Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, cont

Transcription

1 Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at Hye Jin Kim, MD Seong Ho Park, MD Perry J. Pickhardt, MD Sang Nam Yoon, MD Seung Soo Lee, MD Judy Yee, MD David H. Kim, MD Ah Young Kim, MD Jin Cheon Kim, MD Chang Sik Yu, MD Hyun Kwon Ha, MD CT Colonography for Combined Colonic and Extracolonic Surveillance after Curative Resection of Colorectal Cancer 1 Purpose: Materials and Methods: To determine the accuracy of contrast material enhanced computed tomographic (CT) colonography for postoperative surveillance in colorectal cancer patients without clinical or laboratory evidence of disease recurrence. The Institutional Review Board approved this HIPAAcompliant study and waived informed consent. Between January 2006 and December 2007, 742 consecutive patients without clinical or laboratory evidence of recurrence following curative-intent colorectal cancer surgery underwent contrastenhanced CT colonography. Of these, 548 patients who had subsequent colonoscopy and pathologic confirmation of colonic lesions (reference standard) were included in the colonic analysis. All 742 patients were included in the extracolonic analysis. Sensitivity and specificity of CT colonography for nonanastomotic colonic lesions at least 6 mm in size and anastomotic lesions of any size, including performance according to lesion histologic type, were determined. Diagnostic yields of contrast-enhanced CT colonography for colonic cancers and for extracolonic recurrences were obtained. ORIGINAL RESEARCH n GASTROINTESTINAL IMAGING 1 From the Department of Radiology and Research Institute of Radiology (H.J.K., S.H.P., S.S.L., A.Y.K., H.K.H.) and Department of Surgery (S.N.Y., J.C.K., C.S.Y.), University of Ulsan College of Medicine, Asan Medical Center, Asanbyeongwon-gil 86, Songpa-Gu, Seoul , Republic of Korea; Department of Radiology, University of Wisconsin Medical School, Madison, Wis (P.J.P., D.H.K.); and Department of Radiology, University of California San Francisco, San Francisco Veterans Administration Medical Center, San Francisco, Calif (J.Y.). Received February 17, 2010; revision requested April 1; revision received May 29; accepted June 11; fi nal version accepted June 18. Address correspondence to S.H.P. ( seonghopark@paran.com ). q RSNA, 2010 Results: Conclusion: CT colonography depicted all six metachronous cancers and one anastomotic recurrence within the colon in six patients (0.8%; 95% confidence interval [CI]: 0.3%, 1.8%]), for per-patient and per-lesion sensitivities of 100% (95% CIs: 64.3%, 100% and 67.8%, 100%, respectively). All cancer lesions within the colon were amenable to additional curative treatment. CT colonography per-patient and perlesion sensitivity was 81.8% (95% CI: 60.9%, 93.3%) and 80.8% (95% CI: 64.3%, 97.2%), respectively, for advanced neoplasia and 80.0% (95% CI: 68.6%, 88.1%) and 78.5% (95% CI: 68.3%, 88.7%), respectively, for all adenomatous lesions. Negative predictive values for adenocarcinoma, advanced neoplasia, and all adenomatous lesions were 100%, 99.1%, and 97.0%, respectively. CT colonography specificity was 93.1% (95% CI: 90.4%, 95.2%). Contrast-enhanced CT colonography enabled detection of extracolonic recurrences in an additional 11 patients (1.5%; 95% CI: 0.8%, 2.7%). Contrast-enhanced CT colonography is an accurate and practical surveillance tool following colorectal cancer surgery in patients without clinical or laboratory evidence of recurrence, allowing for simultaneous less-invasive evaluation of both colon and extracolonic organs. q RSNA, 2010 Supplemental material: /suppl/doi: /radiol /-/dc1 Radiology: Volume 257: Number 3 December 2010 n radiology.rsna.org 697

2 Surveillance for recurrent and metachronous colorectal cancers after surgery with curative intent is important because early detection and treatment can increase patient survival ( 1 ). Both colonic and extracolonic surveillance are needed because more than half of recurrent tumors will manifest as distant extracolonic metastatic disease, and many local recurrences lack an intraluminal colonic component ( 2 4 ). As a result, current surveillance guidelines generally adopt a combination of clinical assessment, serum carcinoembryonic antigen testing, colonoscopy, and computed tomography (CT) ( 5 7 ). However, intensive colonoscopic surveillance may be somewhat superfluous since the diagnostic yield is fairly low owing to low incidence of anastomotic recurrence or metachronous cancers (according to one review [8], one metachronous cancer per 157 surveillance colonoscopies) ( 3 ). CT colonography has been shown to have high diagnostic accuracy for colorectal neoplasia in