Volume 7, Issue 1 - Proceedings of the CURAC 2021 Annual Meeting, Düsseldorf, September 16-18, 2021

Introduction: Contaminated surgical instruments are manually prepared for cleaning and disinfection in the reprocessing unit for medical devices (RUMED). Manual labour exposes staff to the risk of infection and is particularly stressful at peak times due to the large volume of instruments. Partial automation of processes by a robot could provide a solution but requires a gripper that can handle the variety of surgical instruments. This paper describes the development and first evaluation of an instrument gripper. Methods: First, an analysis of gripping geometries on basic surgical instruments is carried out. Based on the identified common features and a review of the state of the art of gripper technology, the SteriRob gripper concept is developed. The concept is compared with a force closure gripper in a series of tests using seven criteria. Results: Both gripping approaches investigated can be used for handling surgical instruments in a pick-and-place process. However, the SteriRob gripper can transmit significantly higher acting forces and torques. In addition, the gripping process is more robust against deviations from the expected instrument position. Conclusion: Overall, it has been shown that the developed instrument gripper is suitable for about 60% of reusable surgical instruments due to the focus on horizontal cylindrical geometries. Because of the large possible force transmission, this gripping approach is particularly suitable for tasks in which the robot assists with cleaning processes.

A core principle of modern health care is the compliance of hygienic and aseptic techniques in areas that are sensitive to contamination through bacteria, dust, aerosols, and fallout, primarily in operating theatres or around patients with contagious diseases. Keeping track of potentially contaminated surfaces in an environment is a major concern, especially when protecting from COVID-19. This work proposes a novel concept in using 3D sensing technology to track human movement within an indoor area and identifying high-risk contaminated surfaces in real-time. It combines recent Augmented Reality display technology, which allows keeping track of decontaminated surfaces during the cleaning process using an interactive visualization method. The proposed concept of Clean- AR is implemented in a clinical environment used for observation in COVID-19 scenarios. We discuss key challenges and outline further research direction in effectively reducing the risk of contamination using the proposed concept.

Manual control of surgical instruments represents a sensorimotor control task with at least 3-6 degrees of freedom (DoF). The impact of haptic guidance on volumetric navigation tasks, such as milling of planned volumes for prosthesis fits or preserving sensitive tissues, is investigated. Interaction centered studies are performed to evaluate the usability of the assistance modes for navigation within a volume, along the surface of a volume and around forbidden regions. Results show that haptic assistance can reduce the number of constraint violations, if the virtual stiffness is high enough. However, haptic assistance also can increase error rates when counterforces are close to the absolute perception threshold, as a false sense of security can arise. For navigation along complex surfaces bilateral haptic constraints should be preferred, while unilateral constraints are sufficient for simple geometries. This study complements previous publications as a basis for a flexible rule-based selection or adaptation of modular haptic assistance systems.

In endovascular aneurysm repair (EVAR) procedures, the stent graft navigation and implantation is currently performed under a two-dimensional (2D) imaging based guidance requiring X-rays and contrast agent. In this work, a novel 3D stent graft guidance approach based on tracking systems is introduced. A calibration method and the visualization of the stent graft guidance are described. The tracking based stent graft guidance is evaluated by conducting an EVAR procedure on a torso phantom using a stent graft system equipped with an optical fiber and three EM sensors. The physicians were able to navigate the stent graft to the landing zone, and to place and implant it as intended using the introduced guidance. This showed that the application of the stent graft guidance is feasible in a clinical environment and promising for the reduction of radiation and contrast agent.

The distinction between malignant and benign tumors is essential to the treatment of cancer. The tissue's elasticity can be used as an indicator for the required tissue characterization. Optical coherence elastography (OCE) probes have been proposed for needle insertions but have so far lacked the necessary load sensing capabilities. We present a novel OCE needle probe that provides simultaneous optical coherence tomography (OCT) imaging and load sensing at the needle tip. We demonstrate the application of the needle probe in indentation experiments on gelatin phantoms with varying gelatin concentrations. We further implement two deep learning methods for the end-toend sample characterization from the acquired OCT data. We report the estimation of gelatin sample weight ratios [wt%] in unseen samples with a mean error of 1.21 ± 0.91 wt%. Both evaluated deep learning models successfully provide sample characterization with different advantages regarding the accuracy and inference time.

