Virtual colonoscopy—using such imaging techniques as CT or MRI to visualize the bowel and screen for potentially troublesome polyps and lesions—is gaining ground as a supplement or alternative to traditional optical colonoscopy. Virtual colonoscopy has proven about 90% accurate in identifying potentially cancerous polyps of 10mm or more, making it an increasingly viable option to direct visualization with a colonoscope.

With colonscopies, it is not the high-tech scanning equipment but the low-tech bowel-prep procedure that is proving the biggest issue for patient acceptance—and reimbursement issues and lack of training might deter physicians and imaging centers from coming on board.

Techniques
In traditional colonoscopy, patients must complete a fairly harsh preparatory procedure that combines fasting with use of a strong laxative, often polyethylene glycol, to wash all fecal matter from the bowel, ensuring a clean visual field. The bowel is then viewed by means of a video camera that’s inserted via a flexible tube to the proximal end of the bowel, then slowly extracted to allow the endoscopist to scan for polyps and suspicious areas. Any questionable areas usually can be removed for biopsy in the same procedure.

In contrast, virtual colonoscopy does not involve as much preparation or any invasion. Patients consume a low-residue diet for up to 3 days prior to the exam and use a much gentler laxative, often a citrate osmotic preparation, to clean the bowel of nearly all fecal matter; a tagging agent is sometimes used to allow residual matter to appear clearly on scans and be disregarded. The patient’s colon is then insufflated with either room air or carbon dioxide, and a CT scan is taken with the patient in both prone and supine positions. The resulting images can be manipulated by software to provide a variety of views and to subtract the tagged fecal matter from the picture.

The conventional procedure will typically take an endoscopist 20 to 30 minutes to perform, examining the bowel during the procedure. By contrast, virtual colonoscopy could require 10 to 15 minutes to perform and then 10 to 15 minutes for a radiologist to read. Some imaging centers might provide for on-site reads, while some might use the services of a radiologist at a remote location, creating a wait for the results.

Virtual colonoscopy compares favorably with the conventional method in sensitivity of detection of the sorts of polyps that could be the most problematic. For polyps 10mm or greater, virtual colonoscopy shows a sensitivity of 90%. Even slightly smaller polyps are picked up fairly easily. “Seven- to 10-mm polyps, we’re very good at,” says Erik Paulson, professor of radiology and chief of the Abdominal Imaging Division for Duke University (Durham, NC).

For still smaller polyps of 5mm or less, sensitivity drops off to as low as 50%, according to data from the American Cancer Society (ACS) in Atlanta. This might not present much of a problem, as the risk of malignancy decreases along with polyp size. Smaller polyps are much more likely to be hyperplastic and are not likely to become cancerous in the patient’s lifetime, according to the ACS. This finding is not an excuse to be cavalier, however. “Nothing good can come of a polyp,” Paulson says.

In addition to the lower level of accuracy in detecting smaller polyps, virtual colonoscopy has the potential for incorrect readings, says Judy Yee, associate professor of radiology for the University of California, San Francisco, and chief of CT and GI radiology for the VA Medical Center. While residual stool and thick folds of bowel lining and tissue can lead to a false positive, flat lesions and small lesions can be overlooked, leading to a false negative, Yee says. The ACS adds that inaccurate results also can come from diverticular disease and metal or motion artifacts, both of which can produce a false positive.

Virtual colonoscopy has some notable advantages over the more traditional optical method, says Jim Clayton, research analyst with Frost & Sullivan (San Antonio, Texas). First among these is the noninvasive nature of the procedure. “[There is] much less chance of a perforated bowel,” Clayton says. This risk is small but real in conventional colonoscopy, according to Gregory Snyder, medical director for Minnesota Radiology (Edina, Minn). “With endoscopy, one in a thousand will have a ruptured bowel,” he says, adding that 10% to 15% of these conventional procedures will be incomplete, translating to an overall success rate of just 85% to 90%. “As a screening exam, [conventional colonoscopy] doesn’t make sense,” Snyder adds, extolling virtual colonoscopy as “painless, no risk, no morbidity, and no mortality.”

But perhaps the most important advantage from the patient perspective is the bowel prep procedure, which can be less harsh and conducted over a shorter period than that for traditional colonoscopy. “The biggest problem for conventional [colonoscopy],” Clayton says “is the prep solution and time.”

