Sina Hospital to Test AI for Brain Tumor Recurrence

Sina Hospital is preparing to evaluate artificial intelligence solutions for four clinical challenges, including predicting brain tumor recurrence and assessing carotid artery narrowing and plaques, according to ISNA.

Dr Amir Shokri, the hospital’s deputy for research, said participants in a problem-oriented AI-in-healthcare initiative are expected to present initial results in about a month. The solutions will then undergo validation using data from approximately 40–50 new patients to determine whether they meet the expected accuracy target of around 90%. That figure is a target, not a reported result.

For brain tumors, the aim is to analyze medical information, including CT and MRI scans, to help identify patients at greater risk of recurrence after surgery. The vascular applications will examine carotid artery narrowing and plaque characteristics to support clinical assessment and decision-making.

Shokri said approximately 30 research needs had been identified within the hospital before priority challenges were selected in consultation with the Information and Communication Technology Research Institute and organized into a competition for knowledge-based companies.

He stressed that success would depend on clinical usefulness rather than the number of awards or software products. Physicians will assess accuracy, sensitivity, specificity and other relevant measures to establish whether the tools can improve diagnosis without compromising patient safety.

The initiative also looks beyond individual diagnostic applications. Shokri highlighted the potential of AI to organize medical data, coordinate hospitals and manage resources such as blood supplies during emergencies.

Following more than eight months of preparatory work, the hospital hopes to develop the initiative into an annual national program. A permanent secretariat has been proposed to sustain cooperation among healthcare institutions, universities, government bodies and technology companies.
source: ISNA