Abstract

Thyroid nodules are commonly encountered in clinical practice, affecting up to 50% of the population. The large majority of thyroid lumps are benign incidental findings detected by imaging, while approximately 5-15% harbor malignancy. For a target patient’s care, it is of paramount importance to identify and treat thyroid malignancy, while preventing unnec- essary invasive surgery in patients with benign lesions. Although fine needle aspiration (FNA) associated with cytological examination provides malignant risk information, 20-30% of diagnoses fall into the “indeterminate thyroid nodule” (ITN) category. ITN clinical manage- ment remains a challenging issue for physicians since the ITN risk of malignancy varies from 5% to 40% and most thyroid nodules undergo overtreatment with surgery procedures. ITN molecular testing may better define malignant risk in the single nodule and is able to discriminate with accuracy benign from malignant nodules. Nowadays there are different technologies and different molecular panels, each with its own specificity, sensitivity and predictive values. In view of widespread introduction of molecular testing , some outstand- ing questions remain and are addressed in the present review such as the presence of molecular panels acting as “rule in” or “rule out” tools, the effective impact of testing results in the clinical decision-making process, and the prohibitive cost of commercial assays associated with the lack of test reimbursement in national health systems

Introduction

Thyroid carcinoma accounts for more than 13,000 new cancer diagnoses in Italy, with a predominance of women affected, and represents the second most common malignancy in females under 40 years old ¹. The incidence of thyroid tumors has been increasing over the past decades, partly owning to incidentally detected non-palpable small nodules by imaging. Thyroid nodules are very common and their prevalence may involve 20-50% of the general population but only 5-15% of these nodules harbor malignancy ^2,3. Even if the American ⁴ and Italian ⁵ guidelines definitively support a conservative surgery approach since low-risk thyroid nodules can be safely monitored, most thyroid nodules undergo over-diagnosis and over-treatment in routine daily practice ⁶. To date, an accurate identification of the subgroup of clinically relevant nodules is the goal to be achieved to avoid unnecessary invasive procedures in patients with benign and asymptomatic lesions.

In this setting, fine needle aspiration (FNA) cytology provides accurate diagnostic information to evaluate thyroid nodules ⁷. FNA is a rapid, cost effective and reliable tool that can be performed as an outpatient procedure with a complication rate that is very low. Different systems for reporting thyroid cytopathology may be applied, with different diagnostic categories and risk of malignancy. The Bethesda System includes six categories to classify thyroid nodules, and specifically category I is defined as nondiagnostic or insufficient, category II benign, category III includes “atypia of undetermined significance” (AUS) and “follicular lesion of undetermined significance” (FLUS), category IV represents “follicular neoplasm/suspicious for follicular neoplasm” (FN/SFN) while categories 5 and 6 are suspicious for malignancy and malignant, respectively ⁸. The Italian consensus guidelines reported a classification encompassing TIR 1 non-diagnostic, TIR 1C cystic, TIR 2 negative for malignant cells/benign, TIR 3A low-risk indeterminate lesion, TIR 3B high-risk indeterminate lesion, TIR 4 suspicious of malignancy, and TIR 5 positive for malignant cells ⁵. The Royal College of Pathologist (RCP) thyroid cytology reporting system includes Thy1 non-diagnostic for cytological diagnosis, Thy1c non-diagnostic cystic lesion, Thy2 non-neoplastic, Thy2c non-neoplastic cystic lesion, Thy3a neoplasm possible-atypia/ non-diagnostic, Thy3f neoplasm possible, suggesting follicular neoplasm, Thy4 suspicious of malignancy, and Thy5 malignant ⁹. The comparison among the three classification systems is reported in Table I.

Cytological examination has an excellent accuracy in discriminating between benign and malignant nodules, with a negative predictive value of Bethesda class II of 97% and a positive predictive value of 75% for Bethesda category V and up to 99% for Bethesda category VI ^8,10. However the indeterminate thyroid nodule (ITN) category (Fig. 1) defined as Bethesda class III (BIII) and IV (BIV), RCP Thy3a and Thy3f and Italian TIR 3A and TIR 3B counts for 20-30% of all thyroid nodules and commonly requires further evaluations, having a risk of malignancy of 5-25% (BIII, TIR 3A, Thy3a) and 25-40% (BIV, TIR 3B, Thy3f), respectively ^5,9,11.

An important step to better define the risk of malignancy in ITN has been made by the identification of specific gene alterations that are most frequently associated to thyroid malignant behaviour. In this review, we report on the development of different molecular assays, investigating various targets with particular technologies. We also assess the added benefit in clinical management of ITN by coupling molecular testing to FNA cytology along with the critical issue of the high cost for a widespread testing implementation.

Molecular landscape of thyroid cancers

The molecular pathogenesis of thyroid carcinomas mainly involves dysregulation of different pathways that control cellular proliferation and differentiation, in particular the mitogen activated protein kinase (MAPK) and phosphatidylinositol-3 kinase (PI3K)/AKT signalling are the most affected (Fig. 2). The altered genes and the frequency of mutation vary among different histotypes ¹². A mutation in the BRAF gene (mostly V600E) is found in up to 60% of papillary thyroid cancers (PTC) ¹³, 45% of anaplastic thyroid cancers (ATC) ¹⁴ and 33% of poorly differentiated thyroid cancers (PDTC) ¹⁵. Mutations in RAS gene family, including HRAS, KRAS and NRAS genes, affect follicular thyroid carcinomas (FTC) (about 50% NRAS mut) ¹⁶, PTC (6% NRAS mut) ¹³, PDTC (20% NRAS mut) ¹⁵ but also benign follicular adenomas (20-40%) ¹⁷ as well as non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) (30%) ¹⁸. Other common alterations include RET-PTC fusions, mainly found in PDTC or radiation-related PTC, and PAX8-PPARG rearrangement observed in FTC (30-50%), the follicular variant of papillary thyroid cancer (38%) but also in follicular adenomas (2-13%) ^2,19,20. PIK3CA mutations mark predominantly ATC, as found altered in up to 20% of ATC ²¹. Moreover, p53 and Wnt/Beta catenin signalling may be altered in PDTC and ATC, associated with progression, malignant behavior and dedifferentiation. PDTC and ATC frequently harbour mutation in TERT promoter region (up to 40% PDTC and up to 70% ATC) ^14,21. Conversely, TP53 and TERT mutations are reported in less than 10% and 1% of PTC, respectively ²².