Preparation
“[There are] neat new imaging technologies, but the biggest hurdle is low tech: bowel prep,” Paulson says. Conventional and virtual colonoscopy both depend upon the bowel being free or mostly free of fecal matter, and it is this process of cleaning the colon through laxatives and diet that most patients find unpleasant.

According to Snyder, this preparatory procedure is at the heart of patient acceptance. He explains that the standard prep involves the patient ingesting approximately 2 gallons of polyethylene glycol that literally washes fecal matter out of the bowel, a process that patients often find unpleasant due to cramping and diarrhea.

But for virtual colonoscopy prep, Snyder uses products from E-Z-Em (Westbury, NY) that make the preparation less harsh. Patients use a citrate laxative that is osmotic, so ingestion of plenty of water helps to evacuate bowel contents. Then, a barium sulfate stool tag product called Tagitol is used to help the residual fecal matter show up on scans. Once fecal matter is tagged, it can be removed from the image by software programs in a process known as electronic subtraction, or, less accurately, electronic cleansing. This means that residual fecal matter presents less of a problem, allowing patients to undergo less harsh but slightly less complete cleansing procedures prior to the exam.

Although physicians, and certainly patients, may hope for the day that virtual colonoscopy can be performed without preparation, Snyder is not optimistic that this will come. “I don’t know if we’ll ever get to prepless,” he says, noting that insufflation is difficult with a fecal mass in the colon.

Insufflation is another issue where steps can be taken to make the procedure more comfortable and attractive for the patient. Typically, the colon is insufflated with either room air or carbon dioxide. While room air remains in the colon until the patient passes it, carbon dioxide is absorbed into the bloodstream and released through exhalation. Because of this difference, many patients prefer that carbon dioxide be used. Automated insufflation equipment is available to help achieve maximum distention with minimal discomfort.

Reimbursement
From a provider standpoint, the bigger issue is in obtaining insurance, says Snyder, who estimates that, without insurance to defray costs, “nine out of 10 patients won’t go through with [the procedure].”

In general, virtual procedures typically cost around $750. And virtual colonoscopy tends to be less expensive than the conventional procedure. It is this cost differential that makes reimbursement a good deal for most insurers, and it is why Snyder expects to see widespread reimbursement begin within a year. “It’s a no-brainer to sign up for this,” he says. However, some folks anticipate that insurers might see economic advantage in moving deliberately. “If a third-party payor can hold off for 6 to 12 months, [it] can save an awful lot of money [over that period],” says Bryan Westerman, clinical sciences manager for CT for Toshiba (Tokyo).

Training a Key Need
In addition to reimbursement, lack of training opportunities is a key factor limiting radiologists’ acceptance of the procedure, says Abraham Dachman, professor of radiology and director of CT for the University of Chicago and editor of the Atlas of Virtual Colonoscopy (Springer Verlag, 2003). “There’s a long learning curve,” Dachman says. “It’s difficult to learn to interpret [the] examination; [it] takes a lot of skill.” One difficulty arises in learning to view both two- and three-dimensional images. Typically, virtual scans are read in 2D form to identify suspicious areas, then a 3D image is used for closer examination. The 3D images can also be used to create a “flythrough” that mimics the view seen during a conventional procedure.

Opportunities for radiologists to develop this skill are still few and far between, says Dachman, who explains that, currently, no organizations have guidelines mandating frequency or content of training. Some institutions, such as the University of Chicago, offer CME courses infrequently (Chicago offers courses four times a year), and some manufacturers host periodic training sessions (Dachman teaches courses offered by Waukesha, Wis–based GE Medical Systems). “Only a handful of people teach courses,” Dachman says. Alternate means of gaining exposure include consulting books like Dachman’s or using software tutorials. However, the best way of gaining comfort with reading virtual colonoscopy scans is to read practice cases and then compare the results with those gained from a conventional read. As with many techniques, practice builds accuracy.

Target Population
Virtual colonoscopy has certain benefits and advantages that could indicate the best population for this procedure. On one hand, Frost & Sullivan’s Clayton explains that “if you find polyps or lesions greater than 10mm, you can’t take a closer look or remove them.” If such an area is found, “you put the patient through two different procedures,” he continues, by requiring the patient to return for a conventional colonoscopy to examine and perhaps remove the suspicious area.

