An anonymous electrophysiologist, presumably regarding this week’s findings that 12-month AF ablation delays don’t reduce ablation efficacy.

Artificial Intelligence

Cardiology’s AI Optimization Opportunity

Artificial intelligence and machine learning are currently having the biggest impact on imaging-related areas of cardiology, but a trio of new studies highlight AI’s potential to support significantly more cardiology treatment and therapy decisions.

Optimizing HF Treatments – University of Michigan researchers developed an algorithm that analyzes clinical data to identify HFrEF patients who would benefit from guideline-directed medical therapy (GDMT) optimization and then recommends treatment changes. 

• When applied to data from the GUIDE-IT and HF-ACTION trials, the algorithm commonly recommended medication changes (34.9% to 68.1% of patients depending on drug) or dosage increases (48.8% of visits w/ ACR & ARBs; 39.4% of visits w/ beta-blockers).
• The algorithm might have been right, as patients who the AI assigned better baseline GDMT scores had much lower risks of cardiovascular death or HF hospitalization (hazard ratio: 0.41) and all-cause death and hospitalization (hazard ratio: 0.61).

Predicting Post-TAVR PPM – Mayo Clinic researchers used machine learning to predict which patients would require a permanent pacemaker (PPM) after TAVR procedure more accurately than the current standard PPM prediction model.

• Using data from 964 patients without prior PPM who underwent TAVR, the machine learning model identified 167 clinical variables to predict each patient’s PPM risk.
• The model predicted which patients would require PPM at 30 days and one year far more accurately than the standard PPM risk score model (AUROCs: 0.66 vs. 0.55 & 0.72 vs. 0.54; both (P < 0.001).
• Certain variables had greater association with PPMs, including brachiocephalic artery to aortic valve annulus distance to height ratio (the biggest predictor), pre-existing conduction abnormalities, trans-femoral access, and self-expanding valves.

High-Benefit BP Control – Hypertension treatment guidelines focus on high-risk patients, but targeting “high-benefit” patients for intensive BP therapy could have a much bigger population health impact. 

• UCLA researchers used machine learning to analyze data from two BP reduction trials (n=10, 672) to identify patients who would experience the greatest benefits from intensive treatments and assess outcomes.
• They found that providing intensive treatments to “high-benefit” patients prevented one cardiovascular event per 11 patients treated over a three-year follow-up period. 
• That’s more than a five-times better ratio compared to exclusively targeting high-risk patients for intensive BP treatments (~62 per prevented event).

The Takeaway
This batch of studies serves as yet another reminder that AI and high-powered computing could drive a new level of personalization and precision in cardiology care, well beyond the initial diagnostic steps where we’re currently seeing most cardiology AI activities.

Creating A Novice Heart Failure Screening Pathway

We hear a lot about AI’s potential to expand echocardiography to far more users and clinical settings, and a study using Us2.ai’s AI-automated echo analysis and reporting solution showed that echo’s AI-driven expansion might go far beyond what many of us had in mind.

The First Step to Coronary Artery Disease Diagnosis

HeartFlow’s new RoadMap Analysis solution allows CT readers to accurately, efficiently, and consistently identify stenoses in the coronary arteries. See how RoadMap Analysis’ visual and quantitative insights into the narrowing of all major coronary arteries helps readers evaluate coronary CT angiograms before determining the need for an FFRCT.

Precision QRS Detection

QRS detection is essential for any ECG algorithm, and Monebo’s Kinetic QRS ECG Algorithm sets the standard for accuracy. Kinetic QRS accurately detects the QRS complex, no matter the amplitude, waveform, or noise levels.

The Wire

• Later Ablations, Same Outcomes: EHRA 2023 was buzzing this week after one of its presented studies found that delaying AF ablation by 12 months didn’t reduce ablation efficacy compared to treatment within one month, as long as patients were treated with antiarrhythmic drugs. Researchers randomized 100 symptomatic AF patients for early or delayed ablation, finding no significant differences between the two groups in terms of arrhythmia outcomes, AF burden, AF phenotype, or antiarrhythmic drug use at 12 months.

• Cardiologists Still Compensation Leaders: Cardiologists were once again well positioned in Medscape’s 2023 Physician Compensation Report, with the third highest earnings across 29 specialties (+4% to $507k) and the fifth highest share of respondents who would choose their specialty again (93%). The report also highlighted some of cardiology’s problem areas, as it has the third lowest percentage of female physicians (just 17%), the third most working hours per week (56.2), and the eighth most paperwork and admin hours per week (16).

