University of Twente Student Theses


Tracked ultrasound for patient registration in surgical navigation during abdominal cancer surgery

Hiep, M.A.J. (2021) Tracked ultrasound for patient registration in surgical navigation during abdominal cancer surgery.

[img] PDF
Abstract:Surgical navigation is needed for the localization of tumors, malignant lymph nodes and other surrounding structures during abdominal cancer surgery. Currently, a cone-beam computed tomography (CBCT) scan is made on the operating room (OR) for the patient registration with pre-operative imaging and the electromagnetic (EM) tracking system. However, patient movement or tilting the surgical bed after scanning often results in surgical navigation inaccuracies. Additionally, the CBCT is limitedly available, clinical workflow is fully interrupted during scanning and staff and patient are exposed to radiation. Tracked ultrasound (US) might overcome these limitations, since US acquisition is possible on every OR regardless of the patient position, causes minimal workflow interruption and is non-invasive. Therefore, in this thesis, the feasibility and accuracy of EM tracked US was evaluated in a patient study. Firstly, a hybrid magneto-optical tracking setup was developed to examine possible interference of two US devices: BK and Clarius. Static and dynamic measurements were performed, which provided insight to fixate an EM sensor to the US device at a location with minimal interference. Secondly, both US devices were separately calibrated with the EM tracking system using the tracked pointer method, resulting in a calibration accuracy of 1.5 and 3.1 mm for BK and Clarius, respectively. Lastly, an US registration method based on the pelvic bone was applied on patients undergoing navigated abdominal cancer surgery. The accuracy of this method was evaluated at surgical targets compared to the current reference: CBCT registration. The influence of alternate patient positions on the registration was evaluated by acquiring US scans in Trendelenburg and in horizontal patient position. Ten patients were included in total. After exclusion of three unreliable patient measurements, an average target registration error of 2.6 mm was found for the BK US device, which suggest an acceptable navigation accuracy of <5 mm. Comparison of the horizontal with Trendelenburg US registration showed an average target registration discrepancy of 7.0 mm with an average offset in caudal-cranial direction of 6.5 mm. Further research is required to evaluate the true intra-operative navigation accuracy and the development of an automatic US bone segmentation algorithm is recommended. It is concluded that the tracked US registration method is feasible and accurate at surgically relevant targets in the pelvic cavity regardless of the patient position or surgical bed tilting on the OR. This method has the potential to replace the current CBCT-scan in abdominal cancer navigation surgery.
Item Type:Essay (Master)
Faculty:TNW: Science and Technology
Subject:44 medicine, 50 technical science in general
Programme:Technical Medicine MSc (60033)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page