screening patients ( 9 13 ). CT colonography performed with intravenous contrast material inherently serves as both contrast material enhanced diagnostic CT and colonographic evaluations, allowing for simultaneous examination of the extracolonic abdominal organs and the large intestine. Considering that contrast-enhanced Advances in Knowledge n CT colonography is an accurate tool for colonic surveillance following colorectal cancer surgery, demonstrating per-patient sensitivity of 100% (95% CI, 64.3%, 100%) for metachronous or recurrent cancers in the colon and 81.8% (95% CI: 60.9%, 93.3%) for advanced neoplasia. n Negative results at CT colonogra- phy may preclude the need for colonoscopic surveillance in the similar time frame owing to the high negative predictive values of CT colonography for metachronous or recurrent cancers in the colon (100%) and advanced neoplasia (99.1%). abdominal CT is already an accepted postoperative surveillance examination ( 1,5 7 ) and that CT colonography has been validated for detecting colorectal neoplasia in screening patients ( 9 13 ), contrast-enhanced CT colonography could potentially be an attractive standalone examination for combined colonic and extracolonic postoperative surveillance of patients with colorectal cancer ( 14 ). At present, there are limited data regarding the use of CT colonography as a surveillance tool after colorectal cancer surgery ( ). The purpose of our study was to determine the accuracy of contrast-enhanced CT colonography for postoperative surveillance in patients with colorectal cancer who do not have clinical or laboratory evidence of recurrence following curative surgery. Materials and Methods The Institutional Review Board of Asan Medical Center approved this Health Insurance Portability and Accountability Act compliant retrospective study and waived the requirement of informed consent. S.H.P. is a consultant for Infinitt (Seoul, Korea), whose commercial CT colonography system was used in our study. The other authors who were not employees of or consultants for Infinitt had control of inclusion of any data and information that might present a conflict of interest. P.J.P. and D.H.K. are consultants for Medicsight (London, England) and Viatronix (Stony Brook, NY) and are cofounders of VirtuoCTC (Madison, Wis). Patients In January 2006, a clinical protocol change was made in postoperative surveillance in patients with colorectal cancer at Asan Medical Center from routine Implication for Patient Care n Contrast-enhanced CT colonogra- phy could be used as a tool for simultaneous colonic and extracolonic surveillance after colorectal cancer surgery. contrast-enhanced abdominal CT to contrast-enhanced CT colonography, which was related to the potential added benefits of CT colonography. Our routine abdominal CT surveillance following curative colorectal cancer surgery typically consisted of examinations at 6 months after surgery and every year thereafter for 5 years. Contrast-enhanced CT colonography was ordered instead of contrast-enhanced abdominal CT if the following criteria were met: no increased clinical suspicion for cancer recurrence (eg, elevated serum carcinoembryonic antigen or suggestive symptoms and signs) that would generally initiate colonoscopy, thus rendering CT colonography redundant, and no history of prior surgical procedures that would preclude standard CT colonography (eg, total proctocolectomy, abdominoperineal resection, or Hartmann procedure). Between January 2006 and December 2007, 742 consecutive patients (mean age, 58.9 years [standard deviation]; range, years), consisting of 419 men (mean age, 59.1 years ; range, years) and 323 women (mean age, 58.8 years ; range, years), underwent contrastenhanced CT colonography ( Fig 1 ). All patients had one CT colonography examination for the study at the time of their first CT surveillance during the study period. The time interval from surgery Published online before print /radiol Radiology 2010; 257: Abbreviations: CI = confi dence interval NPV = negative predictive value PPV = positive predictive value Author contributions: Guarantors of integrity of entire study, H.J.K., S.H.P., J.C.K.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of fi nal version of submitted manuscript, all authors; literature research, H.J.K., S.H.P., P.J.P., J.Y.; clinical studies, H.J.K., S.H.P., S.N.Y., S.S.L., J.C.K., C.S.Y.; statistical analysis, H.J.K., S.H.P.; and manuscript editing, H.J.K., S.H.P., P.J.P., S.N.Y., J.Y., D.H.K., A.Y.K., J.C.K., H.K.H. See Materials and Methods for pertinent disclosures. 698 radiology.rsna.org n Radiology: Volume 257: Number 3 December 2010

