Diagnostic Imaging
October 2003

COVER STORY

Virtual colonoscopy

Technology makes inroads into imaging practice

Since its introduction by Vining and Gelfand in 1994,1 virtual colonoscopy has been written about extensively in peer-reviewed journals. Rather than review this topic in its entirety, this article will describe the rationale and methodology for introducing virtual colonoscopy into routine clinical practice. Articles published recently by Dachman and Yoshida2 and Hara3 offer comprehensive reviews of this topic.

Concern about cancer of the colon and rectum has yet to reach the level of the public's healthy fear of dietary fat and high cholesterol levels as the precursors of coronary artery disease, and worry about breast and lung cancer. Colorectal cancer is the third most common cause of cancer, however, and the second leading cause of cancer death in the U.S.

Most experts concur that the majority of colon cancers, as many as 95%, begin as adenomas of the colonic mucosa. In asymptomatic patients over the age of 50, the prevalence of adenomas identified by sigmoidoscopy is 10%.4 Colonoscopy studies have demonstrated a higher prevalence rate, reaching as high as 47% in asymptomatic individuals in some studies. Men have a 1.5 relative risk of adenomas compared with women. Age, male gender, and first-degree family history of colorectal cancer are risk factors for adenomas.5

Studies have demonstrated that adenomas grow slowly. Most polyps less than 1 cm in size will be fairly stable over a three-year period. It takes two to three years for an adenoma 5 mm or less to grow to 1 cm, and an additional two to five years for that adenoma to develop into a carcinoma. The probability of cancer grows with increasing size of the polyp. A polyp lesser than or equal to 5 mm in size has almost no chance of being malignant, and those 1 cm in size have only a 1% chance of harboring malignancy.

Screening for colorectal cancer is critical, because colorectal cancer, unlike breast cancer for example, is largely preventable. All available data point to the conclusion that the early detection and removal of colonic adenomas markedly reduces mortality from this disease. These data raise two important questions: Why aren't vast numbers of people in the U.S. being screened? In addition to the traditional options already available, do we need another screening test for colorectal cancer?

Let's look at the second question first. Several options already exist for screening for colorectal cancer: fecal occult blood testing (FOBT), sigmoidoscopy, barium enema, and optical colonoscopy. While FOBT has been proven to decrease the mortality from colorectal cancer, its effectiveness is limited and requires strict adherence to the testing regimen, something accomplished by few individuals. Since polyps bleed intermittently, FOBT has limited sensitivity, and because the blood detected can have myriad causes, limited specificity. Sigmoidoscopy, which has also been shown to reduce mortality from colorectal cancer, examines less than half of the colon and therefore misses 50% of lesions. Studies have also documented that sigmoidoscopy misses 20% of adenomas within the portion it does evaluate.6

We need not dwell on the barium enema. Although studies in the radiology literature have documented that it can detect a high percentage of larger adenomas, an article in The New England Journal of Medicine demonstrated only a 50% sensitivity for polyps greater than 1 cm.7 The barium enema can be difficult for both the patient and the radiologist, and the number of radiologists skilled in this procedure is diminishing.

NOT QUITE GOLD

Conventional colonoscopy, otherwise known as optical or video-assisted colonoscopy, is an excellent diagnostic tool. It theoretically allows evaluation of the entire colon and offers a therapeutic advantage as well in that the examining physician can remove those polyps identified at the time of the diagnostic study. Colonoscopy is considered the gold standard, but is it truly? The true gold standard would be colectomy followed by meticulous analysis of the entire colon by a pathologist, but this is obviously not an option. While colonoscopy is an excellent tool for the detection of polyps, it is not perfect. Colonoscopy miss rates are estimated at 6% for polyps greater than 1 cm, 13% for adenomas 6 to 9 mm, and 27% for adenomas less than or equal to 5 mm.8

Not all colonoscopies are successful, with an incompletion rate of between 5% and 10%. It is not rare for the colonoscopist to incorrectly assume that the cecum has been reached.9 Colonoscopy is also not without risk: Perforation can occur in approximately one in 1000 studies, significant hemorrhage in three in 1000, and death has been very rarely reported. Colonoscopy also requires sedation and therefore loss of a full day of work in most cases.

Compliance with colorectal screening is very low. Perhaps only 15% of the age-appropriate population undergoes any type of colon screening at all, and only a small percentage undergoes colonoscopy. Many patients are loathe to be sedated and undergo an examination that requires the introduction of a sizable instrument into their colon. Screening examinations should have a high sensitivity for detecting the disease in question and should add no risk to the patient. While colonoscopy meets the first challenge well, it comes up short in the second.

