Endoscopy 2018; 50(04): S34-S35
DOI: 10.1055/s-0038-1637129
ESGE Days 2018 oral presentations
20.04.2018 – Esophagus: SCC, Barrett's, GERD
Georg Thieme Verlag KG Stuttgart · New York

IS IT REASONABLE TO PROPOSE AN ENDOSCOPIC MUCOSAL RESECTION FOR BARRET'S OESOPHAGUS WITH HIGH GRADE DYSPLASIA ON THE BIOPSIES?

JM Gonzalez
1   Hôpital Nord, AP-HM, Aix Marseille Univ., Gastroenterology, Marseille, France
,
JP Casanova
2   Aix-Marseille Univ, Hôpital Nord, AP-HM, Gastroenterology, Marseille, France
,
E Bories
3   Paoli Calmettes Institute, UEMCO, Marseille, France
,
S Garcia
4   Aix-Marseille Univ, Hôpital Nord, AP-HM, Pathology, Marseille, France
,
M Giovannini
3   Paoli Calmettes Institute, UEMCO, Marseille, France
,
M Barthet
2   Aix-Marseille Univ, Hôpital Nord, AP-HM, Gastroenterology, Marseille, France
› Author Affiliations
Further Information

Publication History

Publication Date:
27 March 2018 (online)

Also available at
 

Aims:

Endoscopic mucosal resection (EMR) is currently the treatment for high grade dysplasia (HGD) in Barret's esophagus (BE), based on biopsies. The aim was to evaluate the discordance between biopsy sample (BS) and EMR specimen in case of HGD, and cytologists’ inter and intra-observatory variability.

Methods:

Retrospective study including a prospective histological relecture (BS and specimen). Inclusion criteria were: BE with HGD on pre-operative biopsies resected by EMR. The initial biopsies were collected and re-exanimated. The BS discordant with EMR specimens were recorded in a numeric file and a second lecture was carried out by 2 experts and 2 fellows. Five diagnoses were considered: no metaplasia, metaplasia without dysplasia, LGD, HGD, Adenocarcinoma. Concordance statistical tests were performed to assess inter and intra-observatory variability.

Results:

87 patients underwent EMR for HGD. 47% had a discordant result between biopsies and resection specimen. Finally, 33 patients could be analyzed, 29 men and 4 women, with a mean age of 63 years old. The mean length of BE was C3-M5, with relief abnormalities in 18.2%. A mean number of 1.4 endoscopic session was performed, with a mean of 2.7 pieces per EMR. The mean follow-up was 38 months. The Kappa coefficient for the diagnosis of HGD was low on the initial BS, and ranged between 0 and 0.6 for the EMR specimen. The inter-observatory concordance was 0.2 for the diagnosis of HGD. The kappa coefficient regarding HGD between experts was 0.5 for biopsies and 0 for EMR, whereas it was 0.4 and 0.5 for the fellows. The intra-observatory ranged between 0 and 0.6.

Conclusions:

The discordance rate between HGD on BS and results on EMR specimen is high, around 47%. The intra and inter observatory concordance is insufficient, even in expert centers. Thus, the question about random biopsies for BE rather than endoscopic assessment has to be asked.