3 Figure 1 Figure 1: Flow diagram of study patients and results at CT colonography (CTC). Positive colonic fi ndings at CT colonography were defi ned as a lesion 6 mm or greater in the nonanastomotic colon or a lesion of any size at the anastomosis. = fi ndings with respect to colonic status, = results for all histopathologic types of lesions. to contrast-enhanced CT colonography ranged from 6 to 50 months (median, 24 months). Patient characteristics, including initial stage, preoperative chemoand radiation therapy in the cases of rectal cancer, surgical procedure, and serum carcinoembryonic antigen level prior to CT colonography, were recorded. Complete perioperative colonoscopy with clearance of synchronous colorectal lesions had been performed in 711 of the 742 study patients. In the remaining 31 patients, who had had an obstructing cancer, perioperative complete colonoscopic evaluation of the proximal colon had not been achieved. Therefore, although large synchronous cancers in the proximal colon were excluded in these patients by preoperative abdominal CT and surgical exploration, small synchronous lesions in the proximal colon (if they existed) would have been left unremoved. For analysis of colonic findings, we included 548 patients who underwent conventional colonoscopy after CT colonography and pathologic confirmation of colonic lesions (if the lesions existed at colonoscopy) to serve as the reference standard (see Reference Standard, below, for the indications for colonoscopy after CT colonography). All patients were included in the extracolonic analysis ( Fig 1 ). CT Colonography Bowel cleansing, fecal tagging, and colon distention were achieved by using standard methods described in another study ( 19 ) (see Appendix E1 [online]). Both prone and supine scans, one before intravenous contrast material administration and another after administration during the portal phase, were obtained by using a 16-detector CT scanner (Somatom Sensation 16; Siemens Medical Solutions, Erlangen, Germany) (see Appendix E1 [online]). Images were reconstructed with a section thickness of 1 mm at 0.7-mm intervals for colonic evaluation and contiguous 5-mm sections for extracolonic review. The original clinical interpretations of the CT colonography examinations were used for our study. Images were interpreted by one of two radiologists (S.H.P. and S.S.L., with experience reading several hundred CT colonography cases at the beginning of the study). Colonic interpretation was performed by using primary three-dimensional bidirectional endoluminal fly through with two-dimensional problem solving at a CT colonography workstation (Lucion; Infinitt). Assessment of the adequacy of bowel preparation, fecal and fluid tagging, colonic distention, and CT artifacts was made on a per-patient basis according to the CT Colonography Reporting and Data System ( 20 ). Inadequate examinations (scored as C0) (ie, presence of polyps 1 cm or larger could not be excluded) were recorded, along with the cause of each problem. Colonic lesions were primarily measured on three-dimensional views. For the nonanastomotic portions of the colon, lesions at least 6 mm in maximal diameter were reported ( ). For the anastomotic site, all lesions, with the exception of extruded surgical staples, were reported regardless of size since anastomotic recurrences do not develop through the adenoma-carcinoma sequence, and thus, the polyp size based lesion triage may not apply. Extruded surgical staples were not reported since they have no clinical importance ( 14 ). Segmental location and lesion morphology (sessile, pedunculated, flat, or other) were recorded ( 23 ). Flat morphology was defined as lesion height less than 3 mm ( 23,24 ). Location and extenct of extracolonic lesions suspicious for metastatic disease were recorded. Any examination-related adverse effects were recorded. Reference Standard Colonoscopy performed after CT colonography served as the reference standard for colonic findings. A recommendation of colonoscopy was made at CT colonography interpretation if nonanastomotic colonic lesions of at least 6 mm or anastomotic lesions of any size were found, and then the need for subsequent colonoscopy was ultimately determined by the referring surgeons. In the early phase after adopting contrastenhanced CT colonography for the postoperative surveillance, our surgeons Radiology: Volume 257: Number 3 December 2010 n radiology.rsna.org 699