Camera resectioning is the process of finding the intrinsic and extrinsic parameters of the pinhole camera model. A mobile X-ray device (C-arm) can be represented with the pinhole camera model. The C-arm parameters can be determined with a parametrization device and the direct linear transformation (DLT) algorithm. We present the conception process of a parametrization device and further optimizations of the DLT algorithm. The evaluation of the parametrization quality is carried out with an inverse registration approach on a testing device. The seven landmarks in the testing device are compared on an X-ray image and a digitally reconstructed radiograph (DRR). The average distance of the corresponding landmarks between imaging modalities is 2.5 pixels. This deviation indicates an adequate design of the parametrization device for a 2D/3D registration in minimally invasive surgery.

Purpose Computerized medical imaging processing assists neurosurgeons to localize tumours precisely. It plays a key role in recent image-guided neurosurgery. Hence, we developed a new open-source toolkit, namely Slicer-DeepSeg, for efficient and automatic brain tumour segmentation based on deep learning methodologies for aiding clinical brain research. Methods Our developed toolkit consists of three main components. First, Slicer-DeepSeg extends the 3D Slicer application and thus provides support for multiple data input/ output data formats and 3D visualization libraries. Second, Slicer core modules offer powerful image processing and analysis utilities. Third, the Slicer-DeepSeg extension provides a customized GUI for brain tumour segmentation using deep learning-based methods. Results The developed Slicer- DeepSeg was validated using a public dataset of high-grade glioma patients. The results showed that our proposed platform’s performance considerably outperforms other 3D Slicer cloud-based approaches. Conclusions Developed Slicer-DeepSeg allows the development of novel AIassisted medical applications in neurosurgery. Moreover, it can enhance the outcomes of computer-aided diagnosis of brain tumours. Open-source Slicer-DeepSeg is available at github.com/razeineldin/Slicer-DeepSeg.

Elasticity of soft tissue is a valuable information to physicians in treatment and diagnosis of diseases. The elastic properties of tissue can be estimated with ultrasound (US) shear wave imaging (SWEI). In US-SWEI, a force push is applied inside the tissue and the resulting shear wave is detected by high-frequency imaging. The properties of the wave such as the shear wave velocity can be mapped to tissue elasticity. Commonly, wave features are extracted by tracking the peak of the shear wave, estimating the phase velocity or with machine learning methods. To tune and test these methods, often simulation data is employed since material properties and excitation can be accurately controlled. Subsequent validation on real US-SWEI data is in many cases performed on tissue phantoms such as gelatine. Clearly, validation performance of these procedures is dependent on the accuracy of the simulated tissue phantom and a thorough comparison of simulation and experimental data is needed. In this work, we estimate wave parameters from 400 US-SWEI data sets acquired in various homogeneous gelatine phantoms. We tune a linear material model to these parameters. We report an absolute percentage error for the shear wave velocity between simulation and phantom experiment of <2.5%. We validate our material model on unknown gelatine concentrations and estimate the shear wave velocity with an error <3.4% for in-range concentrations indicating that our material model is in good agreement with US-SWEI measurements.

The goal of our research work is the development of a novel endoscopic anastomosis device for the colon. One of the main challenges in this context is the application of forces at the endoscope tip to rejoin the two bowel endings. Thus, we focus on a magnetic two-part compression implant approach. The implant halves are detached from the applicator units by means of electromagnets. In this contribution we present the results of our experiments to determine the implant design with special focus on tissue compression forces and the resultant electromagnet dimensioning to estimate size requirements of the application/detachment system. To achieve the targeted compression forces derived from literature, we used cubic N52 magnetized neodymium magnets1 with a side length of 5 mm and mild steel screws. For these magnets, we evaluated a required electromagnetic repulsion force of 4.1 N. For the electromagnetic detachment system this led to the need for 166 windings for the coils on oral side, and 146 windings for the coils at the aboral side. Based on these requirements, a colonoscope diameter (~14 mm) increase of 10.6 mm on the oral side and of 12 mm on the aboral side due to the application device must be assumed. Nevertheless, this diameter still remains within the size range of other colonoscopic tools, such as e.g., circular staplers.

