Get notified of page updates

Article: Artificial intelligence (AI) may find breast cancer on mammograms sooner

Summary

Artificial intelligence (AI) tools may help doctors read mammograms. This assistance may lead to earlier diagnoses of breast cancer. (Posted 9/29/23)

Printer Friendly Page Read the Original Article
Artificial intelligence (AI) may find breast cancer on mammograms sooner

RELEVANCE

Most relevant for: People interested in the use of artificial intelligence in medicine. It may also be relevant for:

  • healthy people with average cancer risk
  • people with a family history of cancer
  • previvors

Relevance: Medium-High

Research Timeline: Human Research

Relevance Rating Details


What is this article about?

This article focuses on a new way radiologists can use artificial intelligence (AI) to improve the interpretation of mammograms. AI can spot breast cancer at least as well as humans and often earlier than humans, sometimes up to 5 years earlier. In addition, AI can help general radiologists interpret with the expertise of specialized breast imagers. 

Why is this article important?

Mammograms can find breast cancer before you can feel a lump or notice other symptoms. Doctors look for abnormal changes on a . If an abnormality is seen more tests are ordered to determine whether it is or is not cancer.

Breast cancers that are found early are easier to treat and are often curable. has reduced deaths from breast cancer.

But mammograms are not perfect, and they sometimes fail to find cancers. According to the National Cancer Institute, mammograms miss 20 percent of breast cancers, and even more in women with dense breast tissue. Doctors who interpret images cannot always see tiny changes, especially when breasts are dense (see our XRAY review on breast density by clicking here). Still, half of people who have yearly mammograms at some point have a result—a finding that requires more testing but is then found to be not be cancer.

For these reasons, doctors continue to look for ways to improve the accuracy of mammograms. This article focuses on using artificial intelligence (AI) to help read mammograms. Studies show that AI can find breast cancers before they are visible to a highly trained human eye.

How does it work?

Typically a is read by one or two radiologists (doctors who specialize in reading x-rays and other images). AI acts as a second “reader,” identifying areas on the image that should be reviewed or imaged more closely. AI technology will soon be able to  compare mammograms of a person that were taken at different times. This is critical because change is often a sign of cancer. For example, AI might notice a very slight difference between the you had last year and the one you had this year. It would then mark the suspicious area for a doctor to review. AI programs can detect smaller changes that could be missed by the human eye and assist doctors in making a final and more effective diagnosis.

A number of different AI systems are being used in countries around the world, including the United States. A Danish study showed that AI reduced false-positive results by 25 percent. AI does not replace doctors who interpret mammograms. Instead, it helps them by reducing their workload and finding cancers they might miss.

In the United States, the Food and Drug Administration () has approved over 20 AI systems for use with mammograms. Most AI systems that are approved in the US are designed to work with older, two-dimensional (2D) mammograms. Since more than 80 percent of sites that perform mammograms in the US use the newer type of 3D (also called digital breast ) the AI systems designed to work with the older mammograms are not compatible. Nevertheless, there are AI systems that are currently approved and in clinical use for most systems. In fact, every day more and more breast centers are integrating the use of AI and we expect that will result in improved outcomes and probably the detection of earlier, more curable breast cancers.

What does this mean for me?

The use of AI to read mammograms is increasing, both in Europe and  in the United States. In the US, multiple AI systems have been approved by the , so clinics are increasingly using AI to assist in the interpretation of mammograms

Various clinical trials are underway around the world to help improve the use of AI in reading mammograms. These trials are being done to ensure that AI can be used to help read mammograms in people of different ages, races and body types. Research is also being done to see whether AI reduces the number of false positives that are not cancer. There are enough breast centers that are using AI that you might consider asking if your breast center has implemented AI. 

For women at increased risk there are other imaging tests are recommended in addition to , like . Women at high risk might need annual or in addition to to detect the earliest ,most curable breast cancers.  There is AI in development for these imaging approaches as well. Stay tuned for further developments.

References

Satariano A and Metz C, Using A.I. to Detect Breast Cancer That Doctors Miss. The New York Times; published March 5, 2023.

Lauritzen AD, Rodriguez-Ruiz A, von Euler-Chelpin MC, et al. An Artificial Intelligence–based Screening Protocol for Breast Cancer: Outcome and Radiologist Workload. Radiology; 2022; 304; Article number 1. Published online April 19, 2022.

Taylor CR, Monga N, Johnson C, Hawley JR, Patel M. Artificial Intelligence Applications in Breast Imaging: Current Status and Future Directions. Diagnostics. 2023 Jun 13;13(12):2041. Published online June 13, 2023.

National Cancer Institute fact sheet. https://www.cancer.gov/types/breast/mammograms-fact-sheet. Updated February 23, 2023.

Disclosure: FORCE receives funding from industry sponsors, including companies that manufacture cancer drugs, tests and devices. All XRAYS articles are written independently of any sponsor and are reviewed by members of our Scientific Advisory Board prior to publication to assure scientific integrity.

Share your thoughts on this XRAY review by taking our brief survey.  

posted 9/29/23

Questions To Ask Your Doctor
Questions To Ask Your Doctor

  • Am I at high risk for breast cancer?
  • What screening methods are best for me?
  • Does your clinic use AI to help read mammograms?

