CT Colonography versus Colonoscopy for the Detection of Advanced Neoplasia
 

David H. Kim, M.D., University of Wisconsin Medical School,

NEJM 2007;357:1403

Advanced neoplasia of the large intestine consists of both adenocarcinomas and a subgroup of benign neoplasms referred to as advanced adenomas. The advanced adenoma represents the optimal target lesion for strategies to prevent colorectal cancer. This benign lesion is considered to be associated with a relatively high risk of progression to cancer. The advanced adenoma is specifically defined as an adenoma that meets one or more of the following criteria: a size of at least 10 mm, the presence of a substantial villous component, and the presence of high-grade dysplasia. Removal of detected advanced adenomas effectively disrupts the potential pathway to the development of cancer that is believed to be responsible for the majority of colorectal carcinomas.

Most subcentimeter polyps are not adenomatous, and only a small fraction of all adenomas are advanced, suggesting a need for more selective alternatives to the practice of universal polypectomy. The purpose of this study was to compare computed tomographic colonography (CTC) and optical colonoscopy (OC) when applied to the same general screening population. Important outcome measures included detection rates for advanced adenomas and adenocarcinomas for various categories of polyp size and overall polypectomy rates. These observations provided an assessment of CTC as a selective filter for therapeutic OC in the detection of advanced neoplasia.

Methods: We compared primary CTC screening in 3120 consecutive adults (mean [±SD] age, 57.0±7.2 years) with primary OC screening in 3163 consecutive adults (mean age, 58.1±7.8 years). The main outcome measures included the detection of advanced neoplasia (advanced adenomas and carcinomas) and the total number of harvested polyps. Referral for polypectomy during OC was offered for all CTC-detected polyps of at least 6 mm in size. Patients with one or two small polyps (6 to 9 mm) also were offered the option of CTC surveillance. During primary OC, nearly all detected polyps were removed, regardless of size, according to established practice guidelines.

Results: During CTC and OC screening, 123 and 121 advanced neoplasms were found, including 14 and 4 invasive cancers, respectively. The referral rate for OC in the primary CTC screening group was 7.9% (246 of 3120 patients). Advanced neoplasia was confirmed in 100 of the 3120 patients in the CTC group (3.2%) and in 107 of the 3163 patients in the OC group (3.4%), not including 158 patients with 193 unresected CTC-detected polyps of 6 to 9 mm who were undergoing surveillance. The total numbers of polyps removed in the CTC and OC groups were 561 and 2434, respectively. There were seven colonic perforations in the OC group and none in the CTC group.

Conclusions: Primary CTC and OC screening strategies resulted in similar detection rates for advanced neoplasia, although the numbers of polypectomies and complications were considerably smaller in the CTC group. These findings support the use of CTC as a primary screening test before therapeutic OC.

Discussion

Colorectal cancer is a major cause of cancer-related mortality in the United States, accounting for approximately 55,000 deaths per year. However, because this cancer has an identifiable precursor lesion, there is a genuine opportunity for prevention rather than cancer detection alone. In particular, targeted detection and removal of advanced adenomas may be the most effective approach to cancer prevention.OC is an effective screening tool for the detection and removal of advanced colorectal neoplasia and is widely regarded as part of the preferred screening strategy.Our results suggest that primary CTC with selective OC also deserves consideration as a preferred screening strategy because it appears to achieve the same goals of detection and prevention but with the use of substantially fewer resources in terms of OC procedures and polypectomies. Thus, CTC may provide a more targeted screening approach for detection of advanced neoplasia.

In our study, the coexistence of parallel CTC and OC screening programs at a single institution allowed for substantive comparison of diagnostic yields and the use of resources. We observed similar detection rates for advanced adenomas during both CTC and OC screenings. The diagnostic yield for advanced neoplasia was similar in the two groups, despite the fact that small lesions (5 mm) were not reported during CTC. In addition, a subgroup of patients with unresected polyps of 6 to 9 mm were undergoing CTC surveillance, and the frequency of a family history of colorectal cancer was higher in the OC screening cohort. The different handling of diminutive lesions largely accounts for the discrepancies in the overall rates of positive test results (12.9% in the CTC group vs. 37.6% in the OC group) and in the numbers of polypectomies (561 vs. 2434).

Overall, 2006 polypectomies were performed to remove diminutive polyps, which yielded four advanced lesions (0.2%). Such observations reinforce the scarcity of diminutive and small advanced neoplastic lesions and the potential benefits of filtering strategies during CTC. In fact, Markov modeling of large cohorts has shown that the strategy of not reporting diminutive polyps detected during CTC screening is a cost-effective approach that can substantially reduce the rate of polypectomy and complications without any sacrifice with respect to cancer prevention.

Beyond these differences, however, there were also some striking similarities between the two screening strategies. For example, the rates of positive test results at the thresholds of 6 mm and 10 mm were similar, and the characteristics of the advanced adenomas were also quite similar.

