Radiology & Imaging AI
B

Better Medicine

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. Findings open in BMViewer beside the source images and feed an editable report summary, while BMGateway sits beside a hospital's PACS and scanners as a DICOM node, selects and routes studies, and can strip identifiers before export. The kidney product carries a CE mark as a Class IIa device under the EU Medical Device Regulation. A lung module that tracks nodules and volume doubling time is labeled for research use only, iMeasure offers semi automated lesion measurement and reporting, and liver, pancreas, lymph node and bone modules are on the roadmap.

In a retrospective study at Tartu University Hospital published in Communications Medicine in November 2025, six radiologists read 200 CT scans with and without the software. Average reporting time fell 33% and benign lesion sensitivity rose from 79.9% to 86.3%, while the change in malignant sensitivity was not significant. Better Medicine sponsored the study. Tartu University Hospital is moving the tool into routine clinical use, and Leeds Teaching Hospitals NHS Trust and Pärnu Hospital are clinical partners. Priit Salumaa, Dmytro Fishman, Bohdan Petryshak and Martin Reim founded the company in Tartu in 2020, and it has raised about €6 million, including a €1 million round in August 2025 to fund European rollout, FDA preparation and US pilots.

AI Health Index verifiedNovember 19, 2026
Compare Better Medicine with other vendors
Founded
2020
Headquarters
Tartu, Estonia
Categories
radiology-and-imaging-ai
Indexed Products
BMVision Kidney, BMViewer, BMGateway
Buyer Segments
Hospital radiology departments, Radiologists reviewing abdominal CT, Medical imaging platform partners
Assessment

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

AI Capability
AA on AI CentralityThe artificial intelligence is the product. Remove the model and there is nothing left to sell.
Vendor Published

Detection, 3D boundary drawing, measurement and benign or malignant characterization of kidney lesions on CT are the product, and each comes from the trained model. BMViewer presents those findings over the images and turns them into an editable report summary, and BMGateway moves studies to and from the model, so both exist to serve the analysis. iMeasure offers semi automated lesion measurement and reporting, and the lung and roadmap modules extend the same approach to other organs. Without the AI there would be a DICOM router and a viewer and nothing to sell.

BB on Autonomy and Oversight ModelThe oversight structure is described and one part is missing, commonly the threshold at which the system stops or what happens after it is wrong.
Vendor Published

The software proposes lesion marks, 3D boundaries, measurements and a likely benign or malignant class, and the radiologist reviews and edits each finding and the summary before it enters a report. The reader study measured that assisted workflow rather than the model alone, and Better Medicine frames the tool as a triage assistant and safety net that a radiologist confirms. Sites set which studies the gateway selects and where results go. Thresholds for flagging uncertain findings, how a lesion the model misses is surfaced, and whether a second reader is triggered are not described.

BB on Model and Technology TransparencyThe approach or the suppliers are named without the version and update discipline behind them. Naming a supplier is the entry to this band both here and on Model Supply Chain Disclosure, which ask different questions of the same disclosure: who receives the data, and what produces the output.
Peer Reviewed Publication

BMVision Kidney adapts nnU-Net, an open medical image segmentation framework, adds postprocessing and lesion characterization for renal CT, and was trained on 612 CT volumes, 391 from Tartu University Hospital and 221 from public collections. Its Communications Medicine paper sets out the architecture, the training data and a standalone patient level sensitivity of 96% for malignant lesions, 93.4% counted lesion by lesion.

The model is proprietary, with research access only under a collaboration agreement. Better Medicine is also building a multimodal foundation model trained on CT images, reports and metadata for future organ modules. No release history of kidney model versions, or the checks a new version passes before shipping, is described.

BB on Model Supply Chain DisclosureSubstantial partial disclosure, or a chain that is structurally short: an in house build, a cleared model that cannot be quietly swapped, or a deployment where the transfer does not occur at all. Naming the model provider exits the band below into this one; the axis rises from here on the completeness of the party list and on the terms that govern data once it arrives.
Vendor Published

The kidney model is Better Medicine's own, an adaptation of nnU-Net, an open medical image segmentation framework, trained on 612 CT volumes: 391 from Tartu University Hospital and 221 from public kidney cancer collections drawn from the KiTS challenge and The Cancer Genome Atlas. BMViewer is built on the open OHIF imaging framework. Those components run inside the product rather than as outside services receiving images.

BMGateway can send studies to cloud storage or servers a site chooses. Where live kidney inference runs, whether any outside processor receives images or results, and the retention terms on those copies are not stated.

