**3. Screening for colorectal cancer: CTC vs. OC**

**2. Technique**

254 Colonoscopy and Colorectal Cancer Screening - Future Directions

looking for polyps.

distention occurs [16].

hand [20].

fects the quality of CTC [23].

with polyps 0.5 cm in diameter or larger [18].

As advancements in scanner technology and three-dimensional (3D) post-processing helped develop this method to mature into a potential option in screening for colorectal cancer, the fundamentals of the examination remained the same. It is a minimally invasive, CT-based procedure that simulates conventional colonoscopy using 2D and 3D computerized recon‐ structions [10]. CTC utilizes computer virtual-reality techniques to navigate inside a threedimensionalz (3D) patient-specific colon model reconstructed from abdominal CT images,

CTC examination starts by inflating a cleansed colon with room air or carbon dioxide (CO2) introduced through rectal catheter [14]. With the patient in a prone position, air or CO2 is insufflated under gentle pressure to ensure adequate distention of the bowel. The insuffla‐ tion of gas is usually associated with a mild degree of patient discomfort or pain [15]. Al‐ though not common, vaso-vagal reactions can occur, especially if with small bowel

Then abdominal CT slice images are taken in seconds (during a single breath hold) with sub millimeter resolution in both axial and transverse directions resulting in excellent contrast between the colon wall and the lumen. The sliced images are stacked together as a volume image, from which the colon model is constructed. Image segmentation is necessary for the construction of an accurate colon model. Computer graphics are heavily involved to navi‐ gate or fly through inside the 3D colon model. The patient is scanned in both a prone and supine view [17]. Using a second view significantly improves the ability to identify patients

CTC can be performed in patients with prior abdominoperineal resection and sigmoid co‐ lostomy, although increased difficulties with CO2 retention and adequate bowel disten‐ tion exist [19]. The prevalence of transient bacterium after CTC is low therefore it follows that patients with at risk cardiac lesions should not require antibiotic prophylaxis before‐

Most commonly used bowel preparations include sodium phosphates, magnesium citrate and polyethylene glycol (PEG) [21]. Typical oral preparations used for bowel cleansing are: 4 L of PEG solution; 90 ml of phosphosoda; or 300 ml of magnesium citrate. Polyp detection is comparable for all three preparations, although phosphosoda has a significantly higher patient compliance and the least residual stool [22]. Residual fluid coverage negatively af‐

The use of fecal tagging agents and intravenous contrast is not standardized. CTC experts were surveyed regarding their practice patterns [24]. Thirty-eight percent performed fecal tagging regularly and 81% [21/26] believed intravenous contrast was not necessary [24].

Non-operator dependent false positives and false negatives occur with CTC. For exam‐ ple, inadequately tagged stool can have the same density as a polyp, however the two can sometimes be distinguished by comparing prone and supine views, since stool is The primary aim of CTC is the detection of colorectal polyps and carcinomas, however; the precise role of CTC in screening asymptomatic patients is controversial [28]. Studies using patients with known adenomas generally report higher accuracy, while studies employing asymptomatic screening subjects report lower accuracy [28]. Two key areas have held back the widespread application of CTC as a screening test. These key areas are: [1] the variable sensitivity of CTC reported in mass screening programs (see Table) and [2], the expertise re‐ quired to interpret the examination. These two areas are related [29]. Despite these draw‐ backs, the American College of Radiology, has endorsed the use of CTC as a screening tool for colo-rectal cancer stating that the sensitivity and specificity of CTC are high enough and comparable to those of OC [30]. In addition, CTC has received the endorsement of a multisociety task force that included the American Cancer Society and U. S. Multi-society Task Force on Colorectal Cancer [31].

Studies reveal a wide variation in performance measures (sensitivity and specificity) regard‐ ing polyp detection rates, especially for smaller polyps [10]. In an early feasibility study of 44 patients, CTC demonstrated reasonable sensitivity (83%) and specificity (100%) for pol‐ yps larger than 8 mm in size [32]. A second early study performed in 87 patients at high risk for colorectal neoplasia identified 49 patients with a total of 115 polyps and 3 carcinomas [33]. CTC identified all 3 cancers. The sensitivity was 91% for polyps that were 10 mm or more in diameter, 82% [33/40] that were 6 to 9 mm, and 55% [29/53] that were 5 mm or smaller [33]. There were 19 false positive findings of polyps and no false positive findings of cancer. In a larger study of 300 patients CTC demonstrated a sensitivity equal to 90% for polyps 10 mm or larger and 80.1% for polyps at least 5 mm in size [34]. The overall specifici‐ ty for this study was 72.0% [34]. All 8 carcinomas in the study were detected by CTC.

