Comparison of visual assessment (VA) and automated assessment (Digital Image Analysis, DIA) of Ki-67 in breast carcinoma

Abstract

Objective: To determine the agreement between visual assessment and automated assessment of Ki-67 labelling index in breast carcinoma.

Method: The observational, prospective study was conducted at the Histopathology Department, Shifa International Hospital, Islamabad, Pakistan, from November 24, 2018, to November 23, 2019, and comprised female patients regardless of age who had been diagnosed with invasive breast carcinoma. Ki-67 immunohistochemical staining was performed on paraffin embedded sections of the breast. Visual assessment of the proliferative index was done using hotspot method, followed by automated assessment of the digital image by ImmunoRatio software. Microscopic images of the maximally positive areas were taken and uploaded onto the software which gave an assessment of the proliferative index. The outcomes of the two assessments were compared. Data was analysed using SPSS 16.

Result: Of the 200 females with mean age 49±5.6 years (range: 29-83 years), 125(62.5%) showed homogeneous 75(37.5%) exhibited heterogeneous staining for Ki-67 on immunohistochemistry. Overall, high Ki-67 index value was noted for 159(79.5%) cases by visual assessment and 168(84%) by digital image analysis. The corresponding value for low Ki-67 index value were 41(20.5%) and 31(16%). Agreement between the two methods among high Ki-67 index value was in 146(73.0%) cases, and the corresponding value for low Ki-67 value was 18(8.5%). Besides, 36(18%) cases showed disagreement in Ki-67 values.

Conclusion: There was a significant agreement between visual assessment and digital image analysis of Ki-67 labelling index in breast carcinoma, suggesting that both the methods can be used in clinical practice.

Key Words: Ki-67 antigen, Immunohistochemistry, Digital image analysis, Proliferative index, Breast cancer.