Future of Breast Cancer Screening and Artificial Intelligence
There is an urgent need to improve the accuracy of screening mammography. A recent study from the National Cancer Institute concluded that 20% of early-stage breast cancers are not detected on standard mammographic screening (Ref. 1). Recent studies have demonstrated that Artificial Intelligence (AI) assisted breast imaging has the potential to markedly reduce the number of breast cancers that are not detected on routine mammographic screening.
An early report on AI-assisted mammography concluded that 31% of cancers not seen on the standard mammogram could be detected on the AI-assisted mammogram (Ref. 2). A more recent study found that AI detected 79% of the cancers that were not detected on the screening mammogram (Ref. 3). A recent case study reported that in retrospect, a small breast cancer could be detected on AI 4 years before it could be detected on the standard mammogram (Ref. 4). These studies clearly demonstrate that AI-assisted mammograms can detect potentially curable breast cancer before they can be detected on the standard mammogram.
Why Cancers Are Missed on Mammograms (False Negatives)
The standard mammography is very effective at detecting small breast cancers in women with a fatty breast pattern (breast density). Breast cancers are white on the mammogram and fatty tissue is black. The contrast makes it relatively easy to detect small breast cancers. The situation is reversed in women with dense breasts. Dense breast tissue is white on the mammogram making it more challenging to detect small white breast cancers. Detecting a small breast cancer in a dense breast has been compared to the challenge of finding a snowman in a snowstorm.
Young women are more likely to have dense breasts than older women. One recent study found that 74% of women aged 40-49 had dense breasts (Ref. 5). Another important study concluded that 98% of breast cancers were not detected on the screening mammogram occurred in women with dense breasts (Ref. 2). In addition, breast density is a major risk for developing breast cancer and women with dense breasts are at higher risk of developing more aggressive breast cancers (Ref. 6-7).
What is Artificial Intelligence?
The incorporation of AI into the screening process is an emerging field that is very promising in improving the early detection of breast cancer (Ref. 8-9). The goal of AI is to make smart computer systems that perform like humans solve complex problems. The incorporation of AI into the screening process enables a machine to simulate human behavior such as interpreting the findings on a digital mammogram (tomosynthesis).
What is Machine Learning?
The standard mammogram is interpreted by a single radiologist. With AI-assisted mammography, a report is issued from both the radiologist and by the AI-based imaging system. This double reading facilitates the process of machine learning (Ref. 10). For example, assume that the AI-based system fails to identify a small cancer, but the cancer is detected by the radiologist. In this situation, the AI-assisted algorithm can be modified so that it will not make a similar error in the future.
A second limitation of standard mammographic screening is the issue of unnecessary biopsies. When a radiologist identifies an area of concern on the screening mammogram, additional views are typically recommended. If the area of concern persists, a biopsy is recommended. If the biopsy proves to be benign, it is referred to as a false positive biopsy. (Ref. 11). A recent study concluded that half of all women who had regular mammographic screening for a period of 10 years had at least one false positive biopsy (Ref. 12). Recent studies have determined that AI-assisted mammography can reduce false positive biopsies rates by 69% (Ref. 13-14). Both AI-assisted ultrasounds and AI-assisted MRIs can further reduce the rates of unnecessary biopsies (Ref. 15-16).
Other Benefits of AI
Reduce Callbacks: Callbacks refer to the situation in which a woman receives a phone call a few days following her mammogram informing her that she must return to the imaging center for additional views. This phone call usually results in a major spike in anxiety levels (Ref. 17-18). Recent studies conclude that AI-assisted mammography reduces the rate of callbacks. As a result, fewer women go through the anxiety of returning to the imaging center for additional views (Ref. 19-21).
AI can also reduce the time it takes to complete the imaging process (Ref. 22-23). With AI-assisted screening mammograms, a report is issued immediately after the completion of the mammogram. If the AI mammogram determines that additional views are needed, they can be done on the same visit without waiting for the radiologist to issue a report. In most cases, the additional views will be negative, and the patient will not need to wait for the radiologist to issue a final report.
One example of how AI can reduce the cost of breast care is to reduce the rates of false positive biopsies. It is estimated that the cost of covering false positive biopsies in the USA is in the range of $2.2 billion (Ref. 24). A recent study concluded that there is an overall 69% reduction in the rate of false positive biopsies with AI-assisted mammograms (Ref. 25).
The future looks bright for AI-assisted breast imaging. AI has the potential to markedly improve rates of early detection of potentially curable breast cancers. Over time, machine learning will lead to improved rates of early detection and will continue to reduce the rates of unnecessary breast biopsies.
We believe that there is an urgent need to make AI-assisted mammographic screening available to all women regardless of their ability to pay (Underserved Population and Cost/Benefit). This would be a costly investment in the short term, but in the long run, it would lead to a reduction in the cost of breast cancer care and a reduction in breast cancer mortality rate.
The rationale for expanding breast screening to all women is based on the issue of the lifetime cost of breast care. When breast cancers are detected on mammographic screening, the lifetime cost of care is in the range of $50-75 thousand and the survival approach approaches 100% (Cost/Benefit). The cost of care for women who are diagnosed with more advanced breast cancers can exceed $1 million and chances for a cure are low (see internal links: Cost/benefit and Our Screening Guidelines). We are convinced that aggressive screening using AI technology has the potential to be a game-changer in improving breast cancer survival rates and lowering the cost of breast care.
Questions? Contact us
- Mammograms (20% miss rate)
- Mammographically occult breast cancers detected with AI-based diagnosis supporting software: clinical and histopathologic characteristics
- Retrospective Review of Missed Cancer Detection and Its Mammography Findings with
Artificial-Intelligence-Based, Computer-Aided Diagnosis.
- AI detecting breast cancer 4 years before it developed.
- The relationship of mammographic density and age: implications for breast cancer screening.
- Dense Breasts’ Eclipse All Other Known Breast Cancer Risk Factors
- Women with Dense Breasts Have Higher Risk of More Aggressive Cancer
- Artificial Intelligence in Breast Cancer Screening and Diagnosis
- The Promise of Using A.I. Technology to Detect Breast Cancer: Excellent
- Difference between Artificial intelligence and Machine learning
- Limitations of Mammogram
- Half of all women experience false-positive mammograms after 10 years of annual screening.
- Reduction of False-Positive Markings on Mammograms: A Retrospective Comparison Study Using an Artificial Intelligence-Based CAD
- Can an Artificial Intelligence Decision Aid Decrease False Positive Breast Biopsies?
- Artificial intelligence system reduces false-positive findings in the interpretation of breast ultrasound exams.
- Artificial Intelligence Applied to Breast MRI for Improved Diagnosis
- Psychosocial consequences of false-positive results in screening mammography
- False-Positive Mammograms Can Trigger Long-Term Distress
- Getting Called Back After a Mammogram
- What I Wish I’d Known About Mammogram Callbacks
- The Costs of Callbacks: What it Means for Your Radiology Practice and Your Patients
- Artificial Intelligence for Reducing Workload in Breast Cancer Screening with Digital Breast Tomosynthesis
- AI Tool Improves Breast Cancer Detection on Mammography
- Annual Cost of False-Positive Breast Biopsies Exceeds $2 Billion According to Recent Study
- Reduction of False-Positive Markings on Mammograms: a Retrospective Comparison Study Using an Artificial Intelligence-Based CAD
Your donation supports our mission to provide breast cancer education to EVERY woman.