Since its inception nearly a decade ago, virtual endoscopy (VE) has been implemented in a number of diagnostic procedures that stand to compete against more invasive counterparts, including colonoscopy, angiography, bronchoscopy, and cystoscopy.

Several factors have contributed to VE’s accelerated growth: introduction of multislice CT (MSCT), development of inexpensive but powerful personal computers (PCs), and the commercial availability of user-friendly software. Another factor has been the public’s interest in total body screening, which is supported by a number of VE techniques.

Virtual endoscopy is simply an interactive 3D computer technique used to visualize vast amounts of CT or MR image data of hollow viscus organs (eg, blood vessels, stomach, airways) in a manner similar to how a doctor would inspect these organs with a conventional endoscope. A distinct difference is that VE is limited to identifying the morphology (ie, shape and size) of pathology, whereas conventional endoscopes can directly image mucosal detail as well as perform therapeutic procedures. However, VE has several advantages over its counterparts: patient comfort and acceptance, ability to diagnose significant disease outside the primary organ of interest, and potential for cost savings.

Except for CT angiography (CTA) (CPT 75635), shown in Figure 1, a specific VE CPT code does not currently exist, and hence screening of asymptomatic patients with VE procedures, like virtual colonoscopy (VC), is not supported by Medicare or insurers. However, if medical necessity can be demonstrated, VE may be reimbursed by combining the CPT code(s) for the scanning procedure plus a 3D reconstruction code. When facility and anesthesia costs are included in the total cost of traditional endoscopic examinations, VE proves to be significantly less expensive (Table 1, page 14).

Figure 1. (a, b) Surgical planning CT angiography of a renal transplant donor.

Since Medicare and insurers do not currently reimburse for VE screening procedures, many facilities now offer procedures, such as VC, to private-pay patients for prices ranging from $695 to $1,250. It is anticipated that clinical trials studying the efficacy of VE procedures will be supportive of these new techniques, and hence CPT codes will need to be established to address reimbursement.

The Basic STEPS

The VE technique consists of three general steps:

1. Patient Preparation. Some form of patient preparation is usually required, and this can run the gamut from simply coaching patients to hold their breath during virtual bronchoscopy, to more involved bowel cleansing and colon distention for VC. When contemplating the start of a VE service, it is important to consider planning for a well-designed facility, for example, adding a bathroom to a CT scanning suite for VC examinations.

2. Data Acquisition. Volume acquisition of image data with helical CT or MR scanning is the next step. An important caveat to consider when performing a scan for VE applications is that thin slices and fast acquisition times are needed to ensure optimal spatial resolution and to reduce motion artifacts. Both of these requirements are readily met by MSCT, although VE can be performed using single-slice spiral CT technology. The advantage of MR scanning over CT includes its inherent tissue contrast and avoidance of ionizing radiation, but its shortcomings involve limited availability, claustrophobia, and prolonged scanning times. MSCT is emerging as a preferred VE scanning method because of its incredible short scan time (often less than 30 seconds) and potential for reduced radiation. It should be noted that VE could also be accomplished using tomographic digital fluoroscopy data such as that acquired with state-of-the-art angiography equipment.

Some of the challenges associated with VE scanning include its requirement for additional CT or MR image reconstruction time, thus occupying the scanner’s computer and console, but this problem should be solved in the near future with faster or separate reconstruction engines. Another challenge is the need for significant image storage and archival. Whereas a typical CT scan of the abdomen and pelvis might produce 50 images, a VE scan of the same anatomy may produce 1,000 images.

3. Postprocessing. The final step for VE is interactive 3D processing. This task once required exotic and expensive computers, as well as a PhD to process the image data, but today similar power is available on common PCs, and multiple user-friendly software solutions are commercially available.

Unlike MSCT and MR scanners, which are still priced at a a premium, the price of 3D hardware and software has steadily decreased over the last few years and is now generally available for less than $100,000 for most systems.

A factor to consider when setting up a VE service is the amount of time that a technologist and/or radiologist must spend sitting in front of a computer workstation to produce and record VE images for reporting purposes. Some facilities employ knowledgeable technologists to process the data, similar to the way that ultrasound technologists generate images, and then present these VE images or video recordings to a radiologist for final interpretation. More efficient 3D imaging systems are emerging to enable radiologists to perform VE examinations routinely during the course of a busy workday.

Another point to consider when offering VE services is that many private-pay patients undergoing screening examinations (eg, VC) will want to review their results immediately following the examination. Patient consultation by a radiologist or nurse, as well as sending the patient home with a 3D picture or video, might be considered for the support of patient relations.

Today’s health care environment does not readily support novel technologies: technology has to prove beneficial and cost-effective in the practice of medicine. Virtual colonoscopy and CT angiography/angioscopy are apt to fulfill these criteria.