Molecular testing in indeterminate thyroid nodules

Molecular testing of FNA specimens is a potentially practice-changing approach to better define ITN. The American Thyroid Association (ATA) ²³, the NCCN ²⁴, the American Association of Clinical Endocrinologists (AACE), American College of Endocrinology, and Associazione Medici Endocrinologi ²⁵ and the Italian AIOM (Associazione Italiana Oncologia Medica) ²⁶ guidelines suggest that molecular testing may be considered when assessing the risk of malignancy of ITN, most notably when molecular results would influence the choice of surgery. A critical issue is which molecular test to perform. About this, The American Association of Endocrine Surgeons states: ‘To be useful for deciding extent of surgery, a given molecular test needs to first provide reliable prognostic information’ ²⁷. Given that a benign cytologic diagnosis on a thyroid FNA is expected to rule out malignancy in 97% of cases, and the positive predictive value of a FNA malignant result reaches 98%, a reliable molecular test ought to perform with similar accuracy to discriminate between benign and malignant nodules.

“RULE IN” AND “RULE OUT” TESTING

To evaluate the accuracy of a diagnostic test, sensitivity (SN), specificity (SP), positive predictive value (PPV) and negative predictive value (NPV) are all parameters to take into account. To note, according to Bayes’ theorem the disease prevalence influences the predictive values as PPV decrease if decreases prevalence decreases and NPV decreases if the prevalence increases and vice versa. Diagnostic test with a high SN and high NPV are considered valid to “rule out” the presence of disease, whereas a diagnostic test with high SP and PPV are good to “rule in” disease, suggesting that a positive result is consistent with malignancy. An ideal test should accurately discriminate between benign and malignant disease and consequently have a high sensitivity and high specificity ²⁸.

In 2009-2010, the first molecular testing performed on ITN included the evaluation of point mutation in BRAF, KRAS, NRAS, HRAS genes and the rearrangements RET/PTC1, RET/PTC3 and PAX8/PPARG ^29-31. This 7-gene panel applied to thyroid FNA samples with indeterminate cytology is considered as a valid “rule in” test, given the high specificity (SP of 99% and 97% for BIII and BIV nodules, respectively) and positive predictive value (PPV of 88% and 87% for BIII and BIV nodules, respectively). However, the low diagnostic test sensitivity (SN of 63% and 57% for BIII and BIV nodules, respectively) and negative predictive value (NPV of 94% and 86% for BIII and BIV nodules, respectively) show that not all malignancies carry variants or fusions detected by this panel ³². Of note, the malignancy predictive value of molecular alterations varies according to the gene affected and the specific variant or fusion detected. As reported in the systematic review performed by Gordner and colleagues ³³, the BRAF^V600E mutation alone has the highest positive predictive value (PPV) amounting to 98%, whereas other common alterations have lower PPV, such as the fusion PAX8/PPARG (55%), HRAS^Q61R (45%), BRAF^K601E (42%), and NRAS^Q61R (38%). The PPV increases when multiple alterations are found in the same nodule, achieving a cumulative PPV of 77%. On the other hand, about 60% of all ITN harbor a RAS mutation that is mostly associated with benign or low-risk tumors ⁴. In addition, different studies have reported on molecular alterations in benign lesions, with a frequency ranging from 0% to 48% for RAS mutations, from 0% to 68% for RET/PTC rearrangements and from 0% to 55% for PAX8/PPAR-gamma rearrangements ³⁴.

In 2012, the Afirma Gene Expression Classifier (GEC) (Veracyte, San Francisco, CA) was designed as a “rule out” molecular tool. The GEC panel evaluates the expression of 167 genes using microarrays, with a negative predictive value of the “benign” category reaching 95% ^35,36. The validation study ³⁵ involved 265 ITN nodules and was 92% and 52% in overall SN and SP, respectively. The NPV associated with a GEC benign result was 95% and 94% for BIII and BIV nodules, respectively.

Later, several molecular tests investigating both DNA and RNA or miRNA were developed, aiming to perform as both “rule in” and “rule out” tests. In 2013 the ThyroSeq panel v1 (Sonic Healthcare, Rye Brook, NY) is a DNA-based NGS assay that investigates 12-genes for point mutations, namely AKT1, BRAF, NRAS, KRAS, HRAS, PTEN, TP53, TSHR, GNAS, CTNNB1, RET, PIK3CA ³⁷. In 2014 the enriched version ThyroSeq v.2 coupled the DNA-based with the RNA-based NGS, testing simultaneously for point mutations in 13 genes (added TERT gene promoter region) and for 42 types of gene fusions occurring in thyroid cancer ³⁸. The ThyroSeq assay has been additionally expanded in 2015, ThyroSeq v.2.1, with the introduction of EIF1AX gene evaluation, investigating 14 genes for point mutations along with 42 gene fusions, reaching the following performance: SN 90.9%, SP 92.1%, PPV 76.9%, and NPV 97.2% for ITN nodules at cancer prevalence of 22.5% ³⁹.

In 2015 the multiplatform mutation and miRNA test (MPT) was introduced. The ThyGenX/ThyraMIR (Interpace Diagnostics, Parsippany, NJ) combines an 8-gene panel (ThyGenX) with a 10-microRNA (miRNA) expression panel (ThyraMIR) ⁴⁰. The ThyGenX expands the original 7-gene panel with the analysis of PIK3CA gene, mostly altered in FTC and PCTC. The ThyraMIR panel evaluates the expression of 10 mi-RNA, short (~21-23 nucleotides long) non-coding RNAs that negatively modulate the expression of target genes by degradating mRNAs or translational silencing. The ThyraMIR panel subsequently performed tafter a negative result with the 7-gene panel, allows to reach a SN of 89%, SP of 85%, 74% PPV and 94% NPV for BIII and BIV nodules with cancer prevalence of 32% ⁴⁰. Another assay investigating miRNA is the Rosetta GX Reveal assay (Rosetta Genomics Ltd, Rehovot, Israel) that can be performed on a single FNA smear, stained with Papanicolaou or Romanowsky-type stains (Diff-Quik and May Grünwald Giemsa) for cytologic evaluation and stored at room temperature. The assay measures 24 miRNAs to classify cytologically ITNs into benign, suspicious for malignancy, or positive for medullary carcinoma. The assay can be run on smears represented by 1% of thyroid epithelial cells or from which at least 5 ng of RNA have been extracted ⁴¹. The multicenter validation study reported a NPV of 91%, SN of 85% and SP of 72% increased to 99%, 98% and 78%, respectively, evaluating only cases in which all three reviewing pathologists reached an agreement on the diagnosis ⁴².