We present an analysis tool for subgroup identification in medical research based on feature analysis. Our use case is intracranial aneurysms. In the tool, an aneurysm-of-interest’s most similar aneurysms within a database are found. Similarity is defined via user-selected parameters, which can be entirely arbitrary. Different interactive outputs and visualizations include a heatmap view and a graph, which give an intuitive feedback to support researchers in the consideration of research questions, which in the present use case often relate to rupture risk analysis. The tool was evaluated with a pilot study and phantom database and received favorable results for its requirements of reliability and appropriate and clear outputs.

Minimally invasive interventions, e.g., percutaneous needle interventions, have many advantages compared to traditional surgery. However, they may require complex and time-consuming planning with experience-dependent success. Automated access path planning is faster and more consistent but individual preferences and situational circumstances are not considered. To this end, displaying the path planning results directly on the patient’s skin, using projector-based augmented reality (AR), was investigated. A constraint-based path planning was implemented to evaluate the quality of every path, taking into account risk structures and path length. A visualization was developed to display the results on the skin and to allow for path selection. The choice of the path followed by a navigated insertion was evaluated in a pilot study (n=5), considering four levels of the visualization with different amounts of displayed insertion points. Participants stated that they preferred to have multiple potential puncture points displayed. However, the results for the considered variables show only small differences. Overall, it has been shown that projectorbased AR visualization of automated access path planning is possible and enables individual, situation-adapted insertion point selection. More research is required to further explore optimal display of paths.

Fluorescent cardiac imaging can be applied for intraoperative quality control after a coronary bypass grafting surgery to ensure the myocardial perfusion by evaluating the increasing contrast agent enrichment in the heart. The motion due to the beating heart impedes the interpretation of the contrast agent enrichment in the vessels and leads to noisy enrichment curves. We propose tracking of the heart surface features to compensate for the motion of the beating heart and thereby improve the analysis of the contrast agent enrichment. Furthermore, we propose a vessel segmentation pipeline for a local evaluation of contrast agent enrichment directly in the vessels.

Robots from material-based actuators offer high potential for small-scale robots with abilities hardly achievable by classical methods like electric motors. Besides excellent scaling to minimally invasive systems, allowing for omission of metallic components, such robots can be applied in imaging modalities such as MRI or CT. To allow for higher accessibility in this field of research, a facile method for fabrication of such soft actuators was developed. It comprises only two materials: graphene oxide and silicone elastomer. The facile fabrication method does not require specialized equipment. The resulting actuator is biocompatible and controllable by light mediated heat. The bending motion can be controlled by the intensity of applied infrared light and the actuator was experimentally shown to move five times its own weight. Thus, providing capabilities for a medical soft robotic actuator.

Medulloblastoma (MB) is a primary central nervous system tumor and the most common malignant brain cancer among children. Neuropathologists perform microscopic inspection of histopathological tissue slides under a microscope to assess the severity of the tumor. This is a timeconsuming task and often infused with observer variability. Recently, pre-trained convolutional neural networks (CNN) have shown promising results for MB subtype classification. Typically, high-resolution images are divided into smaller tiles for classification, while the size of the tiles has not been systematically evaluated. We study the impact of tile size and input strategy and classify the two major histopathological subtypes-Classic and Desmoplastic/Nodular. To this end, we use recently proposed EfficientNets and evaluate tiles with increasing size combined with various downsampling scales. Our results demonstrate using large input tiles pixels followed by intermediate downsampling and patch cropping significantly improves MB classification performance. Our top-performing method achieves the AUC-ROC value of 90.90% compared to 84.53% using the previous approach with smaller input tiles.

