AI-assisted mammograms show promise in advancing healthcare, according to a study.

AI-assisted mammograms show promise in advancing healthcare, according to a study.

AI-assisted Mammography Shows Promise in Detecting Breast Cancers

Breast Cancer

Artificial intelligence (AI) is revolutionizing the field of medicine, and its potential in breast cancer detection is gaining attention. A recent clinical trial conducted in Sweden has shown that AI programs can effectively assist radiologists in reviewing mammogram images and detecting breast cancers. The trial, which is ongoing, has provided encouraging results, indicating the potential of AI in improving breast cancer detection rates and reducing radiologists’ workload.

Traditionally, mammograms are reviewed by radiologists to detect any abnormalities that could indicate the presence of breast cancer. However, the shortage of breast radiologists in many countries has highlighted the need for accurate AI programs that can assist in the detection process. This clinical trial aimed to directly compare AI-assisted mammography with standard mammography, providing crucial insights into the safety and efficacy of AI-supported screening.

The results from the trial have been promising. A single radiologist, aided by AI, detected approximately 20% more breast cancers from mammogram images than two radiologists working together. This significant increase in detection did not lead to a higher rate of false positives, where a mammogram is incorrectly diagnosed as abnormal and in need of follow-up. Additionally, the AI assistance reduced radiologists’ workload by an impressive 44%.

“Our study showed that mammography screening with AI is safe, since the cancer detection did not decline despite the reduced screen-reading volume for the radiologist,” said Dr. Kristina Lång, the lead researcher of the study. “Women should, therefore, feel safe undergoing AI-supported screening.”

The use of AI-aided mammography is already available in the United States, with more than 20 AI systems approved by the U.S. Food and Drug Administration. However, there are still questions about how to effectively implement and utilize AI programs in the clinical setting. While most studies have been retrospective, evaluating the accuracy of AI programs using past cases, this clinical trial in Sweden is the first to directly compare AI-assisted mammography against standard mammography.

More than 80,000 participants have already been randomly assigned to undergo either standard or AI-assisted mammography at four testing sites in Sweden. The trial aims to enroll and track 100,000 participants for two years of follow-up. The interim safety report, which assesses whether AI-assisted mammography poses any undue risk to the participants, has indicated positive outcomes.

The detection rate of breast cancers for AI-supported screening was 6 per 1,000 screenings, compared to 5 per 1,000 for standard screening. This means that AI-assisted screening detected one additional cancer for every 1,000 women screened. The false-positive rate, where mammograms are incorrectly diagnosed as abnormal, was 1.5% in both groups.

In addition to the improved detection rate, AI-aided screening also resulted in fewer cases requiring further testing. Women undergoing AI-assisted screening were recalled for further testing in about 2.2% of cases, compared to 2% for standard screening. This led to the detection of an additional 41 cancers, 19 of which were invasive.

The reduction in the workload of radiologists is another significant advantage of AI-assisted mammography. Nearly 37,000 fewer screen readings were required by radiologists in the AI-supported group. The estimated time saved amounts to four to six months of workload for the radiologists, assuming an average reading rate of 50 mammograms per hour.

However, it is important to emphasize the complementary role of AI and human intelligence in the screening process. Dr. Lång highlighted that AI should not operate alone and that radiologists still play a central role in reviewing mammograms, even with AI assistance. She also raised concerns about potential downsides, such as over-reliance on AI resulting in increased false positives or overworked radiologists if AI handles simpler cases, leaving more complex cases to be addressed by humans.

The final results from the clinical trial will provide further insights into the effectiveness of AI in early breast cancer detection. The aim is to catch cancers at a stage when they are more easily treatable. Currently, radiologists overlook suspicious early signs of malignancy in about 20% to 30% of breast cancers that are eventually detected. The promising results from this trial suggest that AI can help bridge this gap.

Dr. Lång hopes that further trials will be conducted to validate and expand upon these findings. The consequences and potential challenges associated with widespread implementation of AI need to be thoroughly understood before it can be widely adopted.

Overall, the use of AI in mammography holds great potential for improving breast cancer detection rates while reducing the workload of radiologists. As technology continues to advance, AI-assisted mammography may become a powerful tool in the fight against breast cancer, ensuring earlier detection and more successful treatment outcomes.

More information

The U.S. National Cancer Institute provides additional information about mammograms.

Sources: – Kristina Lång, MD, PhD, Associate Professor, Diagnostic Radiology, Lund University, Lund, Sweden – Robert Smith, PhD, Senior Vice President, Early Cancer Detection Science, American Cancer Society, Atlanta – The Lancet Oncology, August 2023