I think we should just get started. It's about 8 o'clock. So welcome to this oral abstract session. It's nice to have you all here this morning, bright and early. This is a session on physiology, conduction system, pacing, outcomes and arrhythmia management in heart failure patients. The way, if anybody has any questions, I think post, we'll have a couple of minutes for each subject. If people can actually, they'll have to come up here to ask their question into this microphone or else won't be able to distribute the question. There's also the Ask a Question app in the Heart Rhythm HRS 2025 app. All right. So it's my great pleasure to start and introduce the first talk, which is entitled Invasive and Noninvasive Assessment to Determine the Relationship between Patient-Reported Symptoms and HFPEF in Patients with Symptomatic AF. And the speaker is Dr. Jonathan Ariyaratnam. Go ahead, Jonathan. Can you hear me okay? Well thanks very much for the introduction. It's my pleasure to present this data from Adelaide in Australia. I don't have any disclosures for this presentation. So it's now well established that atrial fibrillation and heart failure with preserved ejection are very closely related cardiovascular conditions, and in fact they've been described in the literature as vicious twins. And large epidemiological data shows that the prevalence of AF in cohorts of patients with HFPEF is very high, and similarly the prevalence of HFPEF in patients with AF is also particularly high. In addition, we know that the development of each of these conditions potentiates the development of the other. And these processes are driven by the traditional cardiovascular risk factors which underpin both conditions. Confirming this, we in Adelaide have performed some invasive hemodynamic studies to confirm a high prevalence of patients, a high prevalence of HFPEF in patients with AF. We took 120 AF ablation patients and found that the prevalence of HFPEF based on established hemodynamic criteria was almost 75%. Importantly, we found that this was clinically very significant with an increased burden of AF symptoms in the patients with HFPEF. However, what we don't know is what the precise nature of those symptoms are and whether patients with HFPEF present differently compared to patients without HFPEF in an AF cohort. So the objective of this study was to establish the difference in presenting symptoms between patients with and without HFPEF both before and after AF ablation. So we studied consecutive patients with symptomatic paroxysmal or persistent atrial fibrillation who were due to undergo an AF ablation procedure. And we specifically excluded patients with a reduced ejection fraction, a previous diagnosis of a cardiomyopathy, or moderate to severe valvulopathy. We assessed patient-reported symptoms using the validated AF symptom severity questionnaire. This questionnaire assesses seven symptoms, two of which are exertional symptoms and five of which are non-exertional symptoms. And each of these symptoms were reported by the patient on a scale of zero to five according to how often they had been bothered by that symptom over the preceding four weeks. And we assessed these symptoms both four weeks before an AF ablation procedure as well as 12 months after the AF ablation procedure. We diagnosed HFPEF as we previously reported in the literature using invasive hemodynamic assessment of the left atrial pressure. So following transeptal puncture, we placed a sheath in the left atrium and monitored left atrial pressures. And HFPEF was diagnosed when mean left atrial pressure was greater than 15 millimeters of mercury either at rest or after the infusion of 500 mils of saline. And the remaining patients were diagnosed with no HFPEF. And following the diagnosis of HFPEF, we went on to perform the AF ablation procedure. And this involved isolation of the pulmonary veins as well as the left atrial posterior wall using a single ring isolation technique with radiofrequency ablation. So in total, we screened 172 patients for inclusion. 39 were deemed ineligible according to pre-specified exclusion criteria. Of the remaining 133 patients, 13 declined participation, and a further 13 were unable to satisfactorily complete the questionnaires. So in total, our study cohort included 107 patients. And to date, 57 have completed follow-up at 12 months. Of these 107 participants, the vast majority had HFPEF with 78 demonstrating hemodynamic features of HFPEF and 29 showing no HFPEF. At 12 months, 18 of the no HFPEF patients and 39 of the HFPEF patients are returned for assessment. There were no differences in baseline characteristics between the two groups. And there were also no differences in cardiovascular risk factors between the two groups. In terms of symptom prevalence, we found that the most common symptoms in this cohort of symptomatic AF patients were exertional symptoms, exertional dyspnea, and exercise intolerance. The next most common symptom was palpitations. And this was followed by lightheadedness, resting dyspnea, resting fatigue. And the least prevalent symptom was chest pain. We then compared patients with HFPEF with those without HFPEF and found that exertional symptoms were significantly worse in the HFPEF group compared to the no HFPEF group. We also found that resting dyspnea and fatigue were also worse in the HFPEF group compared to the no HFPEF group. But interestingly, there were no differences in lightheadedness, palpitations, and chest pains between the two groups. Following ablation, we found that the exertional symptoms did improve with AF ablation, as did symptoms of resting dyspnea and fatigue, and palpitations and lightheadedness. Chest pains didn't improve statistically, although this is likely related to a low burden of chest pains at baseline. And as you can see, the symptom prevalence improved across all symptoms, but there was a discrepancy in the amount of improvement in prevalence across the symptoms. So palpitations and chest pains improved by more than 50%. Lightheadedness, resting dyspnea, and resting fatigue improved by around 40%. The symptoms that improved the least with AF ablation procedures were exertional dyspnea and exercise intolerance. And again, when we compared post-ablation symptoms between the HFPEF groups and the non-HFPEF groups, we found that exertional dyspnea and exercise intolerance were greater in the HFPEF group compared to the non-HFPEF group. Same was with fatigue, but resting dyspnea, palpitations, lightheadedness, and chest pains were not different post-ablation. So in conclusion, we've shown that exertional symptoms are significantly worse in patients with HFPEF, both at baseline