AI Revolutionizes Cancer Detection with Unprecedented Accuracy
In a groundbreaking development poised to change the landscape of cancer diagnosis, a novel artificial intelligence model has achieved nearly perfect accuracy in identifying cancer. The new AI, developed by an international team of scientists, significantly outperforms existing methods, including those used by medical professionals.
Using AI to save lives
The AI model, dubbed ECgMPL, was created by researchers, including those from Australia’s Charles Darwin University (CDU). ECgMPL can analyze microscopic images of cells and tissue to identify endometrial cancer with 99.26% accuracy. This marks a significant improvement over current diagnostic methods, which typically achieve an accuracy of around 78.91% to 80.93%.
“The proposed ECgMLP model outperforms existing methods by achieving 99.26 percent accuracy, surpassing transfer learning and custom models discussed in the research while being computationally efficient,” said Dr. Asif Karim, a co-author of the study from CDU.
How the AI Works
ECgMPL examines histopathology images, enhancing image quality to pinpoint early-stage cancerous growths that may be invisible to the naked eye. Endometrial cancer, one of the most common forms of reproductive tumors, is treatable, and the five-year outcome for patients is good if detected early. However, once the cancer spreads outside the uterus, treatment becomes significantly more difficult, emphasizing the critical need for timely and accurate diagnosis.
Sample images of the dataset the AI was trained on featuring normal endometrium (NE), endometrial polyp (NP), endometrial hyperplasia (EH) and endometrial adenocarcinoma (EA)
The researchers note that the AI model is not intended to replace medical professionals, but rather to assist them in making more accurate diagnoses. This technology could offer a more rapid, accessible, and affordable cancer diagnostic method.
“The same methodology can be applied for fast and accurate early detection and diagnosis of other diseases which ultimately leads to better patient outcomes,” said co-author Niusha Shafiabady, an associate professor at ACU.
Expanding Applications
The researchers found that the methodology behind ECgMPL has much broader applications beyond endometrial cancer. The model has shown promising results in detecting other cancers, including:
- Colorectal cancer: 98.57% accuracy
- Breast cancer: 98.20% accuracy
- Oral cancer: 97.34% accuracy
The study, published in the journal Computer Methods and Programs in Biomedicine Update, highlights the potential of AI to revolutionize cancer detection and improve patient outcomes through accurate and timely diagnoses.
