Capsule endoscopy vs. colonoscopy vs. histopathology in colorectal cancer screening: matched analyses of polyp size, morphology, and location estimates

Abstract

Purpose

Colon capsule endoscopy (CCE) is considered a potential alternative to optical colonoscopy (OC) for colorectal cancer screening. However, the accuracy of CCE in polyp size and morphology estimation is unknown.

Methods

A fully paired study was performed (2015–2016), where 255 participants from the Danish national screening program had CCE, OC, and histopathology (HP) of all detected polyps. We developed a new algorithm to match CCE and OC polyps, based on objective measures of similarity between polyps. We performed paired comparisons of size, morphology and location of CCE, and OC- and HP-matched polyps. We used cross-validation to develop a model able to predict HP polyp size, based on CCE.

Results

CCE overestimated size assessed by HP (by 4.3 mm; 95%CI 3.3–5.2 mm) and OC (by 2.7 mm; 95%CI 1.4–3.9 mm). Polyps were more likely to being assessed as “pedunculated” and less likely to being assessed as “flat” in CCE, compared to OC (p < 0.0001). Our model could predict HP polyp size ≥ 6 mm, solely using CCE-assessed size, location, and morphology as model inputs, with a sensitivity = 0.93 (95%CI 0.66–1.00) and specificity = 0.50 (95%CI 0.32–0.68).

Conclusions

If CCE is to be used as a screening test, it is essential: (1) to translate CCE polyp estimations into histopathologic polyp sizes and (2) to consider that, compared to OC, CCE has a higher tendency to assess polyps as pedunculated and a lower tendency to assess them as flat.

Trial registration

Clinicaltrials.gov No. NCT02303756