However, for “normal people over 50 years [old] who want to take preventive measures in their own hands,” Clayton says that virtual colonoscopy might be an attractive option. Paulson agrees and has identified four populations that are particularly suited for virtual colonoscopy: those seeking a simple screening for polyps; patients who have had a failed conventional colonoscopy and need to complete their screening the same day without having to repeat bowel prep; patients with known colon cancer whose tumor obstructs the bowel, preventing complete conventional screening; and patients who particularly request the virtual procedure.

Outlook
The move to virtual colonoscopy for at least some patient screening could be a matter of acceptance rather than technology. “Many hospitals or imaging centers might have the capabilities, but [they’re] not using [them],” Clayton says. He attributes this lacking to the need for adequate training as well as appropriate reimbursement.

Additional research to determine the effectiveness of virtual colonoscopy for a screening population could help both insurance companies and imaging centers to become more comfortable with the procedure. “Studies to date have been based on selected patient groups,” Paulson says. “The efficacy in a screening population is unknown.” He adds that an ongoing study at Duke will be released within the next 3 years. In the meantime, Jacob Sosna and colleagues from Harvard Medical School (Boston) have released a meta-analysis of virtual colonoscopy studies that confirms at least an 81% sensitivity for polyps of more than 10mm.

More research might also indicate whether virtual colonoscopy will remain a CT procedure, or whether MRI will be an alternate modality. “[I have] not seen anybody try this with MRI,” Bryan Westerman says. “The special resolution of CT makes it the preferred modality,” he says, adding that the speed helps CT handle the motility of the bowel. MRI has gained greater acceptance in Europe, says the VA Medical Center’s Yee, who adds that there is more focus on radiation exposure in the international market than in the United States. She explains that MRI colonoscopy lags far behind the CT model in the United States due to a lack of research supporting its efficacy and its higher cost.

Another option to reduce radiation exposure is ultra–low-dose virtual colonoscopy. At the 2003 European Congress of Radiology, Riccardo Iannaccone from the University of Rome, La Sapienza, presented research results that compared same-day virtual low-dose and conventional colonoscopies. Low-dose colonoscopy detected 100% of the carcinomas identified by the conventional screening, and the procedure had an overall 89.1% sensitivity in detecting polyps of 6mm or larger. However, the study also found reduced ability to find extracolonic abnormalities due to lower contrast in such areas as the liver and pancreas.

Improved data collection is also on the horizon. “Data collection is probably pretty good right now with 16-slice scanners,” Westerman says. He adds, “Toshiba has a working prototype of a 256-slice machine.” With dramatic increases in data collection such as this, the medical community will need to reach consensus on cost versus increased collection. “There could be incremental changes that are very expensive,” he says.

Westerman also believes that “hardware is less of a driver [than] data handling.” First among these data-handling methods is computer-aided detection (CAD), a technique that already has found some uses in mammography and is poised to enter virtual colonoscopy within a year or so. In this technique, software looks at the bowel images collected and identifies areas that might be polyps. In combination with electronic subtraction, CAD is poised to become a powerful backup tool. Although it is no replacement for a reading by a human radiologist, CAD could serve as a double check. “It’s a sort of a second read,” says the University of Chicago’s Dachman. And Westerman predicts that CAD might “make life a lot easier for the radiologist; [it] might save an awful lot of time.”

Currently, Dachman says that the University of Chicago, along with the National Institutes of Health (Bethesda, Md) and Stanford University (Palo Alto, Calif), is working on CAD programs that will be available soon—that is, likely within a year or 2. The software will likely include a texture analysis component that analyzes masses to determine which are fecal matter and which might be polyps.

Ultimately, virtual colonoscopy just might be the first step toward a virtual view of nearly any hollow organ. Westerman predicts that “CT endoscopy can be used for anything tubular.” It could open the doors to a variety of virtual, noninvasive procedures that supplement, complement, or even replace a more traditional viewing method. And most experts agree that any procedure that is easier and more attractive to patients will improve acceptance of colon screening, taking a huge step toward widespread earlier detection of colon cancer. s