The Ki67 dilemma: investigating prognostic cut-offs and reproducibility for automated Ki67 scoring in breast cancer

• DOI: 10.1007/s10549-024-07352-4
• Publication Date: 2024-05-26T19:01:26Z
• Authors: Emma Rewcastle; Ivar Skaland; Einar Gudlaugsson; Silja Kavlie Fykse; Jan P. A. Baak; Emiel A. M. Janssen
• Keywords: Adult; Aged, 80 and over; 0301 basic medicine; Research; Reproducibility of Results; Breast Neoplasms; Kaplan-Meier Estimate; Middle Aged; Prognosis; Ki-67 Antigen; 0302 clinical medicine; ROC Curve; Biomarkers, Tumor; Humans; Female; Aged

Abstract Purpose Quantification of Ki67 in breast cancer is a well-established prognostic and predictive marker, but inter-laboratory variability has hampered its clinical usefulness. This study compares the value reproducibility scoring using four automated, digital image analysis (DIA) methods two manual methods. Methods The cohort consisted 367 patients diagnosed between 1990 2004, with hormone receptor positive, HER2 negative, lymph node negative cancer. Manual was performed predefined criteria. DIA QuPath Visiopharm® platforms. Reproducibility assessed by intraclass correlation coefficient (ICC). ROC curve survival identified optimal cutoff values addition to recommendations International Working Group Norwegian Guidelines. Kaplan–Meier curves, log-rank test Cox regression association distant metastasis (DM) free survival. Results hotspot global showed good agreement when compared their counterpart (ICC > 0.780), excellent different platforms 0.781–0.906). Different cutoffs demonstrate significant DM-free (p < 0.05). had greater for 14% (HR 3.054–4.077) than 2.012–2.056). use single all affected distribution prediction outcomes (e.g. false positives negatives). Conclusion demonstrates that superior methods, further required standardize definition cut-off.