Colon: Neoplastic Disease: Virtual Imaging of the Colon: Current Status

Karen M. Horton, MD

Introduction

CT colonography (CTC) is considered by many to be a potential breakthrough in the screening for colorectal polyps and cancer. This technique combines the volume acquisition of CT data, using single or multidetector spiral CT, with a variety of two-dimensional and three-dimensional visualization techniques. The viewing technique which has gained widespread attention is the simulated endoluminal view of the colon (virtual colonoscopy). This method allows the colon to be viewed from the inside, similar to the view through an endoscope as used by the gastroenterologists. Although this method of display can be useful, it is time consuming and computer intensive. In addition, most radiologists are not familiar with this endoluminal perspective, and will thus require considerable training. We have found that in our practice, it is often necessary to utilize a variety of imaging display techniques to best visualize the anatomy and pathology. This lecture reviews the display techniques we utilize when performing CTC.

Technique

Patient Preparation

CTC requires bowel cleansing for optimal results. Most centers use 1 gallon of GoLytely the evening before the study. Other preparations such as Fleets prep can be utilized. Immediately before the exam, an enema tip is inserted into the rectum and either room air or carbon dioxide is instilled to maximum patient tolerance (1-2L). Both prone and supine imaging is necessary. Bowel relaxants such as Glucagon can be administered if necessary. However, a recent study of 60 patients by Yee et al showed than IV Glucagon did not significantly improved colonic distention. Further investigation is necessary.

Scanning Parameters

There are many different protocols being investigated for CT colonography. Most are dependent on the type of scanner and its capabilities. The optimal technique has not yet been determined. With helical scanners, investigators have utilized 3-5mm collimation with a pitch of 1.3-2.0. This technique usually requires a 40-50 second breath hold. We currently have the Siemens Volume Zoom Multidetector Scanner. With this scanner we can obtain 1.25 mm slices by utilizing the 4 X 1 mm collimator configuration. Prone and supine imaging should be obtained, to allow visualization of the entire mucosa. In one study, by Chen et al, 59% of the scans have adequate distention for polyp detection using either the prone or supine acquisition, compared with 87% having adequate distention using a combination of the supine and prone acquisition together.

Data rendering

Data rendering can be performed using shaded surface or volume rendering techniques. Shaded surface technique require less computer power, but are limited by threshold settings. Volume rendering techniques allow interactive manipulation of image thresholds and are considered superior to shaded surface. For endoluminal imaging, a central axis can be created which allows simply navigation through the colon. Volume rendering has been found to be superior to shaded surface for the detection and characterization of poyps.

Our Technique

• Scanner: Siemens Somatom Plus 4 Volume Zoom scanner (Siemens Medical Systems, Iselin , N.J.) with 8 detectors.
• kVp 120
• mA 180
• collimation 4 X 1 mm
• slice thickness 1.25mm
• reconstruction interval 1-2mm
• breath hold approximately 25 seconds

All scan data is then transferred to a free standing Silicon Graphics Onyx Infinite Reality workstation (Silicon Graphics, Mountain View, California) running 3D Virtuoso software (Siemens Medical Systems, Iselin, N.J.). All images are reviewed using the real time viewing and 3D volume rendering technique.

Image Display

There has been an evolution in 3D rendering techniques for CTC over the last several years. Initially, CTC was performed using surface-rendering techniques. However, due to the inherent loss of source data with this method, volume rendering is now considered to be the rendering technique of choice . Although this method requires more computer resources, it has a definite advantage over surface rendering by allowing the entire range of attenuation values within a volume set to be displayed. We use volume-rendering exclusively for CTC.

With volume rendering, parameters can be applied to the volume set to affect the appearance of the data in order to best demonstrate anatomy and pathology. These parameters include window width, level, opacity and brightness and can be adjusted interactively by the user. The window width and level functions are similar to the windowing settings on standard CT scanners . For instance, the window setting can be adjusted to display the colonic air or soft tissue. Opacity refers to the degree to which structures obscure structures behind them. This can be varied from 0-100%. When set at 100% opacity, the appearance of the colon is similar to surface rendering. When low opacity setting are chosen, the colon appears "see through". Brightness affects the appearance of the image by scaling the value of every pixel without changing the apparent diameter of the viewed structures. It can be varied from 0-100%. Selection of brightness setting are largely subjective and are based on individual user preference.