UP-TO-DATE COMMERCIALLY AVAILABLE MOLECULAR TESTS

Nowadays the most commonly used preoperative molecular tests are Afirma (Veracyte, San Francisco, CA), ThyroSeq (Sonic Healthcare, Rye Brook, NY) and the Multiplatform Test Approach (MPT) (Interpace Diagnostics, Parsippany, NJ), with the most updated version of Afirma Genomic Sequencing Classifier (GSC), Thyroseq v3 and MPTX v2, all centrally performed.

Thyroseq v.3 assay

The ThyroSeq v3 Genomic Classifier (GC) is a targeted next-generation sequencing test that analyzes DNA/RNA of 112 thyroid cancer-related genes for point mutations, insertions/deletions, gene fusions, copy number alterations, or gene expression alterations ⁴³. A GC score is calculated according to the genetic alteration detected. Each alteration has an assigned value based on the strength of association with malignancy: 0 (no association with cancer), 1 (low cancer probability), or 2 (high cancer probability). A multicenter study evaluating the performance of Thyroseq v.3 reported in Bethesda III and IV nodules 94% SN, 82% SP, 97% NPV and 66% PPV at 28% cancer/NIPT prevalence ⁴⁴.

The most widely used sample type for Thyroseq assay is FNA material collected directly into a tube with nucleic acid preservative solution at the time of the FNA procedure. However, a recent study validated slide scraping on cytology smears as an alternative source for ThyroSeq analysis in indeterminate thyroid nodules ⁴⁵.

Afirma GCS (Genomic Sequencing Classifier) assay

The Afirma GSC is a RNA-based NGS panel in which whole transcriptome analysis is combined with machine learning algorithms to provide a benign or suspicious result in nodules with ITN. In the validation study ⁴⁶, the AFIRMA GSC panel showed a SN of 91%, SP of 68%, PPV of 47%, and NPV of 96%. The real world implementation of Afirma GSC testing is depicted in the meta-analysis of Nasr and colleagues ⁴⁷, including 13-independent studies, with a pooled data of 97% of SN and 99% of NPV. Moreover, in this meta-analysis are elaborated the observed SP (oSP) and observed PPV (oPPV), excluding from the analysis unoperated patients with suspicious results, and the conservative SP (cSP) and conservative (cPPV), assuming as histology negatives unoperated patients with suspicious results. The corresponding values obtained were 88% and 65% for oSP and oPPV, respectively, and 80% and 49% for cSP and cPPV, respectively.

Of note, the starting material used for the Afirma molecular analyses is exclusively FNA sample collected and shipped at specific conditions, according to the Veracyte Afirma Genomic Testing specimen’s collection instructions.

MPTX v.1 (ThyGeNEXT® / ThyraMIR®) assay

The MPTX v.1 assay is an expanded NGS test (ThyGeNEXT) used in combination with the microRNA risk classifier test (ThyraMIR) ⁴⁸. The expanded ThyGeNEXT mutation panel includes NTRK and ALK fusions as well as TERT and RET mutations. In MPTX testing, cases with no detectable mutations or with weak driver mutations are further risk stratified using the microRNA classifier, which incorporates two thresholds for malignancy risk. MPTX results are reported as one of three categories (negative, moderate, or positive) based on results of the mutation panel and microRNA risk classifier thresholds. In the multicenter study by Lupo and colleagues ⁴⁹ on ITN, the MPTX v.1 had 95% SN, 90% SP, 97% NPV and 75% PPV, at 30% disease prevalence. In the last study, using the MPTX v.2 integrated with the pairwise miRNA expression analysis, the diagnostic accuracy of ITN risk stratification improved, reporting a SN of 98% ⁵⁰.

The MPTX can be successfully performed on FNA specimens either in RNA preservative fluid or on cytology slides, as recently demonstrated by Kumar ⁵¹. The comparison between molecular results of FNA specimens in RNA preservative and matched FNA cytology smears from 47 patients demonstrated 98% concordance between the results of NGS-based mutation sequencing tests and 90% concordance between the results of microRNA expression-based tests.

According to the most recent meta-analysis by Lee and colleagues in 2022 ⁵² comparing the performances of Afirma GCS (Genomic Sequencing Classifier) and ThyorSeq v.3 panels on Bethesda III or IV cases, and no statistically differences were observed. No comparison has been performed with MPTX, due to the small sample size examined with this test. Pooled data obtained by 7 studies including 472 patients and 6 studies enrolling 530 patients, evaluating with Afirma GCS (A) and Thyroseq v.3 (T) assays respectively, showed similar performance in terms of SN (A:97%, T:95%), SP (A:53%; T:50%), PPV (A:63%, T:70%) and NPV (A:96%, T:92%).

Management of ITN: molecular testing to guide surgical decision-making

The introduction of molecular testing in the management of thyroid nodules with indeterminate cytology may significantly decreases the number of surgical procedures carried out (up to 39-68%) ^53-60.

Surgical decision-making influenced by Afirma results

In the study of Jug and colleagues ⁵⁵ the majority of patients with a suspicious Afirma GEC result proceeded to surgical resection (81% underwent surgery) whereas only 13% of patients with a benign result underwent surgery. Samulski ⁵⁶ reported that 107/136 (79%) of suspicious nodules and 23/158 (15%) of benign or quantity insufficient nodules underwent surgical resection. Moreover, adopting the Afirma GSC instead of GEC panel, the rate of surgical intervention in the indeterminate nodule cohort further decreased, mainly due to the improvement in the Benign Call Rate (BCR). According to the meta-analysis of Vuong and colleagues ⁵⁷, the implementation of GSC panel resulted in an approximately 50% relative reduction of surgical interventions and a significant increase of the BCR from the 43.8% of GEC to 65.3% of the GSC. As reported by Endo and colleagues ⁵⁴, the reduction in surgical procedures was statistically significant in both nodules with Bethesda III and Bethesda IV diagnoses. Switching from GEC to GSC they reported that the surgical rate decreased from 51.3% to 14.9% and from 54.8% to 33.3% in Bethesda III and Bethesda IV nodules, respectively.

