BackgroundThe Ki-67 Labelling Index (LI) is used as an ancillary tool in glioma diagnostics. Interobserver variability has been reported and no precise guidelines are available. Nor is it known whether novel digital approaches would be an advantage. Our aim was to evaluate the inter- and intraobserver variability of the Ki-67 LI between two pathologists and between pathologists and digital quantification both in whole tumour slides and in hot spots using narrow but diagnostically relevant intervals.MethodsIn samples of 235 low and high grade gliomas, two pathologists (A and B) estimated the Ki-67 LI (5–10% intervals) for whole tumour slides and for hot spots. In 20 of the cases intraobserver variability was evaluated. For digital quantification (C) slides were scanned with subsequent systematic random sampling of viable tumour areas. A software classifier trained to identify positive and negative nuclei calculated the Ki-67 LI. The interobserver agreements were evaluated using kappa (κ) statistics.ResultsThe observed proportions of agreement and κ values for Ki-67 LI for whole tumour slides were: A/B: 46% (κ = 0.32); A/C: 37% (κ = 0.26); B/C: 37% (κ = 0.26). For hot spots equivalent values were: A/B: 14% (κ = 0.04); A/C: 18% (κ = 0.09); B/C: 31% (κ = 0.21).ConclusionsInterobserver variability was pronounced between pathologists and for pathologists versus digital quantification when attempting to estimate a precise value of the Ki-67 LI. Ki-67 LI should therefore be used with caution and should not be over interpreted in the grading of gliomas. Digital quantification of Ki-67 LI in gliomas was feasible, but intra- and interlaboratory robustness need to be determined.Electronic supplementary materialThe online version of this article (10.1186/s13000-018-0711-2) contains supplementary material, which is available to authorized users.

Survival of glioblastoma patients varies and prognostic markers are important in the clinical setting. With digital pathology and improved immunohistochemical multiplexing becoming a part of daily diagnostics, we investigated the prognostic value of the Ki-67 labelling index (LI) in glioblastomas more precisely than previously by excluding proliferation in non-tumor cells from the analysis. We investigated the Ki-67 LI in a well-annotated population-based glioblastoma patient cohort (178 IDH-wildtype, 3 IDH-mutated). Ki-67 was identified in full tumor sections with automated digital image analysis and the contribution from non-tumor cells was excluded using quantitative double-immunohistochemistry. For comparison of the Ki-67 LI between WHO grades (II-IV), 9 IDH-mutated diffuse astrocytomas and 9 IDH-mutated anaplastic astrocytomas were stained. Median Ki-67 LI increased with increasing WHO grade (median 2.7%, 6.4% and 27.5%). There was no difference in median Ki-67 LI between IDH-mutated and IDH-wildtype glioblastomas (p = 0.9) and Ki-67 LI was not associated with survival in glioblastomas in neither univariate (p = 0.9) nor multivariate analysis including MGMT promoter methylation status and excluding IDH-mutated glioblastomas (p = 0.2). Ki-67 may be of value in the differential diagnostic setting, but it must not be over-interpreted in the clinico-pathological context.

The therapeutic paradigm of gliomas is changing from a general approach towards an individualized and targeted approach. Accordingly, the search for prognostic and predictive biomarkers, as well as the demand for quantitative, feasible and robust methods for biomarker analysis increases. We find that software classifiers can identify and quantify the expression of a given biomarker within different subcellular compartments and that such classifiers can exclude frequently occurring nontumor cells, thereby avoiding potential bias. The use of a quantitative approach provides a continuous measurement of the expression, allowing establishment of new cut-points and identification of patients with specific prognoses. However, some pitfalls must be noted. This article focuses on benefits and pitfalls of novel approaches for quantifying protein biomarkers in gliomas.

Patients with IDH-wildtype glioblastoma (GBM) generally have a poor prognosis. However, there is an increasing need of novel robust biomarkers in the daily clinico-pathological setting to identify and support treatment in patients who become long-time survivors. Jumonji domain-containing protein 6 (JMJD6) is involved in epigenetic regulation of demethylation of histones and has been associated with GBM aggressiveness. We investigated the expression and prognostic potential of JMJD6 tumor fraction score in 184 IDH-wildtype GBMs. Whole-slides were double-stained with an antibody against JMJD6 and an exclusion-cocktail consisting of 4 antibodies (CD31, SMA, CD45, and Iba-1), enabling evaluation of tumor cells only. Stainings were quantified with a combined software- and scoring-based approach. For comparison, IDH-mutated WHO grade II, III and IV astrocytic gliomas were also stained, and the JMJD6 tumor fraction score increased with increasing WHO grade, although not significantly. In multivariate analysis including age, gender, performance status and post-surgical treatment high JMJD6 tumor fraction score was associated with longer overall survival in IDH-wildtype GBMs (p = 0.03), but the effect disappeared when MGMT promoter status was included (p = 0.34). We conclude that JMJD6 is highly expressed in IDH-wildtype GBM but it has no independent prognostic value.