Abstract
Cytology is the gold standard method for the differential diagnosis of thyroid nodules, though 25–30% of them are classified as indeterminate. We aimed to set up a ‘thyroid risk score’ (TRS) to increase the diagnostic accuracy in these cases. We prospectively tested 135 indeterminate thyroid nodules. The pre-surgical TRS derived from the sum of the scores assigned at cytology, EU-TIRADS classification, nodule measurement, and molecular characterization, which was done by our PTC-MA assay, a customized array able to cost-effectively evaluate 24 different genetic alterations including point mutations and gene fusions. The risk of malignancy (ROM) increased paralleling the score: in the category >4 and ≤ 6 (low suspicion), >6 ≤ 8 (intermediate suspicion), and >8 (high suspicion); ROM was 10, 47 and 100%, respectively. ROC curves selected the score >6.5 as the best threshold to differentiate between malignant and benign nodules (P < 0.001). The TRS > 6.5 had a better performance than the single parameters evaluated separately, with an accuracy of 77 and 82% upon inclusion of noninvasive follicular thyroid neoplasm with papillary-like nuclear features among malignant or benign cases, respectively. In conclusion, for the first time, we generated a score combining a cost-effective molecular assay with already validated tools, harboring different specificities and sensitivities, for the differential diagnosis of indeterminate nodules. The combination of different parameters reduced the number of false negatives inherent to each classification system. The TRS > 6.5 was highly suggestive for malignancy and retained a high accuracy in the identification of patients to be submitted to surgery.
Introduction
Cytology is the gold standard method for the differential diagnosis of thyroid nodules, though the 25–30% of them are classified as indeterminate in categories III or IV of the Bethesda System for Reporting Thyroid Cytopathology (Nardi et al. 2014, Cibas & Ali 2017). Indeterminate nodules are often treated with unnecessary surgery, as only 10–40% of them will be confirmed to be a malignant lesion on histological examination (Ho et al. 2014). Several approaches have been proposed to improve cytological diagnosis and to significantly reduce unnecessary procedures. In this context, ultrasound (US) classifications, allowing a better classification of malignant nodules, have developed worldwide. Among them, the European Thyroid Association guidelines introduced in 2017, a US classification for thyroid nodules, the EU-TIRADS, which was confirmed to be a highly sensitive system and is widely used among European Countries (Russ et al. 2017, Yoon et al. 2020). Scoring methods combining US data with clinical assessment have been proposed, too (Ianni et al. 2016, Cozzolino et al. 2020). Moreover, upon the improved knowledge on the genetic characterization of thyroid cancer (TC), several studies focused on the molecular evaluation of nodules in order to optimize the management of cytological indeterminate nodules (Alexander et al. 2012, Duick et al. 2012, McIver et al. 2014, Nikiforov et al. 2014, Nikiforov 2017), as recommended in recent guidelines (Li et al. 2011, Gharib et al. 2016, Haugen et al. 2016). The available molecular diagnostic tools are classified into two different categories: ‘rule-out’ methods, which have the purpose to reduce the avoidable treatment of benign nodules, and ‘rule-in’ approaches that aim to optimize surgical management (total thyroidectomy or diagnostic lobectomy). Among the rule-in methods, the last version of ThyroSeq (TSv3) is able to identify more than 12,000 hotspot mutations and more than 120 fusions, while the rule out Afirma® Genomic Sequencing Classifier (GSC) analyzes the expression profile of 1115 genes and detects single nucleotide variants and fusions. Additional approaches for molecular testing include the analysis of miRNAs expression (revised in Muzza et al. 2020). Although both GSC and TSv3 have been demonstrated to be considerably more cost-effective than diagnostic lobectomy (Nicholson et al. 2019), the major drawback of these molecular tools remains to be the high cost (Sciacchitano et al. 2017) which largely restricts their application in clinical practice, especially in European Countries.
To overcome this limitation, customized 5- or 7-genes rule-in panels have been set up to analyze the most frequent genetic alterations found in TC (reviewed in Muzza et al. 2020). We recently developed a PTC-MA assay that able to evaluate in an extremely cost-effective manner (300 euro/sample), a total of 24 genetic alterations including point mutations and fusions frequently found in TC (Pesenti et al. 2018, Colombo et al. 2019). To further increase the sensitivity and specificity of this molecular tool, we set up an unprecedented ‘thyroid risk score’ system (TRS) based on the combination of clinical, ultrasound, cytological and molecular criteria and prospectively tested it on a large number of indeterminate nodules with histological correlation.