Since colorectal cancer is largely preventable with the identification and removal of adenomatous polyps, the availability of a screening test that meets the requirements of high sensitivity and effectively zero risk is critical. So the answer to the second question-Do we need another screening test for colorectal cancer?-is therefore yes. The answer to the first question, why more people in the U.S. are not getting screened for colorectal cancer, is more complex, but at least part of the answer relates to the lack of an effective, palatable test.

'VIRTUAL' LURES PUBLIC

Virtual colonoscopy has also been termed CT colonography, and this term predominates in many academic centers and published articles. While technically this designation is accurate in describing the technique, its use comes primarily from the fact that in most centers the primary method of interpretation of these studies has remained the 2D axial image. As 3D technology evolves and interpretations become increasingly based on primary interpretation of 3D data sets as well as 2D imagess, the "virtual" in virtual colonoscopy takes on increased relevance.

Another important factor is the public's perception of the procedure. At least one of the reasons for the demise of the barium enema is its unattractive name. No patient would be eager to undergo any study with the term "enema" so prominent. In the 21st century, an examination that includes the term "virtual" is likely to interest a much larger portion of the population than a technical term such as CT colonography. Since the public health goal is increased screening, there is much in a name.

Several studies have investigated the effectiveness of virtual colonoscopy.10 The largest, by Yee et al,11 demonstrated a sensitivity of over 90% for the detection of polyps over 1 cm and a sensitivity of 94% for adenomas over 1 cm. This is the same sensitivity for polyps 1 cm and above as conventional colonoscopy.8 And none of these published studies utilized a 16-row CT scanner with submillimeter collimation and 1-mm slice thickness, which can be expected to increase the sensitivity of polyp detection.

Excellent virtual colonoscopy is dependent on four critical points: optimum bowel cleansing, maximum bowel distention, state-of-the-art hardware and software, and dedication and experience on the part of the radiologist.

Bowel cleansing is critical to remove stool that can mimic colonic polyps, and several choices are available. Polyethylene glycol, phospho soda, and magnesium citrate are the most common options. Polyethylene glycol, the most frequent preparation for conventional colonoscopy, requires the ingestion of 4 L of liquid. While it leaves very little residual solid fecal material, the large amount of ingested water can be problematic. Patients who have taken polyethylene glycol as their bowel cleansing agent tend to have more residual fluid during virtual colonoscopy. Large amounts of retained fluid can hide submerged polyps, although this is mitigated somewhat by scanning in both supine and prone positions. Another alternative is opacifying the residual fluid with another oral contrast agent such as diatrizoate meglumine (Gastrografin). Some virtual colonoscopy software manufacturers have included the ability to then electronically remove the opacified fluid.

Phospho soda preparations yield a drier colon, though some fluid usually remains.12 These preparations can lead to electrolyte imbalance, however, and are contraindicated in patients with renal failure. They must be used with caution in patients with impaired renal function, heart disease, and known electrolyte imbalances. Magnesium citrate preparations have been used for many years in barium enema preparations and can be effective, although they tend to leave a greater amount of residual stool.

All of these preparations are used in conjunction with dietary restrictions, primarily the avoidance of high-residue foods, and either a clear liquid diet or a prepackaged nutritional diet also designed to include only low-residue foods. In our practice, we also administer barium for the purpose of stool tagging. An oral barium mixture commonly used for routine CT examinations is administered at mealtimes beginning at dinner two days prior to the procedure and continued at the three mealtimes the day prior to the study. Stool tagging with barium can help in differentiating residual stool from polyps,13 especially when used in conjunction with a software package designed to identify the high-density barium trapped in residual stool.

Maximum bowel distention is critical for the identification of polyps, and two choices are available to achieve this: room air and carbon dioxide. Room air has the advantage of being free and easily employed, requiring only a small tube and bulb for insufflation. Either the CT staff or the patient can insufflate the air. Some investigators have found better degrees of distention when the patient controls the amount of air introduced. CO2 is generally insufflated by means of an electronic pump. A small tube is placed in the rectum, and the electronic pump infuses the CO2. The amount of CO2 instilled is regulated by a system that limits the pressure that can be reached to a safe level.

The primary advantage of CO2 insufflation is the rapid absorption of the gas through the colonic mucosa and its subsequent excretion through the lungs. By the time the patient has dressed after the conclusion of the procedure, there is no residual feeling of bloating or distention and no significant flatulence, as is the case with room air. Because the pressure of insufflation is limited with current CO2 electronic insufflators, however, episodes of spasm, which are common, may lead to underdistention. In our practice, we begin with CO2 insufflation, and if we witness evidence of spasm or underdistention on the CT scout view, we supplement the CO2 with small amounts of room air. This combination limits the degree of discomfort following the procedure and affords better distention in some patients.

