Optimizing the performance of magnetic-assisted capsule endoscopy of the upper GI tract using multiplanar CT modelling

Background and aims This study used computed tomography modelling of the abdomen to determine the optimal placements of a magnetic capsule endoscope in the stomach that allows complete mucosal visualization and determined the optimal placement of the handheld magnet to aid pyloric traversing. Materials and methods Using multiplanar reformatting, 100 abdominal computed tomography scans were analysed to assess gastric luminal visualization by a capsule endoscope from five fixed stations throughout the stomach. From each station, we assessed the ability of a capsule endoscope to visualize six gastric landmarks. The pyloric canal angles were calculated to create a vector. We mapped the position of this vector on the patient to determine the optimal placement of the magnet that would aid pyloric traversing. Results Complete visualization of the stomach (viewing all landmarks) was only achievable in up to 29% of cases when viewed from a sole station. Maximal visualization required combining stations. Fundal dependent (station 1) and opposite antral dependent (station 4) was the best, achieving visualization in 85% (P<0.01). The optimal positioning of the magnet to aid pyloric traversing was posteriorly between vertebrae T5 and L2 in an area 10 cm to the left and 18 cm to the right (83% cases). Age older than 55 years (P=0.03) and the ability to view the pylorus from station 3 (P=0.04) were associated with an extreme pyloric vector. Conclusion Although there is some extreme variation in the upper gastrointestinal anatomy, the majority of cases will allow the use of a single standard method in performing MACE, which may be very useful for screening purposes.