• Cardiac Imaging in 2040: A recent Radiology editorial forecasted that by the year 2040, cardiac imaging will be more automated and preventive, and even more dominated by CT. The editorial included cardiac CT in seven of its 10 predictions, suggesting that cardiac CT will increasingly be used for preventative screening, allow for more automated interpretations, and achieve far greater image quality. The editorial also projected that artificial intelligence will continue to grow its role in cardiac imaging, streamlining post-processing workflows and further aiding in diagnosis.

• Inflammation Trumps Cholesterol as Predictor of Heart Risks: Analysis of over 31k patients from the PROMINENT, REDUCE-IT, and STRENGTH trials found that among those receiving contemporary statins, inflammation measured by high-sensitivity C-reactive protein (CRP) was a stronger predictor of future cardiovascular events and death than cholesterol levels assessed by LDL-C. Residual inflammatory risk showed significant associations with major cardiovascular events (HR: 1.31) and cardiovascular mortality (HR: 2.68), while residual cholesterol risk showed a neutral or low relationship with these outcomes.

• Women Need More Radiation Protection: Female healthcare providers who work near ionizing radiation need better protection, especially women involved in image-guided procedures. That’s from a new editorial in The BMJ that cited studies showing higher breast cancer rates in female physicians who work with radiation, and proposed that these providers wear additional protection under their standard gowns.

• ChatGPT’s Mixed Cardiology Performance: How does ChatGPT perform answering cardiology questions? It’s complicated. In a preprint study out of the Netherlands, ChatGPT correctly answered 74% of 50 multiple choice questions, and matched in official advice in 9 out of 10 case vignettes. However, when answering questions from GPs regarding complex cases, ChatGPT was only correct 50% of the time, and often provided incomplete or inappropriate recommendations.

• Google AI in Healthcare: While we’re on the subject, Google revealed that it’s opening access to its Med-PaLM 2 large language model (a ChatGPT healthcare competitor) to allow a select group of Google Cloud customers explore healthcare use cases and test features. Google also announced an AI-enabled Claims Acceleration Suite designed to streamline prior authorization and claims processing.

• A Case for FH Genetic Testing: Danish researchers showed that genetic testing could significantly improve familial hypercholesterolaemia (FH) diagnosis rates, while urging more widespread use of genetic FH tests. The researchers performed genetic tests on 1,243 people with suspected FH during a 15 month period, increasing FH diagnosis rates from 22% to 37%. Notably, 20% of participants whose FH was initially believed to be “unlikely or possible” had a pathogenic FH variant that would have gone undetected without genetic testing.

• Help for the Physician Shortage? The number of residency slots for training new physicians would be expanded under federal legislation that’s been introduced in the U.S. House of Representatives. The Resident Physician Shortage Reduction Act of 2023 would expand the number of residency positions by 14,000 over seven years (the 2023 Main Residency Match had just over 40,000 open positions). 

• Clopidogrel Outperforms Aspirin Post-Stenting: Clopidogrel could perform better than aspirin after patients undergo coronary stenting and complete dual antiplatelet therapy. This secondary analysis of the HOST-EXAM randomized clinical trial (n=5,438) found that clopidogrel was associated with lower rates of 24-month all-cause death, myocardial infarction, stroke, readmission due to acute coronary syndrome, and major bleeding among both diabetic and non-diabetic patients (6.3% vs. 9.2%; 5.3% vs. 7.0%).

• PE Still Loves Healthcare: Private equity continues to love healthcare, and it looks like the market turmoil has made it a great time for PE to go bargain hunting. Bain & Company’s annual opus on healthcare PE revealed that 2022 was the second-biggest year on record for PE healthcare investing, with over $90B in transactions and generative AI / healthcare IT called out as hot sectors. That’s way down from $151B in 2021, but still $10B more than the next-closest year.

Optimizing Your Post-Processing Workflow

The growth of cardiac CT and CMRI volumes and continued shortages in the imaging technologist workforce can mean big challenges for imaging organizations. Join this Cardiac Wire Show starring Precision Image Analysis’ Jim Canfield and Cleveland Clinic’s Scott D. Flamm, MD, MBA to see how outsourcing cardiac image post-processing can solve this problem, while improving efficiency, accuracy, and standardization.

The Behavioral Science Behind Change Cardiology Hemo

When Change Healthcare set out to design its next-generation Cardiology Hemo monitoring system, they put behavior science at the heart of its product strategy. See how Change’s UX designers applied its behavioral science team’s findings to improve its Hemodynamics solution to help make physicians and technicians even more efficient.

Personalized Cardiac Care Through Science-Based Plaque Analysis

Check out Cleerly’s latest Cardiac Beat video, where National Jewish Health’s cardiovascular prevention and wellness leader, Dr. Andrew Freeman, and Cleerly founder and CEO, Dr. James Min, discuss addressing the root cause of heart disease and personalizing care through science-based plaque analysis.