Colorectal cancer is a deadly disease that remains a major public health problem and a leading cause of cancer-related deaths because of poor adherence to existing screening strategies. CT colonography (CTC), also referred to as virtual colonoscopy, has high sensitivity and represents a promising screening tool that has the potential to improve overall screening rates, if properly implemented. After a number of early proof-of-concept studies reported encouraging results for polyp detection with CTC, several subsequent studies in low-prevalence cohorts showed disappointing results. However, with improvements in technique, particularly with the use of three-dimensional (3D) polyp detection and oral contrast tagging, the Department of Defense, multicenter, screening trial showed that CTC was equal to, or slightly better than, optical colonoscopy (OC) for the detection of large, advanced neoplasms (that is, large adenomas and cancers; Figure 1).1 Although the actual need for another multicenter trial to validate CTC as an effective screening could be questioned, the American College of Radiology Imaging Network (ACRIN) launched such a study in 2005, the results of which were published in 2008.2
Figure 1 | Large, tubular adenoma detected by CT colonography screening.
Permission obtained from P. J. Pickhardt.
The ACRIN CTC trial evaluated 2,531 patients from 15 clinical centers. The basic structure of the trial was similar to the Department of Defense trial1 in many respects, with some notable exceptions that are discussed below. In the ACRIN trial, asymptomatic adults who were eligible for routine screening underwent CTC, followed by OC with polypectomy as the reference standard.2 The key outcome measure was the detection of large (
10 mm), advanced neoplasms (adenomas and cancers), which represent the primary target for colorectal cancer screening and prevention. In this study, CTC identified 90% of individuals with large adenomas or cancers.2 By comparison, the sensitivity of CTC and OC for the same outcome measure in the Department of Defense trial was 94% and 88%, respectively.1, 3 The authors of the ACRIN study appropriately concluded that their findings support previously published data on the role of CTC in screening patients with an average risk of colorectal cancer.2
Although the ACRIN trial was clearly an overall success, there are a number of limitations related to methodology that deserve some consideration. The lack of segmental unblinding of CTC results at subsequent OC (that is, revealing the CTC results to the endoscopist only after each colonic segment had been evaluated by OC) precludes diagnostic assessment of OC, which was critical for demonstrating equivalence in the Department of Defense trial.3 In the ACRIN trial, 11 patients with large neoplasms were not identified correctly by CTC, whereas 27 patients had large lesions that were identified by CTC but not by OC.2 Among the 15 patients who returned for a second colonoscopy, 5 of the missed lesions were identified; the remaining 12 patients have not undergone repeat OC.2 This finding suggests that the sensitivity for detecting large neoplasms with CTC and OC were similar in the ACRIN trial, but the use of segmental unblinding would have immediately resolved most discordant cases.3 The use of suboptimal 3D CTC software platforms at most of the ACRIN sites probably resulted in the lower overall diagnostic accuracy for polyp detection compared with the Department of Defense trial, as well as longer 3D reading times and only slightly improved sensitivity for primary 3D polyp detection versus primary two-dimensional (2D) detection.1, 2 When more-robust, 3D, fly-through algorithms are employed, the clear superiority of 3D polyp detection over 2D detection becomes readily apparent.4
Several factors contributed to the low, positive predictive value of CTC findings in the ACRIN trial.2 Firstly, nonadenomatous lesions such as hyperplastic polyps, which account for over 30% of patients harboring large polyps,5 were considered as false positives since the main outcome measure was neoplasia. Secondly, the radiologists were apparently not trained with cases containing oral contrast tagging—yet this technique was employed in the trial—which might have led to preventable false-positive interpretations. Thirdly, the lack of segmental unblinding clearly lead to some OC false negatives being mislabeled as CTC false positives. The rate of extracolonic findings that would apparently require further attention in this trial was more than double that of other CTC screening experiences,6 but the methodology was not well described and previously known conditions were not excluded. When extracolonic findings are handled appropriately, CTC has the potential to be both more cost-effective and clinically effective than OC.7
...CT colonoscopy has the potential to be both more cost-effective and clinically effective than optical colonoscopy
Beyond the ACRIN2 and Department of Defense1 CTC trials, which have provided the necessary validation albeit in the necessarily artificial setting of a clinical trial, objective evidence now exists that CTC can more efficiently identify patients with advanced neoplasia in actual clinical practice than OC. Kim et al.8 compared parallel CTC and OC screening arms in over 6,000 asymptomatic adults and found that CTC screening yielded the same number of advanced neoplasms (including more cancers) but with fewer total polypectomies (561 versus 2,434), fewer invasive procedures (246 versus 3,163), and fewer complications (none versus seven perforations) compared with primary OC screening.
The cumulative findings of these three studies1, 2, 8 demonstrate the high performance of CTC, and led to its inclusion as a recommended screening test in the 2008 revised guidelines of the American Cancer Society.9 Importantly, these innovative guidelines emphasized the importance of cancer prevention with structural tests that detect precursor polyps over the stool-based tests that primarily detect cancer but are not preventive. By adding another structural screening option that is equally effective but less invasive than OC, overall compliance should increase, resulting in the detection and prevention of more cancers. For the benefit of society, it is time for CTC to move beyond the validation phase and into the realm of widespread implementation. In order to truly effect mortality from colorectal cancer, however, the remaining challenges for this promising test are to achieve widespread acceptance, reimbursement, and implementation.
Practice point
The ACRIN trial corroborates the findings of the earlier Department of Defense trial that showed CT colonography to be a safe and valuable screening tool, which effectively ends the validation phase and opens the door for widespread implementation.
