Usability of cooperative surgical telemanipulation for bone milling tasks.

PURPOSE: Cooperative surgical systems enable humans and machines to combine their individual strengths and collaborate to improve the surgical outcome. Cooperative telemanipulated systems offer the widest spectrum of cooperative functionalities, because motion scaling is possible. Haptic guidance can be used to assist surgeons and haptic feedback makes acting forces at the slave side transparent to the operator, however, overlapping and masking of forces needs to be avoided. This study evaluates the usability of a cooperative surgical telemanipulator in a laboratory setting. METHODS: Three experiments were designed and conducted for characteristic surgical task scenarios derived from field studies in orthopedics and neurosurgery to address bone tissue differentiation, guided milling and depth sensitive milling. Interaction modes were designed to ensure that no overlapping or masking of haptic guidance and haptic feedback occurs when allocating information to the haptic channel. Twenty participants were recruited to compare teleoperated modes, direct manual execution and an exemplary automated milling with respect to usability. RESULTS: Participants were able to differentiate compact and cancellous bone, both directly manually and teleoperatively. Both telemanipulated modes increased effectiveness measured by the mean absolute depth and contour error for guided and depth sensitive millings. Efficiency is decreased if solely a boundary constraint is used in hard material, while a trajectory guidance and manual milling perform similarly. With respect to subjective user satisfaction trajectory guidance is rated best for guided millings followed by boundary constraints and the direct manual interaction. Haptic feedback only improved subjective user satisfaction. CONCLUSION: A cooperative surgical telemanipulator can improve effectiveness and efficiency close to an automated execution and enhance user satisfaction compared to direct manual interaction. At the same time, the surgeon remains part of the control loop and is able to adjust the surgical plan according to the intraoperative situation and his/her expertise at any time.

Image-to-Image Translation for Simplified MRI Muscle Segmentation.

Deep neural networks recently showed high performance and gained popularity in the field of radiology. However, the fact that large amounts of labeled data are required for training these architectures inhibits practical applications. We take advantage of an unpaired image-to-image translation approach in combination with a novel domain specific loss formulation to create an "easier-to-segment" intermediate image representation without requiring any label data. The requirement here is that the task can be translated from a hard to a related but simplified task for which unlabeled data are available. In the experimental evaluation, we investigate fully automated approaches for segmentation of pathological muscle tissue in T1-weighted magnetic resonance (MR) images of human thighs. The results show clearly improved performance in case of supervised segmentation techniques. Even more impressively, we obtain similar results with a basic completely unsupervised segmentation approach.