AI Model ECgMLP Detects Endometrial Cancer with 99.26% Accuracy, Advancing Cancer Diagnosis
Image Credit: National Cancer Institute | Splash
Researchers from four international universities have developed an artificial intelligence model, ECgMLP, that detects endometrial cancer with a reported 99.26% accuracy using microscopic tissue images. The model, created through collaboration between Daffodil International University in Bangladesh, Charles Darwin University and Australian Catholic University in Australia, and the University of Calgary in Canada, also demonstrates strong performance in diagnosing colorectal, breast, and oral cancers. This development, announced in March 2025, aims to support medical professionals in diagnostic processes and improve patient outcomes.
[Read More: AI Outperforms Humans in Breast Cancer Diagnosis?]
Accuracy in Endometrial Cancer Detection
The ECgMLP model focuses on endometrial cancer, identified by the Cancer Council as the most common gynecological cancer in Australia. It achieves a 99.26% accuracy rate when analyzing histopathological images—microscopic slides of tissue samples—according to the research team. This figure exceeds the typical accuracy range of 78.91% to 80.93% for current automated systems used in endometrial cancer diagnosis. The model employs techniques such as image enhancement and self-attention mechanisms to identify cancerous cells within tissue samples. Dr. Asif Karim from Charles Darwin University, a co-author of the study, noted that the technology could “enhance clinical processes” by providing doctors with more precise diagnostic data.
[Read More: AI Surpasses Radiologists in Prostate Cancer Diagnostics?]
Performance Across Multiple Cancer Types
The model extends its capabilities beyond endometrial cancer, with reported accuracy rates of 98.57% for colorectal cancer, 98.20% for breast cancer and 97.34% for oral cancer. These figures come from tests conducted on various histopathology datasets, as outlined by the research team. Associate Professor Niusha Shafiabady, another contributor, described the model as capable of “fast and accurate early detection” for a range of diseases. The ability to perform consistently across different cancer types highlights the model’s adaptability, based on its design and training methods, which prioritize efficiency and precision.
[Read More: Revolutionizing Cancer Detection: AI Boosts Early Diagnosis Rates by 8%]
Intended Role in Medical Practice
The research team designed ECgMLP to assist healthcare providers in clinical decision-making and enhance patient outcomes. By analyzing tissue images quickly and accurately, the model aims to facilitate earlier cancer detection, a factor often linked to improved survival rates. For endometrial cancer, where timely diagnosis is critical, the tool could offer practical value if integrated into medical settings.
[Read More: Will AI Take Over Radiation Therapy by 2040?]
Source: Medicalxpress, CDU
TheDayAfterAI News
We are your source for AI news and insights. Join us as we explore the future of AI and its impact on humanity, offering thoughtful analysis and fostering community dialogue.