Atrial Fibrillation Is Where Seoul Will Look for Quantum Advantage
Quentir Medicine Monitor
Evidence-based insights for quantum medicine. Published by Quentir Systems LLC · August 25, 2026.
A cardiologist reading a CT scan of a narrowed coronary artery can measure the narrowing and cannot measure the flow. Velocity, pressure and the friction the blood drags along the vessel wall are quantities that inform the risk of a clot forming or a muscle being starved, and none of them is in the picture.
On August 25, 2026 a Korean consortium was given two and a half years and 2.5 billion won to find out whether a quantum computer can supply those numbers on a schedule a clinic could live with. Seoul St. Mary's Hospital of the Catholic University of Korea, the University of Seoul and a medical software company called Flownics were selected for a new 2026 challenge program run by Korea's Ministry of Science and ICT together with the National Research Foundation. Herald Business carried the award on the morning it was announced, with the work running from 2026 through 2028.
Practical takeaway. What this grant buys is a measurement rather than a product. The team has been funded to establish the conditions for quantum gain on one clinical calculation and to report those conditions as numbers, including how many circuits and how many measurements the answer cost.
What a hospital can already see, and what it cannot
Computational fluid dynamics is the standard way to recover the missing quantities. Take the anatomy from a CT or an MRI, apply the physics of a fluid to it, and compute velocity, pressure and wall shear stress everywhere in the reconstructed vessel. Cardiology has already begun buying this. Pressure ratios computed from coronary CT are an established adjunct in stable chest pain assessment, and England's health technology body recommends one such tool for use in the National Health Service, which means the classical version of this calculation has cleared a payer's assessment somewhere.
The trouble is what happens when a clinician asks for more of it. Accuracy in a flow simulation comes from a finer mesh, from honest treatment of the nonlinear terms, and from boundary conditions that reflect a particular patient rather than a textbook. Every one of those improvements multiplies the arithmetic. Past a certain point the answer arrives too late to be part of a decision, and the simulation quietly becomes a research output rather than a clinical one.
The consortium's first target is atrial fibrillation, the most common sustained arrhythmia, present in roughly two to three percent of the population. The choice is a sound one. In a fibrillating heart the left atrium stops emptying cleanly, blood stagnates in its appendage, and stagnation is where thrombus begins. A velocity field inside that chamber is exactly the thing a stroke-risk conversation lacks, and it is also a genuinely hard piece of geometry to simulate.
Four techniques, and the job each one has
The plan starts on classical hardware. A three-dimensional deep learning model will segment the heart, the left atrium and the aorta automatically from CT, because a solver cannot run on an anatomy nobody has outlined. Only then does the quantum hemodynamic model get built on top, and the finished result is checked against real patients imaged with 4D Flow MRI, a sequence that records the speed and direction of the blood itself alongside the geometry of the vessels carrying it. Flownics contributes its own automated analysis platform for that comparison, which puts the measurement of the answer in different hands from the computation of it.
Four quantum techniques are stacked to make the middle step affordable on the noisy processors that exist today. A variational algorithm lets a quantum processor propose a candidate solution while a classical computer scores the error and asks for a correction, looping until the pair converge. A Krylov subspace method narrows the search to the region where the solution plausibly lives instead of exploring the whole space. Classical shadows extract many physical quantities from a comparatively small number of measurements. Non-Markovian error mitigation models the machine's own noise pattern mathematically and subtracts its effect from the result afterwards.
Read together, three of those four attack the same bottleneck. A variational solver for a three-dimensional flow field needs a great many repeated measurements to estimate its expectation values to useful precision. This Monitor reads that as the binding constraint here, alongside circuit depth and qubit count rather than in place of them, because Krylov narrowing and shadow measurement are both tools for spending fewer measurements. The project's own framing points the same way: it states the goal as establishing how far circuit resources and measurement counts can be cut while a target accuracy is held.
Quantum pillar: computing. Technology readiness: TRL 2 of 9. The application has been worked out on paper with its numbers and its accuracy target, and a national funder has agreed the plan is worth building, but no quantum solver for this problem has been run on hardware yet and no patient has been analyzed with one.
Ninety-five percent is a parity target
The headline figure deserves a careful reading. The stated aim is for quantum-based computational fluid dynamics to reach at least ninety-five percent of the precision of the classical method. That is a target to match the incumbent, not to beat it. Any advantage the project finds will therefore have to show up somewhere other than accuracy: in time to answer, in cost, or in problem sizes the classical solver cannot reach at all within a clinical window.
