Artificial intelligence helps scientists better predict, diagnose hard-to-detect form of breast cancer often undetectable on screening

COLUMBUS, OHIO – New research at The Ohio State University Comprehensive Cancer Center – Arthur G. James Cancer Hospital and Richard J. Solove Research Institute (OSUCCC – James) is using artificial intelligence (AI) to devise new ways of predicting which patients will develop an aggressive and difficult-to-detect form of breast cancer called lobular cancer, which represents one in every 10 breast cancers diagnosed in the United States.

Breast cancer is the most common cancer in women and the second-leading cause of cancer-related death in the United States. Invasive lobular cancer makes up about 15% of breast cancers. It grows as a long chain of cells versus a clump of cells, so it often shows up as a subtle thickness of the breast on screening mammograms. Because of this, early signs of this cancer are often not easily detected until it has spread to other parts of the body.

“Lobular breast cancer tends to spread in a hidden but aggressive way and can come back many years after initial treatment. When it comes back late or in patients at high risk, survival outcomes are generally worse,” said Arya Roy, MD, a breast medical oncologist who specializes in the treatment of lobular cancer at the OSUCCC – James.

Even though invasive lobular cancer and the more common invasive ductal carcinoma are different in how they grow, spread and respond to treatment, Roy notes that oncologists currently follow the same treatment guidelines for both diseases.

“Lobular cancer has a puzzling pattern: some patients look high-risk based on their clinical features but low-risk based on their genomics stemming from tissue-testing results. The genomic tests we currently use often give unclear or conflicting results for lobular cancer, which makes it harder for oncologists to decide on the best treatment. We urgently need better tools – specific to lobular cancer – that can predict which patients are truly at high risk,” said Roy.

She is leading a new study that aims to fill this gap by using AI to analyze digital images of tumor tissue along with patient health data.

By doing this, they hope to discover new image-based “biomarkers” that can predict if and when the cancer will return, including early relapses. Roy says combining these findings into a clear and reliable risk-prediction tool designed for patients with lobular cancer could improve both the early detection of this disease and treating it.

Dense breast tissue and lobular cancer

According to the Society for Breast Imaging, about 40% of women over age 40 have dense breast tissue, which can make it challenging to detect all forms of breast cancer. This is because both normal glandular tissue and cancerous growths appear white on screening mammograms, whereas fatty tissue appears dark. This extra (dense) tissue can camouflage early signs of cancer, making it challenging for doctors see them.

Roy notes that individuals with dense breast tissue have a higher risk of developing breast cancer. Additionally, dense tissue makes it especially challenging to detect early signs of lobular cancer on mammograms. She encourages women to talk to their primary care physician about whether additional imaging tests might be appropriate for them, based on personal and family medical history. Breast ultrasound and MRI testing can provide a different view to sift through these tissues and potentially find cancers earlier. The U.S. Preventive Services Taskforce breast cancer screening guidelines for women of average risk say women should have an annual screening mammogram from the ages of 40 to 74. 

For more information about breast cancer screening, treatment and research at the OSUCCC – James, visit cancer.osu.edu/breast.

###