CT Colonography: Techniques, Indications, Findings

European Journal of Radiology 61 (2007) 388–399

CT colonography: Techniques, indications, ﬁndings Thomas Mang a,∗, Anno Graser b, Wolfgang Schima a, Andrea Maier a a Department of Radiology, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria b Department of Clinical Radiology, Grosshadern Campus, University of Munich, Marchioninistrasse 15, 81377 Munich, Germany Received 21 October 2006; received in revised form 29 October 2006; accepted 2 November 2006

Abstract Computed tomographic colonography (CTC) is a minimally invasive technique for imaging the entire colon. Based on a helical thin-section CT of the cleansed and air-distended colon, two-dimensional and three-dimensional projections are used for image interpretation. Several clinical improvements in patient preparation, technical advances in CT, and new developments in evaluation software have allowed CTC to develop into a powerful diagnostic tool. It is already well established as a reliable diagnostic tool in symptomatic patients. Many experts currently consider CTC a comparable alternative to conventional colonoscopy, although there is still debate about its sensitivity for the detection of colonic polyps in a screening population. This article summarizes the main indications, the current techniques in patient preparation, data acquisition and data analysis as well as imaging features for common benign and malignant colorectal lesions. © 2007 Published by Elsevier Ireland Ltd.

Keywords: CT colonography; Polyps; Colorectal cancer; Indications; Bowel preparation

1. Introduction Currently, there are several clinical indications for CT colonography. They include evaluation of the colon after an Computed tomographic colonography (CTC), also known incomplete or unsuccessful conventional colonoscopic exami- as virtual colonoscopy (VC), is a powerful technique for the nation and evaluation of the colon proximal to an obstructing evaluation of the entire colon. It has potential advantages over neoplasm [7–11]. Another potential indication for CT colonog- conventional colonoscopy because of its minimally invasive raphy is in the evaluation of frail and elderly patients or nature and no need for sedation and recovery time. The exami- patients who would have an increased risk with conventional nation is based on a helical, thin-section CT of the cleansed and colonoscopy. The use of CTC to monitor patients after surgery distended colon. Data evaluation is performed with commer- for colorectal cancer is currently under investigation [12,13].In cially available CT colonography post-processing software with addition, CT colonography may contribute to colorectal screen- simultaneously available multiplanar 2D and virtual endoscopic ing by providing a safe, effective, and rapid examination that can 3D image displays. be used to evaluate the entire colon for clinically relevant lesions. The ability of CTC to detect colorectal polyps has been tested This article summarizes the main indications, the current in a multitude of studies. CTC appeared to be promising in high- technique in patient preparation, data acquisition and data anal- risk populations, with a reported sensitivity greater than 90% for ysis, and the imaging features of common benign and malignant polyps ≥10 mm. Recent results in the low prevalence population colorectal lesions. were more heterogeneous and less impressive (34–93.8%) [1–4]. This wide range of results is likely caused by differences in patient selection, examination and evaluation techniques, and 2. Indications reader experience [5,6]. Currently, there are several indications where CT colonography may play an important role in patient care. These include clinical indications, such as the evaluation of patient symptoms, ∗ Corresponding author. Tel.: +43 1 40400 4893; fax: +43 1 40400 4898. evaluation after an incomplete colonoscopy, and screening in E-mail address: [email protected] (T. Mang). asymptomatic high- and average-risk patients.