Many studies demonstrating the efficacy of virtual colonoscopy were performed on single-slice spiral CT scanners and some on four-row multidetector units. Slice thickness ranged from 5 to 3 mm, with overlapping reconstruction. No published study has used a 16-row detector system. With a 16-row CT, we routinely obtain submillimeter collimation with 1-mm slice thickness. Thin-slice acquisition allows much better identification of small polyps; we routinely identify 3-mm polyps. Although the literature agrees that such small polyps are not significant, the ability to resolve very small structures results in a higher degree of confidence to identify polyps 5 mm and larger. With a 16-row CT and very thin sections, 1-cm polyps seem enormous.

INTERPRETATION CONTROVERSY

Interpretation methods are evolving, primarily due to improvements in software and specifically 3D reconstructions for creation of the virtual colon. Although most of the existing literature supports primary 2D reconstruction with 3D problem solving, the existing 3D software was likely not available to those investigators. We currently utilize the V3D-Colon package developed by Viatronix. This software creates an excellent 3D interactive virtual colon that allows simple and rapid navigation in both supine and prone views. Standard axial and multiplanar reformatted images are displayed simultaneously and updated automatically to correlate location.

We routinely image in both supine and prone positions and view each data set in three modes: routine axial images, followed by forward and backward 3D virtual interactive fly-through. In this manner, the colon is examined a minimum of six times: three supine, three prone. Whenever a suspicious area is encountered in any of these reviews, it is bookmarked and then investigated via the three standard axial, sagittal, and coronal 2D projections as well as 3D. The ability to rapidly problem-solve in these multiple projections is critical for proper analysis. A search for barium-tagged stool is also readily performed with an incorporated color-coded density display.

Although controversy continues to surround 2D versus 3D interpretation, this is misguided. With the availability of excellent software that allows rapid investigation in both modes, it is clear that both are necessary and complementary. In a normal patient, total colon review time is less than 15 minutes. A complicated patient with a long, capacious colon, diverticulosis, and multiple polyps can require a longer interpretation time. Colonoscopists state that total exam time for optical colonoscopy, into and out of the colon, is approximately 10 minutes. Anyone who has performed a significant number of virtual colonoscopies and witnessed the need to carefully examine a large volume will likely be amazed that the miss rate in colonoscopy, though substantial, is not even higher. Complete visualization of the colon takes time.

Although most polyps are round or oval in shape, some are flat (height less than half the width). Estimates of the percentage of polyps that are flat range from 8% to 33%. Flat polyps have a higher likelihood of dysplasia and carcinoma, especially those greater than 1 cm in size. Virtual colonoscopy can identify many of these hard-to-detect flat lesions, which are difficult to visualize with conventional colonoscopy, utilizing both 2D and 3D review.14 The use of thin-slice 16-row CT scanning enhances the sensitivity in detecting these lesions.

Because virtual colonoscopy represents CT evaluation of the abdomen and pelvis, it allows for review of the normal anatomic structures visualized in such an examination. Identification of extracolonic findings is common. In a study by Hara et al,15 41% of patients undergoing CT colonography demonstrated extracolonic findings, and 11% had findings that were deemed highly important. These included renal cell carcinomas, lung nodules, significant abdominal aortic aneurysms, and other solid tumors.

PATIENT CRITERIA

Who should have virtual colonoscopy? A study by Macari et al16 demonstrated the usefulness of virtual colonoscopy in patients who have had failed or incomplete colonoscopy, which the study estimated at 10% of the conventional colonoscopies. The major subset of potential patients is the population eligible for screening: men and women over age 50 (or younger in the presence of risk factors such as a strong family history). More than 100 million people in the U.S. are over the age of 50, and each should have a total colon examination at least every 10 years. Many need more frequent examination. Only a fraction of these individuals gets any form of colon screening, and a relatively small number undergo colonoscopy.

Considering the preponderance of evidence that the identification and removal of adenomas of the colon will dramatically reduce the devastating consequences of colorectal cancer, many more millions must be screened. Even those who believe that conventional colonoscopy is superior to virtual colonoscopy must admit that the public has not sufficiently embraced conventional colonoscopy, and even if it did, there are not nearly enough qualified gastroenterologists to begin to handle the demand. Virtual colonoscopy, if done properly, creates a huge tent that can accommodate many more of these individuals at risk. Without question, screening with virtual colonoscopy would save many lives almost immediately.