Materials and methods
Study cohort
In the 2.5 years period here considered (September 2017 to February 2020), 2215 cytological examinations were carried out at our Institution. US-guided FNA procedures were performed using 21 Gy needles under the US guidance with three or more consecutive passes for each nodule. The cytological and histological diagnosis were done by cytotechnologists and a skilled pathologist in accordance with The Bethesda System for Reporting Thyroid Cytopathology 2017 (Cibas & Ali 2017).
Among the 2215 nodules, 271 cases (12.2%) were classified as Bethesda class I (nondiagnostic), 1737 (78.4%) as class II (benign), 28 (1.2%) as class V (suspicious for malignancy), and 41 (2.8%) as class VI (malignant). Moreover, 64 (2.8%) and 74 cases (3.3%) were indeterminate, namely class III (atypia or follicular lesion of undeterminate significance, AUS/FLUS) and IV (follicular neoplasm, suspicious for a follicular neoplasm, FN/SFN), respectively (Supplementary Fig. 1, see section on supplementary materials given at the end of this article).
The pre-surgical thyroid risk score (TRS) including clinical, ultrasonographic, cytological and molecular features of indeterminate thyroid nodules. A different suspicion for malignancy was arbitrarily assigned to the different scores. *Russ et al. (2017); **Cibas and Ali (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
The pre-surgical thyroid risk score (TRS) including clinical, ultrasonographic, cytological and molecular features of indeterminate thyroid nodules. A different suspicion for malignancy was arbitrarily assigned to the different scores. *Russ et al. (2017); **Cibas and Ali (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
The pre-surgical thyroid risk score (TRS) including clinical, ultrasonographic, cytological and molecular features of indeterminate thyroid nodules. A different suspicion for malignancy was arbitrarily assigned to the different scores. *Russ et al. (2017); **Cibas and Ali (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
At the time of FNA, all nodules are routinely submitted to ultrasonographic risk assessment (Esaote, My Lab Class C or MHz Hitachi HI vision Avius) according to the (US) EU-TIRADS score (Russ et al. 2017) by two endocrinologists with more than 20 years of experience.
After approval of the Ethical Committee of the Istituto Auxologico Italiano (#2018_09_25_04) and patients’ approval, 138 indeterminate nodules (64 class III and 74 class IV) were submitted to molecular analyses through PTC-MA assay (Pesenti et al. 2018, Colombo et al. 2019).
For these nodules, a second FNA was performed and needle was washed out with RNA Later buffer (RNAprotect Tissue Reagent, QIAGEN) in a sterilized tube and sent to the genetics lab within 24 h at room temperature.
DNA/RNA extraction, reverse transcription, assessment of thyroid cell content and PTC-MA assay
Genomic DNA and total RNA were extracted from FNA samples using AllPrep DNA/RNA Micro Kit (Qiagen).
Genomic DNA and total RNA were extracted from either frozen or FFPE tissues by commercial methods (Puregene® Core Kit A, Qiagen and Trizol reagent, Thermo Fisher).
One hundred nanograms of each RNA sample was reverse-transcribed using a Superscript reverse transcriptase II Kit (Thermo Fisher), with a hexamer mixture as primers. β-actin amplification from 100 ng of cDNA was performed as an internal quality control (primers: ACTB4F: GCACTCTTCCAGCCTTCCTT, ACTB5R: AATGCCAGGGTACATGGTG).
To test the follicular cell content of the FNA sample, PAX8 expression on cDNA was evaluated by semiquantitative RT-PCR. In particular, 100 ng of cDNA from each sample and 100 ng of control, obtained by retro-transcription of 90% of RNA from HELA cell lines and 10% of RNA from human primary thyroid follicular epithelial cells (Nthy-ori 3-1), were amplified with PAX8 primers spanning exons 3 and 4 (PAX8_3F: TCTGTGACAATGACACTGTG, PAX8_4R: GTCCATAGGGAGGTTGAATG) for 25 cycles. Band intensities were analyzed with the image analysis program NIH ImageJ and the PAX8 signal was normalized to actin signal. According to the sensitivity of the PTC-MA (Pesenti et al. 2018), a thyroid cell content of 10% was chosen to select samples adequate for further molecular analyses (Supplementary Fig. 2).