and after ablation. AF ablation does improve symptoms right across the board, including exertional symptoms, but after ablation, there remains residual symptoms of exertional dyspnea and exercise intolerance, suggesting that these patients remain symptomatic in some way after the ablation. So exertional symptoms and fatigue are more prevalent and more severe in AF patients with HFPEF both before and after AF ablation. And this raises the question as to whether additional therapies targeted at HFPEF may be helpful in these patients with AF and HFPEF even after an AF ablation procedure. Thank you. That was excellent. Any questions from the audience for Jonathan? So I'll start by just asking one quick question. Was there any follow-up echo sort of parameters that you used in addition to the hemodynamic assessments? And do you think that some of the explanation is related in any way to the spectrum of HFPEF within some of these patients? The more severity, obviously, of the disease may correlate with the lack of improvement in symptoms. Yeah. So we have a lot of echo data, quite advanced echo data. We do exercise echo on all these patients. So we have all that information. What we found is that patients with HFPEF have much more advanced left atrial disease. That seems to be the driver of HFPEF in these patients. What we don't know, but what we're going to try and find out is what happens to the left atrial function after ablation. So although the symptoms improve with ablation, we don't actually know whether left atrial function is improving or not after ablation. So that will be the next step of our analysis. And was there any correlation with AF recurrence or AF burden post, or was that looked at? None of the patients presented here had AF recurrence post ablation. So all these patients were in sinus rhythm when they had their symptoms assessed post ablation. But again, we have a cohort of patients who did have recurrence, and we'll be assessing that moving forward to see what their symptoms were like. Thank you. Okay, that was excellent. Okay, moving on to our next discussion. This will be Dr. Douglas Packer, who is well known to everybody in the audience. He'll be talking about the long-term mortality and heart failure hospitalization in reduced ejection fraction versus preserved ejection fraction patients results from a 61,000 patient registry. Dr. Packer. I appreciate two things. One, everyone's awake. And the second thing is, we're doing it with these things. I mean, who got that together? I have potential disclosures. Okay. So we're really looking at long-term outcomes, particularly with mortalities and heart failure, hospitalization. We're looking at HF-PEF and HF-REF. We started with 61,000 group of patients. And I think what you can do is you can take a look at the literature over the last couple of years, the last 10 or 15 years. There's been an apparent increase in the prevalence of heart failure with an accompanying increase in morbidity and mortality, regardless of whether it's HF-PEF or HF-REF. And if you look at preliminary studies, they'll tell you that there's probably also a shift in phenotypes. So whereas there were a larger number in HF-REF, that seems to be shifting to HF-PEF. These findings maybe are limited by the number of patients, but in this kind of a registry, we get a little bit more information. And it also remains unclear what this shift in overall risk might be because of changes to more HF-PEF. So, what we wanted to do is we wanted to validate the apparent appearance of HFPAF over time, and we're looking over a 20-year time period. We wanted to determine if the degree of the shift that's occurring in heart failure phenotypes alters the occurrence of atrial fibrillation. We want to know if we can establish the impact of such changes in terms of morbidity and mortality and determine the overall impact of AF ablation in this patient population. What we did is we worked with the Intermountain population, or registry, and there's 61,000 patients in it. And we looked at patients from January 1st, 2000, up to 20 years of 2019. These patients had to have survival for at least 30 days over time or they're excluded. If you look at the patients in the next group, if you look at the patients where we knew what their ejection fraction was, they had to have had that ejection fraction within three months of the heart failure diagnosis. So, that kicks out 32,000, but it still leaves us with 29,000. And then we wanted to know that we had a robust group with patients with atrial fibrillation actually at the time that heart failure was diagnosed. If that wasn't the case, they were excluded, and that gets us down to 12,000. So, we're really looking at 12,000 patients, but in a registry, 12,000 patients is not bad. If you look at the inclusion criteria, then you get a fairly good idea that it's patients with new heart failure, they have to have that EF, they have to survive, and they have to have AFib at the time of diagnosis. If you look at the phenotypes, we cut at 40 and then 45, less than 40 is HFREF, greater than 50 is HFPEF, and then the middle is in the middle, and we're looking now at five-year time blanks between 2000 and 2004, all the way across. Now, the thing we saw is if you looked at the entire group, then it's about 51% HFPEF and about 30% HFREF, and then there's the middleman of the 16% HFREF. As time went on, what happened is it increased fairly significantly to more patients that had HFPEF. Now, if you want to include HFMREF in the HFPEF patients, then you hit about 60%. There's a decline in the numbers of HFREF patients, but if you were to just look at HFMREF patients alone, that stays pretty much the same. If you look at patient population and characteristics, there's a difference. HFPEFers tend to have more of the comorbidities than do HFREFers, and the significance was fairly high. I don't think we need to spend a lot of time with that. We've done it in a couple of other presentations. Let's look at death. So, this is death in HFREFers over time, and it goes from about 39% to about 29%. That's at three years. At one year, it's kind of 20 to 10%. With HFPEFers, then the mortality at three years was actually higher, pretty much across the board. It dropped. How did it drop? Well, if you take guideline-directed medical therapy, and you treat with that, and you've got four pillars, and you're treating with that, then probably that changes a little bit when you get to the last five years, but not before that. If you look at one-year accounts, then HFPEF still has a bit more of a problem with death. Heart failure hospitalization, which I think is perfectly difficult to quantify, but