Objective: To explore a comprehensive approach for on-site gross visual assessments of liquid-based cytology (LBC) specimens of thyroid nodules and determine morphologic criteria that help predict nondiagnostic rates.
Methods: Two-hundred nodules from 165 patients who underwent fine-needle aspiration (FNA) at our hospital were included in this prospectively designed, retrospective analysis. Specimens were visually assessed on-site for three morphologic categories (specimen color, specimen volume, and particle count) using a 5-point grading.
Results: Twenty-two nodules (11%) showed nondiagnostic results. Regarding specimen color, nondiagnostic rates tended to be higher in grades 1 (75%) and 5 (100%) than in grades 2 (18%), 3 (8%), or 4 (17%), with a significant difference between grade 1 and grade 3 ( P = .003). For specimen volume, nondiagnostic results were significantly more common in grade 1 (33%) and 5 (33%) than in grades 3 (5%) or 4 (1%) ( P<.005). There was a significant negative correlation between the grading of the particle count and the nondiagnostic rate (Spearman ρ = -1.000; P<.001). The sensitivity and specificity in the prediction of nondiagnostic results were 77% and 76%, respectively, at the optimal cutoff value of 2 (grade 2 or lower).
Conclusion: Particle count was an important morphologic criterion that helped predict nondiagnostic rates in LBC specimens of thyroid nodules, and the specimen color and volume were also useful adjuncts. In routine practice, on-site gross visual assessment followed by resampling (if necessary) may potentially help reduce the rates of nondiagnostic results, repeat FNAs, and the number of unnecessary needle passes.
Abbreviations: FNA = fine-needle aspiration; LBC = liquid-based cytology; ROC = receiver operating characteristic; US = ultrasonography.