A Novel Scoring System for AUS Thyroid Nodule

<b>Objective:</b> The primary objective of this study was to further subclassify Bethesda III atypia of undetermined significance (AUS) cytological findings in thyroid lesions using fine-needle aspiration cytology (FNAC). The secondary objective was to propose a novel scoring system to estimate the malignancy risk within these subcategories. <b>Methods:</b> We conducted a retrospective analysis of patients diagnosed with Bethesda III AUS who underwent thyroidectomy at King Khalid University Hospital, Riyadh, Saudi Arabia, from January 2017 to December 2024. Clinical, radiological, and pathological data-including FNAC slides and surgical specimens-were thoroughly reviewed. Cases classified as Bethesda III were further subclassified into six subtypes: III-A (small follicular pattern, poorly cohesive cells, and minimal colloid), III-B (nuclear and/or cellular atypia), III-C (cytological and architectural atypia), III-D (predominantly Hurthle cells), III-E (indeterminate atypia), and III-F (atypical lymphoid cells suggestive of possible lymphoma). A structured malignancy risk scoring system was developed by integrating cytological subtypes, radiologic features (microcalcifications, irregular margins, and hypervascularity), clinical risk factors (family history and prior radiation), and patient age. Statistical analysis was performed using SPSS Version 22, with chi-squared and Fisher's exact tests used to assess associations between variables and malignancy. <b>Results:</b> A total of 338 cases were analyzed, with a mean age of 42 ± 5 years and a female predominance. The malignancy distribution by subcategory was as follows: III-A: 10.65% (<i>n</i> = 36), III-B: 54.73% (<i>n</i> = 185), III-C: 22.49% (<i>n</i> = 76), III-D: 5.03% (<i>n</i> = 17), III-E: 7.69% (<i>n</i> = 26), and III-F: 0.3% (<i>n</i> = 1). Significant associations were found between malignancy and both cytological subcategory and presence of high-risk ultrasound features (<i>p</i> < 0.05). The proposed scoring system stratified patients into three risk groups: low (scores 0-2), moderate (2.5-4), and high (≥ 4.5), offering a predictive framework for clinical decision-making. <b>Conclusion:</b> This study highlights the heterogeneity within Bethesda III AUS nodules and supports subclassification as a meaningful step toward more accurate malignancy risk assessment. The proposed scoring system may serve as a practical tool to guide individualized management decisions. Further prospective validation is warranted.