Virtual Colonoscopy

Table 1. Comparison of virtual endoscopy versus conventional endoscopy reimbursement rates based on the 2002 North Carolina Medicare fee schedule.

Colorectal cancer is the second leading cause of cancer deaths in the United States, and in the past several years, a number of national institutions (eg, National Cancer Institute and American Cancer Society) have launched major public awareness campaigns to urge screening. Traditional screening methods have included fecal occult blood testing, flexible sigmoidoscopy, barium enema, and colonoscopy. VC offers promise, however, because of its minimal invasiveness, comfort, short examination time, no need for sedation, and ability to visualize the entire colon as well as detect disease outside of the colon.

The VC procedure consists of bowel cleansing with a laxative, colon distention with air or carbon dioxide gas, CT or MR scanning of the abdomen and pelvis, and 2D/3D image analysis of that data (Figure 2, page 16). Patient preparation (bowel cleansing and colon distention) remains a major factor in optimizing results. It is essential that the colon be cleared of stool to avoid false positives (ie, small collections of stool simulating polyps), and that the colon be optimally distended to facilitate visualization. New products, including a CO2 insufflator, rectal catheters with fluid traps, and stool tagging agents, could make VC simpler and improve its accuracy.

Early Results

Early clinical studies indicate that VC has a sensitivity and specificity approximating 85% and 90%, respectively, for the detection of significant adenomatous polyps that are precursors to colorectal cancer. Two points need to be made about VC. First, gastroenterologists frequently argue that a positive VC examination necessitates traditional colonoscopy for further evaluation and polyp biopsy or removal, thus patients should undergo colonoscopy anyway. However, the majority of the screening population has a normal colon, meaning that most people could avoid the unnecessary risks associated with traditional colonoscopy if they elect to undergo a VC examination for screening. The second point is that VC’s false positives may actually be colonoscopy’s false negatives. The “gold standard” of colonoscopy is not that golden as shown by Douglas Rex, MD, Indiana University School of Medicine, Indianapolis, in a study in which screening patients underwent back-to-back colonoscopies. In that study, Rex showed that colonoscopy had an overall miss rate of up to 24% for the detection of polyps. An additional benefit of VC is its ability to detect significant incidental findings in approximately 10% of patients, thus supporting the concept of using this procedure during total body CT screening.

Visualizing the Blood Vessels

Figure 2. Example of a virtual colonoscopy examination: (a, b) external views of the colon; (c, d) peering inside the colon; (e) significant colon polyp; and (f) colon polyp identified by computer-assisted diagnosis (CAD) technique.

Another valuable use of VE technology is for the visualization of blood vessels during surgical planning. In many cases, a 30-second MSCT examination following an intravenous bolus of nonionic iodinated contrast can replace more invasive and expensive procedures such as angiography and intravenous pyelography , thus saving time and money, and reducing patient morbidity. Common applications of CTA include the evaluation of suspected aortic aneurysms and presurgical mapping of renal transplant donors (Figure 2).

Virtual Bronchoscopy. Virtual bronchoscopy was the first VE procedure to arise simply because it required the least amount of patient preparation (a patient simply had to hold their breath during a spiral CT scan). The utility of this procedure is somewhat less than that of VC or CTA, but it can play an important role in the evaluation of airway obstruction and visualization of complex anatomy, including identification of surrounding vessels, tumors, and lymph nodes for endoscopic needle biopsy guidance (Figure 3)

Virtual cystoscopy. Virtual cystoscopy may be performed in several ways, including instilling iodinated contrast material and/or gas into the bladder via a Foley catheter prior to CT scanning. Alternatively, MR scanning of a urine-filled bladder with T2-weighted sequences can be used for image acquisition because of MR’s inherent tissue contrast.

Future developments

Figure 3. (a, b) Virtual bronchoscopy can visualize surrounding structures, including lymph nodes (blue), for endoscopic needle biopsy guidance.

A forthcoming development is the combination of structured reporting methods with VE procedures. For example, it does not suffice for a radiologist to simply dictate that “there is a 2-mm polyp in the transverse colon” without being able to more specifically identify its location within the image data. This would be analogous to saying that there is a bald-headed man in America: exactly where is he? Structured reporting allows a radiologist to record specific image coordinates and attach keywords indicating anatomical location and pathological description, thus improving reporting accuracy and efficiency.

The future of VE is promising with continued development of more sophisticated MSCT scanners (16- and 32-slice MSCT scanners are on the horizon), as well as faster and less expensive 3D computers. Improved software with automated functions, like computer-assisted diagnosis (CAD) for polyp detection, will also enhance VE’s capabilities and make it a mainstay in the radiologic armamentarium.

David J. Vining, MD, is associate professor, Department of Radiology, Wake Forest University School of Medicine, Winston-Salem, NC, and founder of a 3D visualization and structured reporting company.