Evaluating Diagnostic Performance of Kaiser Score and BI-RADS in MRI Examination of Suspicious Breast Lesions: a Retrospective Study
Archives of Breast Cancer · 25.07.2026 · 13 (3) · 359-372
DOI: 10.32768/abc.6921483579-046 ↗Kurz gefasst
Fachartikel in der Zeitschrift Archives of Breast Cancer, online erschienen am 25.07.2026. Verfasst von Fahimeh Zeinalkhani und 13 weiteren Autor:innen. Die Arbeit nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Background
The purpose of the study is to compare the diagnostic performance of the Kaiser score (KS) with BI-RADS in breast magnetic resonance imaging (MRI) examination and to evaluate the impact of apparent diffusion coefficient (ADC), background parenchymal enhancement (BPE), and lesion type on its diagnostic efficacy.
Methods
This retrospective study included a total of 259 patients with 284 suspicious breast lesions. All patients underwent MRI examination using standardized protocols. Two blinded radiologists independently evaluated lesions using BI-RADS, KS, and ADC values. Histopathology served as the main reference standard. Diagnostic performance of BI-RADS, KS, ADC, as well as combination of KS and ADC (Kaiser Plus) were evaluated using sensitivity, specificity, and AUC values. Also, subgroup analyses for mass/non-mass lesions and background parenchymal enhancement (BPE) were performed.
Results
The AUC for BI-RADS, KS, ADC, and Kaiser Plus were 0.961 (95% CI = 0.931 to 0.980), 0.955 (95% CI = 0.924 to 0.976), 0.866 (95% CI = 0.807 to 0.913), and 0.959 (95% CI = 0.919 to 0.983), respectively. BI-RADS showed higher sensitivity (100% vs. 88.2%) but lower specificity (75.3% vs. 94.4%) than KS, resulting in comparable AUCs (DeLong p = 0.651). KS reduced unnecessary biopsies by 19.1% (17 benign lesions spared) but missed 11.7% more malignancies (n=23). For mass lesions, KS performance was improved, achieving superior specificity (94.6% vs. 75.7%) with minimal sensitivity trade-offs (93.0% vs. 100%) compared to BI-RADS. Non-mass lesions posed challenges for all methods. Also, KS performance was not different across BPE grades.
Conclusion
The KS showed comparable diagnostic value compared to BI-RADS, reducing unnecessary biopsies, without significant improvement from ADC integration. It also provided consistent performance across lesion types. Our findings support KS's clinical utility for breast MRI interpretation.
Zitieren
APA
Zeinalkhani, F., Hakim, P. K., Mohebbi, A., Zahra, Shayegh, Moradi, Z., Sabet, F. M., Abdi, A., Heydari, A. H., Badr, S., Zeinalkhani, H., Azizinik, F., Katouli, F. S., & Mohammadzadeh, S. (2026). Evaluating Diagnostic Performance of Kaiser Score and BI-RADS in MRI Examination of Suspicious Breast Lesions: a Retrospective Study. Archives of Breast Cancer, 13(3), 359-372. https://doi.org/10.32768/abc.6921483579-046
Vancouver
Zeinalkhani F, Hakim PK, Mohebbi A, Zahra, Shayegh, Moradi Z, et al. Evaluating Diagnostic Performance of Kaiser Score and BI-RADS in MRI Examination of Suspicious Breast Lesions: a Retrospective Study. Archives of Breast Cancer. 2026;13(3):359-372. doi:10.32768/abc.6921483579-046
BibTeX
@article{Zeinalkhani2026,
author = {Zeinalkhani, Fahimeh and Hakim, Peyman Kamali and Mohebbi, Alisa and Zahra and Shayegh and Moradi, Zahra and Sabet, Fatemeh Mahdavi and Abdi, Ali and Heydari, Amir Hosein and Badr, Sharife and Zeinalkhani, Hadise and Azizinik, Fahimeh and Katouli, Fatemeh Shakki and Mohammadzadeh, Saeed},
title = {Evaluating Diagnostic Performance of Kaiser Score and BI-RADS in MRI Examination of Suspicious Breast Lesions: a Retrospective Study},
journal = {Archives of Breast Cancer},
year = {2026},
volume = {13},
number = {3},
pages = {359--372},
doi = {10.32768/abc.6921483579-046}
}
EUVIMED