Endocrine Abstracts

Introduction and aim: Thyroid cancer is the most common endocrine neoplasia affecting 0.2–1.5% of the individuals worldwide and its incidence is increasing [1]. Papillary thyroid carcinoma (PTC) is the main histotype (80–90%). PTC susceptibility is the result of genetic predisposition, environmental factors and lifestyle. We hypothesized that genetic predisposition to PTC is poligenic and characterized by a genetic signature involving combinations of genes previously found to be associated to PTC.

Methods and results: We considered all the genetic variants (SNPs) significantly associated with PTC on the PubMed literature database. 184 informative SNPs were selected after data refining considering linkage disequilibrium. Then, SNPs data were extracted from the 1,000 Genomes database (www.1000genomes.org), comprising genome of 2504 unselected individuals, collected worldwide. The combination of 184 SNPs associated with PTC was used to group individuals in different risk-clusters according to their genetic structure, calculated by Bayesian statistics, as previously successfully performed for polycystic ovary syndrome [2]. Individuals resulted to be distributed among seven groups worldwide, indicating different degree of genetic predisposition to PTC. Then, we repeated the same analyzes considering genetic data from about 1200 individuals (697 PTC versus 497 healthy controls) of Central/South Italian origin registered in a GWAS, specific for PTC, by the German Cancer Research Center (DKFZ) of Heidelberg [3]. This first analysis was refined using the 33 SNPs with highest odds ratio, therefore, reasonably most causative of genetic clustering. We clearly demonstrated that PTC and healthy control groups are genetically different, revealing diverse predisposition to develop cancer. Then, the genetic structure of each subject was indicated as a percentage of affinity to each risk-cluster and re-analyzed together with other risk factors: sex, BMI, area of origin and familiarity (quantified in a growing score as the degree of kinship increases). These data were analyzed together by principal component analysis (PCA) and clustering of the two groups was even more pronounced. The most contributive factors to the diversity between PTC and healthy controls were genetics and familiarity, while sex, body-mass index and area of origin were less relevant.

Conclusion: We demonstrated that PTC risk may be predicted using SNPs and risk factors data. Validation of the model in independent groups of PTC patients (n=200) and healthy controls (n=130) genotyped by iPlex assay is ongoing.

References: 1. Cabanillas et al. Thyroid cancer. Lancet.2016;388:2783–95.

2. Casarini and Brigante. JCEM.2014;99:E2412–20.

3. Köhler et al. Genome-wide association study on differentiated thyroid cancer. JCEM.2013;98:E1674–81.