Comparison of Automated and Manual Assessment of Ki-67 Proliferative Activity in Invasive Breast Carcinomas (IBCs)
Youngmee Kwon, Jungsil Ro, Eun Sook Lee, Han Sung Kang, Seok Won Kim, Kyung Joo Kim, Geon Kook Lee. National Cancer Center, Goyang, Republic of Korea
Background: Immunohistochemical detection of Ki-67 antigen has been widely used to assess the growth fraction in human malignancies, including breast cancer. The Ki-67 proliferative activity has been reported to be associated with clinical response to chemotherapy, relapse, or survival in IBCs. The aim of this study is to determine whether digital image analysis can be adopted in routine pathology practice for evaluating Ki-67 labeling index (LI) in IBCs.
Design: We included 210 IBCs histologically diagnosed and immunostained for Ki-67 between Jun 2012 and July 2012. We assessed Ki-67 LI by manual and automated counts, respectively. We compared Ki-67 values between manual and automated quantitation with St. Gallen criteria (low, intermediate, high). We intended to validate the accuracy of automated quantitation of Ki-67 LI using Aperio ImageScope (Aperio Technologies, Vista, CA, USA).
Results: This study demonstrated relatively high concordance rate (86.7%) with substantial interrater agreement (κ=0.7975, p<0.001) between manual and automated Ki-67 values according to St. Gallen criteria. The difference of mean values between manual and automated Ki-67 LI was statistically significant (1.82±4.91%, p<0.001, 95% CI=1.16-2.49%). However, manual Ki-67 LI was significantly correlated with automated Ki-67 LI with high correlation coefficient (γ=0.9791, p<0.001).
Conclusions: Ki-67 values according to St. Gallen criteria have relatively high concordance between manual and automated quantitation. Therefore, only if pathologists confirm the tumor regions to be analyzed, digital image analysis can be adopted in routine practice for evaluating Ki-67 LI in IBCs.
Tuesday, March 5, 2013 1:00 PM
Poster Session IV # 57, Tuesday Afternoon