Preoperative Typing of Thyroid and Parathyroid Tumors with a Combined Molecular Classifier

SIMPLE SUMMARY: We previously proposed a new diagnostic algorithm that allows identification and classification of malignancy markers of thyroid tumors in cytological preparations of biopsy material through an analysis of several molecular markers. We previously evaluated the diagnostic characteristics of this algorithm on a sample of category III and IV cytological preparations (Bethesda system, 2017) for the detection of malignant tumors. However, in that study, we did not determine the accuracy of classification. Also, the algorithm did not allow discrimination of parathyroid gland nodules. In the present work, our goal was to include the identification of parathyroid cells in the molecular classifier and to evaluate the performance of our algorithm on the typing of thyroid tumors. We demonstrated that the diagnostic panel including the analysis of microRNA and mRNA expression, the V600E mutation in the BRAF gene, and mitochondrial-to-nuclear DNA ratio enables accurate identification of parathyroid and several types of thyroid carcinomas. ABSTRACT: In previous studies, we described a method for detecting and typing malignant tumors of the thyroid gland in fine-needle aspiration biopsy samples via analysis of a molecular marker panel (normalized HMGA2 mRNA level; normalized microRNA-146b, -221, and -375 levels; mitochondrial-to-nuclear DNA ratio; and BRAF(V600E) mutation) in cytological preparations by quantitative PCR. In the present study, we aimed to estimate the specificity of the typing of different thyroid tumors by the proposed method. Fine-needle aspiration cytological preparations from 278 patients were used. The histological diagnosis was known for each sample. The positive and negative predictive values of the method assessed in this study were, respectively, 100% and 98% for papillary thyroid carcinoma (n = 63), 100% and 100% for medullary thyroid carcinoma (n = 19), 43.5% and 98% for follicular carcinoma (n = 15), and 86% and 100% for Hürthle cell carcinoma (n = 6). Thus, we demonstrate that the diagnostic panel, including the analysis of microRNA expression, mRNA expression, the BRAF(V600E) mutation, and the mitochondrial-to-nuclear DNA ratio, allows the highly accurate identification of papillary thyroid carcinoma, medullary thyroid carcinoma, and Hürthle cell carcinoma but not malignant follicular tumors (positive predictive value was below 50%).