Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

THIS WEBPAGE IS DEPRECATED - please visit our new website

J. Weiss, N. Rieke, M. A. Nasseri, M. Maier, C. P. Lohmann, N. Navab, A. Eslami
Injection Assistance via Surgical Needle Guidance using Microscope-Integrated OCT (MI-OCT)
Proc. of International Society of Imaging in the Eye Conference, Association for Research in Vision and Ophthalmology (ARVO), Honolulu, USA, April 2018 (bib)

Injection in ophthalmic interventions requires precise targeting of anatomic layers both in anterior and posterior segment surgery. In the current workflow surgeons mainly rely on the binocular en-face view of the microscope which provides only a limited perception of depth and distance to target. Our work aims at assisting the surgeon by providing the projected intersection point of instrument and anatomy and thereby enabling intuitive and more precise targeting. Methods: The proposed method uses cross-sectional information provided by MI-OCT. The surgical needle and the position of the anatomical target is localized in five continuously acquired parallel MI-OCT B-Scans. To that end, the maximum intensity along each A-Scan is extracted for each OCT image. Given this point set, the target layer and the instrument cross-section is found by a geometry fitting algorithm. In the second step, we find the 3D needle pose based on these segmented cross-sections by using the specific geometry of the surgical needle and a temporal filter. Finally, the estimated injection position is determined by the geometric intersection of the approximated target layer with the needle. By projecting the computed intersection point to the en-face view, we can overlay the estimated position in real-time via a heads-up display on the microscopic view. As a result, the surgeons can infer the distance to the target layer by the distance between needle tip and projected injection point in an intuitive way. Results: We evaluated our method on both anterior and posterior phantoms. In both scenarios, we acquired three sequences of free needle movement. For acquisition, we used a Zeiss Lumera 700 with Resight 700 in 5-line HD OCT mode and the needle is constantly touching the target surface to have ground truth. In the en-face camera view, we manually annotate the ground truth touching/intersection point in each image as the needle tip in 811 images. We report a median error of 0.230mm (mean: 0.299+-0.062mm) for anterior and median error of 0.268mm (mean: 0.358+- 0.090mm) for posterior guidance. Conclusions: We present an approach to provide continuous injection guidance based on conventional MI-OCT BScans. The system provides accurate prediction and visualization of the injection point, thus supporting the surgeon in difficult injection tasks.
This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each authors copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

Edit | Attach | Refresh | Diffs | More | Revision r1.13 - 30 Jan 2019 - 15:16 - LeslieCasas

Lehrstuhl für Computer Aided Medical Procedures & Augmented Reality    rss.gif