Targeted Mutational Profiling and a Powerful Risk Score as Additional Tools for the Diagnosis of Papillary Thyroid Cancer

Abstract

Nowadays, the complementary diagnostics based on the suspicious thyroid lesion specific mutational state analysis is indispensable in the clinical practice. We aimed to test and validate our novel 568-mutational hotspot panel (23 cancer-related genes) on papillary thyroid cancers (PTCs) and their tumor-free pairs to find the most powerful mutation pattern related to PTC. The sequencing method was carried on with Ion Torrent PGM on 67 thyroid tissue samples. The most commonly detected mutation was the BRAF c.1799 T > A in all non-classical PTC cases. We utilized a multivariate statistical method (CVA) to determine a discrimination score based on mutational data array and to assess malignancy risk. Based on variants, the BRAF gene has by far the highest indicative power, followed by TSHR and APC. We highlighted novel aspects of the mutational profile and genetic markers of PTC. CVA has correctly assigned most of the samples based on the mutation frequencies and different variables of the selected genes, with high analytical probabilities. The final goal is to set up a new comprehensive rule-in and rule-out test to support the clinical decision making mainly in inconclusive fine-needle aspiration biopsy cases.

Keywords: Multivariate statistical method; New hotspot panel; Next generation sequencing; Papillary thyroid cancer; Risk score.

Fig. 1

The relationship between the actual state of the tumor-free and PTC samples, analysis of the selected 26 variants which occurred more than twice or more times. a Graphical display of samples separated by CVA of the selected 26 variants which were determined more than 2 or more times in our cohort. The benign tumor-free samples are marked with white lines and PTC samples are represented by black lines. b Table is summarizing the calculated F-ratios and the correlations with the single canonical variate (CV1). F-ratio derived from the training set: the larger this empirical value, the higher the malignancy risk. c Our histologically confirmed PTC or tumor-free samples were grouped according to the 26-variant set. Twenty-one benign samples out of the total 28 were actually assigned into the tumor-free group. Thirty-seven patients with PTC out of the total 39 were considered as malignant

Fig. 2

The relationship between the actual state of the tumor-free and PTC samples, analysis of the selected 18 genes based on mutation frequencies. a Graphical display of samples separated by CVA of the selected 18 genes where alterations were detected. The benign tumor-free samples are marked with white lines and PTC samples are represented by black lines. b Table is summarizing the calculated F-ratios and the correlations with the single canonical variate (CV1). F-ratio derived from the training set: the larger this empirical value, the higher the malignancy risk. c Our histologically confirmed PTC or tumor-free samples were grouped according to the 18-gene set. Twenty-four benign samples out of the total 28 were actually assigned into the tumor-free group. Twenty-five patients with PTC out of the total 39 were considered as malignant