The controversy on Ki67 and mitotic counts as independent prognostic markers is incontrovertible. A “mitotically-dividing” cell and a “proliferating” Ki67-positive cell garner different perspectives from a pathologist and a mitosis researcher. Since a cell that enters cell-cycle may not reach the classical “mitotic-spindle stage” until almost a year, Ki67 offers limited information on mitotic propensity of the tumor. In current breast cancer diagnostic practice, not only are KI and MI measured on different scales, they are also determined from non-overlapping fields; making a direct comparison of the two parameters very challenging. Owing to these issues, the precise relationship between KI and MI has remained poorly defined. We believe it is time to revisit this intricate relationship. Here we propose the rational integration of KI and MI into a “mitotic frequency” index (MFI), which represents the proportion of mitotic cells amongst Ki67-positive cells, and provides better prediction of metastatic risk.
The KI-MI relationship was analyzed retrospectively in 1600 breast carcinoma cases at Northside Hospital, Atlanta, GA. In addition, for an accurate assessment of MFI, we measured MI and KI on the same scale and within the same field by four-color immunofluorescence staining on paraffin-embedded breast tumor tissues (n=100) for α-tubulin, Ki67, and phospho-histone H3, and stain DNA with Hoechst.
Retrospective data analysis shows that although KI and MI increased with tumor grade, grade-wise increase in MI was much slower; consequently, MFI decreased across grades. In effect, we found a striking grade-wise decrease in mitotic frequency amongst Ki67-positive cells. Immunofluorescence data found lower MFI in grade-matched patients with metastasis versus those without metastasis.The KI-MI relationship varies dynamically during tumor progression, with a lower MFI representing a shift in the tumor's agenda from mitosis to migration, indicating elevated metastatic risk. MFI thus reveals a new layer of valuable risk-predictive information.