Two later studies assessing the accuracy of CTC had varying results. Pickhardt et al. evalu‐ ated 1,233 asymptomatic patients with CTC and same-day OC [35]. The sensitivity of CTC for adenomatous polyps at least 10 mm in size was 93.8% and 88.7% for polyps at least 6 mm in size, which was comparable to OC. The specificity of CTC for adenomatous polyps at least 10 mm in size was 96.0% and 79.6% for polyps at least 6 mm in size. These encouraging results were followed a year later by less optimistic findings from a study by Cotton et al., that analyzed 600 participants undergoing both CTC and OC [36]. In the Cotton study, 104 of the participants (17.3%) had lesions sized at least 6 mm. The sensitivity of CTC for detect‐ ing 1 or more lesions sized at least 6 mm was only 39.0% and for lesions sized at least 10 mm, it was 55.0% (95% Cl, 39.9% - 70.0%) [36]. The specificity of CTC for detecting partici‐ pants without any lesion sized at least 6 mm was 90.5% and without lesions sized at least 10 mm, 96.0% (95% Cl, 94.3% - 97.6%). CTC missed 2 of 8 cancers [36]. Lack of adequate radiol‐ ogist training to read CTCs may have resulted in the low accuracy found in this study.

In a subsequent study of 2,531 asymptomatic patients, radiologists trained in CTC reported the accuracy of finding histologically confirmed adenomas [37]. The sensitivity for large ad‐ enomas [10 mm or larger) and medium-sized adenomas (6 – 9 mm) was 90% and 78% re‐ spectively [37]. CTC failed to detect a lesion measuring 10 mm or more in diameter in 10% of patients [37]. Pickhart et al. found that the positive predictive values (PPV) for polyps with threshold sizes 6 mm, 8 mm, and 10 mm are: 92.3%, 93.0%, and 93.1% respectively [38]. Others have also found that for significant adenomas, the PPV of CTC is high and ranges from 96 – 99% [37, 39].

ample, in one study flat polyps were defined as those having a height less than one-half of

Virtual Colonoscopy

257

http://dx.doi.org/10.5772/52544

The sensitivity for flat polyps appears to be lower than non-flat polyps, however, Pickhardt et al. found that the sensitivity of CTC for detecting flat adenomas measuring 6 mm or greater was similar to that of non-flat polypoid lesions [50]. Others have found lower sensi‐ tivities for flat polyps ranging from 15% to 65% [52]. Regardless of reader skills, truly flat or depressed adenomas will most likely pose a challenge for CTC. Also, flat carpet lesions of

CTC readers may not report polyps less than 5 mm in size [54]. The justification for this is that small polyps are usually benign and rarely harbor cancer or have much prognostic sig‐ nificance. In a large OC screening study, advanced histology was found in only 1.7% of pol‐ yps sized 5 mm or less indicating the lack of reporting of these small polyps by CTC may be justified [55]. On the other hand, 6.6% of polyps sized 6 to 9 mm, had an advanced histology implying polyps of this size if found on CTC should be followed up with OC in the near future [55]. This last point is somewhat contentious and some radiologic guidelines suggest surveillance with CTC after a shortened interval as an acceptable option for polyps 6 – 9 mm

We feel a reasonable algorithm for CTC screening might be: 1. follow-up screening in 5 years if no polyps are found: 2. Follow-up CTC or OC in 5 years for polyps smaller than 5 mm: 3. OC if polyps measuring 6 mm or greater are found [57]. This algorithm is consistent with clinical sentiment since 71% of primary care physicians and 86% of gastroenterologists would send patients with polyps 5 mm in size or greater for a follow-up OC [58]. If this ap‐ proach were adopted, a referral rate for OC of about 8% - 14% would be expected in the gen‐ eral population [45, 59]. Using a 6 mm threshold however may be very costly. A decisionanalysis estimated that to prevent one colon cancer death would require over 9,000 OCs,

resulting in 10 additional perforations at an incremental cost of \$327,853 dollars [60].

Sometimes OC can not be completed to the cecum due to technical factors such as prior ab‐ dominal surgery, colon length and number of flexures [61]. An important use of CTC is ex‐ amination of the colon after incomplete OC [62]. In a retrospective study, 88/546 patients had lesions 6 mm or greater on CTC after incomplete OC. OC was repeated if findings on CTC were significant. The PPV of CTC for masses, large polyps, and medium polyps were

It may be valuable to perform a low-dose diagnostic CT before rectal tube insertion in pa‐ tients referred for incomplete colonoscopy. In one study of 262 patients referred for incom‐ plete OC, colon perforation was found on the low-dose CT scans of two of the 262 patients (0.8%; 95% CI, 0.1-2.7%) [64]. One of these patients had no symptoms; the other had mild

their width [50]. In other studies, a definition of 3 mm or less in height is used [51].

the colon can be difficult to detect by CTC [53].

in size [56].

**4. Other uses of CTC**

**4.1. After incomplete colonoscopy**

90.9% and 91.7%, and 64.7% respectively [63].

