Performance analysis of the anti-Ki67 antibody clone 30-9 for immunohistochemical staining of breast cancer.

Background: Although Ki67 has important clinical relevance in breast cancer, its assessment results vary according to assay due to differences in both analytical and interpretation processes. We aimed to validate the performance of anti-Ki67 antibody clone 30-9 by comparison with clone MIB-1 and to investigate utility of the image analysis system in Ki67 assessment using clinical breast cancer samples.

Methods: A series of sequential tissue sections was prepared from formalin-fixed paraffin-embedded blocks of surgically resected breast cancer specimens from 50 patients. The tissue sections were stained immunohistochemically with anti-Ki67 antibodies, 30-9 and MIB-1, as well as with hematoxylin and eosin for morphological analysis. We scanned all the stained slides with Ventana iScan HT and selected the Ki67 counting areas based on morphological findings. Three pathologists independently studied images of the counting areas to determine Ki67-positive rates. In addition, the images of 30-9-stained slides were analyzed using the image analysis system, VENTANA Virtuoso.

Results: Ki67-positive rates by 30-9 showed a strong correlation with those by MIB-1 for all pathologists (pathologist #1: r = 0.985, pathologist #2: r = 0.987, pathologist #3: r = 0.982). Between 30-9 and MIB-1, there was no significant difference of CV%, showing variabilities of Ki67-positive rates among pathologists. Ki67-positive rates showed a strong correlation between the image analytical values and the pathologist-counted median values (r = 0.952).

Conclusions: The performance of 30-9 is equivalent to that of MIB-1 in Ki67 assessment of breast cancer. The image analysis system can substitute for or support visual counting by a pathologist.