in HFREF and in HFPEF, the numbers were lower. There's more heart failure hospitalization in HFREFers, and you can see that goes from about 26% to 18%. If you look at HFPEFers, then we're down between 22% and 12%. The HFPEFers still had an appreciable hospitalization problem, but the bigger problem really is with what's going on with atrial fibrillation and what's going on with absolute mortality. You look at the occurrence, there were more HFREFers ablated over time, over the three-year time frame, up to 16%, about 9% over time, but increasing in both cases with the HFPEFers. It turns out that this correlates pretty well with how the patients did in terms of a decline in atrial fibrillation, which we presented yesterday. If you're looking at hazard ratios, then here are the HFREFers and the HFPEFers, and all of them have fairly significant changes. We're looking at a HR of 0.47, but if you go all the way across the board, then you see with three-year death rates and heart failure hospitalization, you go from the null line, dropping down, and then doing that, heart failure hospitalization doesn't do a lot, but if you look at the morbidity and mortality, the HFPEFers had a greater hazard ratio consideration, and 5-95% confidence intervals were also fairly tight compared to HFREF. About any way you look at this, HFPEF has a problem. We have a problem, because we don't have that much information about them. Now, if you look at the summary, then over the 20 years, HFPEF prevalence increased in this kind of real-life setting, HFREF decreased, patient health characteristics were different. Death rates in three years in HFPEF patients tend to be higher than in HFREFers, but both declined over time, and again, that may be guideline-directed therapy, but it also may be the degree of ablation. Heart failure hospitalization was kind of even-keeled, and I don't know that much needs to be said about that yet, but the comments that were made about quality of life and symptom burden comes in here, and it's quite important. So HFPEF hospitalization was less likely than HFREF, and over the 20 years, then it drops down, but the mortality did decrease dramatically in both if you got the atrial fibrillation part in there, but the patients that were the most surprising is what was going on with the HFPEFers. So we conclude that these data show a significant shift in HFPEF compared to HFREF over 20 years. Guideline-directed medical therapy, atrial fibrillation ablation, a number of things that are important here, but I have to say that we don't know this story completely. AF ablation probably increases, and another paper that we did shows that, but I think there's enough of a question about all of this, that we really need to be poised to do a very large randomized clinical trial, and the one that we're putting forward is the Cabana HFPEF or Cabana HF. That's what it's going to take. Registries are great. They give you kind of a ballpark idea of what's going on, but it's going to take that, like Cabana, to give us the straightforward answers. Thank you. Any questions from the group for Dr. Packer? So Doug, that was really insightful. In terms of the, I'm impressed actually with the degree of not only, okay, improvement in HFPEF over time, even though we don't have great guideline-directed medical therapy, but maybe the underutilization of ablation in that patient group, that there were more ablations being performed in the reduced EF population, and what do you think accounts for that, and do you think there's some degree of maybe unwillingness to ablate? Was it chronicity of atrial fib, or, you know, because these were new heart failure diagnoses with AF, so I'm just curious about that. So important points, everyone. If you looked earlier on, both in HFPEF and HFREF, the morbidity and mortality considerations were greater, dropped off, but there still was a major problem with what's going on at the tail end, and I think that's kind of the atrial fibrillation piece. But when you go backwards and see, I think we did a better job of ablating HFREFs. You know, we've been doing that, you know, if you look at Castle in the Cabana, we've got a 10-year history of doing more in the way of ablations. You have to be careful. That can't be the only reason, but when we get a little bit further down into the HFPEFers, then I think that they do better, but you can't blame it on guideline-directed medical therapy. Which guideline-directed medical therapy do you want to call? SDLT2 inhibitors or GLIPS, or whatever you want to. Right. Okay. Any other questions from the group? Okay. Excellent. Thank you. All right. Thanks for being here, everyone. My name is Anshul. I am a resident Mayo. I'm an incoming cardiology fellow as well. And today, I'd like to talk about the work we've did with the AI ECG, specifically looking at parameters generated by this. In this case, the pre-implant age obtained from an AI-driven ECG or AI-generated ECG and how it can predict response to cardiac resynchronization therapy with defibrillator therapy. As means of an introduction, we know from the past studies so far that predictors of improved survival following CRTD therapy are a female gender, a left bundle branch morphology of the QRS duration, and absence of comorbidities such as atrial fibrillation, advanced kidney disease. Prior studies have been variable in terms of the impact of chronological age at implantation, however. What is clear is that depending on what parameter you're looking at in terms of response, there is about a third of folks that are non-responders, and there is a clear need for improved patient selection. I'm going to jump a little bit here to discussing age. Chronological age, as we know, is just the number of years you and I have in a life. However, there is a concept of biologic age, which is, as best as I can say, the state of someone's overall health in comparison to your average person of the same chronologic age. It encompasses genetic, epigenetic factors, physiologic health, the diseases you have and the treatment for them, and so on and so forth. Our thought was, could this biologic age predict response to CRTD therapy? The question you should ask at this point is, how do you measure biologic age? In the past, the work that has been done includes cellular-level parameters that have been used. You're looking at different genes, like DNA methylation, and so on and so forth, but based on the work that our group had done before, we proposed that the AIECG can be used for this. Using convolutional neural networks, the AIECG can predict an age. The Mayo Clinic AIECG, it was trained on 400,000 folks, it was then tested on 100,000 folks, and it