The last item from the list of requirements for excellent virtual colonoscopy is the dedication and experience of the radiologist. Mastering this procedure requires a definite learning curve. Before offering this examination to the public, we conducted our own study with patients who first underwent virtual colonoscopy in our facility the same day as their conventional colonoscopy. Findings on the virtual colonoscopy were correlated with the colonoscopy findings. This allowed us to optimize our patient preparation and scanning techniques and afforded an opportunity to become facile with the software and method of interpretation.

Accurate interpretation of virtual colonoscopy is time-consuming, with a minimum of 10 to 15 minutes required for comprehensive 2D and 3D interpretation, plus additional time for review of extracolonic findings. Some radiologists have stated that this required time commitment is unacceptable, considering the demands placed on most of them. This concern prompts a question: How much time would any one of us ask a physician to spend to detect a finding that may lead to our escaping cancer? It is, in my view, unacceptable to dismiss performing a worthwhile study because it is too difficult or time-consuming. The public requires a safe, accurate screening test for colorectal cancer, and virtual colonoscopy can meet that requirement.

It is not unreasonable to expect appropriate compensation for time spent, since interpreting or performing any procedure represents an opportunity cost. We currently charge $750 per virtual colonoscopy examination. Assuming the ability to interpret three studies per hour, this level of compensation can support both the technical costs and the professional fees of the radiologist. These studies are not yet reimbursed by Medicare or most other insurers, so patients are required to bear the cost.

Virtual colonoscopy, when performed with state-of-the-art CT equipment and software that allows rapid and thorough 2D and 3D interpretation, is an excellent method to screen for colorectal adenomas, the precursors of the majority of colorectal cancer. The acceptance and proliferation of this technique require the education and commitment of radiologists. Since colorectal cancer is a largely avoidable disease, the public deserves no less from our specialty.

Dr. Klein is a radiologist at Washington Radiology Associates in Washington, DC.

References

1. Vining, DJ, Gelfand DW. Noninvasive colonoscopy using helical CT scanning, 3D reconstruction, and virtual reality. Presented at the 23rd Annual Meeting and Postgraduate Course of the Society of Gastrointestinal Radiologists, Maui, Hawaii, 1994.

2. Dachman AH, Yoshida HY. Virtual Colonoscopy: past, present, and future. Radiol Clin N Am 2003;41:2.

3. Hara, AK. The future of colorectal imaging: computed tomographic colonography. Gastroenterol Clin 2002;31:4.

4. Neugut AL, Jacobson JS, Rella VA. Prevalence and incidence of colorectal adenomas and cancers in asymptomatic patients. In: Rex DK, ed. Gastrointestinal endoscopy clinics of North America. Philadelphia: Saunders, 1997:387-399.

5. Villavicencio RT, Rex DK. Colonic adenomas: prevalence and incidence rates, growth rates, and miss rates at colonoscopy. Semin Gastrointest Dis 2000;11:185-193.

6. Schoenfeld P, Lipscomb S, Crook J, et al. Accuracy of polyp detection by gastroenterologists and nurse endoscopists during flexible sigmoidoscopy: A randomized trial. Gastroenterology 1999;117:312-318.

7. Winauer S, Stewart E, Zauber A, et al. A comparisonl of colonoscopy and double-contrast barium enema for surveillance after poylpectomy. NEJM 2000;342:1766-1772.

8. Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rate of adenomas determined by back-to-back colonoscopies. Gastroenterology 1997;112:24-28.

9. Anderson ML, Heigh RI, McCoy GA, et al. Accuracy of assessment of the extent of examination by experienced colonoscopists. Gastrointest Endosc 1992;38:560-563.

10. Macari M, Bini EJ, Xue X et al. Colorectal neoplasms: prospective comparison of thin-section low-dose multi-detector row CT colonography and conventional colonoscopy for detection. Radiology 2002;224:383-392.

11. Yee J, Akerkar G, Hung R, et at. Colorectal neoplasia: Performance characteristics of CT Colonography for detection in 300 patients. Radiology 2001;219:685-692.

12. Macari M, Lavelle M, Pedrosa I, et al. Effect of different bowel preparations on residual fluid at CT colonography. Radiology 2001;218:274-277

13. Lefere PA, Gryspeerdt SS, et al: Dietary fecal tagging as a cleansing method before CT colonography: initial results-polyp detection and patient acceptance. Radiology 2002;224:393-403.

14. Fidler JL, Johnson CD, MacCarty RL, et al. Detection of flat lesions in the colon with CT Colonography. Abdominal Imaging 2002;27:292-300.

15. Hara AK, Johnson CD, MacCarty RL, et al. Incidental extracolonic findings at CT colonography. Radiology 2000;215:353-357.

16. Macari M, Berman P, Dicker M, et al. Usefulness of CT colonography in patients with incomplete colonoscopy. AJR 1999;173:561-564.