Towards computer-assisted neurosurgery, scene understanding algorithms for microscope video data are required. Previous work utilizes optical flow to extract spatiotemporal context from neurosurgical video sequences. However, to select an appropriate optical flow method, we need to analyze which algorithm yields the highest accuracy for the neurosurgical domain. Currently, there are no benchmark datasets available for neurosurgery. In our work, we present an approach to generate synthetic data for optical flow evaluation on the neurosurgical domain. We simulate image sequences and thereby take into account domainspecific visual conditions such as surgical instrument motion. Then, we evaluate two optical flow algorithms, Farneback and PWC-Net, on our synthetic data. Qualitative and quantitative assessments confirm that our data can be used to evaluate optical flow for the neurosurgical domain. Future work will concentrate on extending the method by modeling additional effects in neurosurgery such as elastic background motion.

This paper presents a proof-of-concept intervention monitoring and outcome assessment tool for thermal ablative strategies like microwave ablation of hepatic tumours. For this purpose, simulated 3D+t temperature maps of the ablation progress are combined with liver MR imaging, as well as corresponding segmentations of the lesion and other relevant anatomical structures. The proposed tool mainly consists of a monitoring and validation part, each comprising 2D and 3D visualisations and various modifiable settings. With the aid of the monitoring features, the heat propagation during the ablation progress could be visually tracked, while the validation part of the tool provides the radiologist with comprehensible and detailed feedback to assess the treatment result. Hence, it could enable an immediate validation of the ablation progress and support interventional decision-making. For evaluation purposes, we used an exemplary patient case to demonstrate the benefits of the proposed prototype. This tool could represent a crucial step towards a suitable computerassisted and more accurate workflow of thermal ablations, with benefits for both, patients and radiologists.

For the percutaneous fixation of scaphoid fractures, navigated approaches have been proposed to facilitate screw placement. Based on ultrasound imaging, navigation can be carried out in a cost-effective and fast manner, furthermore avoiding harmful radiation. For this purpose, a fast and efficient architecture for the automated segmentation of scaphoid bone in ultrasound volume images is needed. Methods: For 2D segmentation of the scaphoid, two architectures are taken into account: 2D nnUNet and Deeplabv3+. These architectures are trained and evaluated on a newly created dataset consisting of 67 annotated in-vivo ultrasound volume images (4576 slice images). Results: In terms of Dice coefficient, the 2D nnUNet achieves 0.67 compared to 0.57 for the Deeplabv3+. In terms of distance metrics, the 2D nnUNet shows an average symmetric surface distance error of 0.66mm, while the Deeplabv3+ achieves 0.55mm. Conclusion: Fast and accurate segmentation of the scaphoid in ultrasound volumes is feasible. Both architectures show competitive results.

Performing minimal invasive interventions under real-time image guidance proves problematic in a closed-bore magnetic resonance imaging scanner. To enable better usability in MRI guided interventions, robotic systems could be used for additional assistance. However, the integration of such devices into the clinical workflow relates to many technical challenges in order to increase precision of the procedure while ensuring the overall safety. In this work, an MR compatible, compact, ultra-light and remotely controllable micropositioning system called μRIGS is presented. The instrument positioning unit can be operated in a 5-DoF range within a working volume of 2100 cm 3 with an instrument feed of 120 mm. The kinematics are actuated with a combination of non-metallic Bowden cables and electric stepper motors from a safe distance inside the scanner room, while their control is initiated from the control room via a custom-fitted GUI. Thereby, the precision of the positioning reproducibility of the respective DoF can be achieved with a mean deviation of 0.12 °. Furthermore, a feed force of 14 N can be provided to puncture various soft tissue.

Smart medical phantoms for training and evaluation of endovascular procedures ought to measure impact forces on the vessel walls worth protecting to provide feedback to clinicians and articulated soft robots. Recent commercial smart phantoms are expensive, usually not customizable to different applications and lack accessibility for integrated development. This work investigates piezoresistive films as highly integratable flexible sensors to be used in arbitrary soft vessel phantom anatomies over large surfaces and curved shapes providing quantitative measurement in the force range up to 1 N with 0.1 N resolution. First results show promising performance at the point of calibration and in a 5 mm range around it, with absolute measuring error of 28 mN and a standard deviation of ±10 mN and response times <500 ms. Future work shall address the optimization of response time and sensor shapes as well as the evaluation with experienced clinicians.