revealed that it could predict an age with an absolute error of about 7 years plus minus 5.6 years, and our score at a point was 0.7. One would say this is just the error, but the thought that was proposed at that point is that that's not actually an error. The AIECG is predicting something that is different than your chronologic age. We know that ECGs change over time. There's different features that can be extracted from them that may be a result of your usual aging and getting diseases, and so the hypothesis was that the AIECG predicted age may be a marker of cardiovascular or biologic age. To further validate this in past studies, delta age was calculated as your chronologic age minus your AIECG age. In theory, if you were healthier than someone else your age, your delta age would be higher, and vice versa if you were less healthy. Our thought was that this could be used to predict multiple things. In this case, the studies that have been done in the past showed that a lower delta age, which is cases in which you're less healthy than folks your age, was shown to predict higher all-cause mortality, higher cardiovascular mortality, and markers of accelerated aging on the genetic level as well. In addition, folks with a lower delta age had a higher burden of hypertension, CAD, and folks with a higher delta age had low coronary artery calcium scores as well. There was something there. What we tried to do then in our methods was just look at this retrospectively in folks with an EF less than 35% and QRS duration less over 120 who underwent CRTD, the thought here being that if these folks have a higher delta age, presumably, they would respond better to CRTD in terms of survival following CRTD therapy. What we did was through the pre-implant ECG, we were able to get the AI ECG age, we were able to get the chronologic age, and then we calculated delta age. Survival analyses were conducted using accelerated feral time models. We used the delta age and baseline characteristics, and variables were chosen with stepwise selection, so both forward and backward selection. This is our table one, pretty loaded slide, but you'll see in the column to the left, there is delta age less than zero, which means unhealthier folks compared to your age-matched cohorts and delta age greater than zero. What I really want to point your attention to at this point is the fact that folks with delta age greater than zero, which means they were healthier compared to age-matched cohorts, they actually had a higher burden of comorbidities to begin with, probably a result of the fact that they were simply older. Now I'm going to jump onto univariate analyses. I'm going to direct your attention to the Kaplan-Meier curve in the top left. The blue line is delta age less than zero, or folks that are unhealthier compared to their age-matched cohorts. Paradoxically, they actually had a higher survival compared to those we thought were healthier. We can see in the table below, which includes accelerated failure time models as well, I'm going to briefly explain time ratios. Time ratios are kind of like hazard ratios, but the exact opposite in that a time ratio greater than one predicts survival. Time ratio less than one predicts worsened survival. So you can see in this case that chronological age, increasing chronological age at the time of implant predicted lower survival, and an increase in delta age, which means as you get healthier and healthier compared to age-matched cohorts, it actually predicted worse survival, contrary to what we thought. And then, thankfully, multivariable analyses came to the rescue. I'm going to again refer to this slide here, again, pointing towards delta age. So firstly, I'll say that the variables listed on this slide are the variables that the model selected through stepwise selection. Delta age was one of them. We see here that actually increasing delta age is predictive of improved survival. What we're seeing in this case is that the lower the AIECG age compared to your chronologic age, the longer you survive post-cardiac resynchronization therapy, as opposed to what we saw in the univariate analyses. Each year, increase in that delta age, which means every year the AIECG age is getting smaller, you survive by about 4% longer. My theory for the improved benefit or for the change in findings compared to the univariate analyses is that now we're accounting for both chronologic age and delta age, which are unique predictors. We also know from the baseline characteristics that folks with a higher delta age were just older in general, and multivariable analyses allows us to account for both of these variables. Other predictors of survival were similar to what's seen in prior studies. We had non-ischemic cardiomyopathy, lower chronologic age, and absence of advanced kidney disease and hypertension. To conclude, what I want to leave you with is that the pre-implant AIECG may be a marker of biologic age. The lower that AIECG age compared to chronologic age, the longer survival post-CRTD. And then the question is where we go from here. Our hope is using these sorts of metrics with blood markers, with imaging markers, can at some point allow us to have a more holistic picture of someone's CRTD candidacy. So a lot to be done, but very interesting findings with this study. And with that, I'll open the floor to any questions. Any questions for Dr. Gupta from the audience? So Dr. Gupta, I'll start just by asking one sort of quick question. As you look into the future now, and maybe give us from a practical perspective in your own center, is this use of the delta age, are you implementing that in any way clinically at this point, or is it only a research tool? In terms of clear clinical guidelines that are provided, no. What we have at Mayo, there is, every time an ECG is done, we have an AIECG dashboard, which does provide us with an assessment of the patient's age, ejection fraction, probability of aortic stenosis, things that have been validated in prior studies. In terms of using it, however, there is no guideline that we follow so far. So you just sort of, you note it and have it there, but it's not used clinically. Okay, excellent. So without any other questions, I guess we can proceed to the next talk. It's my pleasure to welcome Dr. Sandeep Gulati. He's going to talk to us about direct from ECG ventricular activations, so VATs, monitoring in heart failure management in CRT patients. Dr. Gulati. Thank you. So I'll continue again where Dr. Packard and Anshul left in terms of understanding why we're seeing such a high mortality in HPF population and also age, right? So we've been doing the same thing but looking at conduction