Stating it that way is unusually disciplined for this field. A great many quantum medicine announcements assert a capability and leave the comparison unspecified, which makes them impossible to falsify and therefore impossible to buy against. This one fixes the accuracy at parity, names the baseline, names the disease, names the imaging modality that will adjudicate, and puts the burden of proof on resources. A hospital procurement officer reading that has something to hold the team to in 2028.
It is worth being equally clear about what has not happened. The award is a selection for funding. No quantum result on cardiac flow exists yet from this group, the ninety-five percent figure is an objective rather than a finding, and the accuracy of any near-term variational solver on a problem of this size remains an open question that serious people disagree about. The fullest account of the project reports it as a plan, with named milestones and a named end date, which is the correct shape for a story at this stage.
From hypersonic flow to a fibrillating atrium
The most interesting thing about the team is where its mathematics came from. The quantum algorithm work belongs to An Do-yeol, emeritus chair professor at the University of Seoul. The university's own account of the award states that his earlier work was carried out under a United States Air Force Office of Scientific Research project on quantum computing algorithms for hypersonic flow analysis and a Korean program on low-overhead error mitigation for noisy processors, and describes the non-Markovian error mitigation cost function used in the cardiac proposal as one he proposed. Those are the university's claims about its own professor, reported in its supplied release.
Hypersonic airflow over a vehicle and blood swirling in a fibrillating atrium are physically very different problems, with different regimes, different boundary conditions and different validation. What they share is the family of equations underneath, the Navier-Stokes system, and the fact that both grow expensive as the mesh is refined. This Monitor's reading is that a solver technique developed for one is at least a plausible starting point for the other, which would make the earlier defense grant the place the theory was paid for and the medical grant the place it is now to be tested. That is an inference about method provenance, and the project makes no such claim itself. That transfer runs in the direction people usually assume, from a military problem to a civilian one, and it is a reminder that the quantum medicine pipeline is often shorter and stranger than the sector labels suggest.
The group has been checked from outside before. It was the only Korean team selected for a quantum computing challenge organised by the National Center for Advancing Translational Sciences at the United States National Institutes of Health in 2025, and the only Korean team in a 2026 catalyzer program convened by Cleveland Clinic with the investor K5 Global. Neither is a scientific result. Both indicate that people who look at quantum-for-medicine proposals for a living found this one worth backing before a national funder did.
How Quentir Reads It
The Monitor spent yesterday on a shipped product with no quantum technology inside it, and today on a quantum project with nothing shipped. Read as a pair they mark the honest edges of this field, and the readiness ladder is what keeps the two apart. Yesterday's tool sat at the top of it because a company reports the software as finished and in use. Today's project sits near the bottom because the application has been specified and funded and nothing has been built.
What makes the Seoul grant worth watching is the shape of its promise. It commits to a comparison against a classical method that hospitals already own, on a disease with a large population and a clear clinical decision attached, adjudicated by an imaging technique that can independently show whether the computed velocity field is right. Very few quantum medicine programs have all four of those. When the results arrive, they will be checkable, and a negative result will be as informative as a positive one, which is the property that distinguishes a research program from a marketing position.
For a hospital buyer nothing changes for at least two and a half years, and possibly much longer, since a positive finding in 2028 would still face validation, regulatory work and reimbursement. The useful question for anyone approached in the meantime by a vendor claiming quantum methods for cardiovascular imaging is narrow and answerable: against which classical baseline, on which hardware, at what accuracy, and how many measurement shots did it take. Those are the numbers this project has agreed to produce. Ask any other supplier for the same four.
Sources
Primary source: the joint announcement of August 25, 2026 by Seoul St. Mary's Hospital of the Catholic University of Korea, the University of Seoul and Flownics, reporting selection for the 2026 quantum-gain challenge research program of Korea's Ministry of Science and ICT and the National Research Foundation of Korea; the research team, budget, timeline, four quantum techniques, atrial fibrillation target, ninety-five percent precision objective and 4D Flow MRI validation plan are read from that announcement as carried in full by Dailian, with principal investigator Yoon Jong-chan and An Do-yeol of the University of Seoul quoted there. Budget and funding period independently confirmed in Herald Business the same day. Prior selection for the NIH NCATS quantum computing challenge and the Cleveland Clinic catalyzer program, and An Do-yeol's earlier defense-funded hypersonic flow algorithm work and his non-Markovian error mitigation cost function, are claims made by the University of Seoul in its own supplied release as carried by Hankyoreh on August 25, 2026, and are attributed to it in the text. Clinical context for CT-derived coronary pressure ratios is taken from NICE medical technologies guidance MTG32.