In my last article, I discussed how artificial intelligence (AI) has been integrated into the overall healthcare system. That piece included a recent finding from the American Medical Association that AI was already effectively “enhancing diagnostic ability (72%), workflow efficiency (69%) and clinical outcomes (61%).”
This finding is now in focus for breast cancer researchers, as senior editor Jamie DePolo highlights in “AI-Supported Mammogram Reading Detects 20% More Cancers.”
DePolo focuses on a Swedish-based study that observed mammograms using AI to read found 20% more cancers than the routine double-read by two radiologists without increasing the rate of false positive results. This incredible insight into the use of AI for breast cancer diagnoses is further supplemented by the major projects currently underway at the Breast Cancer Research Foundation (BCRF).
One influential example is from Dr. Constance Lehman and Dr. Regina Brazilay, who created a mammography-based deep learning model named MIRAI. The pair tested the model and showed that it “could yield individualized, equitable, and cost-effective improvements in breast cancer risk prediction compared to traditional risk models” while providing “consistent results across mammography sites and machines.” Both doctors are now using MIRAI to find other potential applications.
Dr. Lehman’s latest work aims to shift the norm of breast cancer screenings from age-based to risk-based. Her team has successfully created a prediction model with the software based on patient data to develop personalized risk-based screening plans, leading to earlier detection and decreased costs for patients.
Concurrently, Dr. Brazilay has begun co-authoring a research study to test MIRAI’s capabilities to help decipher and predict which patients are at high risk of developing breast cancer.
These two projects could eventually converge with the widespread adoption of emerging tech tools to completely shift how healthcare providers proactively detect and prevent disease. These are just a few examples of the invaluable work and progress being done in the breast cancer diagnostic field with the adoption of greater AI integration and use.
The employment of AI can improve rates of accuracy, speed, early detection and so much more than we can imagine now. These improved outcomes include faster detection and higher success rates for removing whole tumors, as doctors using the technology are able to identify more than the human eye alone.
These overall benefits could also significantly improve healthcare for underserved populations. As AI is used in detection practices, its greater speed helps healthcare providers reach a wider number of individuals in a shorter time at a lower cost, which could significantly increase the number of services accessible to lower-income individuals and other vulnerable populations.
AI integration into healthcare services would lift patient care across America in many ways, including increasing the quality of care for all while lowering costs and burdens on workers and patients alike. In such a win-win scenario, policymakers should be open to braving the hurdles associated with the early adoption of new technologies, aiding the necessary associated and extensive research, and ensuring that any regulatory actions do not impede such stark improvements to American healthcare, which are already happening right now.