In context of the Corona pandemic, telemedicine acquired a new significance. Whereas previously the aim was to override given barriers, now, in the case of a pandemic, the main idea is to create an intentional distance between patients and healthcare professionals in order to avoid cross-infection. To meet the needs of a fully diagnostic examination, a robotic based system was designed. However, collaborative robotic systems bear new risks, that have to be dealt with. To prepare future physicians for telediagnostics, we developed a training curriculum for the telemedical examinations. It is based upon multiple stages including a skill trainer, healthy volunteers, supervised examinations of real patients and an exam. In a first proof of concept, we demonstrated the existence of a learning curve and significant better performance after the passed curriculum compared to an untrained collective.

Knowing the shape of vascular calcifications is crucial for appropriate planning and conductance of percutaneous coronary interventions. The clinical workflow can therefore benefit from automatic segmentation of calcified plaques in intravascular ultrasound (IVUS) images. To solve segmentation problems with convolutional neural networks (CNNs), large datasets are usually required. However, datasets are often rather small in the medical domain. Hence, developing and investigating methods for increasing CNN performance on small datasets can help on the way towards clinically relevant results. We compared two state-of-the-art CNN architectures for segmentation, U-Net and DeepLabV3, and investigated how incorporating auxiliary image data with vessel wall and lumen annotations improves the calcium segmentation performance by using these either for pretraining or multi-task training. DeepLabV3 outperforms U-Net with up to 6.3 % by means of the Dice coefficient and 36.5 % by means of the average Hausdorff distance. Using auxiliary data improves the segmentation performance in both cases, whereas the multi-task approach outperforms the pre-training approach. The improvements of the multi-task approach in contrast to not using auxiliary data at all is 5.7 % for the Dice coefficient and 42.9 % for the average Hausdorff distance. Automatic segmentation of calcified plaques in IVUS images is a demanding task due to their relatively small size compared to the image dimensions and due to visual ambiguities with other image structures. We showed that this problem can generally be tackled by CNNs. Furthermore, we were able to improve the performance by a multi-task learning approach with auxiliary segmentation data.

Access to clinical information during interventions is an important aspect to support the surgeon and his team in the OR. The OR-Pad research project aims at displaying clinically relevant information close to the patient during surgery. With the OR-Pad system, the surgeon shall be able to access case-specific information, displayed on a sterilepackaged, portable display device. Therefore, information shall be prepared before surgery and also be available afterwards. The project follows an user-centered design process. Within the third iteration, the interaction concept was finalized, resulting in an application that can be used in two modes, mobile and intraoperative, to support the surgeon before/after and during surgery, respectively. By supporting the surgeon perioperatively, it is expected to improve the information situation in the OR and thereby the quality of surgical results. Based on this concept, the system architecture was designed in detail, using a client-server architecture. Components, communication interfaces, exchanged data, and intended standards for data exchange of the OR-Pad system including connecting systems were conceived. Expert interviews by using a clickable prototype were conducted to evaluate the concepts.

Algorithms for automated analysis of intravascular ultrasound (IVUS) images can be disturbed by guidewires, which are often encountered when treating bifurcations in percutaneous coronary interventions. Detecting guidewires in advance can therefore help avoiding potential errors. This task is not trivial, since guidewires appear rather small compared to other relevant objects in IVUS images. We employed CNNs with additional multi-task learning as well as different guidewire-specific regularizations to enable and improve guidewire detection. In this context, we developed a network block which generates heatmaps that highlight guidewires without the need of localization annotations. The guidewire detection results reach values of 0.931 in terms of the F1-score and 0.996 in terms of area under curve (AUC). Comparing thresholded guidewire heatmaps with ground truth segmentation masks leads to a Dice score of 23.1 % and an average Hausdorff distance of 1.45 mm. Guidewire detection has proven to be a task that CNNs can handle quite well. Employing multi-task learning and guidewire-specific regularizations further improve detection results and enable generation of heatmaps that indicate the position of guidewires without actual labels.

