Related Experiment Video
Updated: Mar 18, 2026
Heterogeneity Mapping of Protein Expression in Tumors using Quantitative Immunofluorescence
Published on: October 25, 2011
The Chemotherapy Response Score (CRS): Interobserver Reproducibility in a Simple and Prognostically Relevant System
Ian Said1, Steffen Böhm, Joanne Beasley
1Department of Cellular Pathology (I.S., A.Z.F., S.J., S.L.S.L., J.M., T.O.M., N.S.) Department of Medical Oncology (S.B.), Barts Health NHS Trust Department of Pathology (P.E., G.T.), University College Hospital NHS Trust, London Department of Cellular Pathology (J.B., R.G., L.H., R.M., J.V.), Birmingham Women's NHS Trust, Birmingham Department of Pathology (W.G.M.), Belfast Health and Social Care Trust, Royal Victoria Hospital, Belfast, UK Department of Anatomic Pathology (C.B.G.), Vancouver General Hospital, Vancouver, BC, Canada.
Abstract:
A 3-tier histopathologic scoring system, the chemotherapy response score (CRS), was previously devised for reporting the histologic response to neoadjuvant chemotherapy in interval debulking surgery specimens of stage IIIc/IV tuboovarian high-grade serous carcinoma. This has been shown to predict the outcome and offer additional information to other methods of assessing the treatment response. In the present study, the reproducibility of this scoring system was assessed by determining the interobserver agreement among reporting pathologists. A total of 5 groups each comprising 3 pathologists with different levels of expertise were selected. The participants underwent an online tutorial on how to apply the CRS system. 40 cases (38 cases in 2 appraiser groups) were scored individually by each of the 15 pathologists. The interobserver reproducibility was calculated using Fleiss' κ, Kendall's coefficient of concordance, and the absolute agreement between (a) individual pathologists within 1 group, (b) with the majority score agreement between all groups, and (c) with all individual scores. The CRS system was found to be highly reproducible among all the pathologists' groups (κ=0.761). The agreement in identifying the group of patients with the best response to chemotherapy was exceptionally high (κ=0.926). We conclude that CRS has a high interobserver reproducibility, especially in identifying the subgroup of patients with the best chemotherapy response, justifying its inclusion in clinical trials and reporting practice.