2.1. Clinical indications

The most widely accepted clinical indication is incomplete or failed colonoscopy. An incomplete colonoscopic examination is defined as a failure to intubate the caecum. Incomplete colonoscopy may be the result of poor bowel preparation, redun- dant colon, and patient intolerance to the procedure, spasm, or colonic obstruction caused by a neoplastic or non-neoplastic stenosis. The CTC examination can be performed on the same day directly after conventional colonoscopy, and thus, no additional bowel preparation is needed [14]. CTC can complete the colonic evaluation in the majority of patients, which can also analyze the cause of incomplete endoscopy [7,8]. Patients with a history of incomplete colonoscopy are at higher risk for failure of a second attempt. Therefore, CTC, rather than a second attempt at conventional colonoscopy, may be recommended. Double-contrast barium enema (DCBE), traditionally being used to evaluate the proximal colon in incomplete colonoscopy, has been outperformed by CTC in some institutions [15]. The residual colonic air distention after a failed colonoscopy may be an impediment to a successful, immediate, follow-up DCBE on the same day, but an immediate follow-up CTC is beneficial because the requisite colonic distention is already present. Two recent large prospective clinical trials show a sensitivity for DCBE of only 50% for “significant” polyps, measuring at least 10 mm in diameter [4,16]. At the same time, CTC has shown a better performance in several studies in patients with polyps 10 mm or larger [17]. In addition, patient preference studies are also discouraging for DCBE. In two recent questionnaire studies, patients ranked DCBE a distant last among three alternative colon cancer screening tests CTC, optical colonoscopy, and DCBE [4,18]. Last, but not least, in the future, radiologic experience in performing DCBE studies will fade because of the decreasing volume of investigations and fewer knowledgeable teachers. In cases of an obstructing cancer, CTC offers information about the pre-stenotic colon, local tumor invasion, lymph nodes, and distant metastases [19–21]. CTC reportedly allowed detection of synchronous carcinomas, which occur in approximately 5% of cases, as well as synchronous adenomas, with a reported incidence of 27–55% proximal to the stenosis [7,9,11] (Fig. 1). Fig. 1. Incomplete colonoscopy because of a stenotic cancer of the transverse In this setting, IV contrast is helpful to enable a complete staging colon. (a) Surface-rendered global CT image shows the stenosis of the transverse of the patient [11]. colon (arrow) and a second stenosis in the ascending colon (arrowhead). (b) Another potential indication for CTC is in the evaluation of Supine 2D axial contrast-enhanced CT image shows circular wall thickening and contrast enhancement in the transverse (arrow) as well as in the ascending patients with contraindications to colonoscopy or who refuse colon (arrowhead). Histology revealed a second cancer in the ascending colon other screening options [22]. This includes patients in need of that was not diagnosed by endoscopy. anticoagulation, a history of difficult or incomplete colonoscopy in the past, and patients who cannot be sedated due to medical evaluation of symptomatic patients. In addition, CTC could have conditions. In addition, in cases of advanced patient age, and in an extra benefit for the local staging of tumors (T3, T4 stages) frail or immobile patients and patients with severe comorbidities, because of its high spatial resolution and excellent depiction of CTC can be safely performed to exclude neoplastic or stenotic the colonic wall, as well as adjacent pericolonic fat. conditions [22]. With regard to post-surgical conditions in the colon, there Taylor et al. reported on the use of diagnostic CTC in patients is no general agreement about the use of CT colonography. with positive symptoms of colorectal cancer, such as a change Contrast-enhanced CT colonography has the potential to detect in bowel habits, lower gastrointestinal bleeding, iron deficiency local recurrence, metachronous disease, and distant metastases anemia, or abdominal mass and abdominal pain [23]. The in patients with a history of invasive colorectal cancer [12,13]. authors suggested that CTC may be a valuable tool for the rapid Currently, endoscopy or barium enemas are performed in many 390 T. Mang et al. / European Journal of Radiology 61 (2007) 388–399 cases after colonic surgery for routine surveillance, to detect practical approach is to consider CTC a reasonable alternative tumour recurrence, or to discover a metachronous cancer. After to the other colorectal cancer screening tests when a patient partial colonic resection, some of these follow-up examinations is unable or unwilling to undergo conventional colonoscopy could be replaced by contrast-enhanced CT colonography. In [32]. most cases, CT colonography allows visualization of the entire colon, which is important for the demonstration of post-surgical 2.3. Contraindications anatomic conditions. Two-dimensional views offer information about the wall morphology of the anastomosis [4,16]. Contraindications to CTC include acute abdominal pain, In addition to these main indications, there are other recent abdominal or pelvic surgery, abdominal wall hernia with conditions where the role of CT colonography is not yet entrapment of colonic loops, and acute inflammatory conditions, clearly defined. Some of these conditions may lead to colonic such as acute diverticulitis, acute active stage of ulcerative coli- obstruction, where CTC may be performed after incomplete tis or Crohn’s disease, and toxic megacolon. In these conditions, colonoscopy. However, CTC may also be used for surveillance insufflation of the colon can lead to perforation and widespread of these conditions, as an alternative to colonoscopy or barium peritonitis [33–35]. In addition, there are also general CT con- enema. Diverticular disease is the most common colonic disor- traindications that matter in CTC as well, such as weight and der in the Western world and often leads to diverticulitis. CTC girth limitations of the scanner, artifacts from metal prosthesis, is not only helpful in the assessment of the lumen and the extent pregnancy, and patients with claustrophobia. of diverticulosis, but also for any extramural changes in cases with the chronic stenotic stages of diverticulits [24]. At chronic 3. CT colonography technique stages of inflammatory bowel diseases, CT colonography can provide information on the extent of the disease and about steno- CT colonography is based on a helical, thin-section CT of the sis and prestenotic regions, as well as the extracolonic extent and cleansed and distended colon. The examination consists of three complications of the diseases [25–27]. major steps: patient preparation, including cathartic cleansing At the very least, patients who are unwilling to undergo and distention of the bowel; data acquisition with MDCT; and conventional colonoscopy because of anxiety, fear, or embar- data evaluation. rassment may profit from a relatively painless and safe CTC examination [22]. 3.1. Bowel preparation

2.2. Colorectal cancer screening The key element of a high quality CTC examination is a well- prepared clean, and well-distended colon. Residual stool and In contrast to symptomatic patients, CTC is also considered a fluid may lead to a false-negative or false-positive diagnosis. potential screening tool for colorectal neoplasia in asymptomatic Therefore, CTC, at present, requires full bowel preparation, just high- and average-risk patients. like colonoscopy and DCBE. The goal of colorectal cancer screening is a reduction in the Patients usually follow a clear liquid diet starting about 24 h morbidity and the mortality associated with colorectal cancer prior to the CTC examination. Cathartic cleansing is usually through early detection and resection of adenomas and cancer. performed by oral administration of laxatives. Different com- CTC as a screening tool has the potential for a wider pub- mercially available bowel preparations have been described for lic acceptance compared to conventional colonoscopy. Several CTC, including cathartics, such as magnesium citrate (LoSo studies have shown that patients prefer CTC over conventional Preparation, EZ-Em Inc, Westbury, USA) and phosphosoda colonoscopy [18,28,29]. (Fleet Pharmaceuticals, Lynchburg, VA), and colonic lavage Although a recent study by Pickhardt et al. showed an excel- solutions, such as polyethylene glycol (PEG). Radiologists gen- lent performance for CTC in an average-risk population, with erally need a dry colon because of the inability to aspirate a reported sensitivity of 93.8% for polyps ≥10 mm [3], other residual fluid during the examination. authors achieved less impressive results ranging from 34 to 53% It has been reported that phosphosoda and magnesium cit- [1,2,4]. This wide scatter in sensitivity was attributed to techni- rate result in an adequately prepared and dry colon for the cal differences with regard to the examinations, evaluations, and majority of patients, whereas PEG preparations frequently leave reader experience. large amounts of residual fluid in the colon and, therefore, are Therefore, the current use of VC generally does not include not optimal for CTC [36]. However, it should be mentioned screening of asymptomatic individuals, as suggested by the that phosposoda is contraindicated in patients with known renal American Cancer Society [30] and the American Gastrointesti- failure, preexisting electrolyte abnormalities, congestive heart nal Association [31]. Both of these organizations decided that failure, ascites, or ileus [37]. Therefore, because PEG is not CTC should not yet be used for colorectal cancer screening, contraindicated in patients with renal or cardiac insufficiency, it because data on true screening populations are missing. How- can be used alternatively in these cases [22]. In addition, some ever, further attempts to validate this rapidly evolving technique European centers have reported the use of other laxatives, such are ongoing. Currently, CTC is the second best imaging modal- as Picolax [38]. ity for complete colonic evaluation, being superior to DCBE As opposed to a standard bowel preparation, the addition for detection of colorectal polyps and cancer [79]. Therefore, a of oral contrast agents (small amounts of barium or iodine T. Mang et al. / European Journal of Radiology 61 (2007) 388–399 391

