From a Computational Moonshot to a $585M Medtronic Acquisition: CathWorks and the Science That Made It Possible

Vision Elements
4 hours ago
7 min read

February 2026 | Vision Elements

We are immensely proud to congratulate CathWorks, our portfolio company, on its acquisition by Medtronic (NYSE: MDT) — one of the world's largest healthcare technology companies. This landmark milestone, valued at up to $585 million, is the culmination of over a decade of relentless innovation at the intersection of computational science, artificial intelligence, and interventional cardiology.

As the team that conceived and built the foundational technology behind CathWorks, this moment is deeply meaningful to us at Vision Elements. It is a testament to what happens when deep scientific inquiry meets bold clinical vision — and when a computational challenge that many deemed impossible is met head-on with the right blend of physics, mathematics, and engineering.

The Problem: A Life-Saving Measurement Trapped Behind an Invasive Procedure

Coronary artery disease (CAD) remains the leading cause of death globally. When a physician suspects a blockage in a patient's coronary arteries, the critical clinical question is: does this narrowing actually restrict blood flow enough to warrant intervention?

The gold standard for answering that question is Fractional Flow Reserve (FFR) — a ratio that quantifies the pressure drop across a stenosis to determine its functional significance. An FFR below 0.75–0.80 is a well-established predictor of ischemia, and decades of clinical evidence have demonstrated that FFR-guided treatment decisions lead to better patient outcomes and lower healthcare costs.

But there is a catch. Traditional FFR measurement is invasive. It requires:

Insertion of a guidewire equipped with a miniature pressure transducer, threaded through the patient's coronary arteries and across the lesion
Administration of adenosine (or another hyperemic agent) to induce maximum blood flow — a pharmacologic stress on the heart
Measurements taken at a single transducer location, providing data for only one point in the vessel at a time

The procedure adds risk, time, cost, and complexity to an already delicate catheterization. As a result, despite its proven clinical benefits, wire-based FFR has remained dramatically underutilized in clinical practice worldwide. The medical community was caught in a paradox: everyone agreed FFR was valuable, but the barrier to actually performing it was too high.

The Breakthrough: Extracting Physiology from Anatomy

This is where our story begins.

At Vision Elements, we asked what was, at the time, a radical question: Can we derive the same physiological information — blood flow, pressure gradients, fractional flow reserve — purely from the anatomical images that physicians are already capturing during routine angiography?

In other words: can we turn standard 2D X-ray angiograms — images of structure — into a window on function?

This is, fundamentally, a remote sensing problem. Just as satellite-based remote sensing extracts atmospheric composition, terrain elevation, or ocean temperature from electromagnetic radiation patterns without physically touching the target, we set out to extract hemodynamic physiology from radiographic anatomy — without ever touching the blood vessel with a sensor.

The scientific community was skeptical. Anatomy and physiology, while related, are governed by different physical principles. A vessel can appear severely narrowed yet carry adequate flow. Another may appear only modestly stenosed yet critically impair perfusion. The relationship between the two is nonlinear, patient-specific, and dependent on the entire vascular tree — not just the lesion in isolation.

We had to invent a way to cross that barrier.

The Deep Computational Research

The solution required bridging multiple domains of computational science — a rare convergence of computer vision, fluid dynamics, 3D reconstruction, numerical methods, and artificial intelligence. No single existing technique was sufficient. The problem demanded a new, integrated computational framework built from first principles.

3D Vascular Reconstruction. Standard angiograms are flat, two-dimensional X-ray projections of a beating heart. Extracting the true three-dimensional geometry of the coronary tree from these 2D shadows — while compensating for cardiac motion, respiratory movement, and projection distortion — is an ill-posed inverse problem. Our team developed methods to identify corresponding vessel segments across multiple views and reconstruct the full 3D arterial tree, including diameters, lengths, branching angles, and bifurcation topology.

Hemodynamic Modeling from Geometry. With the 3D model in hand, we computed blood flow dynamics without any physical sensor. The key insight: vascular resistance can be derived from geometry. Using extensions of fluid-mechanical principles and empirical scaling laws, we modeled the coronary tree as a network of resistances and calculated flow through both stenotic and normal pathways.

Virtual Revascularization. We then invented a way to computationally model what each vessel would look like if the stenosis were removed — an astenotic baseline reconstructed from the healthy portions of the vessel. This enabled a direct comparison:

where QS is the computed flow through the diseased vasculature and QN is the flow through the virtually revascularized vessel. The resulting ratio captures the functional impact of the lesion — no guidewire, no drug, no physical measurement inside the vessel.

Whole-Tree Assessment. Unlike wire-based FFR, which yields a single pressure reading at one point, our computational approach models the entire coronary tree simultaneously — producing a complete physiological map with FFR values along every vessel, simulated pullback curves, and lesion-by-lesion impact analysis.

AI Integration. Finally, advanced machine learning automated the most labor-intensive steps — vessel segmentation, image selection, landmark detection — transforming what would otherwise be a lengthy manual workflow into a fast, reliable system that runs in real time during catheterization.

