The study will consist of measuring gases emitted during surgery. Prior to this, the device will be trained in the detection and recognition of the most prevalent gases in cancerous tissue. This will enable the AI to detect and classify cancerous breast tissue from healthy tissue. The study will include patients with different types and subtypes of breast cancer, such as ductal carcinoma in situ invasive ductal carcinoma, invasive lobular carcinoma, triple-negative breast cancer, human epidermal growth factor receptor 2 positive breast cancer, and hormone receptor-positive breast cancer. These types and subtypes reflect the diversity and complexity of breast cancer and may influence the performance and accuracy of the BCGC device.
During the clinical testing phase, the laboratory-acquired tissue detection capability will be evaluated in a real situation. For this, the device will be connected to a sterile PVC hose directly connected to a smoke extractor associated with the electrosurgical unit that will cauterize the tissues. To evaluate the sensitivity, specificity, and accuracy of the device, the AI's detection will be compared with the quick biopsy and the deferred biopsy of the surgical piece. For this, access will be obtained to the histological reports and the surgery will be recorded to identify margins.
The device will be calibrated once a month with gases of known nature, a maximum variability of 5% will be tolerated. If a greater variability is detected at the time of calibration, it will be changed to every 2 weeks. The information will be anonymized and stored on an SSD unit and will be deleted 5 years after the completion of the study.
