Ki67 is a nuclear protein expressed during the active phases of the cell cycle, which makes it a biomarker of cell proliferation. In clinical pathology settings, immunohistochemical (IHC) detection of Ki67 is used to calculate Ki67 proliferation indices (PIs), which have prognostic information and are used to subdivide breast carcinomas and neuroendocrine neoplasias. Calculation of Ki67 PIs is notoriously hard and prone to intraobserver and interobserver variance. In addition, IHC protocol settings [such as primary antibody (Ab) clone, clone format, and stainer platform] can affect the result of the IHC assays and in turn the Ki67 PI. Digital image analysis has been suggested as a useful tool to standardize Ki67 counting. Recently, virtual double staining, a computer algorithm segmenting Ki67+ and Ki67− tumor cells using digitally fused parallel cytokeratin and Ki67-stained slides, has been introduced. In this study, we compare Ki67 PIs obtained by virtual double staining in 41 breast carcinomas stained using the most commonly used commercially available primary Ab clones and formats on the main stainer platforms. IHC protocols for the concentrated (conc) Ab and platform combinations were optimized for the highest analytical sensitivity and optimal signal-to-noise ratio, whereas ready-to-use (RTU) formats were used, as recommended by the vendor. Significant differences in the mean Ki67 PIs (relativized to the mean core Ki67) were observed not only between the different Ab clones but also the different formats and stainer platforms; Ki67 PIs with SP6 conc stained on the Ventana BenchMark ULTRA platform were on average 11.9 percentage points (pp) higher than the mean core average, whereas with Ab 30.9 RTU on the Ventana platform, they were 10.4 pp higher. Mib1 RTU (Dako Autostainer Link 48) and MM1 RTU (Leica Bond) provided 8.6 and 12.5 pp lower Ki67 PIs, respectively. Mib1 conc and SP6 conc on the Dako Autostainer and Leica Bond provided similar results—close to the overall average. Significant variations in the proportion of tumors with Ki67 high-level expression (Ki67 PI ≥20%) were observed among Ab, format, and stainer platform combinations. The results underline the challenges in the comparison of Ki67 PIs across Abs, formats, and platforms. Researchers and clinicians need to account for these differences when reporting Ki67 PIs. To advance the usefulness of Ki67 PIs in the research and clinical setting, standardization of Ki67 IHC assays is needed.
*Institute of Pathology, Aalborg University Hospital
†Department of Clinical Medicine, Aalborg University, Aalborg
‡Institute of Pathology, Aarhus University Hospital, Aarhus N, Denmark
The authors declare no conflict of interest.
Reprints: Rasmus Røge, MD, Institute of Pathology, Aalborg University Hospital, Ladegaardsgade 3, Aalborg DK-9000, Denmark (e-mail: firstname.lastname@example.org).
Received March 5, 2019
Accepted June 26, 2019
Online date: August 26, 2019