UCSF cardiologist Rita Redberg, MD on the Abiomed pump’s latest recall.

Artificial Intelligence

Chest X-Rays’ Cardiovascular Screening Potential

There are over 70 million chest X-rays performed in the U.S. every year, and a pair of new studies highlighted how AI could be used to “opportunistically screen” these exams for undetected cardiovascular disease.

A Mass General Brigham team writing in the Annals of Internal Medicine detailed how their deep learning system was able to identify patients with greater 10-year risks of experiencing major cardiovascular events by analyzing their chest X-rays.

The researchers developed their ‘CXR CVD-Risk’ model using chest X-rays sourced from a cancer screening trial, and then externally validated the AI model against 11,000 patients’ CXRs.

• Among the patients with unknown ASCVD risks, those who received a 7.5% or higher CXR CVD-Risk estimate had a 73% greater adjusted 10-year risk of MACE than patients with estimates below 7.5%.
• Among the patients with known ASCVD risks, the AI model added to the MACE-prediction accuracy of their traditional ASCVD risk scores (adjusted HR = 1.88).

A Columbia University team writing in the European Heart Journal showed that CXR deep learning algorithms can also be used to identify patients with left ventricular structural abnormalities that might be signs of heart failure.

The researchers developed their algorithm using 71,589 unique CXRs from 24,689 patients, along with the patients’ echocardiogram labels for two structural abnormalities: left ventricular hypertrophy (SLVH) and dilated left ventricle (DLV). 

• When tested against 3,667 CXRs from an internal Columbia dataset, the model detected patients with SLVH,  DLV, or either of the abnormalities with relatively high AUCs (0.79, 0.80, 0.80).
• When validated against 8,003 CXRs from an external Stanford dataset, the model detected patients with SLVH,  DLV, or either of the abnormalities with lower but still decent AUCs (0.67, 0.78, 0.76).
• The model also outperformed 15 board-certified radiologists’ visual assessments to detect patients with either SLVH or DLV, with 71% sensitivity (vs. 66%) at a fixed specificity of 73%.

The Takeaway

We’ve already seen significant progress in using chest CT scans to opportunistically screen for patients with undetected coronary artery calcium. But considering the massive volume of chest X-rays performed each year, solutions like these could have a major impact – especially as algorithm accuracy and post-detection workflows continue to evolve.

Advance Your TAVR Planning

Ready to advance your TAVR planning? See how you could leverage TeraRecon’s TAVR (Vessel Analysis) package to streamline and enhance your aortic valve replacement measurements.

Bringing AI Into Echo Workflows

Cardiovascular AI solutions are still far more commonly found in cardiology journal papers than in actual clinical workflows, but that’s changing fast and there’s a lot to learn from AI early adopters. Tune into the latest Cardiac Wire Show with Mount Sinai’s Jeffrey Bander, MD and Us2.ai’s Seth Koeppel to see how they launched and expanded their echo AI partnership, and learn how to “adopt echo AI the right way.”

The Wire

• TriClip Approved: Abbott’s TriClip tricuspid edge-to-edge repair (TEER) system is now FDA-approved. The approval comes less than a month after an FDA expert panel voted that the TriClip’s benefits outweigh its risks, citing its pivotal TRILUMINATE study, which showed clear quality of life benefits, but no benefits for HF hospitalization or mortality. The TriClip’s approval comes just two months after the FDA approved Edwards’ EVOQUE transcatheter tricuspid valve replacement system, giving interventionalists the options to repair or replace the tricuspid valve after years with only surgical TR options.

• A Timely Beat for Early Ablation: A new JAHA study suggests that early ablation for Afib (within 1 year of diagnosis) significantly lowers the risk of recurrence compared to late ablation (>2.9 years). Data from 7,700 patients showed that the 5-year cumulative incidence of Afib recurrence ranged from 42.9% for early ablation to 58.4% for late ablation, regardless of whether Afib was paroxysmal or persistent. The risk of heart failure, stroke, or death was also notably lower in early ablation patients.

• NYT’s Abiomed Takedown: Abiomed was already having a bad week after the FDA announced a labeling recall due to issues with its heart pump cutting through the wall of the left ventricle, but it surely got worse after The New York Times made Abiomed’s struggles national news. The NYT article started with a focus on how Abiomed delayed its notice to the FDA for two years, while also detailing the Abiomed pump’s history of recalls, mixed study results, and hefty industry consulting payments.

• ACC’s Pathway to Success for Chest Pain: A JACC study just validated the ACC’s Expert Consensus Decision Pathway for chest pain, which utilizes a new 0 and 2-hour high-sensitivity troponin. Among over 14,000 patients, the algorithm effectively identified nearly half the patients who were eligible for early discharge while maintaining a high negative predictive value of 99.7% for 30-day death or MI. However, some caution is advised for patients with known CAD, as the NPV fell to 98.5% in this group.

• Camzyos’ Real-World Evidence: BMS added some real-world support for Camzyos’ pre-approval RCT data, in a pair of ACC 2024 studies. Ten-month results from the Camzyos REMS Program showed that just 1% of 1,524 patients reported HF hospitalization and 2.8% reported a decrease in LVEF to <50%. Meanwhile, a 24-week single-center analysis of 53 patients showed that Camzyos led to improvements in cardiovascular symptoms (96%) and one or more NYHA class (49%).