BB on Clinical and Operational EvidenceNamed deployments with dated outcome figures and enough method to test them, or published research short of independent validation.
Peer Reviewed Publication

Six radiologists with 4 to 26 years of experience read 200 CT scans at Tartu University Hospital with and without the software, separated by a washout of three to four weeks, for 2,400 reads. Published in Communications Medicine in November 2025, the study found average reporting time fell 33%, ranging from 18% to 52% across readers, and benign lesion sensitivity rose from 79.9% to 86.3%, while the change in malignant lesion sensitivity was not statistically significant.

Better Medicine sponsored the study, and several authors work for or own shares in the company. The design is retrospective and single site, with appendicitis patients as controls, and the authors call for multicenter studies.

CC on AI Safety and PHI StewardshipGeneral assurances of privacy and security that do not answer the questions artificial intelligence raises: what is retained, what reaches a model, and what happens to it there.
Vendor Published

BMGateway can anonymize or pseudonymize patient information in DICOM headers before images leave the hospital, and each site chooses where studies go. The reader study ran on anonymized scans under ethics approval granted in June 2023, and the Tartu development data is released only to qualified researchers under data use agreements. The privacy policy keeps product support correspondence for up to 10 years for medical device surveillance but covers only website and inquiry data. How long production images and results are kept, whether they are used to train future models, and how a diagnostic safety event is handled are not stated.

Regulatory and Compliance
DD on HIPAA and BAA PostureNo statement of status and no privacy document that reaches the product. Both halves are absent at this grade: a service level privacy document on its own lifts a vendor to the band above even where no status is stated. A website privacy notice, including one that says it does not cover customer data, does not reach the product, so a vendor whose only document is that notice and who states no status grades here.
Vendor Published

Better Medicine's privacy policy applies only where it acts as controller for website visitors and people who contact it, naming Google and Pipedrive as the services that hold that correspondence. It does not mention BMVision, BMGateway, medical images or the hospital data the kidney model processes, and no processor agreement under European data protection law or business associate agreement is described.

The gateway's deidentification is a product feature rather than a contractual position. US pilots and an FDA submission have been in preparation since August 2025, and no business associate status has been stated alongside them.

DD on Security Certifications and Trust CenterControls are asserted with nothing independent behind them, or nothing is published. Read the note before concluding anything: this is the grade most often corrected on a second pass, because assurance material frequently sits on a parent domain or inside an old announcement rather than on the product pages.
Vendor Published

Better Medicine runs an ISO 13485 quality management system for medical devices, shown with the BSI mark, which covers how the device is designed, built and maintained rather than how the company secures its own operations. The gateway adds deidentification and site controlled routing for data leaving the hospital.

No SOC 2 report, ISO 27001 certificate, penetration test or trust center is offered for the service that processes hospital scans, and the cloud destinations a site connects through the gateway carry their own certifications rather than Better Medicine's. The security controls around live inference rest on description alone.

BB on FDA and Regulatory StatusThe pathway is stated and in progress, or a clearance is named without the vintage and scope a buyer needs to match it to the product on offer. A product outside device regulation also sits here when the vendor publishes what it is for and in what setting and the product on offer matches, or names another regime it works under, such as a registration outside the United States. A general clause saying the service is not medical advice does not reach this band.
Vendor Published

BMVision Kidney carries a CE mark as a Class IIa medical device under the EU Medical Device Regulation for detecting, classifying and measuring renal lesions on contrast enhanced CT, certified in early 2025 and backed by an ISO 13485 quality system. The certificate number, the notified body and a registry entry tying the mark to a software version are not given.

The lung module is labeled for research use only while describing nodule detection and scoring, and the roadmap modules carry no regulatory status. The kidney product has no FDA clearance yet, and the August 2025 funding round was raised in part to prepare a US submission.

BB on AI Governance and Bias DisclosureA governance framework with named process behind it, such as certification to an artificial intelligence management standard, or material written for a customer own review committee to evaluate the product with. A framework covering the vendor whole portfolio counts when its scope reaches this product; what it cannot supply is the product specific evaluation the band above asks for. Published results within the one group where the performance of this product is most at risk also sit here: evaluation on the question this axis asks, short of the comparison across groups the band above requires. A medical device clearance or certification, and the quality system behind it, does not reach this band on its own; it is graded under regulatory status. A regulator reviewed plan governing how the model may change does reach it.
Peer Reviewed Publication

The reader study used a test set of 100 men and 100 women, with cancer cases and controls matched for age, and reports sensitivity, specificity and agreement separately for benign and malignant lesions, the split where the software's effect differs most: benign sensitivity rose with AI while the malignant change was not significant.

No results are broken out by sex, age, lesion size, scanner or contrast phase, and the controls were appendicitis patients rather than a routine CT population, a limit the authors name. Several authors are Better Medicine employees, founders or shareholders. No independent audit of model behavior, or monitoring of performance after deployment, is described.

