AI model detects endometrial cancer with near perfect accuracy
Researchers from Charles Darwin University have unveiled a powerful new AI model called ECgMLP that can detect endometrial cancer with remarkable precision. The model achieved an impressive 99.26 percent accuracy in identifying cancerous tissue from microscopic images, significantly outperforming human specialists and existing diagnostic tools.
A leap forward in cancer diagnostics
Endometrial cancer is the most common gynecological cancer in high-income countries. Traditionally, it is diagnosed through histopathological analysis, where human specialists examine tissue samples under a microscope. However, even experienced doctors can miss subtle indicators of disease. Human diagnostic accuracy typically ranges between 78 and 81 percent, leaving room for error and delayed treatment.
The ECgMLP model uses specialized attention mechanisms and deep learning techniques to analyze tissue slides at a level beyond human perception. It learns to recognize complex patterns in cellular structures and can reliably distinguish between healthy and cancerous tissues.
High performance across multiple cancer types
Although the model was developed specifically for endometrial cancer, researchers tested its accuracy on other common cancers as well. The results were equally impressive:
- Colorectal cancer: 98.57 percent accuracy
- Breast cancer: 98.20 percent accuracy
- Oral cancer: 97.34 percent accuracy
This shows that ECgMLP is not only highly effective but also versatile, making it a potential game changer for histopathology across various cancer types.
Transforming healthcare with AI
Early and accurate detection is critical in the fight against cancer. When caught early, many cancers are highly treatable. Tools like ECgMLP have the potential to save lives by accelerating diagnosis and reducing errors. They can also ease the workload of medical professionals, who are often under pressure in resource-limited settings.
Beyond clinical accuracy, AI-driven diagnostics like ECgMLP can help democratize healthcare. In many regions, there is a shortage of trained pathologists. A reliable AI tool can bring expert-level analysis to hospitals and clinics that otherwise lack specialized personnel or equipment.
What comes next
The development team plans to continue improving ECgMLP and test it in real-world clinical settings. The long-term vision is not to replace human doctors, but to assist them with highly accurate and scalable diagnostic tools. As digital pathology gains momentum, AI models like this could become a core component of modern healthcare systems.
ECgMLP represents a major step forward in the integration of artificial intelligence into medicine. With accuracy levels exceeding 99 percent, it offers a glimpse into the future of diagnostic precision and accessibility.
Reference
Original source: Charles Darwin University News