Molecular, ultrasonographic and histological data of the 26 patients with at least one genetic alteration. NH, hyperplastic nodule; FA, follicular adenoma; NIFTP, noninvasive follicular thyroid neoplasms with papillary-like nuclear features; FVPTC, follicular variant papillary thyroid cancer; CPTC, classic variant papillary thyroid cancer; FTC, follicular thyroid cancer. EU-TIRADS classification was done based on Russ et al. (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Molecular, ultrasonographic and histological data of the 26 patients with at least one genetic alteration. NH, hyperplastic nodule; FA, follicular adenoma; NIFTP, noninvasive follicular thyroid neoplasms with papillary-like nuclear features; FVPTC, follicular variant papillary thyroid cancer; CPTC, classic variant papillary thyroid cancer; FTC, follicular thyroid cancer. EU-TIRADS classification was done based on Russ et al. (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Molecular, ultrasonographic and histological data of the 26 patients with at least one genetic alteration. NH, hyperplastic nodule; FA, follicular adenoma; NIFTP, noninvasive follicular thyroid neoplasms with papillary-like nuclear features; FVPTC, follicular variant papillary thyroid cancer; CPTC, classic variant papillary thyroid cancer; FTC, follicular thyroid cancer. EU-TIRADS classification was done based on Russ et al. (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
DNA and cDNA were analyzed using the custom PTC-MA assay, based on matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, and previously set up for the simultaneous identification of 13 known hotspot mutations (BRAFV600E; AKT1E17K; EIF1AX c.338-1G>C; NRASQ61Rand NRASQ61K; HRASG13C, HRASQ61K, and HRASQ61R; KRASG12V and KRASG13C; TERT c.-124C>T and TERT c.-146C>T; PIK3CAE542K and six recurrent fusion genes typical of PTC: RET/PTC1 (RET/CCDC6), RET/PTC2 (RET/PRKAR1A), and RET/PTC3 (RET/NCOA4) and TRK (NTRK1/TPM3), TRK-T1 (NTRK-T1/TPR), and TRK-T3 (NTRK1/TFG) (Pesenti et al. 2018, Colombo et al. 2019). New oligonucleotides were designed to implement the PTC-MA panel with PAX8/PPARγ t(2;3)(q13;p25) fusions. In particular, PAX8_8 (exon 8)/PPARγ (exon 1), PAX8_9 (exon 9)/PPARγ (exon 1) and PAX8_10 (exon 10)/PPARγ (exon 1) fusions were investigated (primers on request) (Placzkowski et al. 2008). Moreover, to further test the follicular content of the sample, the same primers used for semiquantitative RT-PCR (PAX8_3F and PAX8_4R) were added to the PTC_MA panel.
In cases negative for any mutations/fusions analyzed by PTC-MA assay, BRAFK601E mutation was investigated by direct sequencing of exon 15 (Fugazzola et al. 2004), and AKAP9/BRAF and LMO7/BRAF fusions were investigated on cDNA, as previously reported (Ciampi & Nikiforov 2005, He et al. 2018). All molecular alterations found in the FNA samples were confirmed in the resected tissues. Moreover, all the operated cases, regardless of the final histology and either mutated or not mutated, were submitted to the same genetic analyses. DNA and RNA were extracted from the resected tumor tissues, and molecular alterations were investigated as previously reported (Colombo et al. 2019).
The pre-surgical risk score including clinical, US and molecular data
The pre-surgical ‘risk score’ was obtained for each AUS/FLUS or FN/FSN nodule by the sum of the following parameters (Fig. 1):
-
US nodule size: we assigned a score of 1 to nodules with an US largest diameter ≤10 mm, a score of 1.5 for nodules between 10 and 20 mm, score of 2 for nodules between 20 and 40 mm and the highest score of 2.5 if the nodule was >40 mm.
-
EU-TIRADS classification: we assigned a score of 1 if the nodule belonged to EU-TIRADS 2 class (benign lesions with risk of malignancy (ROM) close to 0%), including cystic and spongiform nodules, a score of 1.5 for nodules belonging to EU-TIRADS 3 class (low-risk lesions with ROM of 2-4%), a score of 2 for EU-TIRADS 4 nodules (intermediate-risk with ROM of 6-17%), the highest score of 2.5 if the nodule belonged to EU-TIRADS 5 class (high-risk with ROM of 26-87%) or if cervical suspicious lymph node metastases were documented.