changes. So let's see. Yeah, and again, you know, my colleagues are here to work with Joseph Yu and Dale from PayBridge. Okay, let's take a minute and look back at VAT. I'm sure you're all familiar. Again, it's a deceptively simple metric, okay? All it means is calculate the onset of a QRS to the peak of the R wave in a precordial lead and you can calculate the conduction changes. It's very easily accessible. But having said, but the prolongation of VAT is clinically extremely actionable. And again, you know, in terms of as both a marker for heart failure diagnosis for progression, it's on many pathways from AFib all the way to heart failure, including sudden cardiac death. And again, it's also a treatment variable for, again, diagnosing or identifying candidates for CRT. And again, and tracking CRT outcomes in patients. So what I've seen is by the time people get CRTs, either it's too late, they've had a conduction anomaly for a long time, or again, as we'll talk about the data, we're seeing a large fraction of the CRTs are not tuned. Now the negative consequences of a poorly tuned CRT are worse than any positive good it's going to do for you. And because it accelerates, again, for degeneration in heart failure patients. And for those who have early low burden arrhythmias, it's going to turn them into higher burden ventricular arrhythmias. So being able to track it robustly is clinically very useful. And again, also a lot of the time you hear, again, many of you are treating clinicians, people complaining, even with CRT, two, three, four months later, they still don't have enough, they don't have energy, still have all the same symptoms, cardiac output is very low. And a lot of it comes down to when the CRT is actually working against you. Now as simple as it is, this is the range of the technology we have today. Again, the state of the art is that in the last 60 years, you can get to it, at least you can get to the number. The analytic, it takes very little to get there. You can take a 12 lead ECG and you can get a good estimate of global VAT, which is, it's again qualitative, but not clinically usable. Or, again, in the last few years, body surface mappings come a long way, but I don't know how many of you can get to a 64 electrode system on a regular basis. And then we get to halter monitoring and all the variables, again, even if you could get to a global VAT, the quality is very poor. There's so much excursion of data that there's high variability beat to beat, that the number again becomes non-actionable. And then, you know, that leaves us the catheter, which, so again, the focus for us has been, we have a different halter device, I'll talk in a minute. Can we get to, again, a VAT, a conduction distribution, which is, again, comparable to what one would get from an invasive, from a catheter, and get it with the ease of a 12 lead ECG? Yeah. It's again, so what I'm going to be talking about, we have a number of trials going on, including a pivotal, actually, what started the work is for using VAT. It's a marker component for us to get to non-invasive ejection fraction for ECG, so, yeah. I'll just take a second, and if you haven't seen, this is a device we have, it's just a core device. Basically, you can wear it for seven days, but in the study, and for a purpose of VAT and for core insight, we have a 30-minute protocol, where our subjects are wearing it for 30 minutes, and we ask them to sit for 15 minutes exactly like you, like many of you are sitting in the chairs, right? So that we can get three five-minute windows, so our period is about five minutes as a computational window. We get three of those windows, and, okay, so let's talk about the device first. What even makes it possible? You know, as I say, you need lemons to make lemonade, okay? So let's talk about the core ECG, and again, one, it's an actual realization of Diane Thurman's triangle. It's miniaturized, and again, but it's a little bit more than that. I come from the world of radar imaging, right? I've done a lot of space constellation imaging, so the electrodes, the way they are structured, they actually are like three satellites that let me, again, pulse the electrical field in a certain way, so that I can do a very high-fidelity, a finite grid, okay? And also, it's an immobilized geometry, so when you have 12-lead ECG, it's impossible to be able to get the repeatability. There are very few techs who are going to, even for 10 seconds, be able to put the electrode on the subject and try to get the same data. It's got to go vary from one to the other, so the inter-subject variability is quite high, okay? And in this particular case, having an immobilized construct gives a certain unique properties. Also, if you notice that the round dome is actually sitting on the V2 lead, okay? So for traditional halter studies, we use lead two, lead three, but when it comes to conduction and for our other indications, at that point, we switch into what's called a derived lead mode. So we have six leads with a waveform we validated, actually, on a large Mayo dataset a couple of years back, that our waveforms are, you know, they've got a high end-to-end yield, the waveforms really replicate, and also, we have a high correlation with the precordial leads. So, effectively, I'm working with a 12-lead system with the fidelity and stability of core, okay? And the other part which has been difficult with almost all the halters is the amplitudes and the morphology don't have stability. So the immobilization provides it a tremendous amount of stability, at least for control protocols at this stage, okay? And lastly, as I said, what makes body surface mapping really good is the fact that you have this temporal resolution. You have the fidelity of being able to put 64 electrodes, 128 electrodes, but what core does, as I said, the construct allows us to do a finite element on the electrical field itself. So using four electrodes, we can generate the same, you know, the same resolution through computational oversampling, okay? And in fact, so that's a key, being able to get the information. So let's take a second how we got here. So VAT for conduction defect was not a primary endpoint. A primary endpoint, including for ongoing trials, is to get to ejection fraction severity, get to heart failure diagnosis using our device and as unique property. So that is our, and you can see the system. So we are doing several things. We reported this in last year's HRS is, again, how to get to ejection fraction severity at this stage. We do get a number. We had a 6% error, so it's not good enough to present, but we think our severity is clinically quite robust as the data shows, okay? And