Surgical decision-making influenced by Thyroseq results

In the study of Jug ⁵⁵, the identification of “high risk mutations” applying the ThyroSeq panel was associated to subsequent surgery in 85% of patients, whereas 44% of patients without high-risk mutations underwent surgery. At the same, in the series described by Valderrabano and collegues ⁵⁸, 33/45 (71%) of “high-risk mutations” nodules and 64/137 (47%) of “no high-risk mutations” nodules were surgically resected. In the experience reported by Desai ⁵⁹, 96/121 (79%) Thyroseq positive and only 31/294 (11%) Thyroseq negative nodules underwent surgery.

The randomized controlled trial reported by Livhits ⁶⁰, confirmed that the Afirma GSC and Thyroseqv3 have similar specificity and potential to reduce diagnostic surgery of ITN. In detail, 346 patients with ITN were evaluated either with the Afirma GSC (A) or ThyroSeq v3 (T) panel, according to a monthly block randomization scheme. The prevalence of malignancy was 20% among indeterminate nodules. No significant differences were observed adopting one or the other molecular assay in terms of benign call rate (A: 53%, T: 61%), SP (A: 80% and T:85%) and PPV (A: 53% and T:63%). Diagnostic thyroid surgery was avoided in 87 (51%) patients tested with the Afirma GSC panel and in 83 (49%) patients tested with the ThyroSeq v3 test.

Main limitation in implementing commercially available molecular assays in routine clinical settings: assay cost

In many European countries and in Italy in particular, there is no reimbursement for ITN molecular testing, even if reducing the number of unnecessary diagnostic thyroid surgeries has huge implications for patient quality of life and costs of health care. Notably, the commercially available tests mentioned above require high-throughput technologies and skilled personnel to be performed. They are run in company laboratories and cost about $3000 - $6400 per test, depending on the specific assay ^61,62. Because of the prohibitive assay cost, they are rarely requested in the Italian routine clinical setting. Given the established clinical utility of ITN molecular testing, different strategies may be adopted to overcome this issue:

1. individually testing BRAF, RAS point mutations, TERT promoter region and RET, PPARG fusions using the available in-house technology (quantitative PCR/pyrosequencing/direct Sanger sequencing). Muzza and colleagues ⁶² evaluated the performance of non-commercial assays applied in a clinical setting ^29-31,63-68. In particular, they investigated the application of the most frequently used 7-gene panel (including BRAF, K-N-H-RAS mutations and RET/PTC1, RET/PTC3, PAX8/PPARG rearrangements) and 5-gene panel (including BRAF, K-N-H-RAS, TERT promoter region mutations). They reported pooled values of SN, SP, PPV and NPV of 61.3%, 95.2%, 76.5% and 90.6% for the 7-gene assay and of 46.8%, 86.3%, 66.7% and 73.5% for the 5-gene panel, reaching a higher specificity but lower sensitivity than the commercial Afirma GSC or Thyroseq v.3 assays and posing these non-commercial assays as potential “rule-in” but not “rule-out” tools. For this reason the molecular approach based on single or few genes evaluation may supplement but not replace the cyto-pathological diagnosis in thyroid nodules, as recommended by scientific societies ²⁶.
2. using non-thyroid NGS panels focused but including the most frequently mutated genes in solid tumors (i.e. Ion AmpliSeq™ Cancer Hotspot Panel v2 Termofisher Scientific, TruSight Tumor 170 Illumina, Myriapod NGS Cancer panels Diatech pharmacogenetics, VariantPlex solid tumor panel ArcherDX);
3. developing custom-designed NGS assays. In the last years, different groups have made efforts to set up custom multi-gene panels aiming to provide an additional tool in thyroid nodule diagnosis. Sponziello and colleagues ⁶⁹ developed a dual-component molecular assay composed of a 23-gene NGS panel (both DNA and RNA based) and miRNA (miR-146b-5p) digital PCR evaluation. This approach adopted in the ITN cytology setting reached a high diagnostic accuracy, reducing the number of nodules classified as indeterminate. Considering a cancer prevalence of 22.5%, the NGS panel had a SN of 89% and a NPV of 96% and the miRNA assay had a SP of 93% and PPV of 93%. Kocsis-Deák and colleagues ⁷⁰ designed a 23-gene and 568-mutational hotspot panel encompassing MAPK/PI3K/AKT signalling paired with a mutation risk score to assess risk of malignancy of PTC. Ke and colleagues ⁷¹ developed the tyroline NGS panel, detecting 15 target gene mutations and 2 fusions. They applied the tyroline panels among malignant and benign thyroid lesions and reported a SN of 79.3%, a SP of 93.8%, a PPV of 98.5% and a NPV of 46.9%. More recently, Sgariglia and colleagues ⁷² developed and validated the Nexthyro assay, a 15-gene NGS panel targeting 264 clinically relevant mutations. They reported a 100% SP allowing to detect variants with allele frequency as low as 2%. Validation studies enrolling larger cohorts are needed to ascertain the diagnostic utility of these promising custom panels in routine clinical setting.

Other biomarkers for ITN risk assessment: PET/CT imaging and radiomic features

The impact of ¹⁸F-fluorodeoxyglucose (FDG) PET/CT imaging in ITN risk classification and subsequent surgery sparing has been recently investigated in a multicenter trial enrolling 132 patients with scheduled diagnostic surgery for an ITN nodule ⁷³. The patients were randomized to diagnostic surgery or ¹⁸F-FDG PET/CT, undergoing surgery if ¹⁸F-FDG-positive. The study results showed that 83% of patients in the first group and 42% in the second group (p-value < 0.001) had undergone futile surgery. In particular, for non-Hürthle cell nodules applying ¹⁸F-FDG PET/CT can reduce unnecessary surgery by 48%. The ¹⁸F-FDG driven diagnostic workup had high SN and NPV, 94.1% and 95.1% respectively, with low SP and PPV resulting in a BCR of 31.1%.

