Standardized Assessment of Ki-67 in Breast Cancer Patients Using Virtual Slides and an Automated Analyzer in Comparison to Central/Local Pathological Assessments

Abstract

Purpose: To standardize the methods to measure Ki-67, there is an interest in automating the assessment of Ki-67. Therefore, we reviewed the possibility of introducing an automated analyzer to standardize the Ki-67 evaluation method. Methods: We retrospectively reviewed a clinical database of patients who underwent surgery for early breast cancer at Tokyo-West Tokushukai Hospital. Among them, those who underwent preoperative core needle biopsy (CNB) were enrolled. The concordance rates of estrogen receptor (ER), progesterone receptor (PgR), human epidermal growth factor receptor 2 (HER2), and Ki-67 by local pathologists were reviewed (valuations made by local pathologists ) , and nonmatching cases (from August 2008 to October 2011) were reassessed both by central review and using an automated analyzer with virtual slides. The results were compared with the evaluations made by local pathologists, and we reexamined the concordance rate by using central review and the automated analyzer. Results: The concordance rate of Ki-67 evaluations made by local pathologists in the preoperative CNB and surgical specimens was 78.7% in 287 cases pathologically assessed from October 2008 to March 2013. This rate was significantly lower ( p < 0.01) than that of ER (95.6%), PgR (88.5%), and HER2 (91.6%). Reassessment of the 37 cases of nonmatching Ki-67 values from 2008 to October 2011 using central review and an automated analyzer resulted in clear improvement in matching of 22 (92.1%) and 24 (93.1%) of 37 cases, respectively. Conclusion: The concordance rate of Ki-67 in preoperative CNB and surgical specimens was lower than that of other biological markers; however, they were nearly equal by reassessment using central review and an automated analyzer .