Augmented Reality glasses such as HoloLens 2 may provide visual guidance during surgical interventions. To superimpose the holograms on real world objects (RWO), for instance a patient, spatial registration is required. In this work, we propose an approach to automatically register a hologram to the according RWO. To this end, the framework utilizes the depth camera of HoloLens 2 to acquire the point cloud (PC) of the RWO. A novel and recently published PC registration algorithm allows to register the PC of the RWO and the hologram after a rough initial placement without any need for pre-processing or outlier removal. The approach is evaluated by measuring displacements between certain known positions of the hologram and the RWO. The first metric relies on measuring points using an optically tracked stylus while the second is based on visually perceived positions. The median displacements were 22.3 mm, 35.6 mm, and 13.3 mm for the x-, y-, and z-axes in the first metric and 8.1 mm, 4.3 mm, and 11.9 mm for the second metric. Even though the accuracy is not yet adequate for many surgical interventions, the framework provides an initial step for a convenient marker less registration of holograms to an RWO.

Fractures of the scaphoid bone may be treated in a minimally-invasive fashion. Conventionally, fluoroscopy is required to guide the placement of an osteosynthesis screw. In this work, an alternative method based on volumetric ultrasound is validated. Methods: The fully automatic and fast image processing pipeline involves two machine learning architectures for segmentation and registration. A pre-operatively acquired 3D bone model is registered to the 3D bone surface segmented from the intra-operative ultrasound. Screw positioning is planned in an automated fashion and evaluated in an in-vitro setting: Volumetric ultrasound images of a 3D-printed phantom of a human wrist are acquired for 22 different probe poses. For 220 test runs with different initial displacements, the resulting screw placement within a defined safe zone is evaluated. If the screw lies within the safe zone, its placement is assumed to be successful. Results: An isolated analysis of the registration results in a surface distance error of the registered meshes of 0.49 ± 0.01mm, with successful screw placement in all of the evaluated 220 test runs. The full pipeline, combining segmentation and registration, achieves a mean surface distance error of 0.79 ± 0.37mm, leading to successful screw placements for 149 out of 220 test runs. Poses not suited for the registration could be determined. Excluding these from the analysis, 139 out of 160 test runs are successful. Conclusion: The method proves to be promising when evaluating the registration alone, even given the challenging setup of sub-optimal probe positions. The experiments also demonstrate that further improvement regarding the segmentation is necessary.

Endovascular surgical robotics requires a facile and realistic testbed to validate control algorithms. This work compares methods to manufacture such a testbed. Utilizing animal tissue, a mannequin, and a low-voltage flow pump it is possible to perform catheter-based interventions with X-ray feedback for less than €300. The aim of this paper is to lower the entry hurdle for validating endovascular surgical robots by providing a method to build a facile and low-budget testbed.

Robotic assistance systems for surgery enable fast and precise interventions with reduced complication rates. However, these benefits are accompanied by a more complex operating room (OR) and the risk of collision with robotic assistance systems. Current strategies for collision avoidance and minimizing possible injuries require the adaptation of robotic trajectories and a computational model of the surroundings. In contrast, this work presents a novel companion system for collision avoidance without influencing robotic trajectories. The companion system consists of a preoperative planning application and an augmented reality application for intraoperative support. The companion system visualizes the workflow within the OR and allows robot movements to be seen virtually, before they are executed by the actual robotic assistance system. Preliminary experiments with users imply that the companion system leads to a positive user experience, enables users to follow a predefined workflow in the OR, but requires further refinement to improve accuracy for practical collision avoidance.

