Improving Diagnostic Consistency in Breast FNA Cytology: An Inter-Observer Analysis

By: Aiden Chiu, Stuyvesant High School

Breast fine needle aspiration (FNA) is a minimally invasive, cost-effective method for evaluating breast lesions, but slide interpretation is subjective, leading to variable diagnoses. This study assessed inter-observer agreement in 116 de-identified breast FNA slides reviewed independently by three board-certified pathologists blinded to clinical and radiographic information. Slides were classified as nondiagnostic, negative, atypical, suspicious, or positive for carcinoma. Agreement was measured using Cohen’s kappa (κ), and key cytologic features—architectural disarray, nuclear membrane irregularity, single atypical cells, and myoepithelial cells—were analyzed for their effect on consistency. Results showed good agreement for positive carcinoma (κ≈0.60), fair agreement for negative cases (κ≈0.30), and low agreement for atypical or suspicious slides (κ≈0.11–0.20). Architectural disarray and nuclear membrane irregularity improved concordance, while single atypical cells and myoepithelial cells increased variability. These findings highlight the reproducibility of definitive diagnoses and the challenges of indeterminate cases, suggesting that combining cytology with clinical context and informatics-driven analysis can enhance diagnostic reliability and patient care.

View paper
Previous
Previous

Assessing Biosensor Effectiveness for Early Cancer Detection Within Underrepresented Communities

Next
Next

Code to Care: How Artificial Intelligence Is Rewriting the Geography of Global Healthcare Access