Digital image analysis of Ki67 proliferation index in breast cancer using virtual dual staining on whole tissue sections: clinical validation and inter-platform agreement

Abstract

PurposeThe Ki67 proliferation index is a prognostic and predictive marker in breast cancer. Manual scoring is prone to inter- and intra-observer variability. The aims of this study were to clinically validate digital image analysis (DIA) of Ki67 using virtual dual staining (VDS) on whole tissue sections and to assess inter-platform agreement between two independent DIA platforms.MethodsSerial whole tissue sections of 154 consecutive invasive breast carcinomas were stained for Ki67 and cytokeratin 8/18 with immunohistochemistry in a clinical setting. Ki67 proliferation index was determined using two independent DIA platforms, implementing VDS to identify tumor tissue. Manual Ki67 score was determined using a standardized manual counting protocol. Inter-observer agreement between manual and DIA scores and inter-platform agreement between both DIA platforms were determined and calculated using Spearman’s correlation coefficients. Correlations and agreement were assessed with scatterplots and Bland–Altman plots.ResultsSpearman’s correlation coefficients were 0.94 (p < 0.001) for inter-observer agreement between manual counting and platform A, 0.93 (p < 0.001) between manual counting and platform B, and 0.96 (p < 0.001) for inter-platform agreement. Scatterplots and Bland–Altman plots revealed no skewness within specific data ranges. In the few cases with ≥ 10% difference between manual counting and DIA, results by both platforms were similar.ConclusionsDIA using VDS is an accurate method to determine the Ki67 proliferation index in breast cancer, as an alternative to manual scoring of whole sections in clinical practice. Inter-platform agreement between two different DIA platforms was excellent, suggesting vendor-independent clinical implementability.Electronic supplementary materialThe online version of this article (10.1007/s10549-018-4669-2) contains supplementary material, which is available to authorized users.

Analytical validation of a standardised scoring protocol for Ki67 immunohistochemistry on breast cancer excision whole sections: an international multicentre collaboration

“…NordiQC), involving comparison of laboratory scores with reference scores in periodic assessment challenges, will probably improve interobserver reproducibility further. Recent studies suggest that an even higher level of concordance can be achieved with automated image analysis . The IKWG is actively conducting studies in this area to assess how artificial intelligence may help to standardise Ki67 assessment .…”

“…Recent studies suggest that an even higher level of concordance can be achieved with automated image analysis. [35][36][37][38] The IKWG is actively conducting studies in this area to assess how artificial intelligence may help to standardise Ki67 assessment. 35,38 Also, concordance between Ki67 scores on core biopsies and excision specimens is currently being investigated.…”

Quantitative chromogenic immunohistochemical image analysis in cellprofiler software

“…The data were presented as fitted regression lines with 95% confidence forecast bands and coefficient of determination ( R ‐squared). Inter‐platform agreement between the different methods of manual counting and image analysis was further assessed using Bland–Altman plots . The data were calculated using the difference between the results from Positive Cells‐Manual (combined [counters 1–3] or individually [counters 1–4]) and Positive Cells‐CellProfiler or DAB‐regions‐CellProfiler against the mean of the two measurements .…”

Overview of multiplex immunohistochemistry/immunofluorescence techniques in the era of cancer immunotherapy

“…At the recent 2017 St. Gallen International Expert Consensus Conference, an issue of caution was raised concerning the reproducibility of IHC for Ki‐67 and the implications of this variability for clinical decisions [50]. Multiple groups have previously demonstrated that inter‐observer variability can be negated by scoring Ki‐67 in a reproducible and quantitative manner using digital analysis; for instance, using the Definiens Tissue Studio software (Definiens, Munich, Germany) and the Aperio ePathology image analysis system (Leica Biosystems, Wetzlar, Germany) [51‐54]. This mIHC/IF digital analysis can also be performed using the PerkinElmer Vectra platform (PerkinElmer, Inc., Waltham, MA, USA) followed by inForm software (PerkinElmer, Inc., Waltham, MA, USA) [55‐57], which circumvents this limitation as well.…”