Support & Documentation

AI4CMR stands for Artificial Intelligence for Cardiac Magnetic Resonance.

AI4CMR is an AI-driven software design to automatize the analysis and reporting of CMR exams. The AI conveys the knowledge of CMR experts, automatically providing relevant clinical criteria and a second opinion. AI4CMR operates fully embedded in clinical workflows and it is compatible with different CMR scanners. AI4CMR features offer overarching benefits:

• time gains
• faster reporting
• reduction in CMR costs
• reduction of waiting lists
• improved patient management
• better outcomes

AI4MedImaging is working with different partners to provide seamless integrations. Please contact us to check if AI4CMR is available for your PACs/VNA – sales@ai4medimaging.com

AI4CMR is certified under the European Union Medical Device Regulation 2017/745. The software underwent a carefully design clinical validation using independent multi-center multi-vendor data, analysed by multiple CMR experts and using a balanced mix of healthy and pathological cases (150 subjects in the pre-clinical stage and 150 subjects in the clinical stage)
The studies support that AI4CMR works as intended, with correlation of 0.99 in ventricular volumes computation and 100% sensitivity in detecting suspect cases when compared with experts.

If you are interested in trying out AI4CMR, please reach us on sales@ai4medimaging.com

Virtually any DICOM conformant PACS, but that is assessed case-by-case.

DICOM reports: for visualization – DICOM SC, DICOM Encapsulated PDF; for interoperability – DICOM SR, DICOM GSPS, DICOM SEG, DICOM KOS

Yes. It has been clinically validated for Siemens, GE and Philips scanners. Evidence is being collected for Canon scanners, and it will be supported soon.

Yes, the ranges are customizable.

AI4CMR generates results snapshots that are included on the reports generated for visualization (DICOM SC and DICOM Encapsulated PDF).

Concurrent use is possible depending on the acquired MedDream licenses.

The physician.

Yes. AI4CMR is a cloud service, which means high scalability and elasticity.

CMR analysis by an expert can take around 1 hour, half of which is spent doing the LV/RV functional analysis. The commercially available version of AI4CMR automatizes this task, resulting in a possible reduction from 1h of analysis time to 30 min. Once the roadmap is complete, we aim to reduce it further down to 10 minutes.

Yes, the data is anonymized before it leaves the client premisses. The anonymization process is based on the DICOM standard Attribute Confidentiality Profiles. DICOM attributes containing Individually Identifiable Information are replaced or removed before the DICOM files are sent to the cloud.

DICOM Anonymization on the client’s premises; data encryption in transit and at rest; data deletion upon client request; limited access to client data via RBAC (role-based access control) and PoLP (principle of least privilege).

Average of 86s. Since the processing is automatically triggered right after acquisition, this time is never perceived.

Yes. It has been proved by software testing.