AI Tool Promises Earlier Breast Cancer Detection in Japan
A collaboration between MIT and one of Japan’s premier cancer hospitals could reshape how physicians identify breast cancer risk years before tumors develop.
The partnership brings Mirai, a deep learning model developed at the MIT Jameel Clinic, to Japanese healthcare settings where breast cancer claims approximately 16,000 lives annually. The tool analyzes mammography images to predict breast cancer risk up to five years in advance.
Mohammed Jameel founded Community Jameel, the organization that co-established the MIT Jameel Clinic in 2018 alongside the Massachusetts Institute of Technology. At a ceremony announcing the collaboration with the National Cancer Center Hospital, Jameel emphasized the family’s connection to Japan and the potential impact on women’s health.
“Mirai is a powerful tool that harnesses AI to improve cancer care for women around the world,” he said. “With the Jameel family’s deep and long-standing connection to Japan, we are delighted that the MIT Jameel Clinic and Community Jameel are collaborating with the National Cancer Center Hospital to open the way for Mirai to improve care for Japanese women at risk of breast cancer.”
Validation Across Continents
Mirai has undergone validation on more than two million mammograms across 72 hospitals in 23 countries. The Japanese study will analyze mammography data collected between 2013 and 2024 from the National Cancer Center Hospital and Yotsuya Medical Cube.
Breast cancer accounts for roughly 23% of all female cancer cases in Japan, translating to approximately 98,782 diagnoses each year. When detected at an early stage, the five-year relative survival rate exceeds 90%.
Current screening protocols in Japan recommend mammograms every two years for women aged 40 and above. While these images allow physicians to detect small lumps and microcalcifications invisible through self-examination, interpretation remains primarily visual and depends on clinical experience.
Dr. Kan Yonemori, director of the Department of Medical Oncology at the National Cancer Center Hospital, described the study’s potential. “This study has the potential to contribute to improving women’s health by predicting an individual’s future risk of breast cancer using large-scale mammography screening data from the past and present,” he said.
Personalizing Screening Approaches
If validated in the Japanese clinical context, Mirai could transform breast cancer screening approaches, enabling closer monitoring for higher-risk individuals while reducing unnecessary tests for those at lower risk.
Regina Barzilay, AI faculty lead at the MIT Jameel Clinic, expressed hope the research would inspire new treatment approaches in Japan.
The study represents one of several initiatives supported by Mohammed Jameel through Community Jameel. The organization advances science and learning across areas including climate change, health and education.
Previous breakthroughs enabled by Community Jameel include the discovery of new antibiotics halicin and abaucin at the MIT Jameel Clinic, critical modeling of COVID-19 spread conducted by the Jameel Institute at Imperial College London, and a Nobel Prize-winning experimental approach to alleviating global poverty championed by co-founders of the Abdul Latif Jameel Poverty Action Lab at MIT.
Jameel Corporation provides additional support for the Japanese breast cancer study.