POOR AGREEMENT BETWEEN BIOPSIES AND ENDOSCOPIC SUBMUCOSAL DISSECTION (ESD) SPECIMENS ON GASTRIC LESIONS: A SPANISH MULTICENTRE COHORT STUDY

Aims:

To assess the agreement between the biopsy specimens obtained during the diagnostic upper endoscopy and the ESD specimen for gastric lesions.

Methods:

The demographic and clinical characteristics of the patients and lesions were prospectively collected. Morphological features of the tumors and technical factors were recorded.

Results:

We performed ESD in 55 gastric lesions. Before performing ESD, biopsy specimens were available in 48 cases (87.3%). The recruitment was done in 12 centres from January 2016 to June 2017. Mean lesion size was 33.8 ± 15.8 mm. En bloc resection was achieved in 51 cases (92.7%). The procedure was finished with a piecemeal resection in 3 cases (5.5%) and surgery was needed in 1 case (1.8%) owing to technical difficulties that led to an aborted procedure. Histological results agreed in 22 lesions (45.8%). The biopsy specimen overestimate the final assessment in 7 cases (14.6%). An underestimation of the definitive histological result was observed in 19 lesions (39.6%) as shown in the following table:

Tab. 1:

Inter-observer agreement

	Definitive histology
Biopsy specimen	Non-neoplastic/Low grade dysplasia	High grade dysplasia	Intramucosal carcinoma	Submucosal invasion
Non-neoplastic/Low grade dysplasia	7	7	1	1
High grade dysplasia	3	12	4	5
Intramucosal carcinoma	1	2	3	1
Other	1	0	0	0

The higher inter-observer agreement was observed for non-neoplastic lesions and those with low grade dysplasia (58.3%). All the tumors that showed submucosal invasion in the ESD specimen were underestimated in the previous biopsies (agreement: 0%). Weighted kappa was 0.19 (IC 95%: -0.06 – 0.45; p= 0.13).

Conclusions:

The preoperative biopsy specimens obtained during diagnostic upper endoscopy often underestimates the definitive results for gastric lesions. ESD allows an accurate diagnosis and may help to select patients requiring additional therapies.