A population-based cohort analysis of Barrett’s esophagus patients undergoing surveillance endoscopy suggests that the neoplasia detection rate (NDR) and the rate of missed dysplasia during the index endoscopy may be lower than previously reported in studies of referral-based cohorts. The new results suggest that NDR may be a useful quality control measure for Barrett’s esophagus surveillance.
The finding is welcome. “Just like we’ve done in colonoscopy with the adenoma detection rate, we need to have a quality metric to determine whether or not we’re adequately finding neoplasia while screening our patients with Barrett’s esophagus,” Jeffrey Mosko, MD, a gastroenterologist and interventional endoscopist at the University of Toronto’s St. Michael’s Hospital, said in an interview.
Societal guidelines recommend endoscopic screening in Barrett’s esophagus patients, with the goal of identifying dysplastic Barrett’s esophagus and eradicating it endoscopically before it can develop into esophageal adenocarcinoma (EAC). Despite this, 90% of patients with esophageal adenocarcinoma are diagnosed outside of a surveillance program.
Missed high-grade dysplasia or early EAC could become more invasive or metastasize, potentially leading to greater morbidity, mortality, and cost, although that relationship hasn’t been absolutely established yet the way it has with colonoscopy and colorectal cancer, according to Dr. Mosko.
Variation in endoscopy performance can be caused by the patchy and subtle appearance of dysplasia, and because procedural guidelines are not always closely followed. There is often a significant difference between procedures performed by specialists and nonspecialists. “Endoscopists in general don’t take enough time to examine the segment, they don’t wash appropriately, and when they do look, they may not be well enough trained to know what they’re looking at. The only way to improve on this aside from additional training is to have a metric that measures how you’re doing, and I think [the neoplasia detection rate] is as close as we get to doing that. I think the exact threshold for NDR is not as important as figuring out what your number is and then ways to improve it,” said Dr. Mosko.
A recent meta-analysis estimated NDR to be 7%, but the patient cohort used was derived from referrals to academic centers, where experienced gastroenterologists may register a higher than average NDR. The study also lacked data on patients, providers, or biopsy quality, which prevented assessment of the effects of NDR on subsequent missed dysplasia or predictors of high or low NDR.
To get a better estimate of NDR, researchers led by Lovekirat Dhaliwal, MD, at the Mayo Clinic in Rochester, Minn., analyzed data from the Rochester Epidemiology Project, including patients from 11 counties in Minnesota. They identified 1,066 patients with Barrett’s esophagus, 71.1% of whom were male, with a mean age of 63 years. 77% had surveillance endoscopies performed by gastroenterologists, the remainder by nongastroenterologists such as doctors, surgeons, or internal medicine physicians. About 60% of participants received adequate biopsies per Seattle protocol.
The NDR was 4.9% (95% CI, 3.8%-6.4%), including 3.1% high-grade dysplasia (HGD) and 1.8% EAC. One-quarter of EAC cases had metastatic lymphadenopathy at endoscopy or surgery, and 10.6% had low-grade dysplasia (LGD). Although high-definition monitors and high-resolution endoscopes were added to practices, particularly after 2000, the researchers found no evidence of increasing NDR over time on multivariate analysis. In a separate analysis of targeted biopsies in 54 patients with a visible lesion, 9 had LGD (7.96% of all LGD diagnoses) and 10 had EAC (50.0% of all EAC diagnoses). Visible lesions were more often reported by gastroenterologists than nongastroenterologists (odds ratio, 3.7; P = .0120). Gastroenterologists had a higher rate of NDR on univariate analysis (5.8% vs. 1.7%; P = .0098).
There were 391 Barrett’s esophagus patients with no diagnosis of HGD or EAC at the initial endoscopy underwent another endoscopy at 12 months. At the follow-up procedure, eight patients were found to have HGD/EAC, amounting to 13% of HGD/EAC cases being missed at the index endoscopy. There was no statistically significant association between a missed dysplasia or found dysplasia and segment length (4.7 cm vs. 3.7 cm; P = .4), Seattle protocol adherence (62% vs. 58.7%; P = .8), visibility of lesions (OR, 0.6; P = .55), age, smoking history, or practitioner specialty.
The study was funded by the National Cancer Institute and the National Institute of Aging. Dr. Mosko has no relevant financial disclosures.
SOURCE: Dhaliwal L et al. Clin Gastro Hepatol. 2020 Jul 21. doi: 10.1016/j.cgh.2020.07.034.