When performing CTC using volume rendering, the user is able to adjust any of these parameters interactively in real time. This interactivity allows the user to customize the parameters for each individual case in order to optimize the display of anatomy and pathology. The following is a description of 4 image displays which we use routinely when performing CTC.

Simulated Double Contrast Enema

The display parameters can be manipulated to obtain an edge enhanced view of the colon, which simulates an air contrast barium enema. In order to create a double contrast effect, a percentage classifier is necessary. A trapezoid is created, and only the attenuation values under the trapezoid are included in the image. Attenuation values lying outside the trapezoid appear black . Small polyps can also be detected with this display method and appear as a filling defect or ring shadow, as on conventional double contrast enema.

Simulated Single Contrast Enema

The display parameters can be manipulated to obtain an image which simulates an single contrast enema. This is achieved by creating a negative window level to accentuate the air- filled colon (approximately —350) and a negative window width (approximately —500). The negative window width displays air as white. Therefore, the air-filled colon will appear similar to a single contrast enema. As with conventional single contrast enemas, this display method is especially valuable in patients with strictures or malignant narrowings. As with conventional single contrast enema, small polyps appear as filling defects on a CT simulated single contrast enema. Unlike a conventional single contrast enema where small polyps may not be detected if they are dependent and within the contrast pool, small polyps are still visible on the 3D images, because cut planes can scroll through the volume.

Computer Enhanced Image Display

This display technique accentuates the colonic wall and folds. In can be created by setting the window level at negative 200-300, with a negative window width of approximately 1000. The opacity is set at around 60% and the brightness can be adjusted to according to user preference. With this display method, extracolonic structures appear white, thus highlighting the colon. Polyps and masses are relatively easy to detect with these settings. However, regions of stricture or malignant luminal narrowing may be overlooked if this is the only display method utilized.

Simulated Soft Tissue Window

When a polyp or mass is detected, the display parameters can be manipulated in order to evaluate possible extraluminal extension or regional adenopathy. This is unique to volume rendering and is not possible with surface rendering techniques. To create this display, the window level and width are set similar to soft tissue windows used with routine abdominal/pelvic (i.e. window level = 400, window width = 10). This display method allows visualization of the pericolonic fat and vessels which is essential when evaluating for local tumor extension. In addition this display method is extremely helpful when a colon mass/polyp is detected. Unlike with surface rendering techniques in which a polyp and adherent stool appear identical, volume rendering with the simulated soft tissue window displays all of the attenuation values within the lesion, thus allowing differentiation between a true polyp and stool.

Radiation Dose

Radiation dose is an important issue if CT is considered for widespread screening. In a study by Johnson et al in 1997 using phantoms with simulated polyps at 4 different ma settings (70,140,210,280), radiation dose was assessed. The images obtained at the 70 mA level were slightly degraded compared with other settings, but no difference was found in polyp detection rates. Using 70mA, the effective dose equivalent was 187mrem (males) and 285 mrem (females) which is approximately 505 of the dose of a barium enema, and a 75% reduction in radiation to patients compared to standard CT abdominal examination (280mA). Most investigators uses mA between 100-200.

Results

Many studies have been carried out using different scanning techniques, different software and computer display and real vs. simulated polyps. In a study of 30 endoscopically proven polyps by Hara et al, a combination of 2D and 3D viewing detected 100% of polyps greeter than 1 cm, 71% of polyps between 0.5 and 0.9cm and 28% polyps less than 0.5cm. In a study of 20 proven masses by Royster et al, all were identified on 2D CT colonography and only 19/20 were visualized on the 3D virtual colonoscopy. However, there were 2 false positive using the 2D, and no false positives using the 3D. In the same study, 2D CT colonography detected 14/15 polyps, with 3 false positive. 3D virtual colonoscopy detected 13/15 polyps with no false positives. In a blinded prospective study of 70 patients by Johnson at al, the sensitivity of detection of polyps of 1 cm or more was 75% and the specificity is 90%. In a recent study published by Fenlon et al, comparison was made between virtual and conventionalcolonoscopy in a high risk group of 100 patients. Virtual colonoscopy identified all 3 cancers and 20/22 polyps that were 10mm or more in diameter (91%). There were 19 false postivies. Results were less accurate for smaller polyps 33/40 for polyps 6-9mm, and 29/53 for polyps 5mm or smaller.

