How AI is revolutionizing breast cancer detection

In a recent interview with Contemporary OB/GYN, John Simon, MD, a radiologist with extensive experience in women's imaging and interventional radiology, discussed how artificial intelligence (AI) models are transforming breast cancer detection

Simon began by highlighting the critical importance of early detection in breast cancer, which boasts a cure rate exceeding 99% when identified early. Early detection faces challenges, notably the difficulty in identifying cancer on mammograms, especially in women with dense breast tissue. Dense breasts, which present in approximately 40% of women in the United States, can obscure cancers on mammograms, likened by Simon to finding an object in a cloudy sky.

Mammography remains the first line of imaging, with 3D mammography providing enhanced detection capabilities. For women with dense breasts, additional imaging modalities such as ultrasound and magnetic resonance imaging (MRI) are recommended. Ultrasound acts as a radar, penetrating the "clouds" to detect abnormalities, while MRI, though more costly and minimally invasive because of the use of contrast, is reserved for high-risk patients.

Simon also stressed the importance of personalized breast cancer detection strategies. This personalization involves assessing individual risk factors, such as breast density and family history, using tools such as the Tyrer-Cuzick score. Identifying high-risk individuals allows for tailored surveillance strategies, enhancing early detection efforts.

Another significant issue is patient compliance with annual screening. Despite the life-saving potential of mammography and ultrasound, many patients avoid screenings because of discomfort or fear, with nearly 40% of eligible patients skipping annual exams. Simon emphasized the role of education in overcoming this hurdle.

AI contributes to breast cancer detection by augmenting the radiologist's review of mammograms. AI enhances sensitivity, allowing for the detection of more cancers. False positives, which are defined as instances where cancer is suspected but not present, are also reduced when using AI. Additionally, AI provides consistent and reproducible breast density assessments, minimizing variability in interpretations between different radiologists. This consistency is crucial for informing patients about their breast density, which impacts the success rate of early detection.

Patient history documentation and data organization may also be improved by AI, aiming for a more personalized cancer detection approach. These advances in AI have accelerated in recent years, improving the accuracy and efficiency of breast cancer detection while encouraging patients to stay proactive about their screenings. Simon concluded by urging patients to undergo annual mammograms and, for those with dense breasts, annual ultrasounds, while understanding their personalized risk for enhanced care.