Descriptive Analysis of Common Fusion Mutations in Papillary Thyroid Carcinoma in Hungary

Abstract

Thyroid cancer is the most common type of endocrine malignancy. Papillary thyroid carcinoma (PTC) is its predominant subtype, which is responsible for the vast majority of cases. It is true that PTC is a malignant tumor with a very good prognosis due to effective primary therapeutic approaches such as thyroidectomy and radioiodine (RAI) therapy. However, we are often required to indicate second-line treatments to eradicate the tumor properly. In these scenarios, molecular therapies are promising alternatives, especially if specifically targetable mutations are present. Many of these targetable gene alterations originate from gene fusions, which can be found using molecular diagnostics like next-generation sequencing (NGS). Nonetheless, molecular profiling is far from being a routine procedure in the initial phase of PTC diagnostics. As a result, the mutation status, except for BRAF V600E mutation, is not included in risk classification algorithms either. This study aims to provide a comprehensive analysis of fusion mutations in PTC and their associations with clinicopathological variables in order to underscore certain clinical settings when molecular diagnostics should be considered earlier, and to demonstrate yet unknown molecular-clinicopathological connections. We conducted a retrospective fusion mutation screening in formalin-fixed paraffin-embedded (FFPE) PTC tissue samples of 100 patients. After quality evaluation by an expert pathologist, RNA isolation was performed, and then NGS was applied to detect 23 relevant gene fusions in the tumor samples. Clinicopathological data were collected from medical and histological records. To obtain the most associations from the multivariate dataset, we used the d-correlation method for our principal component analysis (PCA). Further statistical analyses, including Chi-square tests and logistic regressions, were performed to identify additional significant correlations within certain subsets of the data. Fusion mutations were identified in 27% of the PTC samples, involving nine distinct genes: RET, NTRK3, CCDC6, ETV6, MET, ALK, NCOA4, EML4, and SQSTM1. RET and CCDC6 fusions were associated with type of thyroidectomy, RAI therapy, smaller tumor size, and history of Hashimoto's disease. NCOA4 fusion correlated with sex, multifocality, microcarcinoma character, history of goiter, and obstructive pulmonary disease. EML4 fusion was also linked with surgical procedure type and smaller tumor size, as well as the history of hypothyroidism. SQSTM1 fusion was associated with multifocality and a medical history of thyroid/parathyroid adenoma. NTRK3 and ETV6 fusions showed significant associations with Hashimoto's disease, and ETV6, also with endometriosis. Moreover, fusion mutations were linked to younger age at the time of diagnosis, particularly the fusion of ETV6. The frequent occurrence of fusion mutations and their associations with certain clinicopathological metrics highlight the importance of integrating molecular profiling into routine PTC management. Early detection of fusion mutations can inform surgical decisions and therapeutic strategies, potentially improving clinical outcomes.

Keywords: clinicopathological associations; fusion mutations; gene fusions; molecular diagnostics; papillary thyroid carcinoma; thyroid cancer.

Figure 1

Calculated relative distribution of partner genes associated with detected driver fusion mutations (n = 27) in the PTC cohort without representing fusion non-carrier cases (n = 73). The relative frequency of occurrence was significantly different (Chi-square test) between RET and SQSTM1 fusions (p = 0.026) as marked (*) on the plot. The most frequently identified fusion genes were RET (28.57%) and NTRK3 (16.33%) and their common gene partners CCDC6 and ETV6, respectively.

Figure 2

Principal component analysis (PCA) of fusion genes and clinicopathological variables using d-correlation for mixed scale types. Black points (variable positions) are labeled and color-coded (bottom left corner) to reflect the grouping of different individual variables into larger categories. Variables related to gene fusion status are indicated with red rectangles. It is well demonstrated that most fusion mutation-related variables tended to cluster with specific clinicopathological variables (middle right side). Therapy-related and prognostics-related variables (middle left side), however, correlated negatively with gene fusions.

Figure 3

This horizontal bar chart displays the impact of carrying ETV6 and/or NTRK3 gene fusions on having certain comorbidities. The values are derived from a logistic regression analysis of the ETV6 and NTRK3 gene fusion partners (independent variables) and those binary/nominal-type clinicopathological features (dependent variables) that were associated with these fusions in a significant manner (p < 0.05). The length of the bars depends on the strength of the associations relative to other variable constellations in the cohort. All links represented are above the 0 value threshold on the x-axis, indicating that the directions of all the correlations are positive.

Figure 4

Using Chi-square test (p < 0.05), the type of surgical procedure performed was found to be significantly different when NTRK3 and/or ETV6 fusions occurred compared to those cases without these fusions. This vertical bar chart of these two significant fusions, generated by applying multinomial logistic regression, illustrates the potential impact of the NTRK3 (blue column) and ETV6 (orange column) fusion genes on surgical decision-making across three different categories: primary total thyroidectomy, not-total thyroidectomy (usually lobectomy), and secondary total thyroidectomy (completion of a not-total thyroidectomy). The likelihoods of the indications for total thyroidectomies (primary or secondary) are represented relative to not-total thyroidectomies (with a baseline value of 0). PTC patients with both NTRK3 and/or ETV6 fusion mutations underwent total thyroidectomies more frequently than not-total thyroidectomies. The number of NTRK3 and/or ETV6 fusion-positive patients who needed a secondary completion surgery was greater than the number of those with primary total thyroidectomy increasing the risks related to the repeated procedures.

Figure 5

This bar chart illustrates a comparative analysis of the mean age at the diagnosis of PTC between patients carrying those gene fusions occurring at least 6 times in the cohort compared to the mean age of those patients not carrying any gene fusions. The height of the bars along the y-axis represents the mean age of the patients carrying gene fusions. The specific genes are indicated under the corresponding columns with the last column representing an overall positive status for any studied gene fusions (including those mutations with minimal occurrence as well). The red dashed line marks the mean age of the fusion-negative patients. All evaluated fusion mutations were associated with a younger age at the time of diagnosis than the age of patients without any gene fusions, with NTRK3, ETV6, and general fusion-positive status being significant as marked (*) on the plot. Data are presented as mean ± standard deviation (SD).

Figure 6

Heatmap listing significant associations between binary-type clinicopathological variables (x-axis and y-axis) of the PTC study cohort. The color scale illustrates the direction of the correlations ranging from strongly positive correlations (red) to strongly negative correlations (blue). Empty (white) cells mark no significant associations. Significant associations mostly tend to occur as clinically expected (e.g., strong positive correlation between lymphovascular invasion and lymph node dissection surgery). Medical indication of molecular therapies explicitly correlated with variables, such as relapse, thyroid capsule invasion, extrathyroidal extension, lymphovascular extension, or need for EBRT, usually related to a more advanced state of illness.