international group of researchers UT Southwestern Medical Center have developed a model that combines traditional MRI with artificial intelligence (AI) to more accurately detect breast cancer metastasis. The scientists also suggest that the model could eliminate the need for needle or surgical biopsies.. This study was recently published in the journal Radiology: Cancer Imaging.
The non-invasive approach uses traditional magnetic resonance imaging (MRI) in conjunction with machine learning AI to detect axillary metastases (the presence of cancer cells in the lymph nodes under the arm).
Most deaths from breast cancer are due to metastatic disease, and the initial site is usually the axillary lymph nodes. Determining lymph node status is important in determining treatment, but conventional imaging techniques alone are not sensitive enough to exclude axillary metastases. Therefore, patients often have to undergo an invasive procedure that involves injection of radioactive isotopes and dyes, followed by surgery to remove the axillary lymph nodes and test them for cancer cells..
Basak Dogan MD, Professor and Research Leader, Department of Radiology, University of Texas Southwestern Medical Center
Dorgan is also director of breast imaging research and a member of the Harold C. Simmons Comprehensive Cancer Center at UT Southwestern.
This study demonstrated that the AI model was significantly more accurate in identifying individuals with axillary metastases compared to MRI and ultrasound. In clinical practice, the AI model successfully identified 95% of patients with axillary metastases and avoided 51% of benign (non-cancerous) or unnecessary surgical sentinel lymph node biopsies.
This is an important advance because surgical biopsies have side effects and risks. Despite the low probability of a positive result confirming the presence of cancer cells, this model can be used to improve the ability to rule out axillary metastases during routine MRI, thereby improving clinical outcomes. You can reduce your risk..
Basak Dogan MD, Professor and Research Leader, Department of Radiology, University of Texas Southwestern Medical Center
Approximately 350 patients with newly diagnosed breast cancer from the Moody Breast Health Center at the University of Texas Southwestern and Parkland Health’s main campus in Dallas used dynamic contrast-enhanced breast MRI in a retrospective analysis. , everyone knew about nodules. situation.
Using machine learning techniques, the AI model was trained to recognize axillary metastases using photographs and various clinical indicators.
For many patients, this model can also be used in conjunction with regular imaging scans to reduce the anxiety and financial burden of unnecessary testing.
Patients with benign findings on conventional MRI or needle biopsy often undergo sentinel lymph node biopsy. This is because these tests may miss a large proportion of metastases. Our study demonstrates that it is possible to identify non-metastatic patients with high accuracy, which not only benefits patients but also allows physicians to customize treatment. Masu..
Basak Dogan MD, Professor and Research Leader, Department of Radiology, University of Texas Southwestern Medical Center
This research expands on UT Southwestern’s previous work on breast cancer imaging and creating metastasis prediction tools.
Dr. Dougan said:Our research is a testament to UT Southwestern’s commitment to impactful research that addresses real-world medical challenges, and the development and validation of AI models for medical imaging will help improve breast cancer and other cancers. We have great hopes in helping us in the fight against.And this new tool is a big step forward”
The researchers aim to incorporate a wider range of data and refine image analysis procedures to support their conclusions.
Co-authors of the study are lead author Dogan Porat, M.D., a second-year radiology resident; Albert Montillo, Ph.D., assistant professor in the Lyda Hill Department of Bioinformatics and Biomedical Engineering; Keith Halsey, Ph.D., lecturer in radiology; and Liqiang Wang, Ph.D., faculty assistant, and Son Nguyen, Ph.D., postdoctoral research associate, all of the Lyda Hill Department of Bioinformatics.
Dr. Dorgan is also the Eugene P. Frenkel, MD, PhD in Clinical Medicine.
This research was funded by Simmons Cancer Center, the National Institutes of Health (NIH), the National Institute of General Medical Sciences, the NIH National Institute on Aging, the NIH National Cancer Institute, the King Foundation, and the Lyda Hill Foundation.
Journal reference:
Porat, D.S.; other. (2024) Machine learning prediction of lymph node metastasis in breast cancer: Multi-center MRI-based performance of his 4D convolutional neural network. Radiology. cancer image processing. doi.org/10.1148/rycan.230107.