Automated Detection and Tracking of a Twinkling Tissue Marker From Color Doppler Video Stream
Wednesday, April 10, 2024
8:18am – 8:25am
Location: 412
Authors: Luiz Vasconcelos, Mayo Clinic Christine Lee, Mayo Clinic Matthew Urban, Mayo Clinic
Breast cancer is the most common cancer in women with 30-60% incidence of metastasis to axillary lymph nodes. The management of positive nodes is facilitated with the placement of markers, followed by neoadjuvant therapy and surgery. It is known that twinkling artifacts are present when imaging some biopsy markers and, thus, have the potential for inexpensive and non-ionizing lymph node tracking. In this study, we propose a vendor-agnostic twinkling marker detection and tracking method using the Doppler video stream from a clinical ultrasound scanner. A clinical scanner and 9L linear array transducer were used to acquire Doppler images from a polyvinyl alcohol cryogel phantom inside a water tank. The B-mode with color Doppler feed was captured and processed using a conventional laptop. The method was able to identify and localize a polymethyl methacrylate marker in real-time above 30 frames per second with an accuracy of 0.145 mm mean absolute error. In the future, the acquired marker position will be translated to an audio cue so radiologists or surgeons can track the marker position without visually relying on the ultrasound device monitor.