inHEART
inHEART turns cardiac CT and MRI scans into three dimensional digital twins that electrophysiology teams use to plan and guide catheter ablation, mainly for ventricular tachycardia. Hospitals upload DICOM studies through regional portals. An FDA cleared AI module segments adult cardiac CT automatically, trained inHEART staff review and correct the result in Shaper, and the clinician inspects, annotates and exports the model in Explorer before loading it into a mapping system such as Carto, Rhythmia or EnSite. Wall thickness maps help locate the surviving channels inside scar that ablation targets.
The service has supported more than 7,500 cases at over 300 partner hospitals in more than 40 countries, including Hartford HealthCare, the University of Pennsylvania, MUSC Health, CHU Bordeaux, Nippon Medical School and Southlake Regional Health Centre. The randomized InEurHeart trial, published in the European Heart Journal in 2026, assigned 113 patients at 14 European centers to CT guided or conventional ablation. Procedure time fell from 149 to 120 minutes by assigned group, a 19% reduction, while the gain in one year freedom from ventricular tachycardia, 76.8% against 67.3%, was not significant.
FDA clearance K231683 covers the adult CT AI module, with version 3 in commercial distribution, and Circle Cardiovascular Imaging also distributes the product. The company is based at the IHU Liryc institute in Pessac, France, with a United States office in Cambridge, Massachusetts.
Capability Axes
An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check. How grades read
The AI module segments adult cardiac CT automatically before a three dimensional model is built, and staff review and edit its output in Shaper. Around it, the suite accepts MRI, lets specialists adjust models and supports viewing, annotation and export in Explorer, imaging software that works without the model. The trained segmentation is what lets CT processing start automatically and quickly, and the reviewed model is what the clinician uses to read the patient's anatomy and plan the procedure.
The CT module segments automatically during preprocessing, and trained inHEART staff review and edit every automated result in Shaper before anyone relies on the model for planning. The electrophysiology team then inspects the model against the source imaging and annotates or exports it in Explorer, with the clinician responsible for patient management and treatment decisions. No numerical threshold stops an automated result, forces manual correction or escalates a difficult segmentation beyond that routine staff review.
The FDA submission for the automatic CT module names a U Net segmentation model trained on 796 CT cases paired with earlier manual segmentations from 2018 through 2022, drawn from United States and European institutions and several scanner manufacturers. The model's segmentation goes to staff to check and refine in Shaper, and the device register lists version 3 as commercially distributed under the same clearance. No release history shows how weights, training coverage or performance change from version to version.
inHEART runs its own trained CT segmentation and its own staff review the results, and the FDA filing names the model approach and its data provenance. The European patient notice refers to a secure health data host and allows contracted research partners access.
No list names the infrastructure and processing companies in each region, and whether another company takes part in segmentation inference is not stated, so the model chain is in house while the handling chain for uploaded images depends on the regional service.
The InEurHeart trial, published in the European Heart Journal in June 2026, randomized 113 patients with prior myocardial infarction at 14 European centers between June 2022 and January 2024 to CT guided or conventional ventricular tachycardia ablation. Procedure time fell from 149 to 120 minutes by assigned group, a 19% reduction, and by 28% per protocol.
One year freedom from ventricular tachycardia was 76.8% against 67.3%, a difference that was not significant, and major adverse events ran 1.8% against 3.5%. The guided arm used inHEART models built with AI based segmentation, with no software version stated. Studies from 2020 and 2022 sit behind the homepage claims of 60% shorter procedures and 38% less recurrence.
Hospitals upload images after removing direct identifiers, and under the European patient notice scans stay available to physicians for five years and may be kept longer for research. Authorized research staff may use them to improve cardiac imaging and AI algorithms, outside research partners may get access under safeguards, and patients can object to research reuse through their hospital. Training on uploaded images is permitted rather than ruled out, and no clinical incident process covers a segmentation error missed before a procedure.
The image service processes scans after identifying fields are removed. In Europe, the patient notice covers CT and MRI imaging used for care and research under GDPR, names inHEART SAS as controller for its research processing, and sets out who can see the data and the safeguards binding research partners. The United States service is sold as HIPAA compliant through its own regional portal. No standard BAA can be read before contracting, and which United States deployments include one is not stated.
inHEART names SOC 2 Type 2 for its United States operation, with SOC 2 and HIPAA marks on its site, and holds ISO 13485 for its medical device quality system in Europe. The patient notice limits image access to authorized hospital staff, inHEART research staff and contractually bound research partners. No SOC 2 report with its audit period and coverage of the hosted workflow can be retrieved, and ISO 13485 audits device quality rather than service security.