large bowel. Underdistended or collapsed segments may hide intraluminal lesions. Before beginning bowel distention, patients should be advised to empty their rectum. Bowel distention is performed in the left decubitus or supine position with a thin, flexible rubber catheter placed in the rectum (e.g., thin plastic or rubber 14F rectal tube, small gauge Foley catheter) [43]. Larger catheters, such as those used in DCBE, are not required to successfully distend the colon in CTC [44]. For bowel distention, either room air or carbon dioxide (CO2) can be used. The easiest and cheapest method is manual room air distention via a handheld plastic bulb insufflator, an effective technique that can even be performed by patients themselves. Some authors argue that CO2 is superior to air, largely based on colonoscopy and the barium enema literature, which suggests that it causes less discomfort because of its rapid mucosal absorption. CO2 can be administered either manually, over a standard enema bag filled Fig. 2. Fecal tagging with orally ingested barium: prone 2D axial CT image with approximately 3 l of gas (via a gas cylinder) attached to a shows high attenuation of the contrast marked fecal residuals and residual fluid (arrows). rectal catheter over a connecting tube, or automatically, using a dedicated insufflation device (Protocol colon insufflation system, EZ-Em Inc., Westbury, USA). This device electronically contrast) will tag residual stool or fluid (Fig. 2). The resulting controls the flow rate of CO2, the total administered gas volume, higher attenuation of fecal and fluid residues simplifies their and the intracolonic pressure (which is limited up to a maximum distinction from colonic abnormality. Although commercial of 25 mmHg). Advocates of automated insufflation suggest that tagging agents (Tagitol, EZ-Em Inc., Westbury, USA) are not introducing CO2 at controlled flow rates and pressures improves yet available in the EU, there is no general consensus about the patient compliance and overall distention [45,46]. During the agent or the amount preferable for use in fecal tagging. Whereas gas insufflation, the patient is placed in the supine position some authors prefer tagging with barium only [39], others have and gentle insufflation is continuous until the patient feels reported good results with iodine [40] or a combination of uncomfortable or bloated. This generally will take 2–4 l of gas, both to achieve fecal and fluid tagging [3]. In brief, Pickard depending on the patient’s individual colonic anatomy. Fixed et al. achieved excellent results with a combination of 500 ml gas amounts may be impractical because of different colonic of 2.1% barium contrast material for solid stool tagging and volumes. After distention is completed, a CT scout view of the 120 ml of an ionic water-soluble contrast agent (Gastrografin, abdomen is obtained to ensure optimal colonic distention and Schering AG, Berlin, Germany) for fluid opacification [3]. to add additional gas if collapsed segments are identified. Fol- Fecal tagging is now advocated by an increasing number of lowing the supine axial image acquisition, the patient is turned investigators as the method of choice to prepare the colon for CT to the prone position. Before prone image acquisition, another colonography. scout scan is obtained with additional gaseous insufflation Compared to cathartic preparations, so called prepless bowel if needed. preparation is a laxative-free regimen where patients may eat The i.v. administration of antispasmodic agents (hyoscin- and drink normally or are asked to maintain a low residue or N-butylbromide, buscopan, or glucagon) may improve colonic low fiber diet while fecal tagging is achieved by multiple doses distention and reduce spasms. However, the use of spasmolytics of oral contrast (barium or iodinated contrast) over a 24–72 h is controversial; i.v. hyoscin-N-butylbromide likely improves period prior to CTC. With computer-generated electronic distention, but is not licensed in the United States [44], whereas cleansing techniques, either by using thresholding or using spe- glucagon hydrochloride, its licensed alternative, does not appear cific subtraction computer software, the high attenuation tagged to be as beneficial [47,48]. In addition, some authors argue fecal material can be subtracted from the data, leaving only the that administration of an IV drug adds to patient discomfort, colonic wall and colonic lesions [41,42]. Although initial results creates an opportunity for possible new side effects, and are promising for detection of lesions of 1cm or larger [40,41], increases examination time and cost. The general opinion, this method is still at the research stage and must be validated in provided by a recently published, consensus statement is that a screening population. In the future, prepless CTC may increase IV spasmolytics should not be administered routinely, but overall patient compliance by replacing the cumbersome bowel can be used if patients experience pain, discomfort, or spasm preparation. [32]. Bowel distention may lead to perforation of the bowel in 3.2. Bowel distention rare cases. In most of the reported cases, perforation occurred in symptomatic patients with acute inflammatory or stenotic Optimal colonic distention is a fundamental prerequisite for colons, the bowel distention was performed manually, and rectal CTC data evaluation that allows intraluminal evaluation of the balloon catheters were used [33–35]. 392 T. Mang et al. / European Journal of Radiology 61 (2007) 388–399

