Fig 1.
Study design.
Table 1.
Immunhistochemistry staining specifications.
Table 2.
Instrument-specific relative units (RU) for CFX96TM (BIO-RAD®) and the corresponding IHC equivalents.
Fig 2.
Correlation between original (ORI) and paired re-evaluated (REV) or digital image analyzed (DIA) samples for IHC.
Bland-Altman plots showing differences between original, re-evaluated and digital image analysis for ER (A, B), PR (C, D) and Ki67 (E, F). Cut-off value at ±10% is shown by red dotted lines. Scatterplots showing correlation between original, re-evaluated and digital image analysis for Ki67 (G, H). Cut-off value at 20% for Ki67 is shown by black dotted lines. Red dots highlight discordant cases between original and either re-evaluated or digital image analysis.
Table 3.
Median proportion scores for IHC.
Fig 3.
Comparison between IHC and RT-qPCR (MammaTyper®) for estrogen receptor (A, B), progesterone receptor (C, D), proliferation marker Ki67 (E, F) and human epidermal growth factor receptor 2 (G, H). Scatterplots with ORI IHC ER scores (A) or DIA IHC ER score (B) on the y-axis and ESR1 RU values on the x-axis. Cut-off values at 1% for IHC and 0.0 RU for RT-qPCR are shown by black dotted lines. Scatterplots with ORI IHC PR scores (C) or DIA IHC PR score (D) on the y-axis and PGR RU values on the x-axis. Cut-off values at 1% for IHC and 0.0 RU for RT-qPCR are shown by black dotted lines. Scatterplots with ORI IHC Ki67 scores (E) or DIA IHC Ki67 score (F) on the y-axis and MKI67 RU values on the x-axis. Cut-off values at 20% for IHC and 0.0 RU for RT-qPCR are shown by black dotted lines. Scatterplots showing correlation between IHC determination (ORI and REV) and RT-qPCR assessment for ERBB2/HER2 (G, H). Black dotted lines represent the respective cut-off for the RT-qPCR assay (ERBB2: 0.0) and for the positivity of the IHC staining (HER2: 2+). Second dotted line on the x-axis represents a second cut-off for ERBB2 (-0.7) after introduction of an equivocal zone.
Table 4.
Agreement within IHC (ORI versus REV, ORI versus DIA) and between IHC (ORI) and RT-Qpcr.
Fig 4.
Schematic analysis of discrepancies in HER2: a. between the original assessment (ORI) and re-evaluation (REV) and b. between IHC (ORI) and RT-qPCR. Positive calls are indicated by red filled boxes. Negatives calls are presented in green and equivocal calls in grey, according to the 2013 ASCO/CAP guidelines in force at the time of the study.
Fig 5.
Subtype redistribution after re-evaluation (REV) (A), semi-automated analysis (DIA), (B) and RT-qPCR measurement (C).