Reproducibility and validation of tumour stroma ratio scoring on oesophageal adenocarcinoma biopsies

Background: Tumour stroma ratio (TSR) in histological sections of resected oesophageal adenocarcinomas proved to be a prognostic factor for patients' survival. The objectives of this study were to assess inter- and intraobserver agreement for TSR scoring on biopsy material and to validate these biopsy results with the results derived from surgical specimens. Methods: Biopsies and surgical specimens of 91 patients with oesophageal adenocarcinoma were available. TSR was determined on the original haematoxylin-eosin (H&E) tissue sections from primary tumour biopsies. To assess interobserver variation, TSR was scored by three pathologists as 0-25%, 25-50%, 50-75% or 75-100%. A second scoring was done to examine intraobserver variation. The definitive TSR biopsy score was compared with the corresponding resection specimen score. Kappa statistics were applied to evaluate agreement. Results: Biopsies of 10 (11%) patients were rejected because of poor quality. For 81 TSR biopsy scores, interobserver correlations ranged between 0.239 and 0.486 (P < 0.001 for all). By classifying scores into two groups (<50% and >= 50%), interobserver correlations ranged between 0.372 and 0.886 (P < 0.001 for all). Intraobserver agreement was substantial to near-perfect (kappa = 0.780-0.848; P < 0.001 for all). Definitive TSR biopsy score showed moderate correlation with TSR scores on surgical specimens (kappa = 0.506), but it was an independent prognostic factor for survival. Conclusion: Reproducibility of tumour stroma ratio scoring on oesophageal adenocarcinoma biopsies was good. The ease of TSR scoring on H&E sections together with its correlation with patients' survival may have clinical relevance in this era of neoadjuvant therapy. (C) 2010 Elsevier Ltd. All rights reserved.