International Ki67 Reproducibility Study Shows High Intralaboratory but Only Moderate Interlaboratory Reproducibility
In a study reported in Journal of the National Cancer Institute, Mei-Yin C. Polley, PhD, of the National Cancer Institute, and colleagues assessed intralaboratory and interlaboratory variability in Ki67 scoring. They found high intralaboratory reproducibility but only moderate interlaboratory reproducibility, indicating that standardization of scoring methodology is needed.
Study Details
Eight laboratories in North America and Europe participated in the study. In the examination of intralaboratory reproducibility, six laboratories used their own local protocol to stain one section from a 50-case tissue microarray block and scored Ki67 using their own standard scoring method on 3 separate days. In the examination of interlaboratory reproducibility, eight laboratories received two sets of tumor microarray sections containing the same 100 cores, one centrally stained for Ki67 and the other stained by each laboratory following its own local protocol within 2 weeks of cutting.
Staining was performed with antibody MIB-1. All laboratories scored Ki67 as the percentage of positively stained invasive tumor cells. Reproducibility was quantified by intraclass correlation coefficient (ICC). Sources of variation were analyzed using random effects models with log2-transformed measurements.
Intralaboratory Reproducibility
Intralaboratory reproducibility was high, with an intraclass correlation coefficient of 0.94 (95% confidence interval [CI] = 0.93–0.97). The largest sources of total variability appeared to be patient biology followed by the interaction between patient and laboratory effects, laboratory variability, and residual errors.
Interlaboratory Reproducibility
Interlaboratory reproducibility was moderate, with intraclass correlation coefficients of 0.71 (95% CI = 0.47–0.78) for central staining and 0.59 (95% CI = 0.37–0.68) for local staining. The geometric mean of Ki67 values among laboratories across the 100 cases examined ranged from 7.1% to 23.9% with central staining, indicating substantial differences in Ki67 measurement using centrally stained slides from the same cases, and from 6.1% to 30.1% with local staining, indicating large interlaboratory variability in Ki67 measurement. In the interlaboratory evaluations, the largest sources of variability appeared to be biologic variability among patients followed by residual errors and between-laboratory variability.
The laboratories that showed the highest degree of intralaboratory reproducibility tended to generate the most similar Ki67 values. An exploratory analysis showed that interlaboratory variation was lower among laboratories using formal counting techniques compared with laboratories using visual estimation.
The investigators concluded, “Substantial variability in Ki67 scoring was observed among some of the world’s most experienced laboratories. Ki67 values and cutoffs for clinical decision-making cannot be transferred between laboratories without standardizing scoring methodology because analytical validity is limited.”
This study was supported by the Breast Cancer Research Foundation.
Torsten Nielsen, MD, PhD, FRCPRC, of Vancouver Hospital & Health Sciences Centre, is the corresponding author for the Journal of the National Cancer Institute article.
The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.
