Introduction Lesions detected at capsule endoscopy currently can only be viewed by a 2D image and size estimated by subjective image perception. The MiroCam MC4000 uses stereo matching technology to enable reconstruction of selected images in 3D format to enable size calculation. The aim of this study was to assess the reliability of this novel capsule to reconstruct 3D images, and its ability to accurately calculate lesion size within a phantom model. Methods The MC4000 capsule was mounted on a modified plastic shaft with the receiver circuitry system attached. In-vitro measurements at various distances were undertaken on a phantom model with static 2D and 3D lesions with sizes between 5–25 mm. In addition, dynamic movements resembling capsule movement through the bowel was performed. The ability to reconstruct images in a 3D format, in addition to the accuracy of the size calculations within ± 10%, was assessed. Results All lesions within the phantom model were successfully reconstructed in 3D format. For static lesions the MC4000 capsules estimated measurements correlated highly with the known sizes (table 1, 2D lesions). The effective distance for accurate size measurement for the MC4000 capsule was 10 mm to 25 mm. In total 2D & 3D objects, 5 mm to 20 mm in size, were measured with an overall accuracy of 88.4% within this effective distance. Abstract PTH-024 Table There was statistically significant correlation between 2D and 3D lesion size and measured size for each distance (correlation coefficient 0.98–1, p<0.049). Conclusions The MiroCam MC4000 capsule provides reconstructive technology with good size calculations. This novel function may allow more accurate interpretation of lesions identified during capsule endoscopy either for index lesions or when looking for follow up lesion sizes. Further adjustment to the software may be required to bring accuracy to an optimum level to allow further in-vivo studies.

Introduction Advancements in capsule endoscopy (CE) have led to the introduction of double ended capsule endoscopes. With a second camera on the rear end, these capsules can collect double the number of images per second compared to single ended capsules, and provide a field of view of 340 degrees, potentially doubling the area of bowel mucosa visualised. This, however, comes at the cost of a greater size. The clinical benefit of this additional camera has not been confirmed. Methods Prospective randomised cohort study of single v double tip CE from March 2018 to January 2019. Capsule reading speed and reporting was done as per individual preferences. In double-tip studies, readers were advised to alternate between cameras based on which gave the best views. Primary outcome measures were the positive yield (PY) and diagnostic yield (DY). Secondary outcome measures were quality of bowel views, gastric transit time (GTT), small bowel transit time (SBTT) and completion rate (CR). Results 326 CE procedures were performed during the study period: 201 single-tip v 125 double-tip. There was no statistical difference between the PY and DY (147/201= 73.1% v 90/125= 72.0%, p=0.82 and 86/201= 42.8% v 59/125= 47.2%, p=0.44 respectively). Despite the size difference, the transit times and completion rates between the 2 capsules were also similar (Median GTT 25 min v 28 min, p=0.22, Median SBTT 259 min v 247 min, p=0.35, CR 177/201= 88.1% v 114/125= 91.2%, p=0.37). Bowel views was reported as poor in significantly fewer double-tip than single-tip capsules (4.7% v 27.3%, p<0.01). Conclusions Although this study demonstrates that the double-tip capsule provides better views of the bowel mucosa than single-tip, due to the ability to alternate between camera sides, this did not translate into higher diagnostic yields. It may be that a different protocol for reading is required for the double-tip capsule to maximise use of the improved field of view, such as viewing both cameras simultaneously. The size of the capsule was not shown to affect the transit or completion rate.