|Home | About | Journals | Submit | Contact Us | Français|
Surveillance strategies or chemoprevention of clinically diagnosed Barrett's oesophagus are unlikely to alter population mortality from oesophageal adenocarcinoma since it is estimated that only 5% patients with Barrett's oesophagus are diagnosed, despite the increasing use of endoscopy.1 Endoscopic screening for Barrett's oesophagus in white men aged over 50 years with chronic heartburn has therefore been advocated in the US.2 However, the prohibitive cost of endoscopy and the pathological expertise required make this screening programme problematic.
The aim of this study was to develop a screening test that did not rely on endoscopy and would be applicable to primary care. A non‐endoscopic cytological test, which has been successfully used for the diagnosis of oesophageal squamous carcinomas in high‐risk areas, was adapted by combining it with immunocytology. The device (bought from Nanru Technologies, Bellville, South Africa) consists of a sponge contained within a gelatine capsule that dissolves in the stomach within 5 min (fig 1A,B1A,B).3 Liquid‐based cytology was used to create a cell monolayer and the slides created were stained for minichromosome maintenance protein 2 (Mcm2). Mcm2 is a proliferation marker that is abnormally expressed at the luminal surface in Barrett's oesophagus, with expression indices that correlate with the degree of dysplasia.4
Following approval from the Cambridge Local Research Committee, 43 patients with known Barrett's oesophagus (defined by endoscopy and a diagnosis of intestinal metaplasia) and 54 healthy volunteers were recruited (table 11).). Diagnostic endoscopy was voluntary for healthy subjects, and, unfortunately, almost all (95%) refused. The participants rated the acceptability of the capsule sponge as 4.0 (2.0–9.0) using a linear scale of 0 (worst imaginable experience) to 10 (very enjoyable experience), where 5 was neither pleasant nor unpleasant. Of the patients with Barrett's oesophagus, 80% stated that they would choose the capsule sponge in preference to endoscopic surveillance. There were no complications, and 91 of 96 (94.8%) samples were adequate for analysis. The squamous (red arrow) and columnar cells (black arrow) were easily distinguishable (fig 1C,D1C,D)) and good‐quality immunostaining for Mcm2 was obtained (fig 1D1D).
A total of 27 of 40 (67.5%) specimens of Barrett's oesophagus were positive compared with 17 of 52 (32.6%) specimens from healthy volunteers, giving a sensitivity and specificity of 67.5% and 67.3%, respectively. However, patients with Barrett's oesophagus with a positive test had a statistically longer segment than those with a negative test (p=0.01); therefore, if only patients with Barrett's oesophagus with a segment longer than 3 cm are considered as true Barrett's oesophagus, the sensitivity and specificity become 76.0% and 62.7%, respectively. Since most volunteers did not undergo endoscopy, it was impossible to ascertain if they had Barrett's oesophagus. Heartburn is the primary risk factor for the presence of Barrett's oesophagus, and since nine (53%) Mcm2 positive controls were patients with a history of heartburn vs three (8%) for the Mcm2 negative patients, there is a possibility that patients with heartburn might have Barrett's oesophagus. To correct for this possible bias, a post hoc analysis excluding all patients with heartburn was performed and the percentage of controls with a positive test result decreased from 37.5% to 20%. In this subgroup analysis, the sensitivity and specificity were 67.5% and 80% for all patients, and 54.3% and 86.7%, respectively, for patients with Barrett's oesophagus with a segment longer than 3 cm.
In summary, the capsule sponge is acceptable to patients and applicable to primary care, thus bypassing the need for screening endoscopy. The immunostained monolayer specimen is amenable to automated microscopy processing. However, further refinements in technique to increase the sensitivity and specificity of this test are necessary before population‐based studies can be performed, especially since the quoted specificity applies to a population with a 44% incidence of Barrett's oesophagus, and will be much lower if used in the general population. Overall, this proof‐of‐principle study suggests that a non‐endoscopic screening test for Barrett's oesophagus is attainable and warrants further investigation.
Competing interests: None.