3.3. CT scanning media to be of benefit in cases with clinical indications, such as in patients with symptoms of colorectal cancer, or to detect CT scanning is ideally performed on an MDCT scan- local recurrence, metachronous disease, or distant metastasis in ner in both the supine and then the prone positions with a patients with prior colorectal cancer [11,12,20,54]. However, thin collimation. MDCT has several technical advantages over because of the increased cost, the need for intravenous access, single-detector-row CT, including faster imaging times and and the risk of allergic anaphylactic reactions the application acquisition of multiple [4–64] thin sections with nearly isotropic of IV contrast media might be contraindicated in a screening voxels. The higher speed and spatial resolution of MDCT should population and limited to clinical indications. As reported by a offer improved sensitivity and specificity for CTC compared to recently published consensus statement most of CTC investiga- single-detector CT. As mentioned before, acquisition of an initial tors do not routinely administer i.v. contrast media for screening scout view before each scan helps to ensure adequate distention purposes [32]. of the colonic segments with additional CO2 or room air being insufflated, if required. The use of both the supine and prone CT 4. Data analysis datasets helps to differentiate mobile stool from fixed pathology, such as polyps or cancers, allows more even distention of the Image processing and interpretation is performed on a colon because of gas redistribution, and improves visualization commercially available computer workstation equipped with of segments of the colon that may be obscured by intralumi- dedicated CT colonography software. In addition to 2D axial nal fluid [49,50]. To avoid breathing artifacts, which are more and multiplanar reformations in a cine mode, such systems prominent in the upper abdomen, scanning is performed in the provide an interactive, manual, mouse-driven, automated or cephalo-caudal direction. semi-automated, virtual “fly-through” of the surface- or volume- Thin sections are a prerequisite for high-quality multiplanar rendered 3D intraluminal images. reformations (MPR) and 3D reconstructions. As recommended CTC datasets can be evaluated by a primary 2D or a primary by a recently published consensus statement, collimation should 3D approach. In either case, the alternative viewing technique not exceed 3 mm when using MDCT [32]. However, with must be available for rapid correlation and characterization of 16-row and 64-row MDCT, thinner, sub-millimeter, collima- any suspicious findings. The combined use of both, 2D and 3D tions are preferable. Near-isotropic imaging is already provided visualization techniques has been shown to be superior to the on a 4-row MDCT scanner with a detector configuration of evaluation of 3D or 2D views only, with regard to sensitivity 4mm× 1 mm (minimal slice thickness of 1.25 mm), which and specificity [56,57]. allows scanning of the abdomen during a 30-s breath-hold. Primary 2D evaluation is based on “lumen tracking,” by inter- With a 16-row or 64-row MDCT scanner, and a detector con- actively tracing through the 2D dataset, focusing on only the figuration of 16 mm × 0.75 mm or 64 mm × 0.6 mm, scanning air-distended colonic lumen from one end to the other, with is completed in 11–12 s or 6–7 s. Such datasets can be recon- special focus on the cross-section of one colonic segment at structed as 1 mm sections overlapped every 0.7 mm. a time [57]. Primary 2D evaluation provides information about One of the major limitations of CTC is the relatively high the attenuation of findings during the search process and is more radiation exposure, and therefore, increasing attention has been time-efficient [56,58]. Additional 3D views are often used for focused on low-dose protocols. Because a thin collimation is problem solving. Primary 3D evaluation is based on 3D vir- necessary for CTC, dose reduction is widely achieved by reduc- tual endoscopy in an antegrade and retrograde fashion. Primary ing the milliampere-seconds level. A recent MDCTC study 3D evaluation was shown to be sensitive for polyp detection showed excellent results for the detection of polyps >10 mm, because both, the conspicuity, especially of small and medium- with thin beam collimation and an effective 50 mAs [51].How- sized polyps, and the duration of visualization, are increased [3]. ever, lower mAs values (10 mAs, ultra low-dose) have also been Additional 2D views are necessary for characterization of find- reported to be feasible for polyp detection [52]. However, such ings. However, the primary 3D evaluation is time-consuming “ultra low-dose” protocols may not be feasible for the evalua- because it must be performed in an antegrade and retrograde tion of extracolonic structures (because of the increased image fashion for the perception of lesions behind haustral folds. Col- noise,) or for obese patient, but further research is needed. Gen- lapsed segments must be evaluated alternatively, by 2D planar erally useful exposure settings are 120 kVp and 50–100 mAs in images. the prone and in the supine positions. A recent analysis of differ- There is no general consensus on whether a primary 2D or ent CTC research institutions showed a median mAs value of 67 3D approach should be employed. It depends primarily on the at a median collimation of 2.5 mm [53]. Use of automated dose radiologist’s preferences and the capabilities of the workstation. modulation techniques that adapt mAs values to patient anatomy At present, the most commonly used platform for data interpre- should always be used, if these techniques are available on the tation of CTC is a primary 2D interpretation, with 3D used for CT scanner [80]. problem solving, while primary 3D interpretation is preferred by an increasing minority [32]. 3.3.1. Contrast media In 2005, the working group on virtual colonoscopy proposed a Intrinsic colonic lesions and extracolonic pathology can be reporting scheme, “C-RADS—CT Colonography Reporting and enhanced by IV injection of iodine contrast media (IV CM) in Data System,” that includes recommendations for the follow- the second scan [54,55]. Previous studies have shown contrast up of colonic polyps [59]. In brief, this scheme is based on T. Mang et al. / European Journal of Radiology 61 (2007) 388–399 393

Fig. 3. Virtual colonic dissection in a patient with two polyps: virtual dissected CT image shows the colon fully dissected longitudinally and unfolded, similar to a pathological preparation. Two polyps are seen in the sigmoid colon (arrows). lesion size, morphology, location, attenuation, and preliminary recommendations for lesion surveillance. Although this effort needs to be reﬁned with a multi-disciplinary consensus, it repre- sents an important step toward more consistent and reproducible reporting in CTC.

4.1. Improvements in 3D visualization

Current developments in 3D visualization include virtual dissection, or panoramic, as well as unfolded, cube projections and translucency rendering. These views differ from conventional 3D simulation of endoscopic examinations, a technique that has limitations with regard to blind spots, as well as time-consuming, bi-directional evaluation. These new projections have been created to overcome this limitation by increasing surface visibility. With “virtual dissection,” the colon is fully dissected longitudinally and unfolded, similar to a pathological preparation [60] (Fig. 3). The “unfolded cube” projection renders six planar projections at 90◦ viewing angles from points on the central path. The unfolding of such a cube shows the complete field of view at a path position [61]. The panoramic view is a variation of the unfolded cubic view, which renders five faces of a cubic view in Fig. 4. Standard virtual endoscopy versus panoramic view for visualization of the plane in a continuous fashion. To minimize distortions, the two colonic polyps: (a) standard virtual endoscopic CT image shows one 7 mm front view is mapped into a square while the other four faces are polyp in the background of the image. (b) Panoramic view CT image of the mapped around it into a disk [62] (Fig. 4). same area unfolds the colonic wall on a disc, increasing the field of view. Note Initial results for some algorithms tested in small series of an additional 6 mm polyp can be seen with the panoramic view (arrowhead), patients showed a reduced reading time without a significant which was hidden between haustral folds on standard virtual endoscopy. difference in sensitivity [60,61,63]. However, increasing surface visibility, by flattening a 3D structure in a 2D image or by changing viewing angles, is likely to suffer from the fact that assign different colors (e.g., blue, green, red, and white) to areas luminal anatomy is distorted, especially in areas of flexures or of increasing attenuation, and therefore, provide a rapid and in suboptimal distended segments. Therefore, further studies are effective differentiation between soft-tissue lesions and pseu- needed to evaluate the feasibility of such tools. dolesions, such as stool. True colorectal polyps will show a “Translucency rendering” is a 3D tool that has been devel- uniform concentric ring pattern with a homogenous red core, oped to provide information on lesion attenuation beneath the while stool, if tagged, will have a white core. If stool is not rendered surface by an attenuation-dependent color scale super- tagged, because of small gas bubbles, it will display an inhomo- imposed on a 3D endoluminal view [64] (Fig. 5). These views geneous translucent pattern. 394 T. Mang et al. / European Journal of Radiology 61 (2007) 388–399

