Role of cytological and ultrasonographic features in predicting the risk of malignancy in thyroid nodules with indeterminate cytology.Minerva Endocrinol. 2014 Mar; 39(1):43-52.ME
AIM
The aim of this paper was to examine the diagnostic value of several cytological and ultrasonographic features in predicting malignancy in thyroid follicular neoplasms.
METHODS
The sample of the study consisted of 145 patients, who have had the diagnosis of follicular neoplasm on US guided fine-needle aspiration (FNA), and had undergone thyroidectomy. The cytological slides and the ultrasonographic images were reviewed, and several ultrasonographic and cytological features were evaluated and correlated with final histology.
RESULTS
Histological diagnosis of malignancy was obtained in 14.5% of the patients, papillary carcinoma being the most frequent (66% of malignancies). The cytological and ultrasonographic features that have been associated with malignancy were: micro-fragments (P<0.00001), overlapping (P<0.005), hypercellularity (P<0.009), micronucleoli (P<0.013), atypical features (P<0.027), nodule size larger than 2 cm (P<0.029) and micro-calcifications (P<0.0002). Using the features that were statistically independent ones, which included two cytological features: micro-fragments and micronuclei, and one ultrasonographic feature: micro-calcifications, a statistical model for predicting malignancy was constructed. According to this model, it was found that the risk for malignancy is 2.65% in the absence of the three parameters, and amounts to 93.93% in the presence of all three of them.
CONCLUSION
In a thyroid follicular neoplasm, the cytological and ultrasonographic features that were associated with malignancy were: micro-fragments, overlapping, hypercellularity, micronucleoli, atypical features, nodule size larger than 2 cm and micro-calcifications. In an attempt to predict malignancy, we proposed a simple statistical model using only three features derived from cytological and ultrasonographic tests.