Three decades after the first set of work on Medical Augmented Reality (MAR) was presented to the international community, and ten years after the deployment of the first MAR solutions into operating rooms, its exact definition, basic components, systematic design, and validation still lack a detailed discussion. This paper defines the basic components of any Augmented Reality (AR) solution and extends them to exemplary Medical Augmented Reality Systems (MARS). We use some of the original MARS applications developed at the Chair for Computer Aided Medical Procedures and deployed into medical schools for teaching anatomy and into operating rooms for telemedicine and surgical guidance throughout the last decades to identify the corresponding basic components. In this regard, the paper is not discussing all past or existing solutions but only aims at defining the principle components and discussing the particular domain modeling for MAR and its design-development-validation process, and providing exemplary cases through the past in-house developments of such solutions.

The emergence of robotics could enable ophthalmic microsurgical procedures that were previously not feasible due to the precision limits of manual delivery, for example, targeted subretinal injection. Determining the distance between the needle tip, the internal limiting membrane (ILM), and the retinal pigment epithelium (RPE) both precisely and reproducibly is required for safe and successful robotic retinal interventions. Recent advances in intraoperative optical coherence tomography (iOCT) have opened the path for 4D image-guided surgery by providing near video-rate imaging with micron-level resolution to visualize retinal structures, surgical instruments, and tool-tissue interactions. In this work, we present a novel pipeline to precisely estimate the distance between the injection needle and the surface boundaries of two retinal layers, the ILM and the RPE, from iOCT volumes. To achieve high computational efficiency, we reduce the analysis to the relevant area around the needle tip. We employ a convolutional neural network (CNN) to segment the tool surface, as well as the retinal layer boundaries from selected iOCT B-scans within this tip area. This results in the generation and processing of 3D surface point clouds for the tool, ILM and RPE from the B-scan segmentation maps, which in turn allows the estimation of the minimum distance between the resulting tool and layer point clouds. The proposed method is evaluated on iOCT volumes from ex-vivo porcine eyes and achieves an average error of 9.24 µm and 8.61 µm measuring the distance from the needle tip to the ILM and the RPE, respectively. The results demonstrate that this approach is robust to the high levels of noise present in iOCT B-scans and is suitable for the interventional use case by providing distance feedback at an average update rate of 15.66 Hz.

Optical coherence tomography (OCT) is a medical imaging modality that is commonly used to diagnose retinal diseases. In recent years, linear and radial scanning patterns have been proposed to acquire three-dimensional OCT data. These patterns show differences in A-scan acquisition density across the generated volumes, and thus differ in their suitability for the diagnosis of retinal diseases. While radial OCT volumes exhibit a higher A-scan sampling rate around the scan center, linear scans contain more information in the peripheral scan areas. In this paper, we propose a method to combine a linearly and radially acquired OCT volume to generate a single compound volume, which merges the advantages of both scanning patterns to increase the information that can be gained from the three-dimensional OCT data. We initially generate 3D point clouds of the linearly and radially acquired OCT volumes and use an Iterative Closest Point (ICP) variant to register both volumes. After registration, the compound volume is created by selectively exploiting linear and radial scanning data, depending on the A-scan density of the individual scans. Fusing regions from both volumes with respect to their local A-scan sampling density, we achieve improved overall anatomical OCT information in a high-resolution compound volume. We demonstrate our method on linear and radial OCT volumes for the visualization and analysis of macular holes and the surrounding anatomical structures.

No abstract