Breast cancer diagnosed with almost 100% accuracy using AI system

Thank you. Listen to this article using the player above.

Earlier this year, researchers at Northeastern University unveiled a web-based artificial intelligence tool designed to more rapidly and accurately diagnose prostate cancer.

Now, the same group, led by bioengineering professor Saeed Amal, has developed a new AI architecture for breast cancer detection, which the researchers say has achieved an accuracy rate of 99.72%.

According to the American Cancer Society, breast cancer accounts for 30% of new cases of female cancer each year, and is estimated to kill 42,500 women in 2024.

Research on the findings was recently published in the journal cancer,

These projects are part of a larger effort by Amal to create an online framework that doctors can use to diagnose different types of cancers using these innovative AI techniques. Amal says the new tool will “redefine digital pathology.”

He and his team recently submitted an invention disclosure on the idea to the Research Innovation Center.

“The AI ​​will look at high-resolution images and learn from historical data how to identify cancer patterns and make a diagnosis,” he says. “The AI ​​can't miss tumors in biopsies and won't get tired after diagnosing 10 or 20 people.”

Ideally, this framework will enable doctors not only to treat patients more quickly and accurately, but also help in the development of new AI models that can be used to diagnose rare and uncommon cancers for which there is a lack of patient data, he said.

For the breast cancer project, researchers took advantage of a publicly available dataset containing images of malignant and benign breast tissue found on the Breast Cancer Histopathological Database.

Amal said that based on this data, the team developed a deep learning model, in which different models were used to increase accuracy and reduce error and they were trained on breast tissue image data.

“It's like you're taking diagnoses from multiple doctors and you're voting to make the best decision,” he says.

Reference: Balasubramanian AA, Al-Hijawi SMA, Singh A, et al. Ensemble deep learning-based image classification for breast cancer subtyping and diagnosis of aggressiveness from whole slide image histopathology. cancer. 2024;16(12):2222. doi: 10.3390/cancers16122222

This article is republished from the following materials. Note: Material may have been edited for length and content. For more information, please contact the cited source. Our press release publication policy can be accessed Here,