Article Text
Abstract
Introduction An important risk of capsule endoscopy (CE) is capsule retention. Current practice is to proceed with the real CE if the patency capsule (PC) has been excreted at 30 h or if it has entered the colon on abdominal XRay (AXR). Computed Tomographic scanning (CT) may be used to determine device location if it is unclear but involves further irradiation. This study investigated the accuracy with which radiologists can determine the site of the PC on AXR and the frequency with which they would suggest a CT if uncertain. We also assessed the value of a novel single assessment PC screening protocol.
Methods Twelve radiologists (six consultants and six trainees) were asked to locate the PC on 20 AXRs and to comment on the need for CT scanning using the following scoring system: 1-definitely in colon,2-probably in colon but would like a CT, 3-not sure, require CT,4-probably not in colon and would like a CT, 5 definitely not in colon. If the PC was detected by the radiofrequency identification (RFID) hand held scanner the new protocol involved an assessment in the CT scanner using a preliminary scout film avoiding the need for AXR. If the PC was absent, the patient proceeded with the real CE. If however the scout film identified the PC, a limited CT was then performed to identify its exact location.
Results The 20 AXR's consisted of 12 cases where the PC was located in the colon and not reached in seven (six in the small bowel and one in stomach). In the remaining case where the PC was not visible, this was correctly identified by 11/12 radiologists. Of the remaining 228 decisions (19 cases×12 radiologists), only 26% (n=59) were definite (grades 1 and 5) indicating that CT was not necessary but 7% (4/59) of these decisions were incorrect. Grades 2 and 4 (probably know location but would like a CT) comprised 43% of the total but 29% (29/99) of these decisions was incorrect. The remaining 70 decisions (31%) were grade 3 (not sure) indicating a CT was required. 22 patients underwent the new screening protocol; 82% females (n=18%), mean age 45 years (range 18–78 years). In two patients, the PC had been excreted between RFID scanning and the scout film being performed, so no CT was necessary. Of the remaining 20 patients, the device was identified in the stomach in one patient, small bowel in 32% (n=7, two of whom had had a normal small bowel contrast study within the last 12 months) and colon in 54.5% (n=12).
Conclusion AXR cannot be relied on to accurately locate the PC and most radiologists prefer to use CT, which necessitates a further visit to hospital. We recommend a protocol which uses a scout film to perform the same function as an AXR which then guides the use of a limited CT scan at the same examination, to accurately locate the device.