Premium
An audit of residual cancer burden reproducibility in a UK context
Author(s) -
Naidoo Kalnisha,
Parham David M,
Pinder Sarah E
Publication year - 2017
Publication title -
histopathology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.626
H-Index - 124
eISSN - 1365-2559
pISSN - 0309-0167
DOI - 10.1111/his.13054
Subject(s) - concordance , context (archaeology) , medicine , concordance correlation coefficient , breast cancer , confidence interval , reproducibility , cancer , kappa , spearman's rank correlation coefficient , categorical variable , residual , statistics , mathematics , algorithm , paleontology , geometry , biology
Aims The residual cancer burden score ( RCB ) is currently the preferred quantification tool for assessing residual disease following neoadjuvant chemotherapy ( NACT ) in breast cancer clinical trials. This has been shown to be highly reproducible at the MD Anderson Cancer Centre, where it was developed originally. We wanted to evaluate RCB in a UK context, where macroscopic handling of tissue may differ between sites. Methods and results The pathology slides from 90 post‐ NACT patients from Guy's and St Thomas’ NHS Foundation Trust and the Royal Bournemouth Hospital were reviewed independently by two specialist breast histopathologists who recalculated the RCB for each case. Data were collated and analysed statistically for interobserver reproducibility, for both numerical and categorical RCB . Overall, agreement between pathologists was ‘good’ [kappa = 0.775; 95% confidence interval ( CI ) = 0.668–0.882]. The overall concordance for continuous RCB score and for categorical RCB group was statistically significant (Spearman's correlation coefficient = 0.9497; 95% CI = 0.9235–0.9671; P < 0.0001 and Spearman's correlation coefficient = 0.9145; 95% CI = 0.8712–0.9437; P < 0.0001, respectively). Discordance could not be attributed to any one component of the RCB calculation. Conclusions These data suggest that the RCB score is reproducible in a UK context. Further data comparing it to other quantification systems is required, however, before any superiority can be established.