positive finding. Recent CAD algorithms showed a promising performance, with a reported a CAD sensitivity of 89.3% for adenomas ≥1cm [65]. Some of these systems have recently become available on CTC workstations. CAD algorithms can be used as a “first reader,” as a “concurrent reader,” or as a “second reader.” In a first-reader paradigm, the observer only ever reads with CAD-assistance and confines his/her attention to CAD marks alone, ignoring unprompted areas of the colon. However, aside from ethical issues, such a separation is likely to come with an increased number of missed abnormalities [66]. If CAD is used as a concurrent reader, pre-calculated CAD results are displayed in the image data while performing the primary data evaluation. In the second reader scheme, a reader reviews the case, first blinded to the CAD results, and then evaluates the CAD findings to integrate them with his/her own findings. There are relatively little data about the potential benefit of CAD on reader performance. The “potential” contribution of a CAD algorithm to the sensitivity of different readers without CAD was reportedly between 10% and 38% [67,68].

5. Findings

Findings of the colon are characterized by their morphology, by their attenuation characteristics, and by their mobility [69,70]. One of the most common findings detected with CTC is diverticular disease. On 2D CTC images diverticula appear as air-filled outpouchings of the colonic wall. On the 3D virtual endoscopic images, the diverticular orifice can be recognized as a complete dark ring [71] (Fig. 6). Due to the increased risk of perforation CTC has no role in the diagnosis of acute diverticulitis. However, in chronic stages of diverticulitis, CTC may show cone-shaped stenosis with mild wall thickening with involve- ment of a long segment (>10 cm) with pericolic fat stranding, and fluid at the root of the mesentery. Polyps are the most common benign lesions of the colon. The risk of malignant transformation increases with the size of the polyp. Most polyps are sessile although some have a stalk. On 3D virtual endoscopic images, polypoid lesions present as a sessile or stalked, round, oval, or lobulated intraluminal filling defect. Typically, the margin to the normal mucosa is Fig. 5. Colon polyp versus residual stool with translucency view. (a) 3D virtual displayed as an incomplete ring shadow. On 2D plane images, endoscopic CT image shows a polypoid filling defect between two haustral polyps have homogenous, soft tissue attenuation (Fig. 7) [69]. folds (arrow). Translucency view shows a homogenous, uniform, concentric Generally, polyps maintain their position with respect to the ring pattern with a homogenous red core revealing the polyp. (b) 3D virtual endoscopic CT image shows a polypoid filling defect in the transverse colon bowel surface, with the exception of stalked polyps or polyps (arrow). Translucency view shows an inhomogeneous pattern because of small in mobile colon segments [72]. CTC is not able to reliably gas bubbles, revealing untagged stool. distinguish between hyperplastic and adenomatous polyps using morphological features alone, although 50% of polyps 4.2. Role of CAD below 5 mm are hyperplastic [73]. Lipomas are the most common submucosal lesions in the Computer-aided detection (CAD) systems are software pro- colon (especially common on the ileocecal valve). CT colonog- grams that automatically highlight polyp “candidates” and raphy allows confident diagnosis of lipomas based on their therefore support the radiologist by pointing out possible abnor- characteristic fatty attenuation. A lipoma, in general, is 1–3 cm in malities that may otherwise have been missed. Based on size and rarely exceeds 4 cm. On 3D virtual endoscopic images, morphologic and attenuation characteristics, the reader then lipomas are present as a sessile or pedunculated polypoid intra- decides whether the “candidate lesion” is a true- or a false- luminal filling defect, most often with a smooth surface. On T. Mang et al. / European Journal of Radiology 61 (2007) 388–399 395

Fig. 6. Diverticula in the sigmoid colon. (a) 3D virtual endoscopic CT image Fig. 7. Large, 2.6 cm stalked polyp and 1.1 cm sessile polyp in the sigmoid shows a complete dark rings at the diverticula (arrows). (b) Prone 2D axial CT colon. (a) 3D virtual endoscopic CT image shows a large polypoid filling defect image shows multiple gas-filled outpouchings of the colon wall. with a stalk (arrow) adjacent to a smaller, sessile, round, and smooth filling defect (arrowhead). (b) Supine 2D coronal CT image shows a homogenous soft tissue attenuation of the two intraluminal lesions (arrow, arrowhead). Histology 2D plane images, lipomas show a homogenous fatty attenua- revealed two adenomas of the sigmoid colon. tion (Fig. 8). Because lipomas are soft lesions, their shape may change when moving from the prone to the supine position. been a topic of controversy, and recent data from the National In general, small lipomas need no further treatment; only large Polyp Study indicate that the flat lesions that were found are lipomas require endoscopic resection because they can lead to not more aggressive than sessile or stalked polyps [81]. In CTC, intusseption [69]. flat polyps appear as a fairly circumscribed area of mild wall Whereas truly flat adenomas, not arising over the mucosal thickening with homogenous soft tissue attenuation. Sometimes level, are almost impossible to detect at CTC, flat lesions can a mild nodularity is found on the surface by 3D endoluminal be depicted if they are slightly raised over the colonic mucosa images (Fig. 9). level. Flat polyps are defined as lesions with a height less than Colorectal cancer is the most common colonic primary 50% of the lesion width [74]. Their clinical importance has tumour. Most carcinomas show an exophytic, polypous type of 396 T. Mang et al. / European Journal of Radiology 61 (2007) 388–399