4 still ordered routine colonoscopic surveillance even if CT colonography revealed negative findings to reconfirm the CT colonography findings. However, this practice pattern phased out rapidly as the surgeons became convinced of the accuracy of CT colonography and started skipping routine surveillance colonoscopy that was scheduled close to CT colonography. As a result, by the time of data retrieval for our study, colonoscopy after CT colonography had been performed in 111 (91.7%) of 121 patients with positive colonic findings at CT colonography and 439 (70.7%) of 621 patients negative findings ( Fig 1 ). The time interval between CT colonography and subsequent colonoscopy ranged from 0 to 49 months (median, 13 months). A long interval usually represented the next routine surveillance colonoscopy in a patient with negative CT colonography results. Colonoscopy was performed by one of three gastroenterologists and three colorectal surgeons, each with more than 10 years of experience (including J.C.K. and C.S.Y.), by using a video colonoscope (CF 260; Olympus Optical, Tokyo, Japan). The colonoscopists were informed of the CT colonography results prior to colonoscopy. Segmental location, size, and morphology of detected lesions were recorded, and lesions were removed or biopsied for pathologic examination. Lesion matching between CT colonography and colonoscopy required both segmental and size agreement (ie, within the same or an adjacent segment and within 50% of the colonoscopic size) ( 10,12 ). To analyze the diagnostic yield of contrast-enhanced CT colonography for extracolonic recurrences, the follow-up results of those patients with positive extracolonic findings at contrast-enhanced CT colonography were reviewed. Extracolonic recurrences were confirmed with biopsy, positron emission tomography/ct, and/or size increase at follow-up imaging. Statistical Analysis Diagnostic performance measures of CT colonography for detecting nonanastomotic colonic lesions of at least 6 mm and anastomotic lesions of any size included per-patient sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) and per-lesion sensitivity and PPV. Sensitivity and NPV were assessed according to lesion histologic type (ie, all lesions, all adenomatous lesions, advanced neoplasia, and adenocarcinomas). Advanced neoplasia represented both advanced adenomas (defined as 10 mm or by the presence of a substantial villous component or high-grade dysplasia [ 11,25 ]) and adenocarcinomas. The denominator for specificity was the number of patients who had neither nonanastomotic lesions of at least 6 mm nor anastomotic lesions of any size. For the per-patient analysis, at least one relevant match between CT colonography and colonoscopy was required to be considered a true-positive result for a given category. The 95% confidence intervals (CIs) were calculated for proportional data except for PPV and NPV, as they vary with the disease prevalence. Data clustering (ie, multiple lesions in a patient) was accounted for in the calculation of the 95% CI for per-lesion sensitivity. Results The characteristics of the 742 study patients are summarized in Table 1. Colonic Findings Colonoscopy. Five hundred fifty patients underwent colonoscopy after CT colonography. In 98 of these patients, 97 nonanastomotic lesions of at least 6 mm in size and 23 anastomotic lesions of any size were detected at colonoscopy ( Table 2 ). There were 209 additional nonanastomotic lesions that were 5 mm or smaller found at colonoscopy that were excluded from the analysis; none of them were advanced adenomas or adenocarcinomas. Histopathologic findings were available for 118 of the 120 colonic lesions. Two 7-mm sessile nonanastomotic polyps in two patients who had no additional lesions were lost after colonoscopic removal and were excluded from further analysis ( Table 2, Fig 1 ). Seventy-nine (83.2%) of 95 histologically confirmed nonanastomotic Table 1 Characteristics of the 742 Study Patients Characteristic Data AJCC-UICC staging 0 21 (2.8) I 154 (20.8) II 331 (44.6) III 220 (29.6) IV 16 (2.2) Preoperative chemo- and radiation therapy No 654 (88.1) Yes 88 (11.9) Surgical procedure Low anterior resection 302 (40.7) Anterior resection 207 (27.9) Right hemicolectomy 191 (25.7) Left hemicolectomy 42 (5.7) Note. Data are numbers of patients, with percentages in parentheses. All 742 patients had serum carcinoembryonic antigen levels within normal limits (,6 ng/ml) 0 99 days (median, 0 days) prior to CT colonography. AJCC-UICC = American Joint Committee on Cancer International Union Against Cancer. lesions and one (4.3%) of 23 anastomotic lesions were neoplastic ( P,.001, Fisher exact test). Nonanastomotic lesion morphology was sessile ( n = 73), pedunculated ( n = 11), or flat ( n = 11), whereas anastomotic lesion morphology was sessile ( n = 19), flat ( n = 2), ulcerative ( n = 1), or diffusely nodular along the anastomotic rim ( n = 1). Sixty-five patients had at least one adenomatous (either adenoma or adenocarcinoma) lesion. Twenty-two patients had advanced neoplasia (either advanced adenoma or adenocarcinoma). Six (0.8%; 95% CI: 0.3%, 1.8%) of 742 patients had adenocarcinomas, with a total of seven lesions (median size, 12 mm; range, 8 16 mm; stage: Tis [ n = 3], T1 [ n = 3], or T3 [ n = 1]), including six metachronous cancers and one recurrent cancer at the anastomosis. The seven metachronous or recurrent adenocarcinomas were detected months (median, 18 months) after the initial colon surgery. All lesions were amenable to repeat surgical ( n = 3) or colonoscopic ( n = 4) resection with curative intent. 700 radiology.rsna.org n Radiology: Volume 257: Number 3 December 2010

