Turkish Brain Project algorithm flags a glioma mid-scan at Gazi

Turkish Brain Project real-time brain MRI alert system at Gazi University: a region-based convolutional neural network object-detection algorithm integrated into the hospital network, which processes MRI images in real time; Anadolu Agency reports that it evaluates a brain MRI within 5-10 seconds while the patient is still in the scanner and sends the physicians' mobile phones a message with the abnormal regions segmented, through a GSM operator's picture-message service

Not assessed

Not assessed

Gazi University, Ankara (Department of Neurosurgery and Department of Computer Engineering), within the Turkish Brain Project funded by the Presidency of the Republic of Türkiye, with the Presidency's Digital Transformation Office credited for the model's development

Not assessed

Not assessed

Not assessed

Not assessed

A region-based convolutional neural network object-detection algorithm trained on a brain MRI data set built within the Turkish Brain Project. Anadolu Agency: six months of training on stroke, hemorrhage and tumor cases. Integration into the hospital network so MRI images are processed as they are acquired. Anadolu Agency reports evaluation within 5–10 seconds on the hospital's three MRI scanners. An alert channel to the physicians' mobile phones, reported by Anadolu Agency as a GSM operator's picture-message service carrying the segmented abnormal regions. A neurosurgery team that reads the alert at once and can admit and biopsy the patient within a day.

Physicians reviewed the images and admitted the patient. A biopsy and pathology examination established the diagnosis.

The algorithm detected the mass on the real-time MRI data and raised the alert while the patient was still in the scanner. Physicians then read the images, judged the finding consistent with a diffuse glioma and admitted her. The diagnosis itself came from the pathology report after the stereotactic biopsy. What the AI contributed is detection and speed, not a different diagnosis or treatment. The counterfactual is unproven. The authors say radiology would typically take days to report and that further investigations are needed on the advantages of earlier treatment. They give no evidence of what delay this patient would otherwise have faced. The case is reported by the team that built and deployed the algorithm, and the Anadolu Agency coverage carries that team's account. No independent clinical review is cited.

The Turkish Brain Project system detected a mass and alerted physicians while the patient was still undergoing an MRI scan.

Not assessed

Real-time MRI algorithm at Gazi University Hospital alerted doctors to a 46-year-old woman's glioma mid-scan; biopsy next day, Oct 2022.

Improved Physical Health

Not assessed

People.

Not assessed

Not assessed

Not assessed

Not assessed

Yes

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

The developers reported the case. They did not measure how long this patient would otherwise have waited or whether earlier treatment improved her outcome. The cited news coverage repeats the team’s account.

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Developer: Gazi University, Ankara (Department of Neurosurgery and Department of Computer Engineering), within the Turkish Brain Project funded by the Presidency of the Republic of Türkiye, with the Presidency's Digital Transformation Office credited for the model's development. Operator: Gazi University, Department of Neurosurgery, Ankara.

Benefit: Ankara, Turkey. Developer: Turkey. Operator: Turkey.

Not assessed

Not assessed

Not assessed

Not assessed

The lesion is of a kind the model was trained on (the team names stroke, hemorrhage and tumor). The alert reaches a physician who acts on it before the routine radiology report. The hospital can schedule a biopsy or operation promptly. False alerts stay low enough that physicians keep trusting the messages. Anadolu Agency quotes Sağıroğlu that the system had reported over 1,000 abnormal findings and physicians work closely with it because they trust it.

Not assessed

The May 2023 Turkish press coverage reported the published case; no later development in the patient's treatment or the algorithm's deployment was found.

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

AIBD-0117-source-0001–0004

Numbered references below.

Not assessed

Developer, Independent

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

People.

People: 2.

People: One identifiable patient had a diffuse glioma found during her scan, was biopsied the next day and referred for immediate oncological treatment. That was instead of the days the authors say radiology typically needs and the longer wait before treatment. The paper reports no follow-up and no changed outcome. The “saved a life” wording is Anadolu Agency's headline, not the authors' claim.

Not assessed

Not assessed

Not assessed

Not assessed

Independently reported

Not assessed

Not assessed

Healthcare - neurosurgery / brain MRI. Hospital real-time MRI anomaly alerting (academic, non-commercial).

The Turkish Brain Project algorithm was built to detect abnormalities on brain MRI as images are acquired and warn the responsible physicians at once. So the days the authors say radiology typically needs to report, and the longer wait before treatment, are cut to days. Anadolu Agency reports it was trained on stroke, hemorrhage and tumor cases. Flagging this patient's tumor mid-scan is that intended use.

Date basis unassessed: 2022 (precision unassessed). Updated: 22 October 2026.

Not assessed

SHA-256: f2fd1ed0ddbe…

oecd_aim: 2023-05-20-a638.

Not assessed

Not assessed

5 revisions. Latest: 22 October 2026.