-
Cytological classes: a score of 1 was assigned if the nodule belonged to the cytological Bethesda system category ‘III’ (AUS/FLUS) or 2 if the nodule was of Bethesda ‘IV’ category (FN/FSN).
-
Molecular results: according to the prognostic genetic marker-based risk stratification (reviewed in Xing 2019), a score of 1 was given if the nodule was WT at molecular analysis through PTC-MA panel and a score of 1.5 if the molecular analyses revealed the presence of a ‘low-risk’ alteration, usually associated with a low-risk tumor or with a benign lesion (RET fusions, PAX8/PPARγ fusions, N-H-K RAS mutations). A score of 2 was assigned if the nodule harbored a ‘intermediate-risk’ molecular alteration, typical of thyroid cancer but also found in benign diseases (EIFA1X); the highest score of 2.5 for the presence of ‘high-risk’ molecular results usually associated with aggressive thyroid cancer (pTERT mutations, BRAF mutations, BRAF fusions, N-TRK fusions, AKT, PIK3CA, ≥ 2 genetic alterations).
To note that 3/138 samples were not adequate for the molecular analysis due to the insufficient follicular cell content and were thus excluded from the study (Supplementary Fig. 1).
Criteria for surgery
At the time of writing, surgery was performed in 65/135 (48%) patients, either by loboisthmectomy (n = 15) or by total thyroidectomy (n = 50). The indication to surgery was based on the following parameters:
-
Risk score ≥6 (n = 56).
-
Large goiters with compression independently from the risk score (n = 9).
-
Presence of at least one genetic mutation. It is worth noting that all mutated cases were ultimately found to harbor a risk score ≥6.
The remaining cases are still waiting for surgery (n = 20 for risk score ≥6, n = 14 for huge or compressive goiter), or are under ultrasonographic follow-up for risk score <6 (n = 28), or because they refuse surgery despite risk score ≥6 (n = 8).
Surgical specimens found to be malignant were classified according to the 2017 thyroid malignancy World Health Organization classification and the 8th edition of TNM staging (Amin et al. 2017).
Noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP), due to its indolent behavior, is no longer defined as a carcinoma. Nevertheless, limited data are available on long-term follow-up and, to date, this entity is usually placed between clearly benign and overtly malignant tumors (Seethala et al. 2018). Thus, the performance of the TRS score was calculated considering NIFTP as either a benign or a malignant lesion.
Statistical analysis
The specific risk score cut-off corresponding to the highest accuracy in the separation between benign and malignant thyroid nodules was assessed by receiver-operating characteristic (ROC) curves, and their sensitivity, specificity, positive and negative predictive values (PPV and NPV) were defined. In addition to the ROC curves, we used interactive dot diagram to study the accuracy of the studied diagnostic test. In this graph, a horizontal line indicates the cut-off point with the best separation (minimal false negative and false positive results) between the two groups. Statistical significance was defined as P < 0.05. All statistical analyses were performed using MedCalc Software version 11.6.1 for Windows.
Results
Molecular and ultrasonographic data of the indeterminate thyroid nodules with available histology
PTC-MA panel analyses revealed at least one genetic alteration in 26/135 nodules (19%), while 109 nodules (81%) were WT for the molecular alterations covered by this assay. In operated cases, there was a complete concordance between the results obtained, starting from FNA samples and from the corresponding tissues.
Among the mutated nodules which undergone surgery (Fig. 2 and Supplementary Table 1), eight had a RAS point mutation and were, at histology, follicular adenomas in five cases (corresponding EU-TIRADS 3), noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) in two cases (EU-TIRADS 3 and 4) and FVPTC in one case (EU-TIRADS 3). Seven additional cases harbored the BRAFV600E mutation, had a variable EU-TIRADS score (EU-TIRADS 3 in one case, EU-TIRADS 4 in three cases and EU-TIRADS 5 in three cases), and were all PTCs classical variant at histology. Three cases had a TERTG228A mutation, with EU-TIRADS 3 in one case and EU-TIRADS 4 in the remaining two cases; they were all FTCs at histology. One case had an EIF1AX mutation, with an EU-TIRADS 3 score and was a hyperplastic nodule at histology. Fusions were found in two patients: one nodule was PAX8/PPARγ positive, was EU-TIRADS 3 and was NIFTP at histology, whereas the other case had a ret/PTC3 rearrangement, was EU-TIRADS 5 and was a PTC. Finally, a double genetic alteration (NRASQ61R and TERTG228A) was found in 1 nodule, which was classified as EU-TIRADS 3 and resulted at histology to be a follicular adenoma with solid pattern and oxyphilous cell component (Supplementary Fig. 3).
