**11. Cost-effectiveness**

**9. Computer-Assisted Diagnosis (CAD)**

262 Colonoscopy and Colorectal Cancer Screening - Future Directions

trade-offs in terms of sensitivity and specificity [128].

CTC studies, the perforation rate was 0.059% [132].

[122, 123].

**10. Safety**

0.10% of patients [131].

An intensive area of research is the development of computer-assisted diagnosis (CAD) al‐ gorithms. CAD can assist radiologists as a second reader to improve accuracy [117, 118]. It has been shown that CAD can aid trained radiologists in the detection of significant pol‐ yps [119].CAD significantly improved polyp detection by 12% in one study, (from 48 to 60%) with only a moderate increase in interpretation time [120]. Another study demon‐ strated that using CAD in second-read mode increased accuracy in 13 of 19 readers 968%); CAD increased sensitivity of finding polyps but decreased specificity slightly [121]. In general, using CAD increases polyp detection but also increases false positives as well

Using CAD as a primary reader is feasible but early studies showed less sensitivity than hu‐ man readers [124]. The sensitivity of CAD detected polyps 10 mm or greater was 64% (18/28) in one study [125]. In a later study of 1,186 patients undergoing both CTC and OC on the same day, CAD had a sensitivity of 89.3% (25/28; 95% CI: 71.8%-97.7%) for detecting ad‐ enomatous polyps at least 1 cm in size [126]. The false-positive rate was 2.1% (95% CI: 2.0% - 2.2%). CAD detected both of the carcinomas in the study group. In this study, CAD had a per-patient sensitivity comparable to that of OC for adenomas at least 8 mm in size [126]. Another study found a per-patient sensitivity of 96% was for CAD (in patients with a me‐ dian polyp diameter of 6 mm) using external validation [127]. Several CAD polyp detection systems exist such as Polyp Enhanced Viewing (PEV) and the Summers computer-aided de‐ tection (CAD) system (National Institutes of Health (NIH)). These systems vary and have

It is difficult to make a head-to-head comparison of the safety of CTC vs. OC since they are different technologies with varying risks. In one study, CTC screening was performed in 3,120 adults and compared to primary OC screening in 3,163 adults. There were seven co‐ lonic perforations in the OC group and none in the CTC group [45]. Colonic perforation has been reported with CTC but its occurrence is rare [129, 130]. Nine perforations out of 17,067 CTC examinations (0.052%) were reported in one study [131]. In another study of 11, 870

Possible factors that contribute to perforation are presence of an inguinal hernia contain‐ ing colon (n = 4), severe diverticulosis (n = 3), and obstructive carcinoma (n = 1)[132, 133]. In cases of obstructing lesions, gas should be insufflated slowly [133]. Colonic pneumato‐ sis is rarely seen (0.11%) in CTC studies and should not be confused with perforation [134, 135]. Overall, potentially serious adverse events related to CTC occur in less than With a 6-mm size threshold for polyps, the overall referral rate to optical colonoscopy is about 15% [114]. CTC is usually a less expensive test than OC, however the total costs may not be less if one considers all of the variables such as compliance rates and referral rates for OC after detecting lesions. Using a Markov model, screening by CTC costs \$24,586 per lifeyear saved compared to \$20,930 for OC screening [136]. CTC becomes a more cost-effective test as the compliance rate for screening increases or if the cost for CTC is 54% lower than OC. On the other hand in a recent analysis both CTC and OC were more costly and less ef‐ fective than FOBT plus flexible sigmoidoscopy[137].

A Markov model was used to estimate the cost-effectiveness of CTC screening in an Italian population. In this study, colorectal cancer was reduced by 40.9% and 38.2%, with OC and CTC respectively. As compared to no screening, both CTC and OC were shown to be costsaving with CTC being the less expensive option [138]. Since CTC can accurately detect and simultaneously screen for aortic aneurysms, cardiac atherosclerotic risk factors and osteopo‐ rosis, the benefits of CTC screening in an elderly population may be even more cost-effec‐ tive than previously thought [139-141].
