Getty Images

Improving Breast Cancer Imaging with Artificial Intelligence

NYU researchers developed an artificial intelligence tool to improve the accuracy of breast cancer imaging.

A study led by New York University (NYU) researchers created an artificial intelligence tool to improve the accuracy of breast cancer imaging. The computer program was trained to identify patterns among thousands of breast ultrasound images to aid physicians in diagnosing.

When tested on 44,755 already completed ultrasound exams, the artificial intelligence tool increased radiologist’s ability to accurately identify breast cancer by 37 percent. Additionally, the tool helped reduced the number of tissue samples and biopsies necessary to confirm tumors by 27 percent.

The NYU team’s AI analysis is potentially the largest of its kind, involving 288,767 separate ultrasound exams taken from 143,203 patients treated at NYU Langone hospitals between 2012 and 2018.

“Our study demonstrates how artificial intelligence can help radiologists reading breast ultrasound exams to reveal only those that show real signs of breast cancer and to avoid verification by biopsy in cases that turn out to be benign,” study senior investigator Krzysztof Geras, PhD, said in a press release.

Ultrasound exams use high-frequency sound waves passing through tissue to create real-time images of breast and other tissues. While ultrasound is not typically used as a breast cancer screening tool, it has served as an alternative to mammography or for follow-up diagnostic tests for patients.

Ultrasound is a cheaper alternative and more widely available in community clinics. Additionally, it does not involve radiation exposure and is better than mammography in penetrating dense breast tissue to distinguish healthy cells from tumors, according to researchers.

However, the technology has also resulted in many false diagnoses of breast cancer, producing anxiety and unnecessary procedures for patients. Some studies have shown that many breast ultrasound exams indicating signs of cancer turn out to be noncancerous after biopsy.

“If our efforts to use machine learning as a triaging tool for ultrasound studies prove successful, ultrasound could become a more effective tool in breast cancer screening, especially as an alternative to mammography, and for those with dense breast tissue,” said study co-investigator and radiologist Linda Moy, MD.

“Its future impact on improving women’s breast health could be profound,” Moy added.

Geras noted that while the team’s initial results are promising, the researchers only studied past exams in their analysis. Clinical trials of the tool in current patients and real-world conditions are necessary before being routinely deployed.

Geras also plans to refine the AI software to include additional patient information such as family history or genetic mutation associated with breast cancer to determine risk.

For the study, over half of ultrasound breast examinations were used to create the computer program. Ten radiologists then reviewed a separate set of 663 breast exams, with an average accuracy of 92 percent.

With the use of the AI model, the average accuracy increased to 96 percent. All diagnoses were then checked against tissue biopsy results.

According to the latest information from the American Cancer Society, it is estimated that one in eight women (13 percent) in the United States will be diagnosed with breast cancer over their lifetime, with over 300,000 positive diagnoses in 2021.

Next Steps

Dig Deeper on Artificial intelligence in healthcare

xtelligent Health IT and EHR
xtelligent Healthtech Security
xtelligent Healthcare Payers
xtelligent Pharma Life Sciences
Close