Fig. 8. Large flat adenoma of the transverse colon. (a) 3D virtual endoscopic Fig. 9. Large, pedunculated lipoma of the ascending colon. (a) 3D virtual endo- CT image shows a flat filling defect of the colon adjacent to a haustral fold scopic CT image shows a a large polypoid filling defect with a smooth surface (arrow) (b) Prone 2D axial CT image shows a slight focal wall thickening with adjacent to a haustral fold (arrow). (b) Prone 2D axial CT image shows a homogenous soft tissue attenuation. Histology revealed a villous adenoma of submucosal lesion with a homogenous fatty attenuation, revealing a lipoma. the transverse colon. pericolic fat tissue, is readily appreciated by CT. Tumor infiltra- growth with frequent central degeneration. Adenocarcinomas tion to adjacent organs (T4) is most likely if the carcinoma shows tend to infiltrate the bowel wall circumferentially and 50% are a broad-based contact, no intervening fat planes, and indistinct found in the rectum, and 25% in the sigmoid. Colorectal cancer boundaries to other organs. Pericolonic lymph nodes and dis- typically shows extensive focal polypoid, asymmetric, or cir- tant metastases are signs of progression of the disease and can cular wall thickening with short extension (<5 cm), especially be evaluated using 2D axial source images and MPR views. with shoulder formation (Fig. 10) [71,75]. CT differentiation Since the whole abdomen and pelvis is scanned, extracolonic between stages T1 (invasion of mucosa and/or submucosa) and structures can be evaluated. The incidence of extracolonic find- T2 (invasion of the muscularis propria) is not feasible, but ings has been reported to be high (69%), but only about 10% are tumor extension beyond the colon wall (T3), characterized by fat of high clinical importance, such as large aneurysms, masses, stranding, an indistinct boundary, and nodular protrusions into lung nodules, and large lymph nodes (Fig. 10b) [76,77]. Elderly T. Mang et al. / European Journal of Radiology 61 (2007) 388–399 397

studies, which require adequate bowel cleansing and distention, thin-section MDCT scans performed in the prone and supine positions, and a combined evaluation of 2D and 3D images. CTC is widely accepted as a powerful tool for evaluation of the colon after incomplete colonoscopy, or in patients with known or suspected colorectal cancer. The role of CTC in a screening population remains to be determined, with encouraging results published in the last two years. Ongoing research in CAD and improved software products have the potential to increase the accuracy and ease of interpretation of CTC, which would accel- erate its acceptance as a colorectal screening tool.

References

[1] Cotton PB, Durkalski VL, Pineau BC, et al. Computed tomographic colonography (virtual colonoscopy): a multicenter comparison with standard colonoscopy for detection of colorectal neoplasia. JAMA 2004;291:1713–9. [2] Johnson CD, Harmsen WS, Wilson LA, et al. Prospective blinded evaluation of computed tomographic colonography for screen detection of colorectal polyps. Gastroenterology 2003;125:311–9. [3] 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;349:2191–200. [4] Rockey DC, Paulson E, Niedzwiecki D, et al. Analysis of air contrast barium enema, computed tomographic colonography, and colonoscopy: prospective comparison. Lancet 2005;365:305–11. [5] Ferrucci JT. Colonoscopy: virtual and optical—another look, another view. Radiology 2005;235:13–6. [6] Halligan S, Atkin W. Unbiased studies are needed before virtual colonoscopy can be dismissed. Lancet 2005;365:275–6. [7] Fenlon HM, McAneny DB, Nunes DP, Clarke PD, Ferrucci JT. Occlusive colon carcinoma: virtual colonoscopy in the preoperative evaluation of the proximal colon. Radiology 1999;210:423–8. [8] Macari M, Berman P, Dicker M, Milano A, Megibow AJ. Usefulness of CT colonography in patients with incomplete colonoscopy. AJR Am J Roentgenol 1999;173:561–4. [9] Morrin MM, Farrell RJ, Raptopoulos V, McGee JB, Bleday R, Kruskal JB. Role of virtual computed tomographic colonography in patients with colorectal cancers and obstructing colorectal lesions. Dis Colon Rectum 2000;43:303–11. [10] Morrin MM, Kruskal JB, Farrell RJ, Goldberg SN, McGee JB, Rap- topoulos V.Endoluminal CT colonography after an incomplete endoscopic colonoscopy. AJR Am J Roentgenol 1999;172:913–8. [11] Neri E, Giusti P, Battolla L, et al. Colorectal cancer: role of CT colonography in preoperative evaluation after incomplete colonoscopy. Radiology 2002;223:615–9. Fig. 10. Stenotic cancer of the transverse colon. (a) 3D endoluminal image [12] Fletcher JG, Johnson CD, Krueger WR, et al. Contrast-enhanced CT shows an irregular, circular, stenotic filling defect (arrow). (b) Supine sagittal colonography in recurrent colorectal carcinoma: feasibility of simultaneous contrast-enhanced CT image shows circular wall thickening and tumor extension evaluation for metastatic disease, local recurrence, and metachronous neo- beyond the colon wall (T3) (arrow). Note: liver metastasis as an extracolonic plasia in colorectal carcinoma. AJR Am J Roentgenol 2002;178:283–90. finding (arrowhead). [13] Laghi A, Iannaccone R, Bria E, et al. Contrast-enhanced computed tomographic colonography in the follow-up of colorectal cancer patients: a and symptomatic patients are more likely to have significant feasibility study. Eur Radiol 2003;13:883–9. extracolonic findings compared to a younger screening popu- [14] Laghi A. Virtual colonoscopy: clinical application. Eur Radiol lation. It has been noted that the lack of IV contrast, and the 2005;15(Suppl 4):D138–41. [15] Ferrucci JT. Double-contrast barium enema: use in practice and implica- low-dose technique used for VC, limit the evaluation of extra- tions for CT colonography. AJR Am J Roentgenol 2006;187:170–3. colonic findings [78]. [16] Winawer SJ, Stewart ET, Zauber AG, et al. A comparison of colonoscopy and double-contrast barium enema for surveillance after polypectomy. 6. Conclusion National Polyp Study Work Group. N Engl J Med 2000;342:1766–72. [17] Halligan S, Altman DG, Taylor SA, et al. CT colonography in the detection of colorectal polyps and cancer: systematic review, meta-analysis, In summary, CTC is a valuable diagnostic tool for evaluation and proposed minimum data set for study level reporting. Radiology of the entire colon. Accurate diagnosis relies on high quality 2005;237:893–904. 398 T. Mang et al. / European Journal of Radiology 61 (2007) 388–399

