The International Academy of Cytology Yokohama System for Reporting Breast Cytopathology showed improved diagnostic accuracy.

2021 
BACKGROUND The aim of the International Academy of Cytology Yokohama System for Reporting Breast Fine Needle Aspiration Biopsy Cytopathology is to improve cytology practice. This study assessed cytologic diagnoses made with the system and its efficacy when it was applied by pathologists with different levels of experience. METHODS In all, 1080 cases of breast fine-needle aspiration biopsy (FNAB) over a period of 16 years were reviewed and reclassified with the system. The category distribution and the diagnostic performance were compared with the original diagnoses. The concordance rates for diagnoses from pathologists with different levels of experience were also determined. RESULTS The distribution of cytologic diagnoses made with the system was as follows: 11.7% were insufficient, 56.6% were benign, 20.1% were atypical, 6.1% were suspicious for malignancy, and 5.6% were malignant. The rates for the insufficient and atypical categories were lower than the original diagnosis rates (13.1% and 23.8%, respectively). Overall, 120 cases (11.1%) were recategorized. Among those recategorized as benign, suspicious, or malignant with follow-up data, 96.7% were correctly reclassified. A significant improvement in diagnostic performance was found with the system (P < .001). Such improvement was also seen in problematic breast lesions, including fibroepithelial lesions, papillary lesions, and low-grade carcinomas. Pathologists with intermediate experience showed a higher concordance with an expert pathologist in the diagnoses than those with short experience (κ, 0.838 vs 0.634). CONCLUSIONS The system effectively categorized the diagnoses, and the diagnostic performance of FNAB reporting was improved. The structured reporting also enhanced the reproducibility of reporting by pathologists with intermediate experience and, to some extent, those with short experience.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    25
    References
    1
    Citations
    NaN
    KQI
    []