PROJECT SUMMARY
Thyroid cancer incidence is high, with 44,280 new cases diagnosed in the US in 2021. Thyroid nodule
incidence is rising primarily due to increased detection, necessitating more procedures such as fine needle
aspiration (FNA) biopsies to rule out cancer. However, most thyroid nodule biopsies produce benign,
indeterminate, or non-diagnostic results and are potentially avoidable. Our long-term goal is to improve risk
stratification of thyroid nodules, reduce the number of unnecessary biopsies, and minimize the burden of
thyroid cancer diagnosis for patients and the healthcare system. Our central hypothesis is that thyroid cancer
genetic risk estimate will improve risk stratification of thyroid nodules and reduce the number of avoidable FNA
biopsies of benign thyroid nodules. Supported by robust preliminary data, the central hypothesis will be tested
by pursuing two specific aims: Aim 1. Define the genetic architecture of thyroid cancer. Aim 2. Develop and
assess the clinical utility of a genetic thyroid nodule classifier that discriminates between benign and malignant
thyroid nodules. Under the first aim, we will explore genetic associations with thyroid malignancy independent
of benign goiter to discover novel thyroid cancer biomarkers and develop a clinically useful polygenic risk score
(PRS). We will use two approaches: 1) test genetic associations directly using a GWAS meta-analysis with
4,994 thyroid cancer cases and 20,917 patients with benign nodules as controls, and 2) use a computational
GWAS-by-subtraction method to derive summary statistics for the thyroid cancer free from genetic
associations with benign nodular goiter. We will use publicly available genome-wide association studies, such
as from the Global Biobank Meta-analysis Initiative, and perform our meta-analyses using the Colorado Center
for Personalized Medicine Biobank and other Biobanks from around the world. We hypothesize that a thyroid
nodule classifier PRS in combination with standard of care Thyroid Imaging Reporting and Data System (TI-
RADS) ultrasound schema will improve risk stratification of thyroid nodules and ultimately reduce the number
of unnecessary thyroid nodule biopsies. We developed a thyroid nodule classifier PRS that differentiates
malignant and benign thyroid nodules with an area under the receiver operating characteristic curve of 0.61.
We will apply this score to ~600 thyroid nodules from genotyped patients with known cytologic or
histopathologic diagnoses of benign goiter or thyroid cancer. Three expert physicians will estimate TI-RADS
points and categories. We will evaluate the efficacy of the TI-RADS algorithm alone and in combination with
our novel PRS to distinguish benign from malignant thyroid nodules. We will use the precisely defined genetic
landscape of thyroid malignancy (Aim 1) to improve the thyroid nodule classifier PRS. This study will pave the
way for personalized management of thyroid nodules and inform future mechanistic studies aimed at better
understanding the risk of thyroid cancer.