Open clinical trials
Open clinical trials

The following are breast cancer screening or prevention studies enrolling people at high risk for breast cancer.   

Additional risk-management clinical trials for people at high risk for breast cancer may be found here.

Updated: 01/10/2023

Expert Guidelines
Expert Guidelines

The National Comprehensive Cancer Network breast screening guidelines recommend the following for women at average risk for breast cancer: 

  • ages 25-39: 
    • practice breast awareness
    • clinical breast exam every 1-3 years
    • risk assessment, including questions about family and personal medical history should be done during clinical exam to find high risk women who may need additional screening
  • ages 40 and older:
    • practice breast awareness
    • yearly clinical breast exam
    • risk assessment, including questions about family and personal medical history should be done during clinical exam to find high risk women who may need additional screening
    • yearly  - consider if available. 
  • NCCN has a different set of guidelines for individuals who are determined to be at increased risk for breast cancer.

Many other professional societies and organizations have breast cancer screening guidelines that differ slightly. They don't all agree on the age screening should start and how frequently they should be done. 

It is important to note, that all the groups support the opportunity for women ages 40 to 49 to decide it screening is right for them.

Updated: 02/05/2022

Expert Guidelines
Expert Guidelines

The National Comprehensive Cancer Network (NCCN) provides breast cancer risk management guidelines for people with and mutations. We recommend that you speak with a genetics expert who can look at your personal and family history of cancer and can help you to determine the best risk management plan. Note that when we use "men" and "women" we are referring to the sex you were assigned at birth.

Screening for women:

  • Beginning at age 18, be aware of how your breasts normally look and feel. Tell your doctor about any breast changes.
  • Beginning at age 25, have a doctor examine your breasts every 6-12 months.
  • Beginning at age 25, have an annual breast  with contrast (or  if  is unavailable).
  • Beginning at age 30 until age 75, have an annual and an annual breast with contrast.
  • After age 75, speak with your doctor about the benefits and risks of screening.

Risk reduction for women:

  • Speak with your doctor about of the advantages and disadvantages of risk-reducing mastectomy.
    • Research has shown that risk-reducing mastectomy can lower the risk for breast cancer in high risk women by about 90%. Despite this, mastectomy has not been shown to help high risk women live longer.
    • Even after mastectomies, some breast tissue-and therefore some breast cancer risk remains. 
  • Speak with your doctor about the benefits and risks of tamoxifen or other estrogen-blocking drugs to reduce breast cancer risk. The benefits and risks may be different for women with vs.  mutations. Research on the benefit of these drugs to reduce breast cancer risk in women with  mutations has been mixed.

Risk management in men:

  • Beginning at age 35, learn how to do breast self-exams to check for breast changes.
  • Beginning at age 35, have a doctor examine your breasts every 12 months.
  • Beginning at age 50, consider annual  (especially for men with mutations). 

Updated: 07/28/2023

Expert Guidelines
Expert Guidelines

The National Comprehensive Cancer Network (NCCN) provides guidelines for management of breast cancer risk in people with inherited mutations linked to breast cancer. We recommend that you speak with a genetics expert who can look at your personal and family history of cancer and help you to determine the best risk management plan. 

or

  • Begin yearly mammograms at age 40, or earlier based on the family history of breast cancer.
  • Consider yearly with contrast beginning at age 30-35.

, or  

  • Beginning at age 40 (or earlier based on family history):
    • recommend yearly .
    • consider yearly breast  with contrast.

  • No specific breast cancer screening guidelines. Risk management should be based on family history of cancer.

 

  • Beginning at age 30 (or earlier based on family history):
    • recommend yearly .
    • consider yearly breast  with contrast.
    • discuss risk reducing mastectomy. 

  • Beginning at age 30 (or earlier based on family history):
    • recommend yearly .
    • recommend yearly breast  with contrast.
    • discuss risk reducing mastectomy. 

  • Beginning at age 18, learn to be aware of changes in breasts.
  • Begining at age 25:
    • clinical breast exam every 6-12 months beginning at age 25 (or 10 years earlier than the youngest age of onset in the family).
  • Beginning at age 30:
    • yearly and breast with contrast beginning at age 30-35 (or 5-10 years earlier than the youngest breast cancer in the family).  

RAD51C/RAD51D

 

  • Yearly  beginning at age 40 or 5-10 years younger than the earliest age of diagnosis in the family (but no later than age 40).
  • Consider yearly breast with contrast beginning at age 40 or 5-10 years younger than the earliest age of diagnosis in the family.

  • Beginning at age 30:
    • clinical breast examination by a health care provider every 6 months starting at age 30.
    • recommend yearly .
    • recommend yearly  with contrast.

  • Beginning at age 18, learn to be aware of changes in breasts.
  • Beginning at age 20:
    • clinical breast examination by a health care provider every 6 months.
    • recommend yearly breast  with contrast beginning at age 20 or at the age of earliest breast cancer diagnosis if there is a history of breast cancer before age 20 in family.
  • Beginning at age 30
    • recommend yearly
  • Consider risk reducing mastectomy.

Updated: 09/22/2022

Back to XRAY Home