Ultrasonographic and histological data of the 109 patients without genetic alterations at the PTC-MA assay. NH, hyperplastic nodule; FA, follicular adenoma; NIFTP, noninvasive follicular thyroid neoplasms with papillary-like nuclear features; CPTC, classic variant papillary thyroid cancer; SPTC, sclerosing variant papillary thyroid cancer; FVPTC, follicular variant papillary thyroid cancer; FTC, follicular thyroid cancer. EU-TIRADS classification was done based on Russ et al. (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Ultrasonographic and histological data of the 109 patients without genetic alterations at the PTC-MA assay. NH, hyperplastic nodule; FA, follicular adenoma; NIFTP, noninvasive follicular thyroid neoplasms with papillary-like nuclear features; CPTC, classic variant papillary thyroid cancer; SPTC, sclerosing variant papillary thyroid cancer; FVPTC, follicular variant papillary thyroid cancer; FTC, follicular thyroid cancer. EU-TIRADS classification was done based on Russ et al. (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Ultrasonographic and histological data of the 109 patients without genetic alterations at the PTC-MA assay. NH, hyperplastic nodule; FA, follicular adenoma; NIFTP, noninvasive follicular thyroid neoplasms with papillary-like nuclear features; CPTC, classic variant papillary thyroid cancer; SPTC, sclerosing variant papillary thyroid cancer; FVPTC, follicular variant papillary thyroid cancer; FTC, follicular thyroid cancer. EU-TIRADS classification was done based on Russ et al. (2017).
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Among the non-mutated cases which undergone surgery, 33/43 (77%) had a benign thyroid disease at histological examination (hyperplastic nodules or follicular adenomas), the majority of them being EU-TIRADS 3 at ultrasound (Fig. 3). Two additional cases (EU-TIRADS 3) were NIFTP, two (EU-TIRADS 4 and 5) were classical variant PTCs, one(EU-TIRADS 5) was a PTC sclerosing variant, three (EU-TIRADS 3 or 4) were follicular variant PTCs and the remaining two cases (EU-TIRADS 4) were FTCs.
According to these findings, the risk of malignancy (ROM) was of 100% for BRAFV600E mutation, NRASQ61K and RET/PTC3 fusion, and of 75% for TERTG228A mutation. On the other hand, NRASQ61R, HRASQ61K/R, PAX8/PPARγ fusion, and EIF1AX mutations always corresponded to non-malignant cases or to NIFTPs. Still, our PTC-MA panel showed the following parameters, when applied to indeterminate Bethesda III and IV classes: (a) for NIFTPs considered as malignant: sensitivity 63%, specificity 83%, PPV 0.68, NPV 0.79 and accuracy of 75%; (b) for NIFTPs considered as benign: sensitivity 63%, specificity 78%, PPV 0.54, NPV 0.83 and accuracy of 74% (Table 1).
Performance of nodule size, ultrasound, cytological, molecular evaluations or their association (thyroid risk score, TRS) for the identification of malignancy in 65 indeterminate thyroid nodules. Values in square brackets indicate data considering NIFTPs as benign.
| Nodule size (≤ 20 or >20 mm) | Ultrasound evaluation (EU-TIRADS 2–3 vs 4–5) | Cytology result (III vs IV) | Molecular analysis (WT vs mutated) | TRS (≤ 6 vs > 6) | TRS (≤ 6.5 vs > 6.5) | |
|---|---|---|---|---|---|---|
| Sensitivity | 60% [55%] | 62% [71%] | 72% [80%] | 63% [63%] | 84% [85%] | 72% [85%] |
| Specificity | 53% [59%] | 83% [82%] | 33% [36%] | 83% [78%] | 71% [64%] | 80% [80%] |
| PPV | 0.44 [0.32] | 0.69 [0.65] | 0.40 [0.36] | 0.68 [0.54] | 0.