5 Table 2 Colonic Lesions Found at Follow-up Colonoscopy after CT Colonography in 550 Patients Nonanastomotic Anastomotic Type 6 9 mm 10 mm Total 5 mm 6 9 mm 10 mm Total Neoplastic Tubular adenoma Tubulovillous adenoma Serrated adenoma High grade dysplasia Adenocarcinoma Lipoma Subtotal Nonneoplastic Infl ammatory or granulation Hyperplastic Subtotal Undetermined * Total * Colonoscopically resected polyps were lost during retrieval. CT Colonography Performance. The CT colonography performance characteristics are summarized in Tables 3 and 4. Both per-patient and per-lesion CT colonography sensitivity was equal to or greater than 78.5% for all histopathologic categories of colonic lesions. CT colonography specificity was 93.1% (421 of 452). Importantly, CT colonography depicted all seven cancers within the colon. Examples of colonic lesions detected at CT colonography are shown in Figures 2 and 3. CT colonography missed mm lesions among the 53 nonadvanced adenomas, five 8 15-mm lesions among the 19 advanced adenomas, and seven 5 10-mm lesions among the 38 nonneoplastic lesions. Causes for the lesions missed at CT colonography were flat morphology ( n = 10), size discrepancy (ie, the lesion was detected at CT colonography but dismissed as diminutive yet measured 6 mm or greater at colonoscopy) ( n = 3), obscuration by residual stool ( n = 1), suboptimal colonic distention ( n = 1), reader perceptual error ( n = 1), and retrospectively unidentifiable without explicit causes ( n = 8). The PPV and NPV of CT colonography are summarized in Tables 3 and 4. Of the 548 patients in the colonic evaluation cohorts, 110 had positive CT colonographic findings. Of these, 79 had matching colonoscopic findings, resulting in a per-patient PPV of 71.8% ( Fig 1 ). Perlesion PPV was also 71.8% (94 of 131). Per-patient NPV was 96.1% (421 of 438) for all lesions and was 99.1% and 100% for advanced neoplasia and adenocarcinoma, respectively. Extracolonic Findings Extracolonic evaluation identified 11 patients (1.5%; 95% CI: 0.8%, 2.7%) with recurrent cancer manifesting as metastatic disease months (median, 18 months) after the initial surgery. The location of their initial primary colorectal cancer was the rectum ( n = 4), sigmoid colon ( n = 3), descending colon ( n = 1), ascending colon ( n = 1), or cecum ( n = 2). The initial cancer staging of these 11 patients according to the American Joint Committee on Cancer International Union Against Cancer system was I ( n = 1), II ( n = 5), or III ( n = 5), and none of the patients initially had metastatic disease. Location of the extracolonic metastatic disease included the liver ( n = 5), lungs ( n = 4), peritoneum ( n = 1), and adnexa ( n = 1). None of the patients with extracolonic metastatic disease had any cancerous lesions in the colon. Seven patients with hepatic, pulmonary, or adnexal metastasis received repeat curative-intent treatment, whereas the remaining four patients were not amenable to repeat curative treatment. CT Colonography Quality and Complications CT colonography quality was inadequate in 19 (2.6%; 95% CI: 1.6%, 4.0%) of 742 patients owing to either excessive untagged fecal matter ( n = 10) or poor colonic distention ( n = 9). Seven of the nine cases with inadequate colonic distention occurred in the lower abdomen and pelvis, which were the areas of prior surgical resection and/or preoperative chemo- and radiation therapy. The remaining two cases involved inadequate distention of the sigmoid colon following right hemicolectomy. All patients tolerated the examination well. There were no cases of colonic perforation or other substantial complications related to the CT colonography procedure. Discussion The diagnostic performance of CT colonog raphy for detection of relevant colonic lesions in our postoperative surveillance population was comparable to that of well-conducted representative screening CT colonography trials ( 9,10,12,13 ). Importantly, CT colonography demonstrated 100% sensitivity for cancer within the colon. In