Diagnostic Bias in AUS Thyroid Cytology: Role of Perceived Molecular Testing


KURTULAN O., Kilic I., Avcı H., Barkan G. A.

Diagnostic Cytopathology, 2026 (SCI-Expanded, Scopus) identifier identifier

  • Publication Type: Article / Article
  • Publication Date: 2026
  • Doi Number: 10.1002/dc.70092
  • Journal Name: Diagnostic Cytopathology
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, BIOSIS, EMBASE, MEDLINE
  • Keywords: atypia of undetermined significance, bias, FNA, molecular testing, thyroid
  • Hacettepe University Affiliated: Yes

Abstract

Background: Atypia of undetermined significance (AUS) represents one of the most controversial diagnostic categories in thyroid cytology, often subject to interpretive bias. This study evaluates whether perceived knowledge of a “positive molecular test” influences the cytologic interpretation of thyroid fine-needle aspiration (FNA) specimens diagnosed as AUS. Methods: Twenty thyroid FNA cases diagnosed as AUS with benign histologic follow-up following thyroidectomy were selected: eight with nuclear atypia (AUS-N), eight with architectural atypia (AUS-A), and four with oncocytic features. Representative slides (Papanicolaou and Diff Quik stained smears) were chosen and participants were instructed to evaluate the cases based on morphologic evaluation. Randomly selected 10 cases in two groups were presented to 10 participants as having “positive ThyroSeq test” or “no molecular testing information.” No molecular testing was performed in reality, as the aim was to measure the effects of “perceived positive testing.”. Results: Analysis of 200 responses showed AUS remained the predominant diagnosis (64%) in both groups. When informed of a presumed positive molecular test, participants were more likely to assign a neoplastic or malignant diagnosis (24% vs. 12%, p = 0.023). This trend was most pronounced in AUS-N cases (11 vs. 1 reclassified as suspicious/malignant). For AUS-A cases, more were reclassified as follicular neoplasm under presumed molecular positivity (12 vs. 8). Overall interobserver agreement was poor (Fleiss' kappa = 0.046). Conclusion: Awareness of a presumed positive molecular test result increases the likelihood of pathologists reclassifying AUS cases as suspicious or malignant. Recognizing and mitigating such cognitive biases is critical for maintaining diagnostic objectivity and ensuring accurate, evidence-based cytologic evaluation.