BrainScan CT detects and classifies 30 critical brain pathologies, already outperforming average clinicians with 7–10 years of experience in over 30% of cases. The system provides fast, accurate support for emergency radiology teams.

The AI model is fully proprietary, built on BrainScan’s own convolutional neural network architecture. It was trained using a large dataset with ground truth labels automatically extracted from radiological reports by a custom text classification engine.

In the coming year, BrainScan will expand its capabilities to include 20 additional pathologies, such as aneurysms. While highly accurate, the system’s primary role is to serve as an automated second opinion-helping hospitals triage CT scans more efficiently and prioritize urgent neurological cases.

BrainScan CT is a CE MDR 2017/745 certified medical device for AI-assisted interpretation of brain CT scans (Certificate No. TNP MDR/0008/4823/2023). The system is also approved for use in the UK (MHRA), Switzerland, Ukraine, and Morocco—ensuring broad international availability.

In partnership with Mass General Brigham, BrainScan is preparing for FDA certification in the United States. This milestone will not only open access to the U.S. market but also facilitate regulatory expansion into Latin America, where FDA alignment accelerates medical device approvals.

BrainScan CT is trained on a proprietary dataset of over 94 million images from 350,000 brain CT scans—with 50,000 more to be added soon. Each scan includes a radiological report, patient demographics (age, sex), and comes from a wide range of CT scanner models across diverse populations.

The dataset covers both pathological and normal cases, ensuring robust model performance and accurate detection. With scan values ranging from €20 to €50, BrainScan’s dataset represents one of the most valuable and comprehensive image collections in neuroimaging AI.

BrainScan CT uses machine learning to analyze head CT scans without disrupting existing workflows. The system runs in parallel with standard imaging protocols, seamlessly appending its results to the CT series as a DICOM file for easy PACS viewing.

Hospitals can choose between secure cloud deployment – compliant with ISO 27001, SOC 2, and HIPAA – or full on-site installation for maximum data control. Both options ensure fast, automated analysis while maintaining full compatibility with RIS and HIS systems. Results can also be exported in PNG format for reporting or review.

BrainScan CT provides a comprehensive dashboard that integrates AI analysis with the original DICOM scan. It presents all relevant information in a clear, visual format – highlighting abnormalities, lesion locations, and detected brain pathologies to support rapid clinical decisions.

The dashboard is accessible on desktop computers, tablets, and bedside monitors. It includes an automatically generated abnormality score expressed as a percentage, along with visual markers showing the location of suspected lesions. Each finding is accompanied by a confidence score, and pathologies with a probability above 50% are flagged as present. This enables emergency and radiology teams to quickly assess and prioritize critical cases.