in order to compute it, it's actually validating a biomarker of which the conduction and the atrial enlargement are there. So we had a byproduct, a secondary endpoint, and which led us to utility. So let's talk about, before I go to the data. So we, as I mentioned, we have a trial ongoing, there are 12 sites. And so we are sourcing, again, a cohort that is, it is controlled in a sense that in the initial phase, we preferred low ejection fraction subjects, but now it's moved to a phase when it's, again, open enrollment. So out of that, we chose, again, paced patients with, and our focus on, again, so they do have a, there's a high percentage of heart failure patients here already, and combined it with some prior patients, which had both a HFF and a HRF population. And again, on a large number of therapies. So let's take a look at about, you know, how good the VAT is. As I said, so we've taken five minutes of data, we've done all the crunching, we estimated what the conduction profile looks like. So I'll jump into a patient journey. So this is one particular patient, it's a pacemail, the HFF is about 41%, and also has hypertrophic cardiomyopathy. And again, in general, for this entire population, right, so we are past the point of being able to use our segments and our other data to, you know, do they have heart failure or they don't? So they're clinically abnormal, we are past that point, right? So what we are trying to understand is, really, for this clinically abnormal population, okay, how is their conduction system working? Is it getting worse? What pathway are they on, right? So for this particular subject, as an example, all of them are high VAT, it's prolongated in all of them. And so what you notice is, again, I've taken two snapshots from our extra, this is done through signal processing, spectral analysis, and, you know, applied a whole bunch of transforms. But so we're looking at, again, as this figure is, so there's a time axis, and in each, again, on the vertical and the y-axis, what we're seeing is the frequency at each time point, okay? So it's a good way to visualize what's happening to conduction. And then the amplitude shows the conduction velocity, changes in the velocity. What's the variability at that point in time, right? So what you really want is a nice, dark blue surface. If you're a normal person, that's what you get, relatively all cold blue, okay? Consistent conduction, very, again, very low VAT variation, and the mechanical machinery is working robustly. But in this particular cohort, we start seeing these anomalies, right? So let's look at the derived limb changes in the velocity itself, okay? So what you see is, so there is a high variation in the 40 to 60 hertz. What does it mean, okay? So this is a subject where we are seeing changes. Again, there's a prolongation, but it is a signature for us, for most of the patients who have actually left bundle branch, it looks at it, okay? Again, our target is five to 10 milliseconds of VAT changes, and the variability at that level, we think that's the clinical threshold of doing something about it, or at that threshold, if you exceed it, is where we're able to tell the clinicians it may need re-tuning again. And again, so it's preset for a 10 threshold change, and so what we start seeing is that for this particular subject, for example, the spectral power is oscillating, you have these episodes of conduction changes coming and going, right? And now, these are all PACE patients. This should not be happening, to begin with. And if it happens, and if it gets worse over time, that's when the problems start coming in. Suddenly, you put the person on a different morbidity curve, okay? So again, and then we look at the derived precordial lead, so we again see high variation, 40 to 80 hertz. It's, again, consistent for left bundle branch, that's not our primary endpoint. And so we start seeing the fact that, again, it's predictive, as I said, so you're running a core insight trial. In core insight, there are 40 multiplex indications, that means in the one study, you're outputting an entire rhythm stack, like a blood report, of the indication of what's wrong with this particular person, so that the clinician has a chance to very quickly get to a decision point. And so again, we see that the high frequency variations in this, and depending on if this is calibrated to different bands, where we know what it basically means is, that we're going to now expect the ectopic, you know, the ventricular activity to start increasing. We see the burdens changing, okay? Now, if in a HPF, and by the way, this progresses faster for HPF, compared to HREF patients. So that starts explaining some of the data you see, right? They are, again, mostly non-symptomatic, but when things go wrong, and then suddenly, the adverse consequences to them, the changes are much, much higher. The cardiac output goes much faster than HREF patients, even from our small clinical cohorts that we've seen, you know? And a lot of it is because, you know, we never remodel the electrical pathway, we let it get worse, and then it becomes too late, okay? So same thing, let's take another patient, and it's again, like I said, the story is, you know, it's just a little bit different, again, hypertension, it's about the same. We chose a couple of patients in these examples with about the same ejection fraction, okay? I'll take a case with much, much lower as well, and so we start seeing the same, but now you notice the patterns are quite different. The conduction changes, the conduction defect in the two patients is quite different. So their progression is, or the ventricular disc hernia rates are different, okay? One is twice the rate of the other, okay? So this person could have a comorbidity cascade much faster than the one before, okay? And also it's going to lead to a higher risk of supraventricular arrhythmia than ventricular fibrillation, so. So if we just, again, you know, as I said, this is just a picture comparing two, okay? So here's the difference at the end of the day. So if we can make this routine, and with five minutes, like, you know, the time it took you all to sit, if you had just strapped this device, we had the data, we'd be able to tell you what your conduction profile looks like, and where on the spectrum you really fit in. That's the objective, to make it, again, very accessible and robust clinically, okay? So what we can see, the difference in the two is, in one case, the person is actually benefiting, okay? Even though it's a night tarot, again, a very, this yellow area is a very, you know, kind of like a, it's a localized activity, it's high magnitude, but it's a