to 40 transvenous controls. Our patients were similar in age, weight, and BMI, as well as sex, race, and ethnicity, which I will just cautiously note came from the medical record and was not through self- identification. 15 of our Lelus patients fell into the 12 to 17-year-old range. Transvenous patients, as noted here, were more likely to have a history of structural heart disease, which was predominantly ASDs, VSDs, or AV canals. Two of the patients reviewed did have a history of single ventricle, however, both had undergone a heart transplant prior to being included in this study. In terms of indications, device indications were similar and predominantly related to either sinus rest or AV block. Not noted here in the chart, but I think relevant, a majority of both groups were either, had either history of complete congenital heart block or surgical heart block. Patients receiving transvenous devices were notably more likely to have a history of prior device, prior devices, as well as prior complications. The majority of our Lelus patients had micro-AVs, with 20% or four of the patients having either a second-generation AV or VR system. The vast majority were placed via percutaneous access. Only one patient required a cut down for vascular access. The median procedure time was significantly shorter in the Lelus population, while fluoroscopy time was similar. Procedural complications were also similar. The one documented procedural complication for the Lelus patient was a self-resolved retroperitoneal hematoma. Similarly, through 12 months, complications were similar, though notably there were no Lelus complications documented. At lead dislodgement within 30 days of placement was the only complication requiring re-intervention. Patient data was collected at time of placement, at discharge, and at their 30-day, one to four-month, five to eight-month, and nine to 12-month follow-ups. At discharge, Lelus patients were significantly less, had significantly less ventricular pacing. Battery life was similar between both groups, and the mean capture thresholds and impedances were similarly within goal for both. Lelus devices continued to have similar ventricular pacing at each subsequent visit until, as noted here, their nine to 12-month visit, though Lelus patients still trended towards less ventricular pacing. Mean capture thresholds and impedance remained similarly within goal for both groups. So, in summary, compared to a transvenous system, patients receiving Lelus devices were less likely to have a history of structural heart disease, they had shorter procedure times, they had similar fluoroscopy times, they had similar right ventricular pacing at 12 months, they had similar estimated battery life at 12 months, and finally, they had similar procedural and post-procedural complications, though, as noted, the transvenous population had the entirety of the complications through the 12-month follow-up for our patients. We were limited in this being a single-center study. Our patient populations were not equivalent, as the transvenous population tended to have more devices and history of congenital heart disease. We did have patients lost to follow-up or transition to the adult providers, so only 11 or 55 percent of the Lelus patients and 32 or 80 percent of our transvenous patients were able to be reviewed for the study at 12-month follow-up. And again, we only focused on the first 12 months after placement for these both groups. However, I think that our data shows that Lelus devices are a very safe and reasonable option when compared to transvenous devices. There are some technical advancements that will continue and potentially make these even more reasonable options for certain patients, and as we go along, we'll understand better the potential harms and benefits to these devices when compared to the transvenous group. So, thank you all for joining me. I'm just ending with a picture of my sons, Oliver and Henry, who decided to join the world two months early at the beginning of this month, and then our son, Noah, who can't wait for them to come home. So, thank you all. Thank you, Joey, and you're a trooper for coming here and presenting this important data with your kiddo who's thankfully doing now well. I've got a few people in the audience who read my mind and were asking about AV synchrony and the micro-AVs, and so I noted that you said that nobody was discharged with atrial mechanical sensing turned on. Did you attempt to program atrial mechanical sensing in any of those micro-AV patients, and what were the pressure points for that? So, that's an excellent question. I actually don't... we did not collect that data specifically in this, but I'll pass along to Dr. Whitehill who would be able to answer. Did you hear him? Not too well, sorry. I didn't know if you were able to translate Whitehill for us, but okay. We'll come chat later, and then one question from the audience. Yeah, so that's a great question, and I think that, to be transparent, I don't believe I'd be able to comment on that specifically. We didn't have a specific criteria that we were seeing in the documentation that we reviewed for this study. All right, thanks again, Joey. Thank you all. And thank you all again for joining for the early morning session, and thank you to our speakers, and have a wonderful day.

cardiology

children's health

artificial intelligence

Wolff-Parkinson-White syndrome

post-ablation waiting period

ICD lead design

conduction system pacing

leadless pacing

pediatric cardiac care