PURPOSE: We introduce a custom-designed phantom model for the in-vitro evaluation of an augmented reality-based soft-tissue navigation system for ultrasound-guided prostate interventions. MATERIALS AND METHODS: Transrectal ultrasound segmentation of the prostate, navigation aid placement, initial registration, endoscope tracking, and enhanced visualization steps in the navigation procedure were performed to accommodate the actual prostatic motion. In-vitro laparoscopic manipulations simulating surgical procedures were performed by a physician using human prostate specimens. The target visualization error, defining the accuracy of the tracking, is determined by means of a leave-out test strategy by alternately using four navigation aids for endoscope registration and the remaining two navigation aids for accuracy verification. RESULTS: The introduction of the navigation aids lasted approximately 3 minutes. The navigation aids and especially their barbs were visible because of their ultrasound reflecting nature. For each organ, 1000 endoscope registrations were calculated, in which two randomly chosen navigation aids served the purpose of verifying the pose. We were able to demonstrate that the superimposed image could follow automatically the videoendoscopic real-time view. The mean target visualization errors for the respective trials were determined as 0.81 (±0.12) mm, 0.62 (±0.14) mm, and 0.98 (±0.23) mm. CONCLUSIONS: The ultrasound-based inside-out navigation system for laparoscopic prostatectomy overcomes the problem of tissue shift and deformation in an in-vitro model. In case of organ movement, the augmented picture with the detected navigation aids could follow the videoendoscopic image using the navigation aids as landmarks.
PURPOSE: We introduce a custom-designed phantom model for the in-vitro evaluation of an augmented reality-based soft-tissue navigation system for ultrasound-guided prostate interventions. MATERIALS AND METHODS: Transrectal ultrasound segmentation of the prostate, navigation aid placement, initial registration, endoscope tracking, and enhanced visualization steps in the navigation procedure were performed to accommodate the actual prostatic motion. In-vitro laparoscopic manipulations simulating surgical procedures were performed by a physician using human prostate specimens. The target visualization error, defining the accuracy of the tracking, is determined by means of a leave-out test strategy by alternately using four navigation aids for endoscope registration and the remaining two navigation aids for accuracy verification. RESULTS: The introduction of the navigation aids lasted approximately 3 minutes. The navigation aids and especially their barbs were visible because of their ultrasound reflecting nature. For each organ, 1000 endoscope registrations were calculated, in which two randomly chosen navigation aids served the purpose of verifying the pose. We were able to demonstrate that the superimposed image could follow automatically the videoendoscopic real-time view. The mean target visualization errors for the respective trials were determined as 0.81 (±0.12) mm, 0.62 (±0.14) mm, and 0.98 (±0.23) mm. CONCLUSIONS: The ultrasound-based inside-out navigation system for laparoscopic prostatectomy overcomes the problem of tissue shift and deformation in an in-vitro model. In case of organ movement, the augmented picture with the detected navigation aids could follow the videoendoscopic image using the navigation aids as landmarks.
Authors: Ashwin N Sridhar; Archie Hughes-Hallett; Erik K Mayer; Philip J Pratt; Philip J Edwards; Guang-Zhong Yang; Ara W Darzi; Justin A Vale Journal: Nat Rev Urol Date: 2013-06-18 Impact factor: 14.432
Authors: Hannes Götz Kenngott; Anas Amin Preukschas; Martin Wagner; Felix Nickel; Michael Müller; Nadine Bellemann; Christian Stock; Markus Fangerau; Boris Radeleff; Hans-Ulrich Kauczor; Hans-Peter Meinzer; Lena Maier-Hein; Beat Peter Müller-Stich Journal: Surg Endosc Date: 2018-03-30 Impact factor: 4.584