Description of a novel grading system to assess the quality of bowel preparation in video capsule endoscopy

Background and study aims: Inadequate bowel preparation negatively influences the reliability of examinations by video capsule endoscopy (VCE). Currently, only subjective scales are available to describe mucosal visibility. We aimed to design a score that was derived directly from the VCE images. Patients and methods: A computed assessment of cleansing score was developed based on color intensities of the tissue color bar. The feasibility of this score was retrospectively tested in 24 VCE studies. A prospective study was conducted using 40 VCE segments from 10 consecutive VCE studies. The computed scores were compared with three existing methods of assessing small-intestinal cleansing. Correlations between the existing scoring systems and the computed score were evaluated using the intra-class correlation coefficient and Spearman's rho correlation. Results: All computed measurements were obtained twice and resulted in exactly the same results. Both overall and segmental mucosal visibility could be assessed. The computed score and the 10-point quantitative index were significantly associated for both readers (Spearman's rho: 0.68 and 0.75, respectively; P < 0.001). The intra-class correlation coefficient for the 4-point qualitative evaluation and the computed score was 0.67 for reader 1 and 0.64 for reader 2. For reader 1, the mean computed score for segments assessed as either inadequately or adequately cleansed was 5.0 and 6.4 (P = 0.001). For reader 2 these values were 4.0 and 6.3, respectively (P = 0.005). Conclusions: A computed assessment of small-bowel mucosal visibility based on the ratio of color intensities of the red and green channel of the tissue color bar is feasible and more reproducible than existing subjective scales. Such a computed scale could be integrated into VCE reading software. For this novel scoring system we propose the term Computed Assessment of Cleansing (CAC) score.