Moreover, ¹⁸F-FDG PET/CT imaging produce quantitative parameters as SUVmax, SUVpeak, SUVmax-ratio, and SUVpeak-ratio, that can provide information to rule out malignancy in ITN ⁷⁴. Indeed, quantitative ¹⁸F-FDG PET/CT assessment can differentiate preoperative ITN properties since a higher median value for conventional parameters was present in malignant/borderline nodules compared to benign nodules (p < 0.001). Considering that Hürthle cell nodules have a higher FDG avidity resulting in a higher SUV value even if they are benign nodules, higher SUV cut-offs should be applied in Hürthle cell nodules to optimize rule-out ability.

In the last years, radiomics defined as the high-throughput mining of quantitative image features from standard-of-care medical imaging has been investigated as additional tool for identifying malignancies, predicting prognosis or genomic mutation status in different tumors ^75,76. PET/CT imaging can be the source of radiomic features (RFs) reported as histogram-based features, texture-based features, edge features, and shape features. In the study by de Koster and colleagues ⁷⁴ radiomic analysis did not improve the discriminating power of ¹⁸F-FDG-PET/ CT in ruling out malignancy among ITN compared to ¹⁸F-FDG-PET/ CT visual evaluation or its quantitative analysis. However, in the retrospective study by Giovannella and colleagues ⁷⁷ two RFs, shape_Sphericity and glcm_Autocorrelation, showed a significant ability to predict ITN malignancy (AUC = 0.733). These two RFs associated with cytological diagnosis (Bethesda classes) were integrated into a multiparametric model that increases the accuracy of risk stratification compared to Bethesda classification and PET/CT alone. In particular, considering the non-Hürthle cell nodule group, a high diagnostic accuracy was reached with a NPV of 95% and a PPV of 79%. Future studies are needed, possibly investigating non-Hurtle cell nodules and Hürthle cell nodules separately to identify RFs and to integrate quantitative parameters in the predictive model.

In the recent meta-analysis by Ko and colleagues ⁷⁸, the use of ¹⁸F-FDG PET/CT-based RFs applied to thyroid incidentaloma was investigated. The authors included 5 studies, with differences in terms of RFs extraction and selection methods, types and number of features, and modelling methods. They reported a good diagnostic performance of ¹⁸F-FDG PET/CT-based RFs for prediction of malignant thyroid nodules, with a pooled SN and SP values of 0.77 and 0.67, respectively, but they could not recommend the proper radiomics feature selection method or radiomics model method.

These promising results could lead to an expanded use of PET/CT imaging as an additional tool in defining ITN malignant risk.

Liquid biopsy

Liquid biopsy has been emerged as a minimally invasive tool to support diagnosis, refining risk stratification, complementing follow-up, and improving treatment of cancer patients. In the last years, the DNA released by the cell in the blood-stream, the cell-free DNA (cfDNA) and in particular the amount of DNA released by the cancer cells namely circulating tumor DNA (ctDNA), has been evaluated as a surrogate for tumor tissue DNA. In the ITN setting, the potential role of liquid biopsy to discriminate between benign and malignant nodules has been explored. In particular, the most frequently investigated biomarker is BRAF^V600E in cfDNA that correlates with a diagnosis of PTC ^79-81. Other specific biomarkers may be useful in discriminating between PTC and benign nodules including epigenetic traits such as the methylation status of RASSF1 and SLC5A8 ⁸¹ of MGMT ⁸² or CALCA, CDH1, TIMP3, DAPK, and RARB2 genes ⁸³ or serum expression levels of miRNA-95 and -190 ⁸⁴. On the other hand, there are non-specific biomarkers evaluated for refining the diagnosis of indeterminate thyroid nodule such as the quantification of cfDNA or cfDNA fragmentation index ^85,86. In the recent meta-analysis by Hou ⁸⁷ and colleagues all these biomarkers were debated. They found that a combination of multiple cfDNA biomarkers may be more accurate than single biomarker testing. Individually comparing cfDNA mutations, cfDNA methylation, cfDNA integrity index and quantitative analysis (cfDNA level) as potential biomarkers, they reported a high diagnostic accuracy of TC for cfDNA levels, with a sensitivity of 0.84 (0.67-0.94) and specificity of 0.89 (0.85-0.92), even if large-scale studies are needed to ascertain its diagnostic utility. To date, the role of liquid biopsy in complementing an indeterminate thyroid cytology diagnosis remains to be established in a routine clinical setting.

Conclusions

In the clinical setting of patients with thyroid nodules, the challenging issue consists in the management of indeterminate nodules, accounting for 20-30% of nodules, whereas FNA cytological evaluation alone fails to distinguish between benign or malignant features. Coupling FNA cytological evaluation and molecular testing can improve targeted patient care and guide surgical decision making. The commercially available tests are affordable “rule in” and “rule out” tools. The costs of molecular testing and the sustainability in the public health system remains to be addressed beforewidespread clinical implementation.

CONFLICTS OF INTEREST

The authors declare no conflict of interest.

FUNDING

None

ETHICAL CONSIDERATION

None

AUTHORS’ CONTRIBUTIONS

CF and GS: manuscript conception, writing and reviewing.

Figures and tables

Figure 1.Cytological features of the indeterminate thyroid nodules category: (A) Specimen obscured by blood that does not allow an evaluation of cellular details; (B) Few atypical cells; (C) Architectural atypia in poorly cellular specimen; (D) Specimen where sparse colloid is evident and where a definite distinction between a follicular neoplasm and a hyperplastic nodule is difficult; (E) Micro- and macro-follicular pattern; (F) Diffuse and monotonous architectural atypia suggesting follicular neoplasm.

Figure 2.Pathways commonly affected in thyroid carcinoma. The most frequent alterations are depicted. GF: growth factor; RTK: receptor tyrosine kinase.

