A systematic review and meta-analysis by Chinese researchers found that adenomas and advanced adenomas were missed somewhat more frequently than previously reported.
Jun Yao, MD, PhD, of Jinan University in Shenzhen, China, and colleagues reviewed 43 publications involving 15,152 tandem colonoscopies and 10,852 adenomas. The team calculated adenoma miss rates (AMRs) of 26% for adenomas (95% CI 23%-30%), 9% for advanced adenomas (95% CI 4%-16%), and 27% for serrated polyps (95% CI 16%-40%).
Miss rates were also high for proximal advanced adenomas (14%, 95% CI 5%-26%), flat adenomas (34%, 95% CI, 24%-45%), and patients at high risk for colorectal cancer (33%, 95% CI 26%-41%).
The results, published in Gastroenterology, were based on 15 studies from Asia, 13 from the U.S., 10 from Europe, five from multiple countries, and one in Australia; 32 of the studies were randomized controlled trials.
Compared with a previous systematic review of 441 tandem colonoscopies, the current analysis detected a slightly higher miss rate for both polyps and adenomas: 28% versus 21% for polyps and, as noted, 26% versus 22% for adenomas.
Identifying risk factors and indicators related to missed lesions is critical since these lesions drive interval cancers, the authors explained. Even with adenoma resection and surveillance, patients still have a 47% risk of a standardized incidence-based mortality from colorectal cancers, which often develop in the interval between scheduled colonoscopies.
Interval cancers contribute to about 9% of all colorectal cancers, and lesions missed during colonoscopy are responsible for 50-60% of interval cancers, the researchers noted. The incidence of interval cancers from missed lesions is 3.5 per 1,000 screened persons.
In the current meta-analysis, adenoma detection rate (ADR), adenomas per index colonoscopy (APIC), and adenomas per positive index colonoscopy (APPC) emerged as independent predictors of AMR. APPC, however, was the only independent predictor of miss rate for advanced adenomas, and if validated in future studies, this indicator “deserves consideration when designing and/or modifying future recommendations and guidelines on the quality indicators of colonoscopy,” Yao and associates wrote.
An APPC value >1.8 was more effective in monitoring AMR (31% vs 15% for below this value, P<0.0001) than was an ADR of 34% or more (27% vs 17% for below this value, P=0.008). The AMR for advanced lesions with colonoscopies with an APPC value below 1.8 was 35% versus 2% for colonoscopies with an APPC value of 1.8 or more (P=0.0005).
Although most index colonoscopies in the analysis had an ADR of 5%, that increased slightly instead of decreasing, indicating that prior colonoscopy and even tandem colonoscopy probably missed a fair proportion of adenomas, particularly for those behind the folds, the team said. “In addition, these findings also indicated the currently used quality indicators might be insufficient for monitoring AMR.” .
The study also found a 12% false-negative rate for index colonoscopy, which means that 6-10% of patients with adenomas would be diagnosed as normal when the index ADR ranged from 20% to 50%. “In other words, missed adenomas also exerted a direct effect on the ADR, and the latter could partly monitor the AMR of false-negative colonoscopies,” Yao and associates wrote.
They cautioned, however, that ADR has limited value in assessing the number of adenomas per colonoscopy because of the “one-and-done effect,” in which the colonoscopist might not inspect the colon carefully after detecting one adenoma. “Therefore, ADR is necessary but might be insufficient to monitor colonoscopy quality,” the researchers concluded.
“Future efforts should be concentrated to detect advanced adenomas, serrated polyps, sessile or flat adenomas, and adenomas in the high-risk population,” they added, noting that miss rates could be reduced by adequate bowel preparation and auxiliary colonoscopy techniques.
Asked for his perspective, James M. Church, MD, of the Cleveland Clinic, who was not involved with the meta-analysis, called the study “very valuable,” because it highlights the difference between ADRs and AMRs.
“Although ADR has been a widely accepted measure for about a decade, there are problems with it,” he told MedPage Today. “The take-home message here is that colonoscopists should not rely on the ADR for quality assessment but should start using the AMR as well and avoiding checking the box after finding one.”
In addition, he advised colonoscopists to pay special attention to scoping high-risk patients and evaluate their performance, taking steps, if necessary, to improve their skills.
Another interesting finding, Church said, was that while adequate bowel preparation is necessary, high-quality prep is not.
Study limitations, the authors said, included inter-study heterogeneity because of such differences as geographical region, study period, number of centers, study and tandem colonoscopy design, and withdrawal time. Furthermore, the Egger’s test indicated potential publication bias for adenomas, polyps, and small semi-pedunculated and sessile adenomas. In addition, the researchers said, the study-level review could not analyze the characteristics of individual patients and colonoscopists and possibly other variables affecting miss rates.
“Prospective population-based multicenter tandem colonoscopies on large cohorts are needed to clarify these issues,” Yao and co-authors wrote.
The authors reported having no conflicts of interest.
Church disclosed no conflicts of interest relevant to his comments.