New AI system diagnoses brain tumors within minutes

German scientists have introduced a new artificial intelligence-based system designed to diagnose brain tumors rapidly through the analysis of microscopic tissue images. The technology offers a faster alternative to current diagnostic approaches, which rely heavily on complex molecular testing.

According to the researchers, existing methods for diagnosing brain tumors are based on sophisticated molecular analyses that are considered the gold standard for determining tumor type.

However, these procedures require specialized equipment, involve substantial costs, and can take up to two weeks to deliver results. In contrast, the newly developed AI algorithms, known as Hetairos, can provide diagnostic assessments using readily available tissue sample images.

The system was trained on more than 11,000 tumor images collected from 9,606 patients across 11 medical centers worldwide. It is capable of distinguishing among 102 molecular patterns associated with central nervous system tumors, demonstrating broad applicability in neuro-oncology diagnostics.

The developers reported that Hetairos achieves a diagnostic accuracy rate of approximately 87–88 percent, a level comparable to that of experienced medical specialists.

During testing, the system was able to generate diagnostic results in approximately 12 minutes after image upload, whereas conventional molecular diagnostic procedures for similar tumors typically require around 12 days.

Researchers believe the technology could be particularly valuable for hospitals and healthcare institutions that do not have access to advanced genetic testing facilities.

They emphasized, however, that the AI system is not intended to replace existing diagnostic standards. Instead, it is designed to serve as a complementary tool that provides rapid insights into tumor characteristics, enabling clinicians to make faster decisions regarding further testing and treatment strategies.