References

1. I numeri del cancro in Italia 2021 - AIOM.Publisher Full Text
2. Durante C, Grani G, Lamartina L. The Diagnosis and Management of Thyroid Nodules: A Review. JAMA. 2018; 319:914-924. DOI
3. Brito JP, Morris JC, Montori VM. Thyroid cancer: zealous imaging has increased detection and treatment of low risk tumours. BMJ. 2013; 347:f4706. DOI
4. Haugen BR, Alexander EK, Bible KC. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid. 2016; 26:1-133. DOI
5. Pacini F, Basolo F, Bellantone R. Italian consensus on diagnosis and treatment of differentiated thyroid cancer: Joint statements of six Italian societies. J Endocrinol Invest. 2018; 41:849-876. DOI
6. D’Amore A, Patrone R, Docimo L. Thyroid Cancer: Toward Surgical Evolution. J Clin Med. 2021; 10:3582. DOI
7. Sangalli G, Serio G, Zampatti C. Fine needle aspiration cytology of the thyroid: a comparison of 5469 cytological and final histological diagnoses. Cytopathology. 2006; 17:245-50. DOI
8. Cibas ES, Ali SZ. The 2017 Bethesda system for reporting thyroid cytopathology. Thyroid. 2017; 27:1341-1346. DOI
9. Cross P, Chandra A, Giles T. Guidance on the reporting of thyroid cytology specimens. Royal College of Pathologists: London, UK; 2016.
10. Bongiovanni M, Spitale A, Faquin WC. The Bethesda System for Reporting Thyroid Cytopathology: a meta-analysis. Acta Cytol. 2012; 56:333-339. DOI
11. Baloch ZW, Cooper DS, Gharib H. The Bethesda System for Reporting Thyroid Cytopathology: Definitions, Criteria, and Explanatory Notes. Springer: New York, NY; 2017.
12. Prete A, Borges de Souza P, Censi S. Update on Fundamental Mechanisms of Thyroid Cancer. Front Endocrinol (Lausanne). 2020; 11:102. DOI
13. The Cancer Genome Research Network. Integrated genomic characterization of papillary thyroid carcinoma. Cell. 2014; 159:676-690. DOI
14. Kunstman JW, Juhlin CC, Goh G. Characterization of the mutational landscape of anaplastic thyroid cancer via whole-exome sequencing. Hum Mol Genet. 2015; 24:2318-2329. DOI
15. Pozdeyev N, Gay LM, Sokol ES. Genetic analysis of 779 advanced differentiated and anaplastic thyroid cancers. Clin Cancer Res. 2018; 24:3059-3068. DOI
16. Fukahori M, Yoshida A, Hayashi H. The associations between ras mutations and clinical characteristics in follicular thyroid tumors: new insights from a single center and a large patient cohort. Thyroid. 2012; 22:683-689. DOI
17. Nikiforov YE. Molecular diagnostics of thyroid tumors. Arch Pathol Lab Med. 2011; 135:569-577. DOI
18. Nikiforov YE, Seethala RR, Tallini G. Nomenclature revision for encapsulated follicular variant of papillary thyroid carcinoma: a paradigm shift to reduce overtreatment of indolent tumors. JAMA Oncol. 2016; 2:1023-1029. DOI
19. Boos LA, Dettmer M, Schmitt A. Diagnostic and prognostic implications of the PAX8-PPARg translocation in thyroid carcinomas-a TMA-based study of 226 cases. Histopathology. 2013; 63:234-241. DOI
20. Nikiforova MN, Biddinger PW, Caudill CM. PAX8-PPARg rearrangement in thyroid tumors: RT-PCR and immunohistochemical analyses. Am J Surg Pathol. 2002; 26:1016-1023. DOI
21. Landa I, Ibrahimpasic T, Boucai L. Genomic and transcriptomic hallmarks of poorly differentiated and anaplastic thyroid cancers J Clin Invest. 2016; 126:1052-1066. DOI
22. Fagin JA, Wells SA. Biologic and Clinical Perspectives on Thyroid Cancer. N Engl J Med. 2016; 375:1054-1067. DOI
23. Haugen BR. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: What is new and what has changed?. Cancer. 2017; 123:372-381. DOI
24. Haddad RI, Nasr C, Bischoff L. NCCN Guidelines Insights: Thyroid Carcinoma, Version 2.2018. J Natl Compr Canc Netw. 2018; 16:1429-1440. DOI
25. Garber JR, Papini E, Frasoldati A. American Association of Clinical Endocrinology And Associazione Medici Endocrinologi Thyroid Nodule Algorithmic Tool. Endocr Pract. 2021; 27:649-660. DOI
26. Linee guida Tumori Tiroide 2021 - AIOM.Publisher Full Text
27. Patel KN, Yip L, Lubitz CC. The American Association of Endocrine Surgeons guidelines for the definitive surgical management of thyroid disease in adults. Ann Surg. 2020; 271:e21-e93. DOI
28. Roth MY, Witt RL, Steward DL. Molecular testing for thyroid nodules: Review and current state. Cancer. 2018; 124:888-898. DOI
29. Nikiforov YE, Steward DL, Robinson-Smith TM. Molecular testing for mutations in improving the fine-needle aspiration diagnosis of thyroid nodules. J Clin Endocrinol Metab. 2009; 94:2092-2098. DOI
30. Ohori NP, Nikiforova MN, Schoedel KE. Contribution of molecular testing to thyroid fine-needle aspiration cytology of “follicular lesion of undetermined significance/atypia of undetermined significance”. Cancer Cytopathol. 2010; 118:17-23. DOI
31. Cantara S, Capezzone M, Marchisotta S. Impact of proto-oncogene mutation detection in cytological specimens from thyroid nodules improves the diagnostic accuracy of cytology. J Clin Endocrinol Metab. 2010; 95:1365-1369. DOI
32. Nikiforov YE, Ohori NP, Hodak SP. Impact of mutational testing on the diagnosis and management of patients with cytologically indeterminate thyroid nodules: a prospective analysis of 1056 FNA samples. J Clin Endocrinol Metab. 2011; 96:3390-3397. DOI
33. Goldner WS, Angell TE, McAdoo SL. Molecular Variants and Their Risks for Malignancy in Cytologically Indeterminate Thyroid Nodules. Thyroid. 2019; 29:1594-1605. DOI
