MIT Researchers Develop Advanced AI Model ‘Mirai’ for Early Detection of Breast Cancer

It is designed to enhance the accuracy of predicting precancerous changes in high-risk women from mammograms.

Researchers at the Jameel Clinic for Machine Learning and MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed a deep learning (DL) model named ‘Mirai’ for early detection of breast cancer.

This AI system claims to have surpassed the current risk-assessment algorithms in estimating the likelihood of breast cancer. It is designed to enhance the accuracy of predicting precancerous changes in high-risk women from mammograms.

Mirai's advanced capabilities have attracted attention, including from business tycoon Anand Mahindra.

On the social media platform X (formerly known as Twitter), Mahindra shared his thoughts on AI's potential in early cancer detection, stating, "If this is accurate, then AI is going to be of significantly more value to us than we imagined and much earlier than we had imagined."

The Role of AI in Cancer Detection

Numerous studies have highlighted the promise of AI in cancer detection. Researchers at Duke University in the United States have developed an interpretable AI model capable of predicting breast cancer risk over five years from mammograms.

Similarly, a study published in the journal 'Radiology' demonstrated that AI algorithms could outperform standard clinical risk models in predicting five-year breast cancer risk.

While there are still challenges to overcome, AI's potential as a critical tool for radiologists is clear, with the promise of saving lives through earlier and more accurate diagnosis.

Inclusive & Accurate Risk Assessment

A research article in 'Science Translational Medicine' discusses how AI-based risk assessment models can associate mammogram features with future cancer diagnoses.

The article emphasizes the need for risk prediction at various time points, the ability to incorporate non-image data (such as age and family history), and consistent performance across different mammography devices.

Two years ago, MIT's CSAIL and Jameel Clinic scientists introduced a deep learning system designed to assess cancer risk solely from mammograms.

The model has demonstrated impressive accuracy and inclusivity, providing equal predictive accuracy for both white and black women—a crucial advancement, given that black women face a 43% higher risk of dying from breast cancer.

As AI continues to evolve, its integration into medical diagnostics holds the potential to revolutionize cancer detection and treatment, offering hope for earlier intervention and better outcomes for patients worldwide.

In a similar development, a UCLA study found that the AI tool "Unfold AI," developed by Avenda Health, detects prostate cancer with 84% accuracy, compared to 67% accuracy for traditional methods. Cleared by the FDA, Unfold AI uses an algorithm to analyze clinical data, enabling precise treatment planning.