UltraSight Company Spotlight: AI-Guided Cardiac Imaging

UltraSight is building an AI-guided cardiac ultrasound workflow for hospitals that have imaging hardware but not enough specialized acquisition capacity wherever and whenever patients need it. The company's software helps trained medical professionals collect focused echocardiography views, while qualified clinicians retain responsibility for interpretation and patient care.

That distinction is the whole business. UltraSight is not asking a hospital to admire an algorithm in a demo. It is trying to make existing ultrasound fleets more useful across intensive care, emergency, inpatient, and outpatient settings without pretending expertise can be removed from medicine like an unnecessary line item.

UltraSight is led by CEO Davidi Vortman. Its scientific and clinical roots include co-founder and CTO Itay Kezurer, co-founder and Chief Medical Officer Achi Ludomirsky, MD, and co-founder and Chief Scientific Officer Yaron Lipman, PhD. The company has combined geometric deep learning, cardiac imaging, clinical validation, and health-system workflow design into what it now calls the UltraSight EchoSystem.

What UltraSight Does

Cardiac ultrasound has an awkward capacity problem. The machines can be portable. The expertise is not. Capturing the right view requires probe placement, orientation, pressure, image-quality judgment, and enough repetition to know when a picture is clinically usable rather than merely recognizable as a heart.

The UltraSight AI Guidance software analyzes the incoming ultrasound stream and provides real-time instructions for moving the probe toward a standard cardiac view. A quality meter helps the user decide when to save a clip. The broader EchoSystem adds training, workflow integration, analysis, and echo stewardship around that acquisition layer.

The product strategy is deliberately hardware-aware. UltraSight wants to work with compatible systems already present inside health networks, reducing the need for a hospital to replace an entire device fleet just to add guidance software. That creates a harder integration job, but it also places the company closer to a health system's actual purchasing logic: improve the capacity of assets already on the books.

The Regulatory Boundary Matters

UltraSight's regulatory record is meaningful because it also shows where the company's claims must stop. FDA 510(k) K251416 covers software that assists trained medical professionals, excluding expert sonographers, with acquiring ten standard adult two-dimensional transthoracic echocardiography views on compatible ultrasound systems.

The system guides acquisition. It does not turn every user into a cardiologist, and it is not cleared as a roaming oracle that diagnoses patients without supervision. UltraSight's labeling preserves licensed-physician direction and independent clinical review. That governance is not fine print stapled onto the product. It is part of the product.

UltraSight also received clearance for PVAD IQ, software that analyzes cardiac ultrasound clips and supplies measurements relevant to microaxial flow-pump positioning in adults. That expansion moves the company from general acquisition guidance toward a specific intensive-care workflow where timing, consistency, and expert availability can all become constraints at once.

Clinical Evidence Is Becoming Operational Evidence

Medical AI companies love an accuracy number because it fits beautifully on a slide. Hospitals have to live with everything surrounding the number: who acquires the data, how long training takes, what happens when a result is uncertain, and which clinician owns the final decision.

A Stanford randomized trial evaluated internal-medicine residents with no prior echocardiography experience. Residents using UltraSight guidance acquired scans 43% faster than controls and achieved higher overall image-quality scores. The company reported meaningful proficiency after roughly five scans per user. The result supports guidance as a training and acquisition tool, not as a license to erase supervision.

A separate JAMA Cardiology study tested an AI-enabled pathway for moderate-or-greater aortic stenosis. Novice users acquired 1,302 focused cardiac ultrasound exams, 96.6% of which were suitable for automated analysis. The AI-only workflow produced 93% sensitivity and 96% specificity. Expert review of positive and uninterpretable studies, roughly 10% of the exams, increased positive predictive value to 91.1%.

The interesting result is not that experts disappeared. They did not. It is that expert attention could be concentrated on a smaller set of studies. UltraSight's commercial wager is that hospitals will pay for that allocation of scarce attention when it is wrapped in training, compatible hardware, reporting, and accountable clinical review.

Mayo Clinic researchers have collaborated with UltraSight across multiple studies, and Mayo Clinic Ventures is an investor. UltraSight discloses Mayo's financial interest in the technology. Sophisticated buyers should evaluate the peer-reviewed evidence on its merits while keeping the institutional relationship visible.

From Point Solution to Echo Stewardship

UltraSight's recent language has shifted from a guidance application to an Echo Stewardship platform. That sounds like marketing until the operational pieces are unpacked. Hospitals need to decide which focused studies belong at the bedside, who may acquire them, how users are trained, when expert review is required, and how the result enters an established care workflow.

In 2026, Jefferson Health began a pilot evaluating whether advanced practice providers could use UltraSight at the bedside to support right-ventricular assessment in patients with pulmonary embolism. UltraSight also presented Mayo collaboration data across structural heart disease, cardio-oncology, heart failure, and aortic-stenosis screening. Each use case requires its own evidence, but together they show the platform moving toward defined clinical questions rather than generic access to a probe.

UltraSight reports deployments at more than ten commercial sites across intensive-care settings. The company has not disclosed customer names, revenue, renewal rates, contract values, or customer-level outcomes. The traction claim is therefore a useful signal, not a completed commercial scorecard.

The Team Behind UltraSight

UltraSight began with the idea that geometric deep learning could help solve a physical acquisition problem. Yaron Lipman's work spans geometric modeling, shape analysis, and geometric deep learning. Itay Kezurer brought computer vision and graphics expertise. Achi Ludomirsky brought decades of cardiac imaging and pediatric cardiology experience. CEO Davidi Vortman leads the commercialization effort.

That mix matters because cardiac ultrasound is both a geometry problem and an institutional-behavior problem. The software must understand what appears on screen, but the company must also persuade hospitals to change who can acquire a study, under what controls, and with which escalation path. A model can identify the right view. It cannot approve a hospital protocol or win an argument with an integration committee.

Why UltraSight Matters Now

The broader market is moving from AI that reads medical images toward systems that help create better inputs and organize the workflow around them. In ultrasound, acquisition quality determines whether downstream analysis has anything trustworthy to work with. UltraSight sits at that upstream choke point.

The company's opportunity is to turn cardiac ultrasound from a scarce appointment into a governed capability that can travel closer to the patient. The risk is that health systems do not buy software in isolation. They buy implementation, evidence, interoperability, training, accountability, and a credible economic case.

UltraSight's 2026 $24M Series B2 gives it more room to prove that case in the United States. The company already has regulatory clearance, peer-reviewed evidence, specialized clinical workflows, and early commercial deployments. The next milestone is less cinematic: repeatable adoption inside hospitals where every promising technology eventually meets procurement, staffing, governance, and Tuesday morning.