addition, contrast-enhanced CT colonography enabled detection of clinically unsuspected extracolonic metastatic disease in 11 patients, none of whom had cancerous lesions in the colon. The many advantages of contrast-enhanced CT colonography, including excellent diagnostic performance, dual colonic-extracolonic assessment, less invasive nature, and higher safety profile, all support its use in routine postoperative surveillance of colorectal cancer patients and may suggest that negative results at the CT colonography could exclude the need for surveillance colonoscopy during the same time frame. At the same time, the fact that a small number of benign Radiology: Volume 257: Number 3 December 2010 n radiology.rsna.org 701

6 Table 3 Per-Patient Analysis of CT Colonography for Detection of Colonic Lesions Performance Measure All Lesions All Adenomatous Lesions * Advanced Neoplasia Adenocarcinomas Sensitivity (%) 82.3 (79/96) [73.4, 88.7] 80.0 (52/65) [68.6, 88.1] 81.8 (18/22) [60.9, 93.3] 100 (6/6) [64.3, 100] Specifi city (%) 93.1 (421/452) [90.4, 95.2] NA NA NA PPV (%) 71.8 (79/110) NA NA NA NPV (%) 96.1 (421/438) 97.0 (425/438) 99.1 (434/438) 100 (438/438) Note. Data in parentheses are numbers used to calculate percentages, with 95% CIs in brackets, where applicable. NA = not applicable. * Includes both adenomas and adenocarcinomas. Includes advanced adenomas and adenocarcinomas. Table 4 Per-Lesion Sensitivity and PPV of CT Colonography for Colonic Lesions Performance Measure All Lesions All Adenomatous Lesions * Advanced Neoplasia Adenocarcinomas Sensitivity (%) 79.7 (94/118) [70.9, 88.5] 78.5 (62/79) [68.3, 88.7] 80.8 (21/26) [64.3, 97.2] 100 (7/7) [67.8, 100] PPV (%) 71.8 (94/131) NA NA NA Note. Data in parentheses are numbers used to calculate percentages, with 95% CIs in brackets, where applicable. NA = not applicable. * Includes both adenomas and adenocarcinomas. Includes advanced adenomas and adenocarcinomas. adenomas, including five of 19 advanced adenomas, were missed at CT colonography may raise some concerns about the substitution of colonoscopic surveillance with CT colonography. Therefore, concerns regarding the adenomas missed at CT colonography, particularly those with advanced histologic features, need to be further resolved. Another factor to consider with CT colonography as a postoperative surveillance tool is the resulting colonoscopy referral rate. If the referral rate is too high, colonic evaluation at CT colonography will be redundant, and this surveillance approach will become less cost-effective. In our study population, the maximum referral rate would have been 19% (140 of 742) if all patients with either positive CT colonography findings ( n = 121) or inadequate examinations ( n = 19) had undergone subsequent colonoscopy. Not surprisingly, this value is higher than the 9.8% (99 of 1015) to 11.9% (132 of 1110) referral rates reported in general screening populations ( 26,27 ). However, considering that CT colonography can be incorporated into the routine postoperative abdominal CT scan as an add-on examination that incurs only minor additional patient inconvenience and cost, the higher colonoscopy referral rate would seem acceptable. Moreover, the actual colonoscopy referral rate in clinical practice would be lower than the theoretical maximum referral rate owing to several factors. Technical modifications and increasing clinical experience in evaluating the anastomosis may further improve the results we have reported. The rate of diagnostically inadequate CT colonographic examinations in our study was fairly low (2.6%), albeit slightly higher than the 1.1% (12 of 1110) and 2.1% (21 of 1015) rates reported in other studies ( 26,27 ). Seven of the 19 inadequate examinations were owing to poor colonic distention at or near the area of prior surgery and/or radiation therapy. Similarly, researchers in a previous study ( 15 ) reported inadequate sigmoid distention after lower anterior resection. This pitfall is probably related to decreased compliance of the colon and surrounding tissues from posttreatment scarring. A high rate of inadequate colonic distention following right hemicolectomy (eight of 24 patients), presumably caused by colonic decompression through a patulous ileocolic anastomosis, was also reported in that study ( 15 ); however, this seldom occurred in our study (two of 191 patients). The difference can perhaps be explained by the fact that we used an automated colonic insufflator, whereas the previous study used a manual insufflation method. The automated insufflator provides a continuous low-pressure infusion of carbon dioxide, which helps to maintain adequate colonic distention ( 28 ). Given the aforementioned