control. The CRT is kicking in. It's trying to compensate, it's trying to work, and, you know, and also the actual disc hernia rate is quite moderate, okay? On a longer term basis, this is much less than six, seven, I mean, it's not about the 10 millisecond threshold, where we would be alarming, yeah. And the patient actually, now this patient would benefit from a little bit more optimization, if we could do that. So again, the point being, that builder tool, which can be used to, you know, we now have a new class of adaptive pacemaker, pacing. Being able to tune them a little bit more frequently. And take it to a point when, you know, either in the clinic, or the subjects at some point be able to tune it themselves. So you don't have to wait three months, six months, nine months before you can do something about it, okay? And in the other case, it's again, there's a lot of discrepancy, and you need now adaptive pacing, right? So in this case, the pacemaker is not working at all on the right guy, because this is wide dispersion. So we're seeing not big conduction changes, but we are seeing dispersions on a very large time. They're repeatable, they're continuous. So this person is working with, again, a low cardiac output, suboptimally tuned system, right? So here is what surprised us, and as I said, we have a couple of APIs, right? So that nine, you know, for the ejection fraction, for 93% of these patients, and some we had dropouts, you know, again, we also use VAT as a differentiator for cardiomyopathy. So we're using it as part of our markers, how we segment the populations, for again, heart failure, for cardiomyopathy. And so that's, you know, 93%, we could get to the VAT straight away, they're all abnormal, they're abnormal and all the other stuff too, there's some dropouts. But 40% people showed, in a trial with about 50 base patients, or 53 base patients, which have uncontrolled VAT disc hernia. Now, if you don't look at this again, periodically, you would never know. You'd wait till someday something's not right, and you know, get it? So we think that they would benefit from the optimization, and at least try to get off an adverse pathway. Again, so just to summarize, so we think it's a clinically very impactful analytic, it's got a lot of predictive power, VAT in general. We think optimization can drive outcomes, we can get those rates changing, because if we don't do it, we'll be seeing the same numbers. And again, uncontrolled conduction defect can accelerate adverse pathways. So we're going to see again, what we call our, you know, these unexplained effects, as we call them. Suddenly some part of a sub-cohort just drops out. And it is a marker for us, for differentiating the other point. This is how we, in the ejection fraction computation, and heart failure classifier, how we separate populations that reduce ejection and preserve the ejection fraction. The behavior of the two, the conduction variation is quite different. And for cardiomyopathy, I think the device, that is really the key. The device at this time, we have a yield with very high sensitivity and stability, at least, that's what you need. If you don't have it, then the number is not going to be believable. And in different settings, in hospital or home, it's an easy thing, we sit anywhere or everywhere. So you don't have to come and down, lie, get your lead strapped to you, and you know, so it's an easy protocol. And again, it's a powerful, simple tool for accessing effectiveness. And a future is that I think we'll be getting ready to do again, trials against BSM and catheters for routine use. We need to make sure that we are close to the gold standard. And also how to harmonize our VAT back to the exact outputs we are getting from vector cardiography, or from catheter systems. So, thank you. Okay, I think for the sake of time, I'm just gonna allow our next speaker to finish up the series, and then if there are any other questions additionally at the end, we can always ask them. So it's my great pleasure to introduce Dr. Margarita Pujol-Lopez. She's gonna be talking to us about characterization of right bundle branch block with invasive mapping, implications for conduction system pacing. Dr. Lopez. Thanks. Hi, good morning. Thank you for being here. And thanks for the opportunity to present our work. I'm going to present this abstract on behalf of University of Arizona team from Phoenix, and Virginia Commonwealth University. We're going to talk about characterization of right bundle branch block with invasive mapping, site of localization and implications for conduction system pacing. My interest to declare. We have 30 years of experience with CRT and left bundle branch block, but we have less information for right bundle branch block. Conduction system pacing is a promising option for cardiac CRT in patients with complete right bundle branch block and reduced left ventricular ejection fraction. However, we have a problem that it's right bundle branch block correction is infrequent with his bundle pacing in clinical practice. And we have reports about that. The location of the right bundle branch blocking, if we can know what is the block, it will have a lot of clinical implications because we can know if it will be feasible to correct with his bundle pacing. You have seen this figure before. Before it's likely different because it's a figure from circulation. Dr. Tank and Dr. Upadhyay presented it, published it a few years ago with left bundle branch block. Then our question is the same, but for right bundle branch block. For left bundle branch block, they started with a septal mapping, the level of block and what they showed is that in the majority of cases, left bundle branch block, the block is proximal, very proximal and focal. And then it's more easy with a block is proximal. And in the majority of cases of proximal block is a left intrahesion block, very proximal. It's easy to correct with his bundle pacing. What we know about a right bundle branch block, this is a study from 20 years ago by Cecilia Fantoni and Dr. Auricchio. They studied a cohort of 100 patients. Only six patients had right bundle branch block and they studied it with 3D mapping. And they found what was expected is that the right bundle has more delay in activation, but also they found that the degree of left ventricular activation delay is similar between heart failure patients with left bundle branch block and right bundle. And you can see in the image that you have delay in the right ventricle, but also you have delay in the left ventricle. And it has been previously described, it was not novel. Then it's a cold mask left bundle branch block. And you can see it