Measuring depth of invasion of submucosa - invasive adenocarcinoma in oesophageal endoscopic specimens: how good are we?

Aims: Emerging data support that submucosa-invasive (pT1b) esophageal adenocarcinomas are cured via endoscopic resection, provided that invasion measures <= 500 mu m, they lack other histological features predictive of nodal metastasis and have negative margins. Hence, pathologists' measurement of the depth of submucosal invasion in endoscopic resections may dictate further management (i.e. endoscopic follow-up versus oesophagectomy). In this study, we assessed the interobserver agreement in measuring the depth of submucosal invasion in oesophageal endoscopic resections. Methods and results: Six subspecialised gastrointestinal (GI) pathologists from five academic centres independently measured the depth of submucosal invasion in mu m from the deepest muscularis mucosae on 37 oesophageal endoscopic resection slides (round 1 scoring). A consensus meeting with a systematic approach for measuring and discussion of pitfalls was undertaken and remeasuring (round 2 scoring) was conducted. Interobserver agreement was assessed by the intraclass correlation coefficient (ICC) and Cohen's kappa statistics. A lack of agreement was seen among the six reviewers with a poor ICC for both rounds: 1 [0.40, 95% confidence interval (CI) = 0.26-0.56] and 2 (0.49, 95% CI = 0.34-0.63). When measurements were categorised as < or >500 mu m, the overall agreement among the six reviewers was only fair for both rounds: 1 (kappa = 0.37, 95% CI = 0.22-0.53) and 2 (kappa = 0.29, 95% CI = 0.12-0.46). Conclusions: Our study shows a lack of agreement among gastrointestinal pathologists in measuring the depth of submucosal invasion in oesophageal endoscopic resections despite formulating a consensus approach for scoring. If important management decisions continue to be based upon this parameter, more reproducible and concrete guidelines are needed.