Introduction: Minimally invasive devices such as a sponge-on-string device (SoS) combined with discriminant markers are essential to make screening for Barrett's esophagus (BE) a practical and effective strategy. Such approaches must be safe, well-tolerated, and provide sufficient analyte for accurate diagnosis. In a randomized blinded study, we assessed two configurations of a SoS device for the following metrics: success rates for swallowing and retrieval, tolerability, preference relative to endoscopy, DNA yield, and BE detection by assay of selected methylated DNA markers.
Methods: A SoS device (EsophaCap, Capnostics) was swallowed by and withdrawn from BE cases and non-BE controls confirmed by endoscopy done within 24 hours. Participants were randomized to a 25 mm/10 pores per inch (ppi) or a 25 mm/20 ppi sponge. Following SoS retrieval, participants completed a tolerability assessment (rating pain, choking, gagging and anxiety separately plus overall tolerability, on a 1-10 visual analog scale); they also indicated if they would have the procedure again and if they preferred SoS or endoscopy. A mucosal injury score (ranging from 1 (no trauma) to 6 (perforation)) was assessed from video recordings of the subsequent endoscopy. Following retrieval, the sponge was placed in a cell-preservative solution and agitated; the cell suspension was centrifuged and the resulting pellet then treated with lysis buffer. DNA extracted from lysis buffer was quantified, and our previously described methylated DNA markers for BE were assayed by methylation-specific PCR.
Results: 41 subjects (20 BE cases, 21 controls) were randomized: 21 to 25/10 ppi SoS and 20 to 25/20 ppi SoS. The capsule was successfully swallowed and retrieved in 40 subjects (98%). One BE case was found to not meet entry criteria and was excluded. Median age of cases (66) and controls (61) was comparable; 79% of cases and 50% of controls were men. Median BE segment length was 5 cm (IQR 4-8). Results of outcomes assessed are shown (Table). There were no major adverse events. At 100% specificity, sensitivity for BE by a 2-marker panel (methylated VAV3&ZNF682) was 100% for both sponge configurations.
Conclusion: Non-endoscopic detection of BE with a SoS device and assay of selected methylation markers appears to be safe, well-tolerated, and accurate. The 25/10 ppi sponge was better tolerated and yielded highest DNA amount. Further studies are warranted to corroborate and extend these encouraging findings.