on the ECG on the right that you have anterior ME block. Then if you have right bundle only, isolated, will be more easy to correct. But if you have something plus the right bundle, will be more difficult. And we know that with right bundle, the axis should be normal. Then if we have something on top of the right bundle, the axis will not be normal. And this on top could be concomitant intraventricular conduction disease or maybe diffuse conduction system disease. And it has been previously showed by Dr. Villamaran that not all the patients with right bundle branch block and heart failure had normal axis. High percentage of patients had a deviation of the axis. We want to locate the block. It has been previously done with a population that is slightly different than our population. Patients with congenital defect. And they mapped the right ventricle before the surgery and after the surgery. And they found that in the majority of cases, the block was distal. And they also showed the pattern of activation. And the activation, if the block is very proximal, all the right ventricle is delayed. But if the block is more distal, the apex is activated normally. Then there is a lack of both the detailed electrophysiological and mechanical data in patients who present with right bundle branch block. And right bundle block physiology needs further investigation. Our objectives were to characterize right bundle branch physiology with invasive mapping. Identify the site of block. Assessment of the percentage of correction with his bundle pacing. And drawing clinical implications for his bundle pacing and left bundle branch pacing. We had 33 patients. It's a recentric study. We had data also from Virginia, thanks to Dr. Ellen Bowen and Dr. Ajay P.J. for sharing with us the data. We had consecutive patients with right bundle branch block mapping during EPS study. And we had nearly 80% of patients with high-density mapping. Then we characterized the level of block. And then we measured the distance from his, one-to-one his, to last recorded Purkinje. And we assessed the correction with his bundle pacing. And we have here some figures about the methodology. Then you can see what we have done. This is the more extreme case. It's a block six centimeters. It is in the right ventricle. We have measured from the his to the last Purkinje. This is the more extreme case. We have another very distal block. You can see here the measurement from the his to the last recorded Purkinje. We have also assessed if we can correct with his bundle pacing. You can see here right bundle branch block with parachesium pacing. We have here a spike R of 107 milliseconds. And when we correct the right bundle and we capture the his, you can see that we had a delta of 30 milliseconds. And we know because of the Jastrzewski criteria that if a patient had spike R of less than 100 milliseconds and also we do not have notching or slurring, we are pacing the his. And here we have shown two examples of very distal block. Here we have the propagation map of another distal. This is four centimeters in the right ventricle. And you have the propagation map on the left of an example of a proximal block. We had 33 patients, six were women and the median QRS was 151 milliseconds. As I have said, we had 80 person with high density mapping and the median points that were used for the right ventricle were nearly 1,000. The distance from one to one his to the last recorded Purkinje was the median of three centimeters. Not all the patients had three centimeters. But you can see here the distributions of the distance. We have seen these more extreme cases but the median was three centimeters. And this has a lot of clinical implications because if we have a very distal block, this is the opposite to what we have with left bundle range block, that the block is very proximal, then it would be very difficult to correct with his bundle pacing. And we have started with this slide. Here you have the summary for right bundle range block. We can see that we have the opposite compared to left bundle. With left bundle was the block very proximal and here we have the majority of cases the block was distal. Then would be very difficult to correct with his bundle. In our series, 12% of patients we corrected with his bundle pacing. And in this context that the block is very distal, maybe the better options are left bundle range pacing or playing with fusion to correct the right bundle range block. And let me conclude with main ideas of this work. The site of right bundle range block may be more distal than left bundle range block. Left bundle it's very proximal. The implication is that correcting his bundle pacing is unlikely. And then the correction will be more with septal fusion over right bundle range block recruitment. And finally, left bundle range pacing with fusion may be a more suitable mechanism of correction in patients with distal disease. Thank you so much to all the team. Thank you. Thank you, Dr. Lopez. That was excellent. If I could just ask one quick question, which is in your patient population, in your centers now, do you foresee or should we be considering doing more invasive studies at the onset or just declaring, instead of declaring that right bundle branch doesn't benefit from any kind of re-synchronization or left bundle, do you think that we should be mapping or getting better characteristics of the right bundle branch? Or do you think because the applicability of more distal block in a sense just tells you immediately to go to the left bundle branch pacing? Thanks for the question. We're including more patients, but if we can demonstrate that in the majority of patients, the block is distal, maybe we can go for left bundle branch pacing because we have seen that if we have data that left bundle branch pacing will have less complications, then maybe it's more useful to go for left bundle branch pacing. With left bundle branch pacing, we are also working with this to characterize because with left bundle branch block, it's difficult to know with the ECG if it's true left bundle or not. Maybe for CRT and CRM cases for left bundle, it's needed to do EPA study, but maybe if we can demonstrate with right bundle that the block is very distal, we can go for left bundle branch pacing. Thanks. Okay, I think that concludes the session for this morning. If there are any other questions, certainly feel free to come and ask. Otherwise, have a good rest of your Saturday at Heart Rhythm.

arrhythmia management

heart failure

atrial fibrillation

HFPEF

AF ablation

AI-generated ECGs

cardiac resynchronization therapy

VAT mapping

right bundle branch block

personalized medicine