34. Najafian A, Noureldine S, Azar F. RAS Mutations, and RET/PTC and PAX8/PPAR-gamma Chromosomal Rearrangements Are Also Prevalent in Benign Thyroid Lesions: Implications Thereof and A Systematic Review. Thyroid. 2017; 27:39-48. DOI
35. Alexander EK, Kennedy GC, Baloch ZW. Preoperative diagnosis of benign thyroid nodules with indeterminate cytology. N Engl J Med. 2012; 367:705-715. DOI
36. Kloos RT. Molecular Profiling of Thyroid Nodules: Current Role for the Afirma Gene Expression Classifier on Clinical Decision Making. Mol Imaging Radionucl Ther. 2017; 26:36-49. DOI
37. Nikiforova MN, Wald AI, Roy S. Targeted next-generation sequencing panel (ThyroSeq) for detection of mutations in thyroid cancer. J Clin Endocrinol Metab. 2013; 98:E1852-1860. DOI
38. Nikiforov YE, Carty SE, Chiosea SI. Highly accurate diagnosis of cancer in thyroid nodules with follicular neoplasm/suspicious for a follicular neoplasm cytology by ThyroSeq v2 next-generation sequencing assay. Cancer. 2014; 120:3627-3634. DOI
39. Nikiforov YE, Carty SE, Chiosea SI. Impact of the multi-gene ThyroSeq next-generation sequencing assay on cancer diagnosis in thyroid nodules with atypia of undetermined significance/follicular lesion of undetermined significance cytology. Thyroid. 2015; 25:1217-1223. DOI
40. Labourier E, Shifrin A, Busseniers AE. Molecular Testing for miRNA, mRNA, and DNA on Fine-Needle Aspiration Improves the Preoperative Diagnosis of Thyroid Nodules With Indeterminate Cytology. J Clin Endocrinol Metab. 2015; 100:2743-2750. DOI
41. Benjamin H, Schnitzer-Perlman T, Shtabsky A. Analytical validity of a microRNA-based assay for diagnosing indeterminate thyroid FNA smears from routinely prepared cytology slides. Cancer Cytopathol. 2016; 124:711-721. DOI
42. Lithwick-Yanai G, Dromi N, Shtabsky A. Multicentre validation of a microRNA-based assay for diagnosing indeterminate thyroid nodules utilising fine needle aspirate smears. J Clin Pathol. 2017; 70:500-507. DOI
43. Nikiforova MN, Mercurio S, Wald AI. Analytical performance of the ThyroSeq v3 genomic classifier for cancer diagnosis in thyroid nodules. Cancer. 2018; 124:1682-1690. DOI
44. Steward DL, Carty SE, Sippel RS. Performance of a Multigene Genomic Classifier in Thyroid Nodules With Indeterminate Cytology: A Prospective Blinded Multicenter Study. JAMA Oncol. 2019; 5:204-212. DOI
45. Nikiforova MN, Lepe M, Tolino LA. Thyroid cytology smear slides: An untapped resource for ThyroSeq testing. Cancer Cytopathol. 2021; 129:33-42. DOI
46. Patel KN, Angell TE, Babiarz J. Performance of a Genomic Sequencing Classifier for the Preoperative Diagnosis of Cytologically Indeterminate Thyroid Nodules. JAMA Surg. 2018; 153:817-824. DOI
47. Nasr CE, Andrioli M, Endo M. Real World Performance of The Afirma Genomic Sequencing Classifier (GSC) - A Meta-analysis. J Clin Endocrinol Metab. 2022. DOI
48. Jackson S, Kumar G, Banizs AB. Incremental utility of expanded mutation panel when used in combination with microRNA classification in indeterminate thyroid nodules. Diagn Cytopathol. 2020; 48:43-52. DOI
49. Lupo MA, Walts AE, Sistrunk JW. Multiplatform molecular test performance in indeterminate thyroid nodules. Diagn Cytopathol. 2020; 48:1254-1264. DOI
50. Finkelstein SD, Sistrunk JW, Malchoff C. A Retrospective Evaluation of the Diagnostic Performance of an Interdependent Pairwise MicroRNA Expression Analysis with a Mutation Panel in Indeterminate Thyroid Nodules. Thyroid. 2022; 32:1362-1371. DOI
51. Kumar G, Timmaraju VA, Song-Yang JW. Utility of microdissected cytology smears for molecular analysis of thyroid malignancy. Diagn Cytopathol. 2019; 47:289-296. DOI
52. Lee E, Terhaar S, McDaniel L. Diagnostic performance of the second-generation molecular tests in the assessment of indeterminate thyroid nodules: A systematic review and meta-analysis. Am J Otolaryngol. 2022; 43:103394. DOI
53. Glass RE, Marotti JD, Kerr DA. Using molecular testing to improve the management of thyroid nodules with indeterminate cytology: an institutional experience with review of molecular alterations. J Am Soc Cytopathol. 2022; 11:79-86. DOI
54. Endo M, Nabhan F, Porter K. Afirma Gene Sequencing Classifier Compared with Gene Expression Classifier in Indeterminate Thyroid Nodules. Thyroid. 2019; 29:1115-1124. DOI
55. Jug R, Parajuli S, Ahmadi S. Negative Results on Thyroid Molecular Testing Decrease Rates of Surgery for Indeterminate Thyroid Nodules. Endocr Pathol. 2019; 30:134-137. DOI
56. Samulski TD, LiVolsi VA, Wong LQ. Usage trends and performance characteristics of a “gene expression classifier” in the management of thyroid nodules: An institutional experience. Diagn Cytopathol. 2016; 44:867-873. DOI
57. Vuong HG, Nguyen TPX, Hassell LA. Diagnostic performances of the Afirma Gene Sequencing Classifier in comparison with the Gene Expression Classifier: A meta-analysis. Cancer Cytopathol. 2021; 129:182-189. DOI
58. Valderrabano P, Khazai L, Leon ME. Evaluation of ThyroSeq v2 performance in thyroid nodules with indeterminate cytology. Endocr Relat Cancer. 2017; 24:127-136. DOI
59. Desai D, Lepe M, Baloch ZW. ThyroSeq v3 for Bethesda III and IV: An institutional experience. Cancer Cytopathol. 2021; 129:164-170. DOI
60. Livhits MJ, Zhu CY, Kuo EJ. Effectiveness of Molecular Testing Techniques for Diagnosis of Indeterminate Thyroid Nodules: A Randomized Clinical Trial. JAMA Oncol. 2021; 7:70-77. DOI