The main limitations to using these techniques include, retained colonic fluid and feces, underdistended segments, and long interpretation times.

Most investigators use a combination of 2D and 3D displays and find them complementary. However several recent studies suggest that the 2D images alone, may be adequate for lesion detection and would therefore decreased the time necessary for evlaution.

Cost Effectiveness

Is CT colonography a cost —effective option for colorectal cancer screening? In a study by Sonneberg et al, a hypothetical population of 100,000 subjects aged 50 undergoes a screening procedure (CT colonography vs. conventional colonoscopy) every 10 years. Suspicious findings of CT colonography are worked up by conventional colonoscopy. After polypectomy, conventional colonoscopy is repeated every 3 years until no adenomatous polyps are detected. In the analysis, the cost of CT colonoscopy was $478, conventional colonoscopy $728, and colonoscopy with polypectomy $1,139. Computer analysis by the Markov process revealed that screening by CT colonography cost $24,586 per-life saved event, compared with $20,930 spent on screening conventional colonoscopy. For the two screening procedure to become similarly cost-effective, CT colonography would have result in a 15-20% better compliance rate than conventional colonoscopy or a procedural cost 54% less than conventional colonoscopy.

Future Directions

Despite marked technical advancement in the last several years, standardization and validation of CT colonography techniques in necessary before widespread implementation. Groups are currently working on "intelligent soft ware" and artifical intelligence to potentially triage the data-set and identifiy only suspicious areas for review. This may increase accuracy and reduce exam time. Groups are also working on perfecting the bowel prep, including stoll tagging regimes.

Conclusions

Although CT colonography is still a very new technique, initial reports and results for its use in colorectal polyp detection is encouraging. Its exact role in the widespread screening for colorectal cancer will not be determined for years. It is yet to be determined if this technique will be cost —effective.

References

Calhoun PS, Kuszyk BS, Heath DG, Carley JC, Fishman EK. Three-dimensional volume rendering of spiral CT data: theory and method. Radiographics 1999;19:745-764.

Fenlon HM, Nunes DP, Schroy PC, Barish MA, Clarke PD, et al. A comparison of virtual and conventional colonoscopy for the detection of colorectal polyps. N Eng J Med 1999;341:20.

Felon HM, Calrke PD, Ferucci JT. Virtual colonoscopy: imaging features with colonoscopic corrrelation. AJR 1998;5:282-288.

Hara A, Johnson D, MacCarty R, et al. CT colonography: single vs. multi-detector row imaging. Radiology 2001:219:461-465

Hopper KD, Iyriboz AT, Wise SW, Neuman JD, Mauger DT, Kasales CJ. Mucosal detail at CT virtual reality: surface versus volume rendering. Radiology 1000;214:517-522.

Johnson CD, Hara AK, Reed JE. Computed tomographic colonography (virtual colonoscopy): a new method for detecting colorectal neoplasms. Endoscopy 1997;29:454-461.

Kay CL, Evangelou HA. A review of the technical and clinical aspects of virtual endoscopy. Endoscopy 1996;28:768-775.

Macari M, Milano A, Lavelle M, Berman P, Megibow AJ. Comparison of time-efficient CT colonography with two- and three-dimensional colonic evaluation for detecting colorectal polyps. AJR 2000;174:1543-1549.

McFarland EG, Brink JA, Pilgrim TK, et al. Spiral CT colonography: reader agreement and diagnostic performance with two- and three-dimensional image-display techniques. Radiology 2001;218:375-383.

McFarland EG, Brink JA. Helical CT colonography (virtual colonoscopy): the challenge that exists between advancing technology and generalizability. AJR 1999;173:549-559.

Paik DS, Beaulieu CF, Jeffrey RB, Karadi C, Napel S. Detection of polyps in CT colonoscopy: a comparison of a computer-aided detection algorithm to 3D visualization methods. 1999 RSNA Conference in Chicago, IL.

Royster AP, Gupta AK, Felon HM, Ferrucci JT. Virtual colonoscopy: current status and future implications. Acad Radiol 1998;5:282-288.

Tarjan Z, Zagoni T, Gyorke T, et al. Spiral CT colonography in inflammatory bowel disease. Eur J Radiol 2000;35:193-198

Vining DJ. Virtual endoscopy: is it reality? (editorial). Radiology 1996;200:30-31.