[18] Gluecker TM, Johnson CD, Harmsen WS, et al. Colorectal cancer screening [41] Callstrom MR, Johnson CD, Fletcher JG, et al. CT colonography without with CT colonography, colonoscopy, and double-contrast barium enema cathartic preparation: feasibility study. Radiology 2001;219:693–8. examination: prospective assessment of patient perceptions and prefer- [42] Zalis ME, Hahn PF. Digital subtraction bowel cleansing in CT colonogra- ences. Radiology 2003;227:378–84. phy. AJR Am J Roentgenol 2001;176:646–8. [19] Chung DJ, Huh KC, Choi WJ, Kim JK. CT colonography using 16- [43] Dachman AH. Advice for optimizing colonic distention and minimizing MDCT in the evaluation of colorectal cancer. AJR Am J Roentgenol risk of perforation during CT colonography. Radiology 2006;239:317–21. 2005;184:98–103. [44] Taylor SA, Halligan S, Goh V, et al. Optimizing colonic distention for [20] Filippone A, Ambrosini R, Fuschi M, Marinelli T, Genovesi D, Bonomo multi-detector row CT colonography: effect of hyoscine butylbromide and L. Preoperative T and N staging of colorectal cancer: accuracy of rectal balloon catheter. Radiology 2003;229:99–108. contrast-enhanced multi-detector row CT colonography—initial experi- [45] Burling D, Taylor SA, Halligan S, et al. Automated insufflation of carbon ence. Radiology 2004;231:83–90. dioxide for MDCT colonography: distension and patient experience com- [21] Iannaccone R, Laghi A, Passariello R. Colorectal carcinoma: detection pared with manual insufflation. AJR Am J Roentgenol 2006;186:96–103. and staging with multislice CT (MSCT) colonography. Abdom Imaging [46] Shinners TJ, Pickhardt PJ, Taylor AJ, Jones DA, Olsen CH. Patient- 2005;30:13–9. controlled room air insufflation versus automated carbon dioxide delivery [22] Macari M, Bini EJ. CT colonography: where have we been and where are for CT colonography. AJR Am J Roentgenol 2006;186:1491–6. we going? Radiology 2005;237:819–33. [47] Rogalla P, Lembcke A, Ruckert JC, et al. Spasmolysis at CT colonography: [23] Taylor SA, Halligan S, Saunders BP, et al. Use of multidetector-row CT butyl scopolamine versus glucagon. Radiology 2005;236:184–8. colonography for detection of colorectal neoplasia in patients referred [48] Yee J, Hung RK, Akerkar GA, Wall SD. The usefulness of glucagon via the Department of Health “2-week-wait” initiative. Clin Radiol hydrochloride for colonic distention in CT colonography. AJR Am J 2003;58:855–61. Roentgenol 1999;173:169–72. [24] Lefere P, Gryspeerdt S, Baekelandt M, Dewyspelaere J, van Holsbeeck [49] Chen SC, Lu DS, Hecht JR, Kadell BM. CT colonography: value of B. Diverticular disease in CT colonography. Eur Radiol 2003;13(Suppl scanning in both the supine and prone positions. AJR Am J Roentgenol 4):L62–74. 1999;172:595–9. [25] Biancone L, Fiori R, Tosti C, et al. Virtual colonoscopy compared with con- [50] Yee J, Kumar NN, Hung RK, Akerkar GA, Kumar PR, Wall SD. Com- ventional colonoscopy for stricturing postoperative recurrence in Crohn’s parison of supine and prone scanning separately and in combination at CT disease. Inflamm Bowel Dis 2003;9:343–50. colonography. Radiology 2003;226:653–61. [26] Tarjan Z, Zagoni T, Gyorke T, Mester A, Karlinger K, Mako EK. Spiral CT [51] Macari M, Bini EJ, Xue X, et al. Colorectal neoplasms: prospective com- colonography in inflammatory bowel disease. Eur J Radiol 2000;35:193–8. parison of thin-section low-dose multi-detector row CT colonography and [27] Ota Y, Matsui T, Ono H, et al. Value of virtual computed tomographic conventional colonoscopy for detection. Radiology 2002;224:383–92. colonography for Crohn’s colitis: comparison with endoscopy and barium [52] Iannaccone R, Laghi A, Catalano C, Mangiapane F, Piacentini F, Pas- enema. Abdom Imaging 2003;28:778–83. sariello R. Feasibility of ultra-low-dose multislice CT colonography for [28] Svensson MH, Svensson E, Lasson A, Hellstrom M. Patient acceptance of the detection of colorectal lesions: preliminary experience. Eur Radiol CT colonography and conventional colonoscopy: prospective comparative 2003;13:1297–302. study in patients with or suspected of having colorectal disease. Radiology [53] Jensch S, van Gelder RE, Venema HW, et al. Effective radiation doses in 2002;222:337–45. CT colonography: results of an inventory among research institutions. Eur [29] van Gelder RE, Birnie E, Florie J, et al. CT colonography and colonoscopy: Radiol 2006;16:981–7. assessment of patient preference in a 5-week follow-up study. Radiology [54] Morrin MM, Farrell RJ, Kruskal JB, Reynolds K, McGee JB, Raptopoulos 2004;233:328–37. V. Utility of intravenously administered contrast material at CT colonog- [30] Smith RA, Cokkinides V, Eyre HJ. American Cancer Society guidelines raphy. Radiology 2000;217:765–71. for the early detection of cancer, 2003. CA Cancer J Clin 2003;53:27–43. [55] Oto A, Gelebek V, Oguz BS, et al. CT attenuation of colorectal poly- [31] Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and poid lesions: evaluation of contrast enhancement in CT colonography. Eur surveillance: clinical guidelines and rationale-update based on new evi- Radiol 2003;13:1657–63. dence. Gastroenterology 2003;124:544–60. [56] Dachman AH, Kuniyoshi JK, Boyle CM, et al. CT colonography with three- [32] Barish MA, Soto JA, Ferrucci JT. Consensus on current clinical practice of dimensional problem solving for detection of colonic polyps. AJR Am J virtual colonoscopy. AJR Am J Roentgenol 2005;184:786–92. Roentgenol 1998;171:989–95. [33] Burling D, Halligan S, Slater A, Noakes MJ, Taylor SA. Potentially seri- [57] Royster AP, Fenlon HM, Clarke PD, Nunes DP, Ferrucci JT. CT ous adverse events at CT colonography in symptomatic patients: national colonoscopy of colorectal neoplasms: two-dimensional and three- survey of the United Kingdom. Radiology 2006;239:464–71. dimensional virtual-reality techniques with colonoscopic correlation. AJR [34] Pickhardt PJ. Incidence of colonic perforation at CT colonography: review Am J Roentgenol 1997;169:1237–42. of existing data and implications for screening of asymptomatic adults. [58] Macari M, Milano A, Lavelle M, Berman P, Megibow AJ. Compari- Radiology 2006;239:313–6. son of time-efficient CT colonography with two- and three-dimensional [35] Sosna J, Blachar A, Amitai M, et al. Colonic perforation at CT colonog- colonic evaluation for detecting colorectal polyps. AJR Am J Roentgenol raphy: assessment of risk in a multicenter large cohort. Radiology 2000;174:1543–9. 2006;239:457–63. [59] Zalis ME, Barish MA, Choi JR, et al. CT colonography reporting and data [36] Macari M, Lavelle M, Pedrosa I, et al. Effect of different bowel preparations system: a consensus proposal. Radiology 2005;236:3–9. on residual fluid at CT colonography. Radiology 2001;218:274–7. [60] Juchems MS, Fleiter TR, Pauls S, Schmidt SA, Brambs HJ, Aschoff AJ. [37] Fass R, Do S, Hixson LJ. Fatal hyperphosphatemia following Fleet Phospo- CT colonography: comparison of a colon dissection display versus 3D Soda in a patient with colonic ileus. Am J Gastroenterol 1993;88:929–32. endoluminal view for the detection of polyps. Eur Radiol 2006;16:68–72. [38] Taylor SA, Halligan S, Goh V, Morley S, Atkin W, Bartram CI. Optimiz- [61] Vos FM, van Gelder RE, Serlie IW, et al. Three-dimensional display modes ing bowel preparation for multidetector row CT colonography: effect of for CT colonography: conventional 3D virtual colonoscopy versus unfolded Citramag and Picolax. Clin Radiol 2003;58:723–32. cube projection. Radiology 2003;228:878–85. [39] Lefere PA, Gryspeerdt SS, Dewyspelaere J, Baekelandt M, Van Holsbeeck [62] Geiger B, Chefd’hotel C, Sudarsky S. Panoramic views for virtual BG. Dietary fecal tagging as a cleansing method before CT colonog- endoscopy. Med Image Comput Comput Assist Interv Int Conf Med Image raphy: initial results polyp detection and patient acceptance. Radiology Comput Comput Assist Interv 2005;8:662–9. 2002;224:393–403. [63] Hoppe H, Quattropani C, Spreng A, Mattich J, Netzer P, Dinkel HP. Virtual [40] Iannaccone R, Laghi A, Catalano C, et al. Computed tomographic colonog- colon dissection with CT colonography compared with axial interpretation raphy without cathartic preparation for the detection of colorectal polyps. and conventional colonoscopy: preliminary results. AJR Am J Roentgenol Gastroenterology 2004;127:1300–11. 2004;182:1151–8. T. Mang et al. / European Journal of Radiology 61 (2007) 388–399 399

