Fully Automated Postlumpectomy Breast Margin Assessment Utilizing Convolutional Neural Network Based Optical Coherence Tomography Image Classification Method

ConclusionIt is feasible to use CNN based algorithm to accurately distinguish cancerous regions in OCT images. This fully automated method can overcome limitations of manual interpretation including interobserver variability and speed of interpretation and may enable real-time intraoperative margin assessment.
Source: Academic Radiology - Category: Radiology Source Type: research