findings, we believe that contrast-enhanced CT colonography can be used as an effective tool for simultaneous colonic and extracolonic surveillance after colorectal cancer surgery, which will help reduce negative postoperative colonoscopy studies, even though our study design does not allow us to specifically suggest how to optimally integrate CT colonography with other surveillance methods. Considering the imperfect sensitivity of CT colonography for advanced adenomas, it may be prudent to first ensure complete colonoscopic evaluation and clearing of the large intestine in the perioperative period before employing CT colonography for surveillance to 702 radiology.rsna.org n Radiology: Volume 257: Number 3 December 2010

7 Figure 2 Figure 3 Figure 2: Images in a 66-year-old man with a 16-mm ulcerating anastomotic recurrence (arrowheads) after low anterior resection. U = ulcer. (a) Three-dimensional endoluminal CT colonographic image obtained 10 months after surgery shows an ill-defi ned ulcerating lesion. (b) Digital photograph from follow-up colonoscopy shows corresponding ulcerating lesion. Surgical resection was performed and pathologic analysis confi rmed recurrent adenocarcinoma. minimize the potential risk of advanced adenomas that are missed at CT colonography being found later as metachronous cancers. Given that colonoscopic blind areas exist, whereas CT colonography more readily covers the entire colonic surface ( 29 ), it may also be worthwhile to investigate whether the alternating use of colonoscopy and CT colonography could capitalize on their complementary strengths and contribute to improved patient survival. Our study had several limitations. First, since our study was a clinical practice-based retrospective study, there may be some potential for case selection bias even though we used predefined criteria for performing the surveillance CT colonography. Second, a small portion of patients did not undergo follow-up colonoscopy after CT colonography, particularly in the group with negative findings at CT colonography. As a result, we may have slightly overestimated the sensitivity of CT colonography. Third, there was a long time interval between CT colonography and colonoscopy in some patients, predominately those with negative CT colonography findings. However, if anything, this prolonged interval would presumably underestimate the sensitivity and NPV of CT colonography, since some missed lesions could have conceivably developed or increased in size since the time of CT colonography. Last, our study included a relatively small number of metachronous or recurrent Figure 3: Images in a 52-year-old woman with an 8-mm metachronous cancer, seen as a fl at-elevated lesion (arrowhead), and a 6-mm tubulovillous adenoma, seen as a sessile polyp (arrow), in ascending colon after low anterior resection. (a) Threedimensional endoluminal CT colonographic image obtained 3 years after surgery shows both a fl atelevated lesion and a sessile polyp in ascending colon. (b) Two-dimensional transverse CT colonographic image at a colon window setting (width, 2000 HU; level, 0 HU) shows profi le view of fl at-elevated lesion. (c) Digital photograph from follow-up colonoscopy shows corresponding fl at-elevated lesion and sessile polyp. Colonoscopic submucosal dissection of the fl at lesion found an adenocarcinoma limited to submucosa, while polypectomy of the sessile lesion found a tubulovillous adenoma. cancers, despite the large patient cohort. However, this low incidence of cancers in the colon underscores the value of a less invasive surveillance tool, such as CT colonography, that maintains accuracy. In conclusion, contrast-enhanced CT colonography is an accurate and practical surveillance tool following colorectal cancer surgery in patients without clinical or laboratory evidence of recurrence that allows for simultaneous lessinvasive evaluation of both colon and extracolonic organs. References 1. Jeffery M, Hickey BE, Hider PN. Follow-up strategies for patients treated for nonmetastatic colorectal cancer. Cochrane Database Syst Rev 2007; ( 1 ):article CD Radiology: Volume 257: Number 3 December 2010 n radiology.rsna.org 703