61. Sciacchitano S, Lavra L, Ulivieri A. Comparative analysis of diagnostic performance, feasibility and cost of different test-methods for thyroid nodules with indeterminate cytology. Oncotarget. 2017; 8:49421-49442. DOI
62. Muzza M, Colombo C, Pogliaghi G. Molecular markers for the classification of cytologically indeterminate thyroid nodules. J Endocrinol Invest. 2020; 43:703-716. DOI
63. Beaudenon-Huibregtse S, Alexander EK, Guttler RB. Centralized molecular testing for oncogenic gene mutations complements the local cytopathologic diagnosis of thyroid nodules. Thyroid. 2014; 24:1479-1487. DOI
64. Eszlinger M, Krogdahl A, Münz S. Impact of molecular screening for point mutations and rearrangements in routine air-dried fine-needle aspiration samples of thyroid nodules. Thyroid. 2014; 24:305-313. DOI
65. Eszlinger M, Böhme K, Ullmann M. Evaluation of a Two-Year Routine Application of Molecular Testing of Thyroid Fine-Needle Aspirations Using a Seven-Gene Panel in a Primary Referral Setting in Germany. Thyroid. 2017; 27:402-411. DOI
66. Bongiovanni M, Molinari F, Eszlinger M. Laser capture microdissection is a valuable tool in the preoperative molecular screening of follicular lesions of the thyroid: an institutional experience. Cytopathology. 2015; 26:288-296. DOI
67. Censi S, Cavedon E, Bertazza L. Frequency and Significance of Ras, Tert Promoter, and Braf Mutations in Cytologically Indeterminate Thyroid Nodules: A Monocentric Case Series at a Tertiary-Level Endocrinology Unit. Front Endocrinol (Lausanne). 2017; 8:273. DOI
68. Macerola E, Rago T, Proietti A. The mutational analysis in the diagnostic work-up of thyroid nodules: the real impact in a center with large experience in thyroid cytopathology. J Endocrinol Invest. 2019; 42:157-166. DOI
69. Sponziello M, Brunelli C, Verrienti A. Performance of a dual-component molecular assay in cytologically indeterminate thyroid nodules. Endocrine. 2020; 68:458-465. DOI
70. Kocsis-Deák B, Árvai K, Balla B. Targeted Mutational Profiling and a Powerful Risk Score as Additional Tools for the Diagnosis of Papillary Thyroid Cancer. Pathol Oncol Res. 2020; 26:101-108. DOI
71. Ke Z, Liu Y, Zhang Y. Diagnostic value and lymph node metastasis prediction of a custom-made panel (thyroline) in thyroid cancer. Oncol Rep. 2018; 40:659-668. DOI
72. Sgariglia R, Nacchio M, Migliatico I. Moving towards a local testing solution for undetermined thyroid fine-needle aspirates: validation of a novel custom DNA-based NGS panel. J Clin Pathol. 2022; 75:465-471. DOI
73. de Koster EJ, de Geus-Oei LF, Brouwers AH. [18F]FDG-PET/CT to prevent futile surgery in indeterminate thyroid nodules: a blinded, randomised controlled multicentre trial. Eur J Nucl Med Mol Imaging. 2022; 49:1970-1984. DOI
74. de Koster EJ, Noortman WA, Mostert JM. Quantitative classification and radiomics of [18F]FDG-PET/CT in indeterminate thyroid nodules. Eur J Nucl Med Mol Imaging. 2022; 49:2174-2188. DOI
75. Lambin P, Leijenaar RTH, Deist TM. Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017; 14:749-762. DOI
76. Rizzo S, Petrella F, Buscarino V. CT Radiogenomic Characterization of EGFR, K-RAS, and ALK Mutations in Non-Small Cell Lung Cancer. Eur Radiol. 2016; 26:32-42. DOI
77. Giovanella L, Milan L, Piccardo A. Radiomics analysis improves 18FDG PET/CT-based risk stratification of cytologically indeterminate thyroid nodules. Endocrine. 2022; 75:202-210. DOI
78. Ko WS, Kim SJ. Prediction of Malignant Thyroid Nodules Using 18 F-FDG PET/CT-Based Radiomics Features in Thyroid Incidentalomas. Clin Nucl Med. 2023; 48:497-504. DOI
79. Patel KB, Cormier N, Fowler J. Detection of Circulating Tumor DNA in Patients with Thyroid Nodules. Int J Endocrinol. 2021; 2021:8909224. DOI
80. Pupilli C, Pinzani P, Salvianti F. Circulating BRAFV600E in the diagnosis and follow-up of differentiated papillary thyroid carcinoma. J Clin Endocrinol Metab. 2013; 98:3359-3365. DOI
81. Khatami F, Larijani B, Heshmat R. Hypermethylated RASSF1 and SLC5A8 promoters alongside BRAFV600E mutation as biomarkers for papillary thyroid carcinoma. J Cell Physiol. 2020; 235:6954-6968. DOI
82. Khatami F, Teimoori-Toolabi L, Heshmat R. Circulating ctDNA methylation quantification of two DNA methyl transferases in papillary thyroid carcinoma. J Cell Biochem. 2019; 120:17422-17437. DOI
83. Hu S, Ewertz M, Tufano RP. Detection of serum deoxyribonucleic acid methylation markers: a novel diagnostic tool for thyroid cancer. J Clin Endocrinol Metab. 2006; 91:98-104. DOI
84. Pilli T, Cantara S, Marzocchi C. Diagnostic Value of Circulating microRNA-95 and -190 in the Differential Diagnosis of Thyroid Nodules: A Validation Study in 1000 Consecutive Patients. Thyroid. 2017; 27:1053-1057. DOI
85. Higazi AM, El Hini SH, El-Sharkawy EA. Diagnostic Role of Cell-free DNA Integrity in Thyroid Cancer Particularly for Bethesda IV Cytology. Endocr Pract. 2021; 27:673-681. DOI
86. Dutta S, Tarafdar S, Mukhopadhyay P. Plasma Cell-Free DNA to Differentiate Malignant from Benign Thyroid Nodules. J Clin Endocrinol Metab. 2021; 106:e2262-e2270. DOI
87. Hou F, Sun XD, Deng ZY. Diagnostic value of cell-free DNA in thyroid cancer: a systematic review and meta-analysis. Medicine (Baltimore). 2023; 102:e32928. DOI