FDA clearance K231683 covers inHEART MODELS for reading, interpreting, reporting and planning with medical images, and the device listing ties version 3 of inHEART MODELS AI, in commercial distribution, to that clearance. The AI module is indicated for adults and for CT images, while the suite takes MRI through other modules under the earlier clearance K220727. The labeling leaves the final patient management decision to qualified professionals and excludes primary mammography interpretation, and the technology carries the CE mark in Europe.
The FDA submission sets out where the model's CT data came from, 796 previously segmented cases from United States and European sites across several scanner manufacturers, and notes that demographic details were unavailable for that training data. Staff review in Shaper checks each output before use. No results compare the current model's performance by sex, age, ethnicity or scanner, and ISO 13485 certifies the device quality system rather than any evaluation of model behavior across patients.
Trained staff can correct a segmentation in Shaper, and the physician can verify or annotate the model before transferring it into a procedure, with the FDA labeling leaving final patient decisions to qualified users. The European patient notice lets patients access or correct their personal data through the hospital that can link the record to the pseudonymized image, a right over data rather than over the model's output. No warranty or remedy covers a planning error that survives review, and patients have no direct route to dispute an AI output.
The service takes DICOM images from CT and MRI scanners and sends the prepared model to cardiac mapping systems, with Carto, Rhythmia and EnSite named as compatible, while Explorer works as the viewer and export tool so the physician reviews the model before it enters the procedure.
Integration runs through imaging and mapping rather than the hospital record: no bidirectional connection to a named record system is described, and planning models are not written back into the chart through a clinical API.
Hospitals reach the hosted service through separate portals for the United States and for Europe and the rest of the world, submitting a study and receiving the prepared model back for clinician review. The European patient notice commits to storing and processing personal data in the European Union, with GDPR safeguards for any transfer, and gives physicians five years of access. Where United States cases are processed and stored is not committed, and isolation of hospital data within each region is not described.
Hospitals and electrophysiology teams buy access to the model service through a commercial relationship with inHEART and a regional portal, covering preparation of a cardiac model from each submitted study and clinician access to the result in Explorer. A demonstration comes before purchase, and Circle Cardiovascular Imaging also distributes the product. The workflow takes DICOM images and exports into mapping systems such as Carto and Rhythmia. No license amount, price per case or volume tier is stated.
The CT module serves adult cardiac imaging, with electrophysiology and imaging staff reviewing the models, and the clinical workflow plans ventricular tachycardia ablation at centers that import the model into their mapping system. InEurHeart validated that use at 14 centers in Austria, France, Germany and Switzerland in patients with prior myocardial infarction, and the service has reached more than 300 hospitals in over 40 countries. The suite takes CT and MRI, with the automatic module confined to CT, and its validation is deliberately narrow: ventricular tachycardia after infarction, in electrophysiology labs.
Alternatives to inHEART
Other products that do the same kind of work as inHEART, and what each one does.
Circle Cardiovascular Imaging, based in Calgary and working in cardiovascular imaging since 2007, makes cvi42, vendor-neutral software for reading and reporting cardiac MR, cardiac and vascular CT, structural heart and electrophysiology planning, with StrokeSENS for acute stroke.
AISAP sells AI software for cardiac ultrasound at the point of care, built at the Heart Center of Sheba Medical Center in Ramat Gan, Israel.
Better Medicine builds AI for reading abdominal CT, starting with BMVision Kidney, which detects kidney lesions as small as 5 mm, draws their boundaries in 3D, measures the longest dimension and three orthogonal axes, and separates likely malignant from benign findings for a radiologist to review.
MediView XR builds XR90, an augmented reality visualization and navigation system for image guided interventional procedures, and markets OmnifyXR for interventional suites.
Optellum, based in Oxford with a US office in Houston, sells the Virtual Nodule Clinic, which identifies patients with pulmonary nodules, triages and tracks them, and scores malignancy risk with its Lung Cancer Prediction AI, a convolutional neural network applied to the CT scan.
Exo, pronounced echo, makes the Exo Iris handheld ultrasound probe and describes itself as a medical imaging software and devices company.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
| Entry Price | Pricing Basis | BAA Tier | Implementation | Source |
|---|---|---|---|---|
|
By quote
|
Hospitals purchase access to the cardiac model service. The scope includes preparation of a model from a submitted study and clinician access through Explorer. Standard rates for a case or license are unavailable. | The United States service states HIPAA compliance. A standard BAA and an applicable commercial tier are unavailable. | Setup involves regional portal access, scan transfer and compatibility with the hospital's mapping workflow. Fees depend on the purchase agreement. | Vendor Published |
The service includes model preparation and review by trained staff. The clinician accesses the finished model through Explorer and may send it to a compatible mapping system. Distribution can also be arranged through Circle Cardiovascular Imaging.