Abstract：
BACKGROUND: Ki67 reflects the proliferation activity in breast cancer (BC). However, an optimal method for its assessment in clinical settings has yet to be robustly defined. In this study we compared several methods to score Ki67 to identify a reliable and reproducible method for routine practice.
METHODS: Sections from luminal BC cohort (n = 1662) were immunohistochemically stained with Ki67 and were assessed for the percentage, pattern, and intensity of expression. Ki67 positivity was evaluated using three methods: (i) quantification of Ki67-positive cells among 1000 invasive tumour cells within hotspot, (ii) average estimation of Ki67 within a defined hotspot, and (iii) average estimation of Ki67 positivity within the whole section. Time required for scoring, interobserver agreement and association with outcome were determined.
RESULTS: The mean percentage of Ki67 expression per 1000 cells method was 16%, while the mean value of Ki67 scores using the average estimation within hotspot and whole slide were 14% and 12%, respectively. Quantification of Ki67-positive cells within 1000 cells had the highest degree of consistency between observers, and the highest hazard ratio predicting patient outcome when compared to using different common Ki67 cutoffs, which was independent of the other two methods. Granular pattern of Ki67 expression was associated with poorer outcome as compared to the other patterns.
CONCLUSIONS: Assessment of Ki67 expression using quantification positive cells among 1000 tumour cells is an optimal method to achieve high reliability and reproducibility. Comment on the predominant Ki67 expression pattern would add prognostic and predictive value in luminal BC.
METHODS: Sections from luminal BC cohort (n = 1662) were immunohistochemically stained with Ki67 and were assessed for the percentage, pattern, and intensity of expression. Ki67 positivity was evaluated using three methods: (i) quantification of Ki67-positive cells among 1000 invasive tumour cells within hotspot, (ii) average estimation of Ki67 within a defined hotspot, and (iii) average estimation of Ki67 positivity within the whole section. Time required for scoring, interobserver agreement and association with outcome were determined.
RESULTS: The mean percentage of Ki67 expression per 1000 cells method was 16%, while the mean value of Ki67 scores using the average estimation within hotspot and whole slide were 14% and 12%, respectively. Quantification of Ki67-positive cells within 1000 cells had the highest degree of consistency between observers, and the highest hazard ratio predicting patient outcome when compared to using different common Ki67 cutoffs, which was independent of the other two methods. Granular pattern of Ki67 expression was associated with poorer outcome as compared to the other patterns.
CONCLUSIONS: Assessment of Ki67 expression using quantification positive cells among 1000 tumour cells is an optimal method to achieve high reliability and reproducibility. Comment on the predominant Ki67 expression pattern would add prognostic and predictive value in luminal BC.
摘要：
背景：Ki67反映了乳腺癌（BC）的增殖活性。然而,在临床环境中评估其最佳方法尚未得到可靠定义。在这项研究中，我们比较了几种对Ki67评分的方法，以确定一种可靠且可重复的常规方法。
方法：用Ki67对来自管腔BC队列（n=1662）的切片进行免疫组织化学染色，并评估其百分比，模式，和表达的强度。使用三种方法评估Ki67阳性：（i）在热点内1000个侵袭性肿瘤细胞中的Ki67阳性细胞的定量，(Ii)定义热点内Ki67的平均估计，和(iii)整个切片内Ki67阳性的平均估计。评分所需时间,确定了观察者之间的共识和与结果的关联。
结果：每1000个细胞中Ki67表达的平均百分比为16％，而使用热点和整个幻灯片内的平均估计，Ki67得分的平均值分别为14%和12%，分别。1000个细胞内的Ki67阳性细胞的定量在观察者之间具有最高程度的一致性，与使用不同的常见Ki67截止值相比，预测患者预后的风险比最高，它独立于其他两种方法。与其他模式相比，Ki67表达的颗粒模式与较差的结果相关。
结论：使用1000个肿瘤细胞中的定量阳性细胞评估Ki67表达是实现高可靠性和可重复性的最佳方法。对主要Ki67表达模式的评论将增加管腔BC的预后和预测价值。
方法：用Ki67对来自管腔BC队列（n=1662）的切片进行免疫组织化学染色，并评估其百分比，模式，和表达的强度。使用三种方法评估Ki67阳性：（i）在热点内1000个侵袭性肿瘤细胞中的Ki67阳性细胞的定量，(Ii)定义热点内Ki67的平均估计，和(iii)整个切片内Ki67阳性的平均估计。评分所需时间,确定了观察者之间的共识和与结果的关联。
结果：每1000个细胞中Ki67表达的平均百分比为16％，而使用热点和整个幻灯片内的平均估计，Ki67得分的平均值分别为14%和12%，分别。1000个细胞内的Ki67阳性细胞的定量在观察者之间具有最高程度的一致性，与使用不同的常见Ki67截止值相比，预测患者预后的风险比最高，它独立于其他两种方法。与其他模式相比，Ki67表达的颗粒模式与较差的结果相关。
结论：使用1000个肿瘤细胞中的定量阳性细胞评估Ki67表达是实现高可靠性和可重复性的最佳方法。对主要Ki67表达模式的评论将增加管腔BC的预后和预测价值。