[64] Pickhardt PJ. Translucency rendering in 3D endoluminal CT colonography: [73] Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screen- a useful tool for increasing polyp specificity and decreasing interpretation ing: clinical guidelines and rationale. Gastroenterology 1997;112:594– time. AJR Am J Roentgenol 2004;183:429–36. 642. [65] Summers RM, Yao J, Pickhardt PJ, et al. Computed tomographic virtual [74] Rembacken BJ, Fujii T, Cairns A, et al. Flat and depressed colonic neo- colonoscopy computer-aided polyp detection in a screening population. plasms: a prospective study of 1000 colonoscopies in the UK. Lancet Gastroenterology 2005;129:1832–44. 2000;355:1211–4. [66] Yoshida H, Dachman AH. CAD techniques, challenges, and controversies [75] Taylor SA, Halligan S, Bartram CI. CT colonography: methods, pathology in computed tomographic colonography. Abdom Imaging 2005;30:26–41. and pitfalls. Clin Radiol 2003;58:179–90. [67] Summers RM, Jerebko AK, Franaszek M, Malley JD, Johnson CD. Colonic [76] Gluecker TM, Johnson CD, Wilson LA, et al. Extracolonic findings at CT polyps: complementary role of computer-aided detection in CT colonog- colonography: evaluation of prevalence and cost in a screening population. raphy. Radiology 2002;225:391–9. Gastroenterology 2003;124:911–6. [68] Taylor SA, Halligan S, Burling D, et al. Computer-assisted reader software [77] Hara AK, Johnson CD, MacCarty RL, Welch TJ. Incidental extracolonic versus expert reviewers for polyp detection on CT colonography. AJR Am findings at CT colonography. Radiology 2000;215:353–7. J Roentgenol 2006;186:696–702. [78] Pickhardt PJ. Virtual colonoscopy: issues related to primary screening. Eur [69] Macari M, Bini EJ, Jacobs SL, Lange N, Lui YW. Filling defects at Radiol 2005;15(Suppl. 4):D133–7. CT colonography: pseudo- and diminutive lesions (the good), polyps [79] Johnson CD, MacCarty RL, Welch TJ, et al. Comparison of the relative (the bad), flat lesions, masses, and carcinomas (the ugly). Radiographics sensitivity of CT colonography and double-contrast barium enema for 2003;23:1073–91. screen detection of colorectal polyps. Clin Gastroenterol Hepatol 2004;2: [70] Pickhardt PJ. Differential diagnosis of polypoid lesions seen at CT colonog- 314–21. raphy (virtual colonoscopy). Radiographics 2004;24:1535–56, discussion [80] Graser A, Wintersperger BJ, Suess C, Reiser MF, Becker CR. Dose 1557-1539. reduction and image quality in MDCT colonography using tube current [71] Fenlon HM, Clarke PD, Ferrucci JT. Virtual colonoscopy: imaging features modulation. AJR Am J Roentgenol 2006;187:695–701. with colonoscopic correlation. AJR Am J Roentgenol 1998;170:1303–9. [81] O’Brien MJ, Winawer SJ, Zauber AG, et al. Flat adenomas in the National [72] Laks S, Macari M, Bini EJ. Positional change in colon polyps at CT Polyp Study: is there increased risk for high-grade dysplasia initially or colonography. Radiology 2004;231:761–6. during surveillance? Clin Gastroenterol Hepatol 2004;2:905–11.