What is the Utility of Genetic Testing in Indeterminate Thyroid Nodules?

BACKGROUND

The prevalence of clinically palpable thyroid nodules in the general public is estimated between 4% and 7%.^1-3 Ultrasound can identify additional nodules in up to one third of the population.^{2, 3} Despite their high incidence thyroid nodules, most often harbor benign or indolent pathology and frequently require no surgical treatment.^1-5

A growing body of information continues to guide one's workup of thyroid nodularity with sonographic and histologic findings often spearheading management strategies. For example, the Thyroid Imaging Reporting and Data System used in conjunction with the Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) can identify a majority of nodules as either benign, or malignant. Unfortunately, up to 30% of fine needle aspirate (FNA) samples can return an indeterminant diagnosis – “Atypia of Undetermined Significance/Follicular Lesion of Undetermined Significance” (AUS/FLUS/Bethesda III) or “Follicular Neoplasm/Suspicious for Follicular Neoplasm” (FN/SFN/Bethesda IV). In this scenario, one's options are fourfold; observation, repeat FNA, diagnostic surgical excision, or to offer molecular testing.

Molecular testing is an increasingly utilized adjunct when evaluating indeterminant thyroid nodules in an effort to avoid unnecessary surgical or diagnostic risk to a patient.^1-5 Its use is supported by the American Thyroid Association and the National Comprehensive Cancer Network (NCCN) among other medical societies.^{3, 4} The potential utility of molecular evaluation of thyroid nodules is broad, with a possible future role in staging, treatment, disease monitoring, and identifying familial risk among other uses.^3-5 This review seeks to evaluate the existing literature to determine the utility and applicability of molecular testing in indeterminate thyroid nodules.

LITERATURE REVIEW

Molecular testing can increase one's understanding of the malignant potential of an indeterminant thyroid nodule. For example, nearly 70% of differentiated thyroid cancers demonstrate BRAF, or RAS mutations or PAX8/PTC translocations.⁵ Nodules that are histologically indeterminant via TBSRTC can be queried to further understand their molecular profile. Low risk molecular profile nodules can be observed with more assurance, and potentially help avoid further evaluation by repeat FNA or surgical lobectomy.

A panel of experts' 2018 update to the NCCN Guidelines for Managing Thyroid Carcinomas commented on the utility of molecular testing. Their recommendation reached a uniform consensus and suggested consideration of molecular testing for indeterminant thyroid nodules, and for evaluation of advanced, progressive, or threatening differentiated thyroid carcinomas.³

Interspace Biosciences (formerly Interspace Diagnostics) offers genetic and mutational analysis for a variety of cancers, including pancreatic, lung, and thyroid. Thyroid molecular testing options offered by Interspace include the ThyGeNEXT, and the ThyraMIR. ThyGeNEXT uses next-generation sequencing to identify DNA mutation variants and messenger RNA fusion transcripts. The sequencing aims to identify known high and low risk driver mutations to help stratify malignancy risk. The ThyraMIR microRNA analysis is combined with ThyGeNEXT to constitute the Multiplatform Test (MPTX) assessment (Interspace Biosciences [formerly Interspace Diagnostics]). When using the MPTX, samples without or only containing low risk driver mutations as identified by ThyGeNEXT are further assessed with ThyraMIR to determine a sample's overall malignant potential. In a 2020 assessment performed by Lupo et al, 197 AUS/FLUS and FN/SFN cytologically indeterminant thyroid nodules were retrospectively assessed with MPTX. The authors identified a negative MPTX result reduced one's risk of malignancy to only 3% to 4%, whereas positive MPTX results significantly elevated one's malignancy risk to 75% to 82%.¹ MPTX characterized a nodule's malignant potential with sensitivity (SN) 93%, specificity (SP) 90%, negative predictive value (NPV) 95%, and positive predictive value (PPV) 74% at disease prevalence of 30% (Table I). The authors further concluded that “ancillary use of the three category MPTX approach is expected to accurately inform the need for surgery in four out of five indeterminate nodules tested”.

ThyroSeq v3 is a next-generation DNA, and RNA sequencing test released by the University of Pittsburgh Medical Center and Sonic Healthcare USA. The evaluation identifies mutations, gene fusion, gene expression alterations, and copy number variations to provide a “positive” or “negative” result prognosticating 59% to 99% cancer probability with a positive result, or <5% risk of malignancy with a negative result. Steward et al evaluated the utility of the ThyroSeq v3 test in 2018.² The group performed a prospective multicentered blinded study evaluating 247 cytologically indeterminant thyroid nodules treated with surgical excision. They identified SN 94%, SP 82%, NPV 97%, and PPV 66% in a cohort of patients with disease prevalence of 28% (Table I). The authors suggested surgery could be avoided in up to 61% of AUS/FLUS/Bethesda III or FN/SFN/Bethesda IV thyroid nodules when ThyroSeq v3 testing is performed. Additionally, the group reported a robust performance across multiple types of thyroid malignancy, offering surgical avoidance in over 50% of ThyroSeq v3 tested Hürthle cell nodules. Authors suggested a potential role of molecular profiling to guide treatment as more information is discovered regarding each specific mutation's potential to impact disease severity or clinical manifestation.

The Afirma GSC (Veracyte, South San Francisco, California) is a machine learning derived classification algorithm utilized to evaluate FNA indeterminant thyroid nodules. The model detects RNA transcriptome expression, mitochondrial RNA, and genomic copy number alterations including loss of heterozygosity to identify samples with suspicious pathology.⁵ In 2018, Patel et al evaluated the Afirma GSC in a cohort of patients treated surgically for cytologically AUS/FLUS/Bethesda III and FN/SFN/Bethesda IV thyroid nodules, demonstrating a SN of 91.1%, SP 68.3%, NPV 96.1%, and PPV of 47.1% in a cohort of 190 patients with malignant incidence of 23.7% (Table I).⁵ The model appropriately identified FNA samples as suspicious for medullary thyroid carcinoma, Hürthle cell carcinoma, poorly differentiated carcinoma, follicular carcinoma, and papillary thyroid carcinoma. Veracyte's Afirma Xpression Atlas adds additional information regarding mutational composition in suspicious nodules in an effort to offer personalized treatment options for known high risk genomic variants and gene fusions.⁵

BEST PRACTICE

Molecular testing can enhance knowledge of the malignant potential of AUS/FLUS/Bethesda III and FN/SFN/Bethesda IV (cytologically indeterminant) thyroid nodules. In addition to avoiding unnecessary diagnostic surgery, these assessments may provide critical insight as well as offer contemporary personalized treatment options for advanced disease. Molecular testing has no clear role in the management of benign or low-grade/stage malignant lesions. The information attained from these assessments encourages further study into the clinical and genetic characteristics of thyroid malignancy and may help to guide personalized medical care in the future.

LEVEL OF EVIDENCE

References 1, and 2, contain Oxford Centre for Evidence-Based Medicine (OCEBM) level 2 evidence. Sources 3, and 4 are consensus statements detailing OCEBM levels 2, 3, 4, and 5 evidence. Reference 5 is a review detailing OCEBM level 2, 3, and 4 evidence.