8 2. Manfredi S, Bouvier AM, Lepage C, Hatem C, Dancourt V, Faivre J. Incidence and patterns of recurrence after resection for cure of colonic cancer in a well defined population. Br J Surg 2006 ; 93 ( 9 ): Renehan AG, Egger M, Saunders MP, O Dwyer ST. Impact on survival of intensive follow up after curative resection for colorectal cancer: systematic review and meta-analysis of randomised trials. BMJ 2002 ; 324 ( 7341 ): Tsikitis VL, Malireddy K, Green EA, et al. Postoperative surveillance recommendations for early stage colon cancer based on results from the clinical outcomes of surgical therapy trial. J Clin Oncol 2009 ; 27 ( 22 ): Engstrom PF, Arnoletti JP, Benson AB 3rd, et al ; National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: colon cancer. J Natl Compr Canc Netw 2009 ; 7 ( 8 ): Engstrom PF, Arnoletti JP, Benson AB 3rd, et al ; National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: rectal cancer. J Natl Compr Canc Netw 2009 ; 7 ( 8 ): Desch CE, Benson AB 3rd, Somerfield MR, et al ; American Society of Clinical Oncology. Colorectal cancer surveillance: 2005 update of an American Society of Clinical Oncology practice guideline. J Clin Oncol 2005 ; 23 ( 33 ): Rex DK, Kahi CJ, Levin B, et al. Guidelines for colonoscopy surveillance after cancer resection: a consensus update by the American Cancer Society and US Multi-Society Task Force on Colorectal Cancer. CA Cancer J Clin 2006 ; 56 ( 3 ): Graser A, Stieber P, Nagel D, et al. Comparison of CT colonography, colonoscopy, sigmoidoscopy and faecal occult blood tests for the detection of advanced adenoma in an average risk population. Gut 2009 ; 58 ( 2 ): Johnson CD, Chen MH, Toledano AY, et al. Accuracy of CT colonography for detection of large adenomas and cancers. N Engl J Med 2008 ; 359 ( 12 ): Kim DH, Pickhardt PJ, Taylor AJ, et al. CT colonography versus colonoscopy for the detection of advanced neoplasia. N Engl J Med 2007 ; 357 ( 14 ): Pickhardt PJ, Choi JR, Hwang I, et al. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Engl J Med 2003 ; 349 ( 23 ): Regge D, Laudi C, Galatola G, et al. Diagnostic accuracy of computed tomographic colonography for the detection of advanced neoplasia in individuals at increased risk of colorectal cancer. JAMA 2009 ; 301 ( 23 ): Choi YJ, Park SH, Lee SS, et al. CT colonography for follow-up after surgery for colorectal cancer. AJR Am J Roentgenol 2007 ; 189 ( 2 ): Fletcher JG, Johnson CD, Krueger WR, et al. Contrast-enhanced CT colonography in recurrent colorectal carcinoma: feasibility of simultaneous evaluation for metastatic disease, local recurrence, and metachronous neoplasia in colorectal carcinoma. AJR Am J Roentgenol 2002 ; 178 ( 2 ): Laghi A, Iannaccone R, Bria E, et al. Contrastenhanced computed tomographic colonography in the follow-up of colorectal cancer patients: a feasibility study. Eur Radiol 2003 ; 13 ( 4 ): You YT, Chang Chien CR, Wang JY, et al. Evaluation of contrast-enhanced computed tomographic colonography in detection of local recurrent colorectal cancer. World J Gastroenterol 2006 ; 12 ( 1 ): Amitai MM, Fidder H, Avidan B, et al. Contrast-enhanced CT colonography with 64-slice MDCT compared to endoscopic colonoscopy in the follow-up of patients after colorectal cancer resection. Clin Imaging 2009 ; 33 ( 6 ): Park SH, Kim SY, Lee SS, et al. Sensitivity of CT colonography for nonpolypoid colorectal lesions interpreted by human readers and with computer-aided detection. AJR Am J Roentgenol 2009 ; 193 ( 1 ): Zalis ME, Barish MA, Choi JR, et al ; Working Group on Virtual Colonoscopy. CT colonography reporting and data system: a consensus proposal. Radiology 2005 ; 236 ( 1 ): Bond JH. Clinical relevance of the small colorectal polyp. Endoscopy 2001 ; 33 ( 5 ): Togashi K, Shimura K, Konishi F, et al. Prospective observation of small adenomas in patients after colorectal cancer surgery through magnification chromocolonoscopy. Dis Colon Rectum 2008 ; 51 ( 2 ): Dachman AH, Zalis ME. Quality and consistency in CT colonography and research reporting. Radiology 2004 ; 230 ( 2 ): The Paris endoscopic classification of superficial neoplastic lesions: esophagus, stomach, and colon: November 30 to December 1, Gastrointest Endosc 2003 ; 58 ( 6, Suppl ): S3 S Winawer SJ, Zauber AG. The advanced adenoma as the primary target of screening. Gastrointest Endosc Clin N Am 2002 ; 12 ( 1 ): Pickhardt PJ, Taylor AJ, Kim DH, Reichelderfer M, Gopal DV, Pfau PR. Screening for colorectal neoplasia with CT colonography: initial experience from the 1st year of coverage by third-party payers. Radiology 2006 ; 241 ( 2 ): An S, Lee KH, Kim YH, et al. Screening CT colonography in an asymptomatic averagerisk Asian population: a 2-year experience in a single institution. AJR Am J Roentgenol 2008 ; 191 ( 3 ): W100 W Burling D, Taylor SA, Halligan S, et al. Automated insufflation of carbon dioxide for MDCT colonography: distension and patient experience compared with manual insufflation. AJR Am J Roentgenol 2006 ; 186 ( 1 ): Pickhardt PJ, Taylor AJ, Gopal DV. Surface visualization at 3D endoluminal CT colonography: degree of coverage and implications for polyp detection. Gastroenterology 2006 ; 130 ( 6 ): radiology.rsna.org n Radiology: Volume 257: Number 3 December 2010