Model finds polyp surveillance effective for cancer prevention

By Eric Barnes, AuntMinnieEurope.com staff writer

January 21, 2008 -- The risk of developing colorectal cancer from polyps found at virtual colonoscopy (CT colonography [CTC]) but left unresected is not precisely known, and the practice of polyp surveillance is still controversial. But a new study concludes that the low likelihood of advanced neoplasia combined with the high cost of colonoscopy argue against the immediate removal of small lesions.

To evaluate the associated risks and costs of surveillance, Dr. Perry Pickhardt and Dr. David Kim from the University of Wisconsin at Madison constructed a decision analysis model to evaluate the cancer risk associated with polyp surveillance versus the cost and risks of immediate referral to colonoscopy and polypectomy. Joining in the study were Dr. Cesare Hassan and colleagues from Italy's Nuovo Regina Margherita Hospital in Rome, and Dr. Andrea Laghi and Dr. Franco Iafrate from the University of Rome "La Sapienza."

"Some authors have proposed that large polyps (≥ 10 mm) detected at CTC should trigger an immediate polypectomy, small polyps (6-9 mm) could either be referred for colonoscopy or undergo CTC surveillance, and potential diminutive polyps (≤ 5 mm) should be ignored," the researchers wrote (American Journal of Roentgenology, January 2008, Vol. 190:1, pp. 136-144).

The CT Colonography Reporting and Data System (C-RADS) management model (Radiology, July 2005, Vol. 236:1, pp. 3-9) recommends surveillance of one or two 6-9-mm lesions, and has generally been well received in the radiology community but remains more controversial among gastroenterologists, the group noted.

"The debate primarily centers around the potential risks of a patient developing or even harboring a malignancy from an unresected lesion, and the potential ethical and medicolegal concerns over the practice of not reporting possible diminutive lesions," the authors explained.

Aiming to estimate the risks of polyp surveillance versus resection, the decision analysis incorporated expected polyp distribution, advanced adenoma prevalence, colorectal cancer risk, virtual colonoscopy performance, and costs related to the screening and treatment of colorectal cancer, they wrote.

"The aim of this decision analysis was to assess the relative yield of referring patients to colonoscopic polypectomy for diminutive, small, and large CTC-detected polyps," the team wrote. "Specifically, we sought to quantify the relative benefits, harms, costs, and resource utilization of polypectomy referral for lesions detected at CTC screening."

The model calculated the number of small, medium, and large polyps needed to be removed to detect a single advanced adenoma over a 10-year time period. The average age of the asymptomatic screening population was 58 years. The researchers also calculated the cost-effectiveness of polypectomy in life years gained.

According to the results, the estimated 10-year risk of colorectal cancer for unresected diminutive, small, and large polyps was 0.08%, 0.7%, and 15.7%, respectively. "The number of diminutive, small, and large polyps that needed to be removed to avoid leaving behind an advanced adenoma was 562, 71, and 2.5, respectively; similarly, 2,352, 297, and 10.7 polypectomies would be needed, respectively, to prevent one CTC over 10 years," Pickhardt and colleagues wrote.

The incremental cost-effectiveness ratio of resecting all diminutive and small polyps found at VC was $464,407, and $59,015 per life year gained, respectively, while polypectomy for large polyps saved $151 per individual screened.

"Our study shows that the expected yield of colonoscopic referral for polyps detected at CTC screening is strongly related to polyp size, even though we made a conservative assumption that small advanced adenomas had the same (colorectal cancer) risk as large advanced adenomas, which very likely overemphasizes the importance of small lesions," they wrote. "For large polyps, the yield for polypectomy referral was high, needing only 2.5 lesions removed for each advanced adenoma and 10.7 lesions removed for each CRC prevented. In comparison, more than 500 diminutive lesions would need to be referred to detect one advanced lesion and more than 2,300 diminutive polypectomies would be needed to prevent one case of CRC."

Despite using the conservative assumption that small and large advanced adenomas harbored the same cancer risk, the study showed that the expected cancer yield for colonoscopy referral is strongly related to polyp size, the authors noted.

"Despite only minimal gains in projected CRC prevention, the large number of polypectomies for diminutive lesions corresponded to much higher procedural costs and higher complication rates due to the increased utilization of colonoscopy," they wrote." In fact, the accepted perforation rate for therapeutic colonoscopy clearly exceeds the projected number of procedures needed to prevent one (colorectal cancer) over a 10-year time horizon by removing diminutive polyps, and even approaches the rate for small 6- to 9-mm polyps."

A sensitivity analysis showed that the size-specific prevalence of advanced neoplasia was the key variable in the study, the group reported. And although advanced histology was rare for diminutive lesions (0.6%) based on the study cohort, other authors have reported higher rates that could potentially affect the results significantly.

Conversely, the performance of virtual colonoscopy had little effect on outcomes in the sensitivity analysis, and hence a relatively small effect on the cost of additional life years.

The authors noted several limitations of the study, including an "unavoidable simplification of the adenoma-carcinoma sequence and of CRC screening in general." Due to a paucity of longitudinal information on the natural history of polyps, the model also assumed equivalent cancer risk regardless of adenoma size, which likely overestimates the actual risk for small and diminutive lesions because villous histology is rare for these lesions.

"Our analysis shows that colonoscopic referral for small (6-9 mm) and diminutive (≤ 5 mm) polyps detected at CTC screening is an inefficient strategy in terms of advanced adenoma removal and (colorectal cancer) prevention," Pickhardt and colleagues concluded. The results suggest that polyp surveillance may be a reasonable alternative to polypectomy for small polyps due to the high cost of polyp removal and the low yield for cancer prevention, they added.

"Further investigation of the natural history of small colorectal polyps may provide more robust input data and lead to greater insight into this critical management issue," they wrote.

By Eric Barnes
AuntMinnie.com staff writer
January 21, 2008