Meta-analysis is a tool that attempts to summarize varying results across multiple studies. Meta-analysis of data suggests CTC has excellent per-patient average sensitivity and aver‐ age specificity for detection of adenomatous polyps and cancer [40]. In one meta-analysis, 2,610 patients were included for study [41]. Large polyps (10 mm or greater) had a per-pa‐ tient average sensitivity of 93% (95% CI,73% - 98%) and specificity of 97% 9(95% CI, 95% - 99%) [41]. The sensitivity and specificity decreased to 86% (95% CI: 75% - 93%) and 86% (95% CI, 76% - 93%), respectively, when the threshold was lowered to include medium sized polyps (6 mm to 9 mm). These findings are similar to another more recent meta-analysis us‐ ing average risk patients that found a sensitivity of 87% and specificity of 97.6% for polyps at least 10 mm in size [42].

One of the problems of evaluating the test performance of CTC is the use of OC as the gold standard because OC has a miss-rate for polyps and cancer as well. In one study, Pickhardt et al. compared 1,233 asymptomatic adults who underwent same-day CTC and blinded seg‐ mental OC [43]. Polyps that were detected by CTC but initially missed by OC were consid‐ ered missed polyps for OC. It was found that OC had a miss rate of 12% for adenomas 10 mm or greater. Of the missed polyps on OC, 14/15 (93.3%) non-rectal neoplasms were locat‐ ed on a fold. Five of 6 (83.3%) missed rectal lesions were located within 10 cm of the anal verge [43].

Another study analyzed 286 tandem colonoscopies [44]. The OC miss rates for adenomas 5 mm and larger and advanced adenomas (≥ 10 mm or high grade dysplasia) were: 11% and 9% respectively [44]. Therefore, OC does have a significant miss rate for adenomas 5 mm and larger and/or advanced adenomas. In fact, the OC miss-rate is similar to the CTC missrate for polyps 6-9 mm in size [27]. It should also be pointed out that in screening studies, CTC and OC have similar detection rates for advanced neoplastic polyps and cancer, 3.2% vs. 3.4% respectively [45]. OC detects significantly more adenomas less than 5 mm of size although the benefit of this remains to be seen [46]. In summary, CTC appears to have simi‐ lar sensitivity to OC in detecting polyps 5 mm or greater when performed by readers with high experience [47].

The detection of flat adenomas are a major concern for colo-rectal cancer screening since these polyps are at a higher risk of harboring advanced pathology and are more difficult to detect by CTC as well as OC [48, 49]. In the general population, there is wide variation in the reported incidence of flat lesions, which may in part be due to the lack of a uniform defini‐ tion of flat polyps. Various definitions of flat polyps have been used in CTC studies. For ex‐ ample, in one study flat polyps were defined as those having a height less than one-half of their width [50]. In other studies, a definition of 3 mm or less in height is used [51].

The sensitivity for flat polyps appears to be lower than non-flat polyps, however, Pickhardt et al. found that the sensitivity of CTC for detecting flat adenomas measuring 6 mm or greater was similar to that of non-flat polypoid lesions [50]. Others have found lower sensi‐ tivities for flat polyps ranging from 15% to 65% [52]. Regardless of reader skills, truly flat or depressed adenomas will most likely pose a challenge for CTC. Also, flat carpet lesions of the colon can be difficult to detect by CTC [53].

CTC readers may not report polyps less than 5 mm in size [54]. The justification for this is that small polyps are usually benign and rarely harbor cancer or have much prognostic sig‐ nificance. In a large OC screening study, advanced histology was found in only 1.7% of pol‐ yps sized 5 mm or less indicating the lack of reporting of these small polyps by CTC may be justified [55]. On the other hand, 6.6% of polyps sized 6 to 9 mm, had an advanced histology implying polyps of this size if found on CTC should be followed up with OC in the near future [55]. This last point is somewhat contentious and some radiologic guidelines suggest surveillance with CTC after a shortened interval as an acceptable option for polyps 6 – 9 mm in size [56].

We feel a reasonable algorithm for CTC screening might be: 1. follow-up screening in 5 years if no polyps are found: 2. Follow-up CTC or OC in 5 years for polyps smaller than 5 mm: 3. OC if polyps measuring 6 mm or greater are found [57]. This algorithm is consistent with clinical sentiment since 71% of primary care physicians and 86% of gastroenterologists would send patients with polyps 5 mm in size or greater for a follow-up OC [58]. If this ap‐ proach were adopted, a referral rate for OC of about 8% - 14% would be expected in the gen‐ eral population [45, 59]. Using a 6 mm threshold however may be very costly. A decisionanalysis estimated that to prevent one colon cancer death would require over 9,000 OCs, resulting in 10 additional perforations at an incremental cost of \$327,853 dollars [60].
