AI Outperforms Sonographers in Assessing Heart Function

A study published in the peer-reviewed journal "Nature" found that artificial intelligence (AI) is better at assessing heart function than sonographers.

The study was conducted by investigators at the Smidt Heart Institute and the Division of Artificial Intelligence in Medicine at Cedars-Sinai and was a first-of-its-kind, blinded, randomized clinical trial of AI in cardiology. Cardiologists more frequently agreed with the initial assessment made by AI and made corrections to only 16.8% of the AI initial assessments. Cardiologists made corrections to 27.2% of the sonographers' initial assessments, and the physicians were unable to distinguish between AI and sonographer assessments. AI also saved cardiologists and sonographers time. The hope is to save clinicians time and minimize the more tedious parts of the cardiac imaging workflow. However, the cardiologist remains the final expert adjudicator of the AI model output. Regulatory approvals are now needed for such tools to be considered for use in clinical practice.

The success of AI in assessing cardiac function presents promising implications for the future of cardiac imaging. The research team is optimistic about the potential for wider adoption of AI in both the clinical and regulatory arenas. Another study conducted by the same investigators revealed that the AI algorithm from CT angiography images accurately predicted heart attack risk within five years, leading the team to believe that AI can help build better prediction models that will quickly get interventions to those who need them, ultimately saving lives.

REFERENCE

He B, Kwan AC, Cho JH, Yuan N, Pollick C, Shiota T, Ebinger J, Bello NA, Wei J, Josan K, Duffy G, Jujjavarapu M, Siegel R, Cheng S, Zou JY, Ouyang D. Blinded, randomized trial of sonographer versus AI cardiac function assessment. Nature. 2023 Apr 5. doi: 10.1038/s41586-023-05947-3. Epub ahead of print. PMID: 37020027.


Comments

Post a Comment

Popular posts from this blog

AI and Acoustic Health

Image
Doctors have been listening to the sounds our bodies make for centuries. Before the invention of stethoscopes, they simply put their ears to their patients' chests or abdomens, to listen to the sounds produced by our body's internal workings. Today, the practice of auscultation, using a stethoscope to examine a patient, has evolved into a sophisticated diagnostic tool that provides valuable insights into various aspects of our health: lungs, heart, bowel, vocal cords, joints, arterial conditions, sleep quality and more. Moreover, by amalgamating biomarkers and clinical data with a myriad of acoustic cues like sneezes, breath patterns, speech, throat clearings, wheezes, and more, the healthcare decision-making process stands poised for remarkable enhancement.  Auscultation is a valuable diagnostic tool because it is safe and noninvasive. Doctors place their stethoscopes at various positions on the chest to listen to different heart valves. The sounds they hear include the closur
Read more

AI for Predicting Heart Transplant Outcomes

Image
A recent review has highlighted the growing use of artificial intelligence (AI) in managing post-heart transplantation (HTx) and mechanical circulatory support (MCS) clinical care. This technology is particularly promising in the field of heart failure practice, where many decisions are often made based on expert opinions in the absence of high-quality data-driven evidence.  The review specifically focused on studies that examined post-HTx care, including post-operative management and long-term outcomes. The most common data sources used for development, training, and validation of AI algorithms are the United for Organ Sharing (UNOS) and the International Society of Heart and Lung Transplantation (ISHLT) registry. The primary outcomes studied include 1-year mortality or recipients, survival time, dependence on chronic dialysis, and graft survival. The machine learning models used include neural networks, support vector machines, random forests, and gradient-boosted machines. The predi
Read more

Smart Wearables for Cardiac Monitoring

Image
Medical devices marketed directly to consumers (DTC) are becoming increasingly more complex, capable to diagnose a variety  of cardiovascular issues.  Remote patient follow-up by monitoring heart rate is most popular, as it does not require complex analyses of cardiac waveforms such as ECG and PPG.  Most known devices - such as Apple watch - are used for the surveillance or “spot checking” of arrhythmias. The amount and quality of data recorded and stored depend on the battery capacity, storage capabilities, size of the device and cost of the device. The management of non-atrial fibrillation arrhythmias is done using  devices placed at a care   provider's  office during a scheduled appointment - like  Holter, event recorder, and electrocardiogram (ECG) patches. Wristbands and textiles are also used by some providers. New lightweight designs such as a single-patch 12-lead ECG (SP-ECG) are also being researched . Implementation in other patient cohorts, such as ST-elevation myocardi
Read more