HJF Collaborates with Google Health and DoD to Use AI to Understand and Treat Breast Cancer Tumors

Bethesda, MD - Researchers with Google Health, Naval Medical Center San Diego (NMCSD), and The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc. (HJF) have developed artificial intelligence (AI) models to predict status of three important biomarkers from routine breast cancer haematoxylin and eosin-stained slides, enabling better understanding of the links between biological factors for treatment and the structure (morphology) of a tumor. 

A recently published paper in Communications Medicine titled, “Determining breast cancer biomarker status and associated morphological features using deep learning,” unveils that the AI models can identify if a tumor has the important estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) biomarkers.

Examples of saliency maps
Saliency analysis for patch-level biomarker status prediction. Examples of saliency maps for patches representing each biomarker status prediction; ER (a), PR (b), and HER2 (c). Inset regions highlight higher magnification on both H&E and IHC for the most relevant regions of saliency analysis. Warm colors on the overlay indicate that the underlying pixels from the H&E slide were salient to the model’s predicted probability for the indicated biomarker status. ER Estrogen Receptor, PR Progesterone Receptor, HER2 human epidermal growth factor receptor 2. 


Treatment plans for breast cancer, one of the most common forms of cancer in the United States, rely on identification of biological molecules within the body, called biomarkers. Having an AI model that can interpret slides and predict biomarkers could help doctors determine the biomarker status of breast cancer and the right treatment more quickly and accurately,” said James Wren, an HJF employee and epidemiologist who is a co-author on the paper. “It could also help scientists better understand how biomarkers impact tumor morphology.”

With more validation, these models could eventually be used to help doctors determine if the tumor histology, its microscopic structure, is consistent with initial biomarker test results, or if additional testing could benefit the patient before treatment is initiated.

This research is the product of an ongoing collaboration between Google Health, NMCSD, and HJF. This paper builds on research done in 2017 and 2018 where AI models were trained to identify metastatic breast cancer. More information about the previous work can be found here.


About HJF

The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc. (HJF) is a global nonprofit that administers more than $500 million in medical research funds annually. For more than 35 years, HJF has partnered with researchers and clinicians to provide bench to bedside to battlefield research support. More than 3,000 HJF teammates perform as trusted and responsive partners by providing scientific, administrative, and program operations services to researchers in the military, academia, and private industry. For more information, visit