Mo1669 DETECTION OF ≥10MM POLYPS BY COLON CAPSULE ENDOSCOPY VERSUS CT COLONOGRAPHY: A POST-HOC ANALYSIS

Colon capsule endoscopy (CCE) is a new modality that allows examination of the gastrointestinal lumen under physiological conditions, while CT colonography (CTC) is an established screening test. In a previous prospective trial, we demonstrated that the detection rate of large (≥2cm) flat tumors by CCE was superior to CTC. We conducted this post hoc analysis to compare the detection rates for all ≥1cm colorectal polyps by CCE and CTC, using data from the previous prospective trial. Data for patients referred for endoscopic submucosal dissection of colorectal tumors in two tertiary-care hospitals since September 2014, were reviewed. All patients underwent CCE, then colonoscopy, followed by CTC on the same day. Magnesium citrate solution with Gastrografin was used both as a booster for CCE and fecal tagging of CTC. An experienced gastroenterologist (MT) in a third hospital evaluated the CCE studies and recorded the location, size and morphology of all lesions detected, blinded to colonoscopic findings but aware of the presence of, at least, one large (≥2cm) flat tumor. An experienced radiologist (MK) read the CTC studies under the same conditions. Colonoscopic findings were defined in a standard manner. If the lesion was in an adjacent segment within 50% of the reference standard measure, the diagnosis was considered correct. Per-lesion detection rates (sensitivity) are expressed with 95% confidential intervals (95%CI) comparing CCE and CTC, using the McNemar test for paired nominal data. Thirty patients (16 females, 14 males; median age 70 years, range 36-86) were enrolled. The capsule excretion rate within eight hours was 87% (26/30), but all capsules went beyond all lesions ≥1cm. Cleansing levels were excellent in nine, good in 14 and fair in seven. In addition to the 30 lesions rated in the previous trial, 21 more lesions were evaluated, A total of 51 lesions (median size 20mm, interquartile range 10-25; proximal colon 39, distal colon eight, rectum four; sessile serrated adenoma eight, adenoma 26, mucosal cancer 11, invasive cancer 4) were evaluated. Per-lesion detection rates (sensitivity) were 88% (45/51) by CCE and 69% (31/51) by CTC (P=0.0063). Subgroup analyses based on morphology, anatomical location, lesion size and histology, showed detection rates for CCE significantly higher than for CTC for flat morphology (p=0.0063), proximal colon (p=0.011), small size (10≤diameter<20mm) (p=0.031) and adenomatous histology (p=0.031). In a post hoc analysis of all polyp types, the detection rate by CCE was superior to CTC. This trend was observed for flat morphology but not polypoid. This may be attributed to the fact that flat lesions are depicted as polypoid on CCE without air insufflation.