Ki67 is a simple and valuable tool in breast cancer management, whereas its interpretation can vary due to differences in scoring methods, such as hotspot and global scoring. Standardization efforts by IKWG aimed to improve the scoring consistency and to enhance clinical utility of Ki67 in routine practice resulting in global scoring is recommended by both ASCO and ESMO international breast cancer guidelines.

Therefore, we investigated the variability of Ki67 scores using two different scoring methods in routine practice across two clinics and explored how the transition from hotspot to global scoring has impacted clinical decision-making guidelines. To our knowledge, this is the first study that investigates the recommendations of IKWG and the variability of different scoring methods in real-world data, however there are previous studies focusing on the analytical validity of Ki67 in breast cancer [8, 19].

Before starting this study, we hypothesized that variability between pathologists would be higher with the hotspot scoring due to subjective nature of the method [20]. Still, we found acceptable consistency between individual pathologists regardless of the scoring method.

Our findings suggest that variability between individual pathologists is not as pronounced as expected, even in resection specimens that encompass a larger area than biopsies. Standardization of pre-analytical variables, even though cases were stained at two laboratories and evaluated by different pathologists working at those centers, may have contributed to this result. Furthermore, our findings suggest that pathologists at the same department have the same scoring tendencies.

In the current study, while the pathologists’ scores are consistent with those of their colleagues with both scoring methods, a significant variability between laboratories exists. This finding aligns with the results from our earlier study, which observed inter- and intra-laboratory variability in pathological parameters, including Ki67, within a national cohort in Sweden [15]. After transitioning to global scoring, the distribution of Ki67 categories shifted significantly, with an increase in the number of cases categorized as intermediate, resulting in a more similar distribution between the centers. For this reason, global scoring is recommended for better alignment across centers.

While our study highlights notable differences in Ki67 risk categories depending on the method used, not all studies have reported such discrepancies. For instance, in the study of Jang et al., no significant differences were observed in the categorization of 493 tumors (Ki67 low vs. high) when comparing hotspot and global scoring methods [20]. However, their study is not directly comparable with this current study due to differences in the definition of global scoring, the cut-off values applied, and patient population involved.

From a biological perspective, global scoring provides a more accurate reflection of the overall tumor proliferation status. Zilenaite-Petrulaitiene et al. observed lower intratumoral heterogeneity in cases at both ends of the Ki67 proliferation index scale when using a study-specific cut-off value [21]. Moreover, establishing a universal cut-off for dichotomization is challenging for a continuous marker as Ki67, because it may not apply consistently across pathology departments due to pre-analytic and analytic variability. Therefore, defining an intermediate category with the proposed cut-off values from the IKWG can be valuable, as it may identify a subgroup with higher intratumoral heterogeneity and uncertain behavior, highlighting the need for additional information about tumor biology, potentially supported by multigene tests or AI-based analysis [22, 23].

Half of the studied breast cancers fall into the intermediate grade category that provides limited clinical value for guiding patient management [15, 24]. Furthermore, the expansion of the intermediate Ki67 group may present a new challenge: an increased number of cases with uncertain treatment decisions, which may lead to greater reliance on molecular testing to guide adjuvant therapy. This raises concerns about fair access to health care, as patients who cannot get these tests may have less treatment options. On the other hand, high variability in Ki67 scoring between pathology labs can also cause inequality in cancer care. To ease these concerns, other established clinicopathologic factors (e.g.: age, tumor size, grade, lymph node status, hormone receptor status) considered together with Ki67 may aid the decision-making.

Ki67 scores are clinically relevant particularly for early-stage luminal/HER2-negative breast cancer in post-menopausal patients, whereas treatment choices for triple-negative and HER2-positive cancers are mostly independent of Ki67 [25]. In the latest St. Gallen Conference, Ki67 and change of Ki67 by treatment were mentioned as prognostic biomarkers with limited value in influencing adjuvant chemotherapy decisions [26]. Ki67 is also a companion diagnostic for the FDA-approved abemaciclib therapy in the USA, specifically for patients with a Ki67 score of ≥ 20% measured by a specific FDA-approved test. However, there are criticisms regarding the predictive value of this cut-off in real-world scenarios, particularly concerning the sparse evidence on reproducibility [27, 28]. In Stockholm, the clinical guidelines changed in 2022 and global Ki67 has been incorporated together with grade, tumor size, ER/PR/HER2 status in the decision-making on molecular profiling test and or on adjuvant chemotherapy. Molecular profiling test is recommended for patients with T1c/T2, grade 2, Ki67 intermediate and ER positive/Her2 negative breast cancer. For all the other ER positive/Her2 negative breast cancer patients, the above-mentioned clinicopathologic factors with global Ki67 score shall be used to decide on recommending adjuvant chemotherapy. This context highlights the clinical relevance of refining and standardizing Ki67 scoring methods.

As a limitation of our study, we did not account for the clinical- and molecular subtypes of breast cancer. Furthermore, our study was not prospectively designed to demonstrate clinical utility. However, only randomized clinical trials can provide evidence on clinical utility and the international guidelines already recommend assessing Ki67 in clinical practice. Therefore, we believe that our study offers valuable real-world data about the analytical validity of implementing global Ki67 scoring in clinical routine.