AUDIT OF COLONOSCOPIES CARRIED OUT FOR THE INDICATION OF ABNORMAL RADIOLOGICAL IMAGING OVER A 4 YEAR PERIOD IN A LARGE TERTIARY REFERRAL CENTRE

Aims A frequent indication for colonoscopy referral is abnormal findings on cross-sectional imaging or PET. Colonic abnormalities may also be incidentally detected in patients undergoing imaging for other reasons. Our aim was to determine the yield of endoscopy in the investigation of intestinal abnormalities detected on cross-sectional imaging.

Methods Colonoscopy reports from the 1^st of July 2015 to the 1^st of July 2019 which had ‘abnormal imaging’ listed as the indication were retrieved from the St James Hospital endoscopy database. 158 patients were identified from this initial search. Duplicates, patients with external imaging, and patients with incomplete colonoscopies were excluded resulting in a final study cohort of 134 patients. Comparison between sensitivity of different cross-sectional imaging modalities was made.

Results 39 patients underwent PET-CT and the remainder underwent either CT or MRI. 57% (n = 76) of patients had an endoscopic abnormality correlate with abnormal imaging findings. Comparing imaging techniques, 74% (29/39) of PET-CT versus 48% (47/95) CT/MRI patients had an endoscopic-radiological correlate p = 0.01. 22 colorectal cancers were diagnosed (16%); 4 (18%) described as “colonic thickening’ with the other 18 (82%) described as a ‘mass/tumour/FDG-avid abnormality with lymph nodes or concerning features’. Where imaging demonstrated ‘colitis/fat stranding (n = 19)’, 84% (n = 16) of patients had no correlating findings at colonoscopy. No patients with this CT finding had a colorectal cancer.

Conclusions Radiological abnormalities are moderately correlated with endoscopic findings. Endoscopy should generally be performed to investigate radiological abnormalities. There is a significantly higher yield for endoscopy performed to investigate PET-CT abnormalities compared abnormalities detected on standard cross-sectional imaging. Colitis on cross sectional imaging is poorly correlated with endoscopic findings.