A new AI model could enhance the precision of breast cancer tumour removal during surgery.
Microscopic cancer cells, which cannot be seen with the naked eye, often pose a challenge for surgeons.
Now, a team of medical and engineering experts have developed an AI tool that predicts in real-time if cancerous tissue has been completely removed.
The innovation results from a collaboration between the University of North Carolina School of Medicine, the Joint UNC-NCSU Department of Biomedical Engineering and the UNC Lineberger Comprehensive Cancer Centre.
Kristalyn Gallagher, DO, section chief of breast surgery in the Division of Surgical Oncology and UNC Lineberger member, explained: ‘Some cancers you can feel and see, but we can’t see microscopic cancer cells that may be present at the edge of the tissue removed. This AI tool would allow us to more accurately analyse tumours removed surgically in real-time and increase the chance that all of the cancer cells are removed during the surgery. This would prevent bringing patients back for a second or third surgery.’
To train their AI model, the researchers used hundreds of specimen mammogram images matched with final specimen reports from pathologists. Demographic data from patients, such as age, race, tumour type and tumour size, were also incorporated to refine the model's accuracy.
The study revealed that the AI model matched – or even slightly outperformed – human interpretation in identifying positive margins, showing promise for use in breast cancer surgery.
The AI model’s potential is particularly significant for patients with higher breast density, where distinguishing between cancerous and healthy tissue can prove challenging because of their similarities in appearance when seen on mammograms.
Shawn Gomez, EngScD, a professor of biomedical engineering and pharmacology and co-senior author of the paper, noted: ‘It is like putting an extra layer of support in hospitals that maybe wouldn’t have that expertise readily available. Instead of having to make a ‘best guess’, surgeons could have the support of a model trained on hundreds or thousands of images and get immediate feedback on their surgery to make a more informed decision.’
Although the AI model is in its early stages, the researchers plan to expand its capabilities by incorporating more images from different patients and surgeons.
Further studies will be conducted to validate the model before it can be utilised in clinical settings.
The team is optimistic that continuing to refine their model will play a pivotal role in enhancing the accuracy of breast cancer tumour removal.
This potentially reduces the need for additional surgeries and improves patient outcomes.
The research findings are published in the Annals of Surgical Oncology.