• Fortified Eggs “Best” for CV Safety: In an Eggland’s Best-funded study of 140 people with CVD or at high risk of CVD, participants who ate 12 or more fortified eggs per week over four months had numerically lower LDL and HDL (-3.14 mg/dL and -0.64 mg/dL) than a group randomized to eat only two eggs or fewer per week. Levels of other CV biomarkers were also statistically similar.

• Cleerly’s Ischemia Advantage: A new JACC analysis highlighted Cleerly’s advantages for identifying coronary ischemia in CT scans and predicting long-term MACE risk. Using data from the CREDENCE and PACIFIC-1 trials, Cleerly’s AI-QCT detected ischemia with far higher AUCs (0.80 & 0.85) than FFRCT (0.69 & 0.78), MPI (0.65), and SPECT (0.72)… but below PET (0.89). After adjusting for clinical risk factors and coronary CTA-determined obstructive stenosis, a positive Cleerly AI-QCT ISCHEMIA test was associated with a 7-fold higher risk of MACE over 8 years.

• Overworked Cardiologists: It’s not always good to be number one, and that’s certainly true in a new Doximity poll of 2.7k physicians, which found that cardiologists were the most likely of 26 specialists to view themselves as “overworked” (93% of cardiologists). Next up on the list were neurologists and pediatricians (91% & 90%). The poll found that 81% of overall physicians viewed themselves as overworked in 2023, down from 86% in 2022 but higher than 73% in 2021.

• Eko’s Low EF AI: Smart stethoscope company Eko Health took a key step towards addressing heart failure’s early detection and diagnostic access challenges, announcing the FDA clearance of its Low Ejection Fraction AI algorithm (Low EF AI). The solution expands heart failure detection to stethoscope exams at primary care offices (vs. specialist-performed echo exams at clinics/hospitals), while identifying patients with low EF within 15 seconds. Eko’s Low EF AI solution will be added to its SENSORA Platform, alongside FDA-cleared algorithms for Afib and structural murmurs.

• Time to Notch a New LBBB Criteria: Researchers at the University of Chicago developed a simple ECG criterion—time to notch longer than 75 milliseconds in lead I of the 12-lead ECG—which significantly improved accuracy in identifying left bundle branch blocks (LBBB). With 71% sensitivity and 74% specificity when used in conjunction with standard LBBB criteria, the new method produced far fewer false positives than the Strauss criteria’s 33% specificity.

• Adagio’s CE Mark: Adagio Medical announced the European CE Mark approval of its ultra-low temperature cryoablation (ULTC) system for the treatment of monomorphic ventricular tachycardia. The ULTC system includes Adagio’s vCLAS ventricular cryoablation catheter, which is uniquely designed to address the challenges of VT ablations, and its cryoablation console. Adagio also has US studies in the works, suggesting that an FDA approval might come in the future.

Can AI Prevent Heart Disease Progression?

Tune in to Cleerly’s on-demand webinar where study leaders will discuss how the landmark TRANSFORM randomized controlled trial will test whether an AI-personalized care strategy can outperform traditional risk factor management and prevent cardiovascular events.

PIA’s Post-Processing Solution

Advanced cardiac imaging often calls for a time-consuming post-processing step, requiring costly software, hardware, and training. See how PIA provides this post-processing at lower cost, improved consistency, and greater efficiency.

The Resource Wire

• Experience the future of learning: Medtronic Academy 2.0 is here! Unlock your ultimate destination for structural heart medical education with the newly redesigned Medtronic Academy 2.0. Gain access to expert-led courses, webinars, and a wealth of resources to stay ahead in cardiovascular care. Visit now!

• Merge and Duly Health Streamline Cardiology Reporting: Over the last 10 years, Dr. Sujith Kalathiveetil of Duly Health and Care has seen a significant evolution in cardiovascular imaging and experienced a similar evolution with Merge’s cardiology solutions. See how Merge Cardio has helped make cardiology reporting more consistent, accurate, and easier to obtain for Dr. Kalathiveetil and his colleagues.

• HeartFlow FFRCT Catches Missed Blockages: Karen Moore had always been diligent about her heart health, which is why she was concerned when she began showing symptoms of heart disease but all of her tests came back negative. See how Karen and her physician used HeartFlow FFRCT to catch a 90% blockage and place a stent in the right location, before it was too late.

• Better Cardiovascular Care, With a Better Bottom Line: Twenty million chest CTs are acquired in the U.S. each year, but CAC is typically unreported. See how you can leverage Bunkerhill Health’s Incidental CAC algorithm to screen for incidental coronary calcium on routine chest CTs in real-time, improving care and your bottom line.

• ECG Data You Can Trust: Noise and artifacts can make automated ECG analysis less reliable than what’s required for the exacting standards of cardiac safety trials. Monebo’s Kinetic Intervals ECG Algorithm provides precise interval measurements between any two points on the ECG waveform, allowing clinicians to utilize data they can trust.