Associação entre os achados citopatológicos e ultrassonográficos dos nódulos tireoidianos através da comparação entre as classificações de Bethesda, TI-RADS e ATA 2015

Abstract

Thyroid ultrasonography and cytopathological analysis from fine-needle aspiration (FNA) specimens have become the main diagnostic weapons addressing thyroid nodule investigation. OBJECTIVE: To verify the association between the ultrasonographic characteristics (USC) of thyroid nodules and their respective FNA results, based on: TI-RADS, ATA 2015 and Bethesda classifications. MÉTODOS: A single-institution prospective evaluation was conducted by blind examiners, comparing the USC and FNA of thyroid nodules. Several scenarios were tested, allocating intermediate categories of USC differently in each scenario, comparing TI-RADS and ATA 2015 systems. Generalized linear models were developed for calculation of: confidence interval (CI), sensitivity, specificity, PPV, NPV, LR positive and LR negative. ROC curves were also used to evaluate area under the curve. USC Statistical analysis tests did not include Bethesda I, III, and IV cases. RESULTS: A total of 339 nodules were examined. Bethesda II represented 83.6% of satisfactory cytopathological samples. Bethesda III or IV represented only 7.9% of the cases. The presence of microcalcifications in the USC increased the risk of cytology malignancy by more than 18 times (odds ratio = 18,70 with CI = 4.79 - 73.00). Excellent NPV results (94.1% - 99.3%) were found regardless of USC intermediate categories distribution, but the PPV results were bad (11.2% - 50%). The best area under the ROC curve values were obtained when intermediate suspicion categories of ATA 2015 and TI-RADS (4A and 4B) were considered malignant: 0.792 for ATA 2015 (CI = 0.67 - 0, 90) and 0.786 for TI-RADS (CI = 0.68 - 0.88). CONCLUSION: TI-RADS and ATA 2015 appear to be statistically equivalent assessing risk of malignancy in FNA, with an area under the ROC curve value close to 0.8. The risk of malignancy associated with the presence of microcalcifications was high. Excellent NPV results and poor PPV results were found. More studies are needed to investigate intermediate suspicion and inconclusive categories: TIRADS 4A / 4B; ATA 2015 intermediate suspicion and Bethesda III / IV.