The Intellectual Property: Protecting a Decade of Innovation

The depth of the computational research is reflected in the intellectual property portfolio: approximately 42 patent families assigned to CathWorks, spanning worldwide jurisdictions including the US, EU, Japan, China, Korea, and Israel.

The patents protect multiple layers of interrelated inventions — from the core method of computing FFR from 2D angiographic images in real time, through 3D vascular reconstruction techniques, correspondence modeling across projections, virtual revascularization algorithms, and automated disease severity scoring. More recent filings extend into angiographic image selection optimization, post-PCI analysis, four-dimensional coronary motion analysis, and machine-learning-based vascular segmentation — reflecting a platform that has continued to evolve well beyond its initial conception.

This is not a single patent wrapped around a product. It is a dense, interlocking body of IP that protects every major computational layer of the system.

From Lab to Cath Lab: The Clinical Validation

Groundbreaking science means little without rigorous clinical evidence. CathWorks invested heavily in demonstrating that computational FFR could match the invasive gold standard:

Pooled Analysis of 5 Prospective Cohort Studies (Witberg et al., JACC Cardiovascular Interventions, 2020): Demonstrated the highest diagnostic accuracy among non-hyperemic indices and angiography-based technologies, with excellent positive and negative predictive values compared to wire-based FFR.

PROVISION Trial (Randomized Controlled Trial, presented at TCT 2024 and PCR 2025): This landmark study demonstrated non-inferiority of FFRangio to wire-based FFR, with similar revascularization rates and comparable major adverse cardiovascular events (MACE) at one year. This is the strongest level of clinical evidence — a randomized head-to-head comparison showing equivalent outcomes.

ALL-RISE Global Clinical Study: Completed enrollment of 1,927 patients — one of the largest studies of angiography-derived physiology, further building the evidence base for real-world clinical use.

Real-World Outcomes Studies (2022–2023): Published and presented data showing excellent one-year outcomes in clinical practice, consistent with data from large wire-based FFR real-world registries.

The FFRangio System received US FDA 510(k) clearance (2018), CE Mark (2017), Japan PMDA approval (2019), and EU-MDR approval (2024), with CathWorks holding ISO 13485 and ISO 27001 certifications.

The Medtronic Acquisition: A Validation of Deep Tech

When Medtronic announced on February 3, 2026, its intent to exercise the acquisition option for CathWorks — following three years of a co-promotion partnership that began in 2022 — it was a validation not just of a product, but of a philosophy.

As Jason Weidman, Senior Vice President and President of Medtronic's Coronary & Renal Denervation business, stated: "This acquisition allows Medtronic to transform the cath lab with a technology that provides real-time data, informs individualized treatment approaches, and drives new standards of care."

CathWorks CEO Ramin Mousavi added: "Bringing Medtronic and CathWorks together will create a best-in-class organization focused on driving new standards of care to transform the cath lab."

For us at Vision Elements, this outcome embodies our core mission: accelerating and de-risking scientific, computer vision, and artificial intelligence missions. When CathWorks was co-founded in 2013, it drew on the deep computational expertise that the Vision Elements team had forged over years of tackling the hardest problems in computer vision, physics-based modeling, and applied mathematics. Our scientists and engineers — specialists in 3D reconstruction, fluid dynamics, numerical methods, and machine learning — provided the foundational research and development that turned an audacious idea into a working system. The goal was bold: replace an invasive medical device with pure computation, deployed in real time, at the point of care.

That device — the pressure guidewire — has been a pillar of interventional cardiology for decades. To replace it with software required not incremental improvement but a fundamental scientific leap: proving that the physics of blood flow could be faithfully captured from the physics of X-ray imaging, and that the gap between anatomy and physiology could be bridged through mathematics and computation alone.

What This Means for the Future

CathWorks' integration into Medtronic's vast global infrastructure — spanning 95,000+ employees across 150+ countries — means that FFRangio will reach exponentially more patients, in more cath labs, in more countries. The technology that was born in our research labs will now be available to transform cardiovascular diagnosis and treatment on a truly global scale.

But beyond this specific success, the CathWorks story carries a broader message for the deep tech community: computational science can replace physical devices. When the underlying physics is understood deeply enough, and the mathematical and algorithmic machinery is powerful enough, software can do what hardware once monopolized — less invasively, more comprehensively, and at greater scale.

At Vision Elements, this is what we do. With over 300 completed projects, 150+ patent applications supported, and now $2B+ in total M&A value contributed across our portfolio, we continue to partner with visionary teams tackling the hardest computational challenges in healthcare, autonomous systems, defense, and beyond.

To the entire CathWorks team — past and present — and to our co-founders Ran Kornowski, MD, and Ifat Lavi, PhD: congratulations on this extraordinary achievement. The science you helped build is now in the hands of a global leader, poised to improve the lives of countless patients worldwide.

The best computational challenges are the ones everyone says can't be solved. This one was.

Guy Lavi is the Founder and Managing Partner of Vision Elements and Co-Founder and former CEO of CathWorks.

Vision Elements provides expertise in computational sciences, including computer vision, deep learning, physics-based modeling, and AI. Learn more at www.vision-elements.com.

For more information about CathWorks and the FFRangio System, visit cath.works.