Computer Vision, AI & Robotics on Medical Applications – Invited Special Session

B4L-D: Computer Vision, AI & Robotics on Medical Applications - Invited Special Session

Session Type: Lecture
Session Code: B4L-D
Location: Room 4
Date & Time: Thursday March 23, 2023 (14:00-15:00)
Chair: Richard Jaepyeong Cha
Track: 12

Paper No.	Paper Name	Authors	Abstract
3142	Artificial Intelligence-Assisted Laparoscopic Cholecystectomy in a Preclinical Swine Model	Khalid Mohamed Ali{1}, Michele Saruwatari{1}, Kochai Jawed{1}, Yoseph Kim{3}, Seihoon Park{2}, Seongjun Cha{2}, Bo Ning{1}, Richard Jaepyeong Cha{1}	Iatrogenic bile duct injury during laparoscopic cholecystectomy has an incidence rate of 1-2%. Their small size, multiple anatomical variations and inflammatory conditions make it challenging for the surgeon to identify or protect during dissection. Despite the difficulties and rising concerns of identifying bile duct during surgeries, surgeons do not have a specific modality to identify bile duct except intraoperative cholangiography, that disrupts workflows. Here, we present our new study of laparoscopic cholecystectomy using a quasi-real-time artificial intelligence assisted biliary specific fluorescence imaging in a swine model in vivo.
3193	Body-Mounted MR-HIFU Robotic System: Mechanical Design and Accuracy Evaluation	Ayush Nankani, Pavel Yarmolenko, Karun Sharma, Kevin Cleary, Reza Monfaredi	Background: Magnetic Resonance Image-Guided High-Intensity Focused Ultrasound (MR-HIFU) is a non-invasive procedure for the thermal ablation of benign and malignant tissues and pain-causing nerves using high-intensity ultrasound waves under MRI guidance. MRI is preferred due to its superior soft tissue contrast to other imaging modalities. Over the last decade, MR-HIFU has been used for various applications, including uterine fibroids, bone metastases, and prostate diseases. However, current MR-HIFU systems have some limitations; for example, systems with table-imbedded transducers severely limit patient positioning. Also, systems that offer a more flexible approach by integrating the HIFU transducer with a manual positioning arm suffer from a lack of accuracy and can be cumbersome. In addition, current clinical systems are not compatible across all MRI scanner manufacturers. In this study, we propose to address these limitations by developing a body-mounted MRI-compatible robotic positioning system. Methodology: We propose a body-mounted MRI-compatible robotic system with 5 DOF to enable in-bore procedures specifically for treating musculoskeletal lesions in the extremities. The robotic system was designed using SolidWorks keeping all the design requirements/constraints in mind. The initial mechanical prototype was manufactured using 3D printers and other manufacturing tools. Workspace and kinematics analysis of the robot was conducted using MATLAB & Simulink. The accuracy of the mechanical system was measured using an optical tracker through a benchtop study. Results: The complete mechanical robotic system with 5 degrees of freedom (DOF) was designed, prototyped, and tested in a benchtop study. The system\'s average position and angular error, determined using the accuracy study, were ±0.2 mm and ±1.68 degrees, respectively. Conclusion: The accuracy is good enough for surgical procedures treating musculoskeletal lesions. Due to its high modularity, the proposed system can also be used for other probe/needle-guided, MRI-based medical interventions. We will continue to develop and automate the proposed body-mounted MRI-compatible robotic positioning system to address the limitations of current clinical MR-HIFU systems and use different MRI scanners from different manufacturers.
3149	Development of Multifunctional Hyperspectral/Near-Infrared Imaging Camera System for Intraoperative Tissue Characterization and Assessment in Vivo	Christy Lee, Aneesh Naik, Sailee Naik, Michele Saruwatari, Kochai Jawed, Khalid Mohamed Ali, Bo Ning, Richard Jaepyeong Cha	Intraoperative and precise tissue characterization such as oxygenation and perfusion levels during general and abdominal surgery is challenging. Using a newly developed multifunctional camera system that combines real-time snapshot hyperspectral imaging with laser speckle contrast imaging, this study presents a real-time visualization of operative field with precise tissue characteristics to assist with surgical decision makings. The imager’s performance was evaluated to show versatile imaging modes through in vivo animal studies involving rodent and swine models.