CT colonography of colorectal polyps: a metaanalysis

For proper evaluation of the accuracy of CT colonography, prospective multiinstitutional trials would be ideal. Until these trials are available, data can be collectively analyzed. The purpose of this study is to use metaanalysis to assess the reported accuracy of CT colonography compared with conventional colonoscopy for detecting colorectal polyps.

Articles comparing CT colonography and conventional colonoscopy were identified, and a standardized form was used to extract relevant study data. Fisher's exact test and the Mantel-Haenszel test were used for pooling of data. A 95% confidence interval (CI) was selected to determine sensitivity and specificity, and the Kruskal-Wallis exact test was used to identify trends relating to polyp size. Meta-analysis methods were used to test strength of results. Comparisons were made for the percentage of polyps detected grouped by size (> or = 10 mm, 6-9 mm, < or = 5 mm) and the percentage of patients identified who had polyps of the same size.

Fourteen studies fulfilled all the study inclusion criteria and gave a total of 1,324 patients and 1,411 polyps. The pooled per-patient sensitivity for polyps 10 mm or larger was (sensitivity [95% CI]) 0.88 (0.84-0.93), for polyps 6-9 mm it was 0.84 (0.80-0.89), and for polyps 5 mm or smaller it was 0.65 (0.57-0.73). The pooled per-polyp sensitivity for polyps 10 mm or larger was 0.81 (0.76-0.85), for polyps 6-9 mm it was 0.62 (0.58-0.67), and for polyps 5 mm or smaller it was 0.43 (0.39-0.47). Sensitivity for detection of polyps increased as the polyp size increased (p < 0.00005). The pooled overall specificity for detection of polyps larger than 10 mm was 0.95 (0.94-0.97).

The specificity and sensitivity of CT colonography are high for polyps larger than 10 mm.