- Summits
- Sponsorship
- Get Tickets
- Retreats
- Media Opportunities
- News
- About
Ask most people outside cardiology what an echocardiography or heart ultrasound report looks like, and few would guess it can take a specialist twenty or thirty minutes to produce, measurement by measurement, guideline by guideline, for every single patient. It’s this unglamorous but clinically vital bottleneck that MediRapp set out to solve with EchoRapp, a platform that has quietly grown from a one-cardiologist side project into one of the more closely watched AI-assisted reporting tools in cardiac imaging. MedTech World’s Wara Samar sat down with the company’s co-founders, CMO/CTO Schlomo Aschkenasy, MD and CEO Paola Daniore, PhD, to talk through how the platform came to be, and where they believe AI genuinely belongs in echocardiography.
“If EchoRapp were a person, it would be the very experienced and friendly colleague you can always turn to when you have a complex echocardiography case,” says CMO/CTO and founder Schlomo Aschkenasy. “It takes care of the routine measurements to save you time, helps you interpret the findings and reach the right diagnosis, and then writes the complete report in the required language and format.”
That framing, colleague rather than replacement, turns out to be the throughline of everything MediRapp builds.
MediRapp CMO/CTO and founder Schlomo Aschkenasy, who moved from anesthesiology into cardiology before eventually running his own practice, built the first version of the software simply to solve his own problem: wasted time on echo reports.
“The first version of EchoRapp was actually very simple,” Aschkenasy recalls. “It imported measurements from the ultrasound machine through structured reports and put those measurements into a written report. There was very little interpretation, no integrated viewer, and no possibility to review the images or perform measurements within the software itself.”
What forced the software to evolve wasn’t ambition so much as necessity. Echocardiography itself has become significantly more complex over the past two decades. Aschkenasy points to diastolic function assessment as an example: once a relatively simple read based on mitral inflow patterns, current guidelines run to roughly 40 pages, with around 15 different diagnostic pathways depending on the patient’s condition and available measurements. Add in reference data for some 700 different prosthetic heart valves, figures no clinician could realistically memorize, and the case for structured software support becomes obvious.
Today’s EchoRapp interprets more than 300 measurements against a wide body of guidelines and scientific literature, with a fully integrated image viewer, configurable workflows, and multi-language report generation built around how sonographers and cardiologists actually work, whether that’s a single cardiologist running measurements and reporting alone in their own practice, or a larger hospital setting where a sonographer or junior fellow captures the images and initial measurements, and the supervising cardiologist reviews everything, makes amendments directly in EchoRapp, and finalizes the report.
Aschkenasy’s earlier career, writing software for banks and airlines long before he trained as a physician, left a mark on how MediRapp approaches product design, arguably more than medicine itself did.
“Practicing medicine did not change my philosophy of building software very much,” he says. “It has always been extremely important to me that software is written for the user, not for the programmer.” He recalls digitizing paper forms for bank secretaries decades ago by recreating the exact layout they already knew on screen, so the underlying technology changed completely while the person’s workflow barely did.
That same instinct now shapes EchoRapp: rather than asking clinicians to adapt to the software, the software is built to mirror how a cardiologist already thinks through a case. “As a cardiologist, I constantly ask myself: if I had to solve this clinical problem, how would I want the software to help me?” Aschkenasy says. Being both the developer and a practicing physician, he adds, gave him a depth of understanding of the underlying clinical problem that would be difficult to build secondhand.
For a platform designed to work across every ultrasound machine and PACS system on the market, the hardest technical problem isn’t the AI; it’s DICOM interoperability. “DICOM is officially a standard,” Aschkenasy notes, “but when you get into the details, I have rarely seen a standard that is interpreted in so many different ways by different manufacturers.” He offers a telling example: most systems flag an image as containing color only when it holds color Doppler data, but some vendors trigger the same flag simply because a logo somewhere on the image happens to be blue.
Genuine vendor neutrality, in his view, isn’t about supporting DICOM on paper; it’s about absorbing the accumulated, idiosyncratic ways each manufacturer bends the standard, and adapting quickly without breaking what already works elsewhere.
On the business and regulatory side, Daniore brings a background spanning chemical engineering, roles at Microsoft, Bosch, On Running and BlackRock, a PhD in Digital Health Epidemiology, and a Responsible AI in Healthcare fellowship, a combination that, she says, shows up daily in decisions about what’s genuinely in the interest of both patients and clinicians, and in the culture of the team building the product.
Daniore is candid about what she sees as the industry’s biggest misconception: that clinicians will simply absorb new technology into their existing workflow. MediRapp’s answer has been to keep the clinical need in the driver’s seat. “We don’t start by asking, what can AI do? and then try to build a clinical product around it,” she says. “We start by asking, what does the clinician actually need, and what is the best and most responsible way for technology to help?” AI, notably, was added to EchoRapp only after pilot clinicians asked for automated measurements; it was never the product’s starting point. Reporting itself remains deterministic; AI is scoped specifically to measurement support, with the clinician retaining control throughout.
That discipline extended to MediRapp’s path through MDR certification, completed by a three-person team over roughly seven months. Daniore’s advice to other founders facing the same process is direct: “Make sure you have the right regulatory experts by your side.” Just as important, she says, is building a quality management system the company can actually live by afterward, not one designed merely to pass an audit.
Both founders are notably measured about AI’s role. Daniore sees its clearest future value not in speeding up measurements, but in closing the gap between objective data and the morphological judgment cardiologists still perform manually, particularly in early disease detection and recognition of rare conditions, where subtle patterns across images can escape even experienced clinicians. Where she draws a firm line is autonomy: “We feel very strongly that AI in medicine must be implemented in a way that keeps the physician in the loop… not to make that decision for them.”
Aschkenasy echoes that boundary when addressing skeptical cardiologists directly: “If you removed all the AI from EchoRapp tomorrow, it would still be a very powerful echocardiography platform. AI is not the reason for the software to exist; it is a technology that makes an already strong system even better.” The goal, he says, is software that makes physicians more informed and more capable, never less involved.
It’s a stance that positions MediRapp somewhat against the current of AI-in-healthcare hype: less concerned with what the technology can automate, and more focused on what it can responsibly hand back to the clinician’s judgment.