Polyps of at least 10 mm appear to represent a very useful surrogate for advanced adenomas, accounting for the great majority of all advanced lesions in our study. Large polyp size has already been singled out by some observers as the most important criterion for advanced neoplasia. Only 20 subcentimeter polyps in our study were histologically advanced, which corresponded to an overall prevalence of 0.2% (15 of 6283 patients). Only four advanced adenomas were identified in the diminutive category. Furthermore, only 3 patients had four subcentimeter polyps with high-grade dysplasia (0.05%), and there were no subcentimeter cancers in more than 6000 patients. These observations suggest that a 10-mm threshold for polypectomy at asymptomatic screening would probably capture the vast majority of clinically relevant lesions.

The overall prevalence of advanced neoplasia in this healthy screening cohort of 6283 adults was 3.3%, which is somewhat lower than the prevalence of 4 to 6% reported in several other studies and substantially lower than the prevalence of 10.5% in a population of male veterans.These differences are probably multifactorial, but variations in age, sex, ethnic background, family history, and frequency of symptoms may all play a role. A recent colonoscopy study showed that the prevalence of advanced neoplasia was less than 3% in certain low-risk cohorts. In our study, the frequency of advanced histologic findings among subcentimeter lesions, particularly high-grade dysplasia and invasive carcinoma, was also generally lower than previously reported. However, the inclusion of 10-mm lesions in the category of small polyps in some previous studies substantially increased the reported prevalence.Furthermore, a recent colonoscopy series evaluating a large screening population reported low cancer rates among resected adenomas measuring 6 to 9 mm (0.07%) and 1 to 2 cm (2.4%). In our study, the exclusion of a subgroup of patients who had small, unresected polyps and were undergoing CTC surveillance probably had only a small effect on the prevalence of advanced adenomas.

The clinical management of small polyps of 6 to 9 mm that are detected during CTC is controversial. One approach is to offer OC for polypectomy to all patients with CTC-detected polyps of at least 6 mm. However, an option of short-term CTC surveillance for patients with one or two small CTC-detected polyps has also been suggested. Short-term CTC surveillance for small polyps allows for more efficient detection and removal of the uncommon advanced neoplasms because only the enlarging lesions are removed. As discussed previously, potential benefits include the decreased use of resources, procedural risks, and cost. Potential drawbacks mainly involve the possibility of following a polyp that harbors a focus of cancer or transforms to cancer during the surveillance period, resulting in a lost opportunity for cancer prevention. The presumed low risk for this subgroup of polyps is echoed by the low prevalence of subcentimeter lesions harboring high-grade dysplasia or invasive carcinoma in the population we studied. In addition, the limited natural-history data from several older longitudinal studies that monitored lesions with the use of barium enema and endoscopic examination support the practice of short-term CTC follow-up. Ultimately, more investigation will be needed to determine which strategy is more beneficial during CTC. Such a surveillance strategy for small polyps that are detected during primary OC would clearly be less appealing because the scope is already in place and the only incremental costs and risks that are incurred are related to the polypectomy itself.

Adverse events were uncommon during OC screening, and no serious complications were reported in the CTC group. The perforation rate of 0.2% (7 of 3163 patients) in the OC group was within the expected range reported in previous colonoscopy series. The absence of perforations in the CTC screening group was largely due to both the minimally invasive nature of CTC and the decreased numbers of OC studies and polypectomies, as compared with the primary OC group. Concern has been raised regarding the potential risks associated with radiation exposure from CTC. Some observers contend that the risk is too small to quantify.Proponents of the linear, no-threshold model argue that a small risk exists, but even members of this group agree that the benefits of screening for colorectal cancer appear to outweigh these small theoretical risks.

A major limitation of our study was the lack of randomization. Thus, a potential exists for selection bias affecting the composition of the study population for each program, leading to different prevalences of advanced adenomas. Although most patients in both cohorts were being screened for the first time, it is possible that some of them had undergone previous colorectal screening elsewhere. However, the groups were similar in several important respects, including a relatively young age and a predominance of women. Age and sex have been shown to be strong predictive factors for the prevalence of adenomas and high-grade dysplasias. The percentage of patients with a positive family history was higher in the OC group, which should have resulted in more advanced adenomas in that group. The fact that similar numbers of advanced adenomas were seen in the two groups further reinforces the potential of CTC for screening.

In conclusion, CTC and OC screening methods resulted in similar detection rates for advanced neoplasias within the same general population. This finding is important because advanced neoplasms represent the primary target of colorectal screening and cancer prevention. The marked decrease in the use of OC and total rates of polypectomies in the CTC group suggests that this technique is a safe, clinically effective, and cost-effective filter for therapeutic OC. Furthermore, by combining primary CTC and primary OC screening efforts, with the choice between tests driven by patient preference, the overall screening compliance for total colonic examination could substantially increase.