[Intraobserver reproducibility of Ki-67 assessment of breast cancers based on digital slide]

Objective: To investigate the intra-observer reproducibility of Ki-67　　　assessment in breast cancers using three methods based on digital slide. Methods: Thirty cases of invasive breast cancer tissues were immunostained for Ki-67 by automatic stainer, and then scanned into digital pathological slides. Ki-67 positive index was measured individually by three pathologists using size-set visual assessment of hot spot (SSVAHS), size-set semi-automatic counting of hot spot(SSSACHS), and size-set automatic counting of hot spot (SSACHS), respectively, and repeated for 10 times. Intraclass correlation coefficient (ICC) of each assessment method was calculated, and the intraobserver reliability was classified as excellent, good, fair and poor according to ICC. Results: The ICC by 3 pathologists using SSVAHS was 0.832, 0.843 and 0.826, respectively, The ICC using SSSACHS was 0.926,0.938,0.929, and the ICC using SSACHS was 0.964, 0.971 and 0.968.The intraobserver reliability level of all three methods was excellent. Conclusion: The three methods of Ki-67 assessment achieve satisfactory intraobserver reproducibility, and the order of reproducibility from high to low is SSACHS, SSSACHS, and SSVAHS.