CC on AI Liability and RecourseMechanisms exist that let someone challenge an output, such as audit trails, source traceability or review before commit, with nothing standing behind the output and no route for the harmed party.
Vendor Published

Radiologists see every proposed lesion mark, measurement and classification beside the source CT and can edit the summary before it enters a report, so an error can be caught before a clinician acts on it. Support questions and complaints go to Better Medicine, and its privacy policy keeps product support correspondence for up to 10 years to meet postmarket surveillance duties for medical devices. Those are routes to review and report a problem. No accuracy warranty, indemnity or remediation period attached to the kidney output is stated, so remedies for a missed or misclassified lesion sit in each hospital's contract.

Integration and Deployment
CC on EHR and Interoperability DepthIntegration is claimed through standards or a middleware layer with no system named and nothing to verify.
Vendor Published

BMGateway connects to local PACS and scanners as a DICOM node, filters which studies reach the model, removes or replaces identifiers and routes results to destinations each site defines. BMViewer, built on the open OHIF framework, shows findings over the images and produces an editable summary the radiologist carries into the report. Deepc and CARPL, two platforms that bring third party imaging AI into hospital workflows, appear among its partners.

Tartu University Hospital said in December 2025 that the tool was moving into its clinical workflow. No named PACS vendor, RIS or EHR integration is described, and it is not stated whether structured findings are written back to the PACS or reporting system.

CC on Deployment Model and Data ResidencyA single hosted option with location implied rather than committed.
Vendor Published

A deployment pairs the BMViewer web viewer with BMGateway, which sits beside a hospital's PACS or scanners as a local DICOM node and sends studies over DICOM, SFTP or to cloud storage the site chooses. Installations run in Germany and Estonia, with pilots in the United Kingdom and South Africa, and the August 2025 funding round was raised partly to start pilots in the United States. Where the kidney model itself runs, on premises or in a hosted environment, which region holds image copies and results, and how one hospital's data is kept apart from another's are not stated.

Commercial
CC on Commercial TransparencyNo price is published and the posture is discoverable: a buyer can establish how the product is sold and what drives the cost before contacting the vendor. Most of the index sits here.
Vendor Published

BMVision is sold to hospital radiology departments, scoped to a site's CT workflow and its existing PACS and viewer, through direct sales and distributors. First installations are paid from hospital site budgets, and Better Medicine describes a later model funded by insurers. Demos are booked by request. The billable unit, whether a site license, a subscription or a charge per scan, is not stated, and neither are hosting, integration or support fees. Funding of about €6 million, including a €1 million round in August 2025, is directed at European rollout and US pilots, which shapes where the sales effort goes next.

BB on Setting and Specialty CoverageCoverage is named with validation behind part of it.
Vendor Published

Radiologists reading contrast enhanced abdominal CT are the users, including on scans ordered for other reasons such as trauma or abdominal pain, where kidney lesions surface incidentally. Validation comes from a single hospital, Tartu University Hospital, using scans from 2010 to 2020 read under instructions focused on kidney cancer, which the authors say may overstate sensitivity for incidental findings.

Installations in Germany and Estonia and pilots in the United Kingdom and South Africa extend the settings in use beyond that site, with no results reported from them. Liver, pancreas, lymph node and bone modules remain on the roadmap, and the lung module is for research use only.

Alternatives

Alternatives to Better Medicine

Other products that do the same kind of work as Better Medicine, and what each one does.

Romanian imaging AI company based in Timisoara whose Rayscape CXR analyzes chest X-rays, detecting and localizing 148 findings across 17 clinical classes, including tuberculosis, with bone suppressed views, an automatic cardiothoracic ratio and a normal or abnormal flag for worklist…

contextflow reads chest CT and shows the radiologist comparable cases alongside its findings.

MosaicOS is a cloud radiology platform from Mosaic Clinical Technologies, the technology subsidiary Radiology Partners launched in July 2025.

Seoul based MRI AI company with US operations. SwiftMR is a deep learning enhancement tool that reduces noise and sharpens MRI images after acquisition, so sites can scan faster or get better images from the same scan.

New York company, founded in 2024, building deep learning tools for PSMA PET/CT in prostate cancer and theranostics.

CARPL is a vendor neutral platform that lets radiology departments and imaging groups buy, test, deploy and monitor other companies' imaging AI through one integration and one viewer.

Commercial

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
Hospital site quote
BMVision Kidney is sold to hospital radiology departments for their CT workflow through direct sales and distributors, with first installations paid from hospital site budgets. No business associate agreement or tier is stated. Terms for processing hospital images sit in each commercial agreement. Deployment connects BMGateway to the hospital's PACS or scanners and sets up BMViewer for the CT workflow. Installation, hosting and support fees are set in each agreement. Vendor Published

No site price, per scan rate or subscription amount appears. Better Medicine also describes a later model funded by insurers for the wider imaging offer.