Paper Submission & Registration
8th Dutch Bio-Medical Engineering Conference
13:40   Endo & Laparoscopy
Chair: Joris Jaspers
15 mins
Multimodal markers for augmented surgical navigation
Mohamed Benmahdjoub, Wiro J. Niessen, Eppo B. Wolvius, Theo van Walsum
Abstract: Background: Augmented reality (AR) has been used in surgery for navigation purposes[1]. AR potentially solves the depth perception issue, the hand-eye coordination issue, and the requirement to continuously switch focus between the screen and the patient compared to conventional navigation systems. AR for surgical navigation has been implemented in two ways: a standalone method which uses only AR devices to track anatomical structures and visualize 3D data, and an integrated method, where AR devices are combined with conventional navigation systems. In the latter method, visualizing 3D data at the right location requires an alignment procedure between the AR device’s coordinate system (CS) and the CS of the tracking system of the navigation. A common way to achieve this alignment is by attaching markers to the AR device which can be tracked by the tracking system[2]. The current abstract presents an alignment method to align an AR device and a surgical tracking system which does not require markers attached to the AR device. Methods: The alignment is established by utilizing a multimodal marker (MM). The MM consists of a 2D pattern (QR-code) to which an EM sensor is rigidly attached. By calibrating the spatial relationship between the 2D pattern and the EM sensor, the MM is used as a bridge to link the patient and the various instruments, which are known to the navigation system, to the AR device’s CS. The AR overlay accuracy on a 3D printed phantom skull was assessed for different camera poses (20, 45, and 90 degrees) based on 2 metrics: intersection over union (IoU) and point-based annotation. Results: For a calibration achieving a target registration error (TRE) of 1.2 mm, the IoU had a mean accuracy of 94% for 20 and 45 degrees, and 88% for 90 degrees. The point-based annotation showed a mean accuracy of 2.5 mm for 20 and 45 degrees, and 10.4 mm for 90 degrees (due to an implementation limitation). Conclusion: Despite the technical improvement and user evaluation required to use the setup clinically, the approach successfully aligned a virtual 3D skull model on a physical one. References [1]M. Zhu et al., “Effectiveness of a novel augmented reality-based navigation system in treatment of orbital hypertelorism,” Ann. Plast. Surg., vol. 77, no. 6, pp. 662–668, Dec. 2016, doi: 10.1097/SAP.0000000000000644. [2]K. Gavaghan et al., “Evaluation of a portable image overlay projector for the visualisation of surgical navigation data: Phantom studies,” Int. J. Comput. Assist. Radiol. Surg., vol. 7, no. 4, pp. 547–556, Jul. 2012, doi: 10.1007/s11548-011-0660-7.
15 mins
The subjective perception and objective measurement of image quality of laryngoscopes
Geert Geleijnse, Laura Veder, Marieke Hakkesteegt, Mick Metselaar
Abstract: Ear Nose and Throat (ENT) specialists apply flexible laryngoscopes to examine the throat. Although these flexible type of endoscopes are an important tool, hospitals do not apply standardized tests to assess the image quality of the endoscopes they use or consider to purchase. Recently we evaluated the Rez Checker Target Nano Matte, a test chart that was designed by Image Science Associates for this purpose. This test chart can be used to measure image quality in terms of sharpness, noise and colour fidelity. The purpose of this study was to explore the relationship between the subjective perceived image quality of ENT-specialists and the objectively measured sharpness, noise and colour fidelity. We assessed the subjective ranking of six types of endoscopes by imaging the larynx of one subject. The images were presented to 30 ENT-professionals for a forced pairwise comparison. The objectively measured parameters were MTF50 for sharpness, RMS for noise and CIE dE 2000 for colour fidelity. ENT-professionals ranked the sharpest endoscopes best. There was no relation found between the subjective ranking and noise or colour fidelity. Whether better image quality also improves the diagnostic accuracy of the ENT-specialist seems plausible, but still has to be established.
15 mins
External force compensation in a tendon-driven flexible robotic endoscope using Cosserat Rod modeling
Willem Hoitzing, Yoeko Xavier Mak, Momen Abayazid
Abstract: Tendon-actuated endoscopic camera is used extensively in minimally invasive surgery and can be robotically controlled to improve its positioning accuracy. Numerous studies have been conducted to develop and investigate the feasibility of flexible endoscopic tool for laparoscopy procedures. While long and slender flexible endoscope has advantage to safely maneuver and reach difficult targets compared to the rigid scopes, it lacks stiffness and can easily deform due to external loads such as gravity and interaction forces. This deformation of flexible endoscopes can be estimated using a Cosserat based model. This will enable the system to compensate for such deformations and thus achieve reliable positioning during minimally invasive procedures. A prototype of a flexible tendon-actuated endoscopic system that can be robotically controlled while compensating for external load is developed in this study. The prototype uses 2 pairs of antagonistic tendon actuation mechanism with force measurement on the tendon. The position of endoscope tip is measured using an electromagnetic tracker. The designed actuation system shows good tracking performance according to the design goal, using PD position controller on each tendon. Secondly, the deformation prediction due to external loads was validated by comparing the simulated Cosserat-rod based model to the experimental result using the developed prototype. Identification of stiffness and mass parameters of the prototype with movable tip length of 110 mm were performed and the difference in deformation between simulated model and physical case were measured at 6 different bending, 12 axial rotation and loading amount m\in {0, 5, 10} grams. Significant prediction errors were observed for model-based compensation: 12.2\pm\ 6.4 mm error norm for case of 10 grams of loading. This is due to the unmodeled friction between tendons and the endoscope’s spacer disks. In conclusion, the design and modeling of flexible endoscope were successfully achieved in this study, while the deformation prediction has on average 14.6 mm (13.3% of the endoscope length) error over all tested configurations. Further work in improving the accuracy of the prediction is required to progress towards clinical use.
15 mins
The influence of Trocar design on gas leak
Daniel Robertson, Frank Sterke, Tim Horeman
Abstract: During laparoscopic minimally invasive surgery, the abdominal cavity is inflated with carbon dioxide (CO2). The gas in this peritoneum could become contaminated with micro-organisms such as viruses and surgical smoke. Because gas leaks into the surroundings through, for instance, the instruments and past trocar valves, the abdomen continuously has to be insufflated, potentially exposing the medical staff to contaminated gas. However, no studies have been performed that relate trocar and instrument design to the escape of CO2. In this study we have studied the performance of several types and sizes of trocars in terms of CO¬2 leakage. For the experiment, we used a rigid container with a silicone interface for airtight clamping of trocar, instrument or combination of both. We used a custom built pressure generator to pressurize the container, along with a pneumotachograph to measure flow at the inlet of the container. A protocol with several use-cases was designed to simulate the use of the trocar during surgery. This protocol simulates the motions the surgeon exerts on the trocar as well as some ‘worst-case’ scenarios, such as keeping an instrument tip between the two valves of the trocar. In total, we included twenty-three individual trocars, reusable and disposable, of sizes 5 mm, 5/12 mm and 12 mm, of which, we recorded the valve types and other key geometries. The results showed that for an average laparoscopic procedure of 55 minutes, the 5 mm trocars leaked between 2.3 L and 70.0 L of CO¬2 and the 5/12 mm and 12 mm trocars leaked between 2.3 L and 144 L. These leaks had different causes for different trocars, some of which could be clearly linked to factors in trocar design. Based on this study, we propose trocar redesigns to overcome some of the causes of gas leaks. Also we recommend an international standard for testing CO2 leakage for all new trocars and instruments so that surgical teams can avoid a potential health hazard when selecting new equipment.

end %-->