ABSTRACT
BACKGROUND: Precise laparoscopic liver resection requires accurate planning and visualization of important anatomy such as vessels and tumors. Combining laparoscopic ultrasound with navigation technology could provide this. Preoperative images are valuable for planning and overview of the procedure, while intraoperative images provide an updated view of the surgical field. PURPOSE: To validate the accuracy of navigation technology based on preoperative images, we need to understand how much the liver shifts and deforms due to heartbeat, breathing, surgical manipulation and pneumoperitoneum. In this study, we evaluated liver tumor shift and deformation due to pneumoperitoneum in an animal model. METHODS: Tumor models were injected into the liver of the animal, and 3D CT images were acquired before and after insufflation. Tumor shifts and deformation were determined. RESULTS: The results showed significant tumor position shift due to pneumoperitoneum, with a maximum of 28 mm in cranio-caudal direction. No significant tumor deformation was detected. Small standard deviations suggest rigid body transformation of the liver as a whole, but this needs further investigation. CONCLUSION: The findings indicate a need for anatomic shift correction of preoperative images before they are used in combination with LUS guidance during a laparoscopic liver resection procedure.
Subject(s)
Laparoscopy/adverse effects , Liver Neoplasms/surgery , Liver/surgery , Pneumoperitoneum/surgery , Surgery, Computer-Assisted/methods , Animals , Disease Models, Animal , Intraoperative Care , Laparoscopy/instrumentation , Laparoscopy/methods , Liver Neoplasms/pathology , Pneumoperitoneum/pathology , Preoperative Care , Surgery, Computer-Assisted/instrumentation , SwineABSTRACT
We identified errors in the software implementation of the mathematical model presented in: Sá Couto CD, van Meurs WL, Goodwin JA, Andriessen P. A model for educational simulation of neonatal cardiovascular pathophysiology. Simul Healthcare 2006;1:4-12. Simulation results obtained with corrected code are presented for future reference. All but one of the simulation results do not differ by more than 9% from the previously published results. The heart rate response to acute loss of 30% of blood volume, simulated with corrected code is stronger than published target data. This modeling error was masked by errors in code implementation. We improved this response and the model by adjusting the gains and adding thresholds and saturations in the baroreflex model. General considerations on identification of model and code errors and model validity are presented.