Interdisciplinary exchange of medical datasets between clinicians and engineers is essential for clinical research. Due to the Data Protection Act, which preserves the rights of patients, full anonymization is necessary before any exchange can take place. Due to the continuous improvement of image quality of tomographic datasets, anonymization of patient-specific information is not sufficient. In this work, we present a prototype that allows to reliably obscure the facial features of patient data, thus enabling anonymization of neurological datasets in image space.

The documentation of surgeries is usually created from memory only after the operation, which is an additional effort for the surgeon and afflicted with the possibility of imprecisely, shortend reports.The display of process steps in the form of checklists and the automatic creation of surgical documentation from the completed process steps could serve as a reminder, standardize the surgical procedure and save time for the surgeon. Based on two works from Reutlingen University, which implemented the creation of dynamic checklists from Business Process Modelling Notation (BPMN) models and the storage of times at which a process step was completed, a prototype was developed for an android tablet, to expand the dynamic checklists by functions such as uploading photos and files, manual user entries, the interception of foreseeable deviations from the normal course of operations and the automatic creation of OR documentation.

Dizziness is one of the most frequent symptoms in outpatient practices. For the differentiation of peripheral or central pathogenesis of vertigo, history taking and clinical examination with the detection of nystagmus is elementary. The aim of this study was to investigate the effect of lighting for the detection of horizontal vestibular nystagmus while utilizing a conventional webcam. In the proof-of-concept study, caloric induced vestibular nystagmus was recorded with a conventional video-nystagmography and mobile webcam in addition to an external light source. The analysis of recorded data was performed with a self-developed software using computer vision techniques. The self-designed algorithm detected the existence of nystagmus and its direction in several cases. The experimental webcam-based vestibular nystagmography could be enhanced by improving lighting conditions. Currently, a clinical application for this technique is not approved. Further software improvements are necessary to increase its accuracy.

Currently, various AI-based systems for computerassisted adenoma- and polyp-detection during colonoscopy have been brought to the market and are under clinical investigation. With these systems available to be used during routine screening colonoscopy and first results published about experiments and findings, it has become of interest how and to which extend such systems are used during the examination. Specifically, similarly to automotive navigation, it is of interest of how much visual focus is put onto the augmented image of the above-mentioned devices, signalling possible hypothesis of adenomas or polyps, and how much time-of-attention remains on the original colonoscopic video data. Thus, within a study, N = 36 participants using a prototype of a polypdetection system have been observed with an eye-tracker-system, to capture and evaluate the relative time of attention with respect to the original and augmented video data and differentiate these values between various sub-groups based on experience, education and gender. T-tests were conducted to identify potential significant differences. Based on the obtained data, the augmented video data is used with a very high attention (up to 75%) depending on the regarded sub-group. Experienced as well as less-experienced users (with > 500 colonoscopies) both preferred looking at the original data. In contrast, gastroenterologists (in contrast to nurses, students, engineers) were more interested in the outcome of the novel AIsystem. The female group preferred looking at the unobstructed data, while the male group was highly interested in the AI-based data.

Background and objectives: Liver lesions are a relatively common incidental finding in computer tomography (CT) of the abdomen. The current gold standard is liver biopsy, which has the downside of respecting only a small part of the total lesion volume. Furthermore, this invasive method carries interventional risks like bleeding or infection. Therefore, an image-based biomarker would be highly desirable. Conventional “radiomics” methods have often been utilized for similar problems, but the results are often not reproducible. This is mainly due to sampling errors and interobserver variability, but also the seemingly complex nature of the problem. We present a new approach that implements cutting-edge research in machine learning which is nevertheless cheap and easily applicable in a routine clinical setting. To achieve this, we use convolutional neural networks (CNN) to predict the histopathological findings from liver lesions from preoperative liver CT. Methods: After splitting the study population into a training and test set we trained a CNN to predict the histopathological tumor type from CT data. Results: The developed CNN workflow is able to predict liver tumor histology from routine CT images. We also evaluated in how far transfer learning and data augmentation can help in solving this problem and implemented the developed workflow in a clinical routine setting. Conclusion: We propose a robust semiautomatic end-to-end classification workflow for the prediction of the histopathological type of tumor lesions based on abdominal CT and a deep convolutional neural network model. In our cohort, the model shows reliable and accurate results even with limited computational resources.

