Importance of hot spot definition for Ki-67 quantification using digital image analysis

D. Omanovic, A. Schønau
Visiopharm A/S, Hoersholm, Denmark

Objective: Assessment of biomarker expression in automatically identified hot spots rather than manually assessed hot spot or as an average of the entire tumour area is currently being discussed and implemented into scoring guidelines, but a consistent definition of hot spots has not yet been defined. In this analytical study, we investigate the influence of the way the hot spot is defined and the consequences it has on the quantification of Ki-67.
Method: Digital images of serial whole tissue sections, stained with panCK and Ki-67 respectively, were aligned using the Visipharm VirtualDoubleStaining™ technique. The pan-CK image was used to separate the tumour component from the stromal tissue, and restrict the remaining analysis to tumour regions. Ki-67 positive and negative nuclei were identified and a heatmap was generated. Based on this heatmap four types of hot spots were created: circle, square, heat map count and heat map area. For each hot spot type the adjustable parameters where varied to assess the impact of the hot spot definition on the Ki-67 quantification
Results: Ki-67 scores on analytical samples as a function of the hot spot definition.
Conclusion: Proper application of hot spot definition can significantly affect the Ki-67 proliferation index in clinically relevant samples.