How the “AI Effect” Energizes Ultrasound

From the Curie brothers’ discovery of the piezoelectric effect in 1880, to the introduction of AI-guided diagnostic software, ultrasound has earned the reputation of a well-established workhorse in today’s healthcare. And all signs point to a paradigm shift as AI agents and algorithms enhance ultrasound access, automation and accuracy. With the resulting significant improvements in image acquisition, as well as the subsequent analysis and diagnosis, the ultrasound process will see dramatic improvements in quality, speed and consistency.

Enhanced Image Capture. AI-guided image capture provides real-time, on-screen guidance and view identification that aids ultrasound technologists, and less experienced operators, to position the probe and capture reliable clinical quality images.

Adaptive Optimization. AI automatically adjusts imaging configurations such as gain, depth, focus and time-gain compensation based on patient’s anatomy. Manual corrections or “knob time” are reduced, and consistent quality is achieved across different patients and operators.   

Standardized Acquisition. AI automatically recognizes and helps to acquire standardized scan planes such as standard cardiac views or fetal biometric planes. It can also provide automated image quality assessments that reduce variability across different staff and sites.

Improved Image Quality. Image preprocessing, including denoising, artifact suppression and enhancement benefit from algorithmic improvements, particularly Convolutional Neural Networks (CNNs). The challenges of acquiring reliable scans from difficult targets, or low-contrast anatomy, are minimized as a result of significant improvements in signal clarity, contrast and quality.

Automated Measurements. AI can identify structures and suggest caliper placement for measurements like fetal biometry, bladder volume and vascular diameters. It will also automatically calculate complex metrics in echocardiography, lowering inter-operator and intra-operator variabilities. We’re now seeing wide acceptance of automated measurements of Left Ventricular Ejection Fraction (LVEF) and Global Longitudinal Strain (GLS) among clinicians in echocardiography labs.

Assisting Diagnostic Accuracy. By identifying visual patterns and features often missed by the human eye, AI can identify symptomatic areas, segment structures, or categorize various pathologies.

Autonomous Ultrasound: The Future

When AI replaces the primary operator, with clinicians providing supervision and clinical judgement, ultrasound will be a fully autonomous system, moving beyond a scarce, costly procedure to become a scalable and truly accessible modality. Autonomous ultrasound will improve patient outcomes, reduce costs, dramatically improve access, and increase labor efficiency. 

The key to ultrasound autonomy is the implementation of intelligent, wearable sensors in conformable patches that replace both the handheld probe and its operator. Continuous monitoring of vital signs enables earlier identification of complications, often in the pre-symptomatic phase, by tracking subtle anatomical or physiological changes over time. 

Advancing from episodic imaging to frequency or continuous, AI-interpreted data streams, autonomous ultrasound in obstetrics, for example, provides the benefits of more frequent and consistent maternal and fetal assessments, without repeated specialist visits. All pregnant women will applaud. 

Autonomous ultrasound also supports cardiology care, as more frequent and broader screening initiatives become possible. Routine, remote echocardiographic screening for older adults would lead to earlier detection of undiagnosed heart valve disease, along with many other pathologies, in the pre-symptomatic phase. Clinicians can track changes in cardiac function, recording heart failure and changes in volume status over time, and receive real-time alerts triggering earlier intervention. Cardiac patients could receive automatic smartphone updates on their status as well, encouraging them to follow their prescribed self-care programs.

Remote Care via AI 

As “Physical AI” in ultrasound eliminates the need for a dedicated operator in a hospital or clinic environment, the present bottleneck of scheduling around technician or equipment availability is eliminated. Ultrasound’s reach can be extended beyond the metropolitan area and into non-urban and rural areas. Soon, a patient’s home may be the ideal setting for diagnosis and monitoring. That would be a true step towards the democratization of healthcare.

Advances in healthcare are often hailed as "breakthroughs", but few will match the breadth and wide-reaching potential of autonomous ultrasound. With fewer adverse events, faster diagnosis, fewer repeat exams, and lower cost of care, AI will prove to be a valuable partner, helping humans heal our healthcare system.