Background and objectives: Both hepatic functional reserve and the underlying histology are important determinants in the preoperative risk evaluation before major hepatectomies. In this project we developed a new approach that implements cutting-edge research in machine learning and nevertheless is cheap and easily applicable in a routine clinical setting is needed. Methods: After splitting the study population into a training and test set we trained a convolutional neural network to predict the liver function as determined by hepatobiliary mebrofenin scintigraphy and single photon emission computer tomography (SPECT) imaging. Results: We developed a workflow for predicting liver function from routine CT imaging data using convolutional neural networks. We also evaluated in how far transfer learning and data augmentation can help to solve remaining manual data pre-processing steps and implemented the developed workflow in a clinical routine setting. Conclusion: We propose a robust semiautomatic end-to-end classification workflow for abdominal CT scans for the prediction of liver function based on a deep convolutional neural network model that shows reliable and accurate results even with limited computational resources.

This work compares three different approaches to automatically segment the femoral artery from 2D ultrasound images. Two of the architectures follow a sequential structure, where each ultrasound image is considered a slice of the whole vessel volume, and its previous segmentation result will be part of the input, thus leading to a spatial prior. The Dice score on test data show a better performance on the baseline U-Net (0.819) compared to the sequential U-Net approaches (0.633, 0.725) for the femoral artery segmentation. This could be attributed to the misalignment of the slices being used in those networks. A possible improvement could be assumed in the implementation of a spatially calibrated and tracked ultrasound probe. Overall, these results indicate promising approaches for an automatic segmentation of the femoral artery using 2D ultrasound data.

Under-staffing of nurses is a significant problem in most countries. It is expected to rise in the coming years, making it challenging to perform crucial tasks like assessing a patient's condition, assisting the surgeon in medical procedures, catheterization and Blood Transfusion etc., Automation of some essential tasks would be a viable idea to overcome this shortage of nurses. One such task intended to automate is the role of a 'Scrub Nurse' by using a robotic arm to hand over the surgical instruments. In this project, we propose to use a Collaborative Robotic-arm as a Scrub nurse that can be controlled with voice commands. The robotic arm was programmed to reach the specified position of the instruments placed on the table equipped with a voice recognition module to recognize the requested surgical instrument. When the Surgeon says "Pick Instrument", the arm picks up the instrument from the table and moves it over to the prior defined handover position. The Surgeon can take over the instrument by saying the command "Drop". Safe pathways for automatic movement of arm and handover position will be predefined by the Surgeon manually. This concept was developed considering the convenience of the Surgeon and the patient's safety, tested for collision, noisy environments, positioning failures and accuracy in grasping the instruments. Limitations that need to be considered in future work are the recognition of voice commands which as well as the returning of the instruments by the surgeon in a practical and safe way.

Three-dimensional visualizations and 3D-printed organs are used increasingly for teaching, surgery planning, patient education, and interventions. Hence, pipelines for the creation of the necessary geometric data from CT or MR images on a per-patient basis are needed. Furthermore, modern 3D printing techniques enable new possibilities for the models with regard to color, softness, and textures. However, to utilize these new features, the respective information has to be derived from the medical images in addition to the geometry of the relevant organ structures. In this work, we propose an automatable pipeline for the creation of realistic, patientspecific 3D-models for visualization and 3D printing in the context of liver surgery and discuss remaining challenges.

For minimally invasive drilling processes, the temperature development in the drilling ground is of crucial importance for patient safety. To monitor the temperature during drilling, a drill prototype was developed by BREDEMANN ET AL. which can record the drill temperature in parallel to the process and in real time. The measurement principle of the thermistor (temperature sensor) integrated in the drill could be validated. [1] The prototype must be refined for use in the operating room, as the drill does not yet meet all the medical requirements that need to be fulfilled. In further development, the recorded temperature data in particular must be processed and communicated to the surgeon in order to provide added value for the surgical procedure.