Massachusetts General Hospital Boston, Massachusetts
Professor of Radiology, Harvard Medical School Chief of Breast Imaging and Co-Director of AVON Breast Center
Using advanced tools of artificial intelligence to develop methods to improve breast cancer screening and identify women most at risk.
The first step in breast cancer prevention is an accurate assessment of individual risk so that personalized prevention strategies can be implemented. Breast mammography has served as the gold standard for breast cancer screening. However, mammography and other imaging technologies used in health care require highly specialized training and are also subject to human interpretation. Furthermore, they each have limitations that can either result in false positives (reporting a cancer that is not there), or false negatives (reporting no cancer when, in fact, cancer is present). With the rapid advances in artificial intelligence (AI), there is a great opportunity to improve upon current age-based risk models, which are limited in scope for both personalized risk assessment and predicting risk of specific breast cancer subtypes. Dr. Lehman and her multidisciplinary team of scientists use AI to extract complex information contained within a breast image—such as breast density—to develop risk models for breast cancer. The shift from age-based to risk-based screening will benefit significantly from the predictive power of AI with the potential application to different cancer types and patient populations.
Dr. Lehman and her team created an advanced breast cancer risk prediction model combining imaging, biologic, and behavioral data to personalize screening regimens and improve early detection while lowering the overall cost of screening. They confirmed that advanced AI methods applied to mammogram images support more accurate breast cancer risk assessment across diverse races and ethnicities as compared to traditional risk models. Dr. Lehman and her team have expanded their infrastructure and AI-assessed mammogram database by increasing the number of sequential mammograms to assess AI-based risk scores over time and increasing the number of exams from women who identify as Asian, Hispanic, and Black. These advances will allow for better assessment of the performance of the AI risk model over time and across diverse populations of patients. The team’s next phase of this study will inform best practices for screening recommendations across diverse races/ethnicities.
Dr. Lehman’s team is now poised to proceed into a previously unexplored domain—the potential of the AI risk model to measure changing risk levels in individual women over time. Currently, there are no effective methods to assess if a woman’s changes in exposures (diet, weight, alcohol intake, smoking, hormone use, chemoprevention) are having an impact and to what degree. Dr. Lehman and her team will measure changes in risk scores over time utilizing 486,584 mammograms from 118,252 women obtained between 2009 and 2022. The team will then compare cancer occurrence in women with stable, increasing, and decreasing AI risk scores. This will demonstrate an AI risk model that changes over time and provides a more accurate method of predicting future breast cancer events.
Constance “Connie” Lehman, MD, PhD is a Professor of Radiology at Harvard Medical School, Chief of Breast Imaging, and co-Director of the Avon Comprehensive Breast Evaluation Center at the Massachusetts General Hospital. She is a physician scientist who received her MD and PhD from Yale University after graduating Phi Beta Kappa from Duke University. She is a change agent, innovator and pioneer in the domain of Artificial Intelligence implementation in clinical medical practice. Her current work applies artificial intelligence and advanced methods of deep learning to improve breast cancer detection, diagnosis and treatment. Her current efforts focus on applying AI to deliver higher-quality health outcomes at lower costs for the full diversity of patients at risk for and with cancer. This is a continuation of her career-long commitment to expand access for all patients globally to highest quality, patient-centered, affordable care.
With over 250 peer-reviewed scientific publications, she has led careful studies of advanced imaging tools to identify breast cancer at its earliest stages—when it can be cured. Collectively, her philosophy embodies the notion that we improve the health of our community by delivery the highest quality patient-centered care in a setting of active innovation and education.
In her prior leadership roles at the University of Washington and Seattle Cancer Care Alliance, she developed an internationally-recognized program of excellence in patient care, education and research. She collaborated with partners in industry, development, and multiple clinical and basic science disciplines to foster new opportunities for faculty to pursue their goals. As division chief of breast imaging and co-Director of the AVON Comprehensive Breast Center at MGH, she rebuilt the service care model as well as the education and clinical research programs to stimulate new growth and productivity in times of scarce resources.
In her career, Dr. Lehman has developed novel patient care pathways, which substantially improve efficiency and increase access for patients and increase satisfaction of patients, faculty and staff. Her methods are used widely by her former fellows and faculty, many who lead their own breast imaging programs.
Her research in clinical applications of breast MRI shaped the American Cancer Society and NCCN recommendations for screening MRI in high-risk patients. Dr. Lehman’s research on breast ultrasound updated the ACR appropriateness guidelines for women under 40, and this work continues to support best practices worldwide for women with palpable breast lumps.
Dr. Lehman has inspired many others to pursue careers in academic medicine and has served on a number of key national committees, including the National Cancer Institute’s Breast Cancer Steering Committee and the American College of Radiology’s (ACR) Committee on Breast Imaging for Appropriateness Criteria and Guidelines. She has given over 400 national and over 60 international invited talks spanning diverse topics in advanced imaging for more effective diagnosis and treatment of breast cancer.
2019
The Lampert Foundation Award
Please remember BCRF in your will planning. Learn More
