A BC Cancer research team received provincial funding in September 2024 through the Ministry of Health's Innovation Pathway Program to explore a new expedited breast cancer triage system for faster initiation of appropriate treatment, chemotherapy or surgery.
Called CAN-TRI-NLP, or a “CANcer TRIage system to expedite care with Natural Language Processing” the project team included researchers and clinicians from BC Cancer, surgical programs, the Data Science Institute and Data Analytics Reporting and Evaluation (DARE).
Project funding has now concluded. The scope included the development, application and evaluation of Natural Language Processing (NLP) models in a triage system that automated extraction and collection of cancer data.
"Our team developed an advanced pipeline to detect breast cancers from an electronic feed of pathology reports,” says Dr. Alan Nichol, Radiation Oncologist Co-Chair of the Breast Cancer Outcomes Unit at BC Cancer.
The team found that this pipeline reduced workloads and improved the speed of filtering reportable pathology reports in comparison with manual review of each report for patients in BC.
“We have developed NLP models to identify the high-risk breast cancer subtypes that have good responses to chemotherapy,” notes Dr. Nichol. Using NLP, these high-risk breast cancer subtypes can be identified with 98-100% accuracy.
Currently, about a third of patients with high-risk stage 2 and 3 breast cancers who can benefit from chemotherapy before surgery (called neoadjuvant chemotherapy) receive surgery before chemotherapy.
Using this new triage system, patients will meet medical oncologists promptly after diagnosis and will be more likely to benefit from the advantages of neoadjuvant chemotherapy, such as reduced use of mastectomies, breast reconstructions and axillary lymph node dissections.
“With our new ability to automatically identify high-risk breast cancer subtypes from breast biopsy reports, the next stage of the project is to communicate with primary care providers and surgeons about patients with stage 2 and 3 breast cancers to arrange early medical oncology referrals,” Dr. Nichol says.
The project team is also exploring ways to expand this novel method of cancer triage to more types of breast cancers, and eventually, all cancers.
