CONFOCAL LASER ENDOMICROSCOPY IN COMBINATION WITH MOLECULAR BIOMARKERS FOR THE DIAGNOSIS OF INCONSPICUOUS DYSPLASIA IN BARRETT'S ESOPHAGUS: A RANDOMISED CROSS-OVER TRIAL
Presentation Number: 227View Presentation
AuthorBlock: Mathew Vithayathil1, Ines Modolell2, Jacobo Ortiz Fernández-Sordo3,4, Dahmane Oukrif5, Apostolos Pappas1, Wladyslaw Januszewicz1, Maria O'Donovan2, Andreas V. Hadjinicolaou1,2, Michele Bianchi1, Adrienn Blasko1, Jonathan White3,4, Philip Kaye4, Marco Novelli5, Krish Ragunath6,7,3, Massimiliano Di Pietro1
1MRC Cancer Unit, University of Cambridge, Redhill, United Kingdom; 2Cambridge University Hospitals NHS Foundation Trust, Cambridge, Cambridgeshire, United Kingdom; 3Nottingham Digestive Diseases Biomedical Research Centre, Nottingham, United Kingdom; 4Nottingham University Hospitals NHS Trust, Nottingham, Nottingham, United Kingdom; 5University College London Hospitals NHS Foundation Trust, London, London, United Kingdom; 6Curtin University, Perth, Western Australia, Australia; 7Royal Perth Hospital, Perth, Western Australia, Australia;
Dysplasia in Barrett’s esophagus (BE) is often invisible on white-light endoscopy. The combination of probe-based confocal endomicroscopy (pCLE) and a molecular biomarkers had high accuracy for dysplasia in a previous pilot study (1). We aimed to compare the diagnostic accuracy for dysplasia of the combination of autofluorescence-imaging (AFI) guided pCLE and molecular biomarkers versus high-resolution white-light endoscopy (HRWLE) and random biopsies in BE patients without visible lesions.
This prospective multi-centre randomised cross-over trial included patients with BE at least 2cm in length and no evidence neoplastic lesions on HRWLE. Eligible patients were randomised to standard endoscopy (HRWLE with Seattle protocol biopsies) or AFI-guided pCLE and targeted biopsies for molecular biomarkers. Patients crossed over to the other arm after 6-12 weeks. Procedures were performed by 6 endoscopists, who were blinded to referral histology and results of previous trial endoscopy. Expression of p53 and Cyclin A was assessed by immunohistochemistry and aneuploidy was measured by image cytometry on AFI-targeted biopsies. Histological endpoint was diagnosis from all study biopsies (trial histology). Sensitivity analysis was performed including referral diagnoses within the 12 months of enrolment (overall histology). The primary outcome was dysplasia diagnosis by real-time by pCLE and standard endoscopy. The secondary outcome was the added value of molecular biomarkers.
Of the 153 patients recruited, 134 completed both arms, of which 19 (14.2%) received a diagnosis of low-grade dysplasia (LGD) and 18 (13.4%) were diagnosed with high-grade dysplasia (HGD) or intramucosal cancer (IMC). AFI-guided pCLE had a sensitivity and specificity for all grades of dysplasia of 73.0% and 67%. Standard endoscopy had equal sensitivity for dysplasia (73.0%). When including referral histology, pCLE and standard endoscopy had similar sensitivity for dysplasia (61.8% vs. 49.1%, p=0.09). Multivariate logistic regression showed that optical dysplasia on pCLE, aberrant p53 and aneuploidy had the strongest correlation with overall histology (p<0.05). A panel comprising these 3 biomarkers showed a sensitivity and specificity for dysplasia of 89.2% and 56.7%, respectively, with an AROC of 0.83. This 3-biomarker panel had a higher sensitivity than Seattle protocol in detecting dysplasia in both trial (89.2% vs 73.0%, p=0.06) and referral (80.0% vs 49.1%, p<0.001) histology.
Inconspicuous dysplasia in BE is missed by random Seattle protocol biopsies in approximately 50% of cases. AFI-targeted pCLE detects inconspicuous dysplasia with similar accuracy to the current gold standard. The addition of molecular biomarkers further improves the diagnostic accuracy.
1. Di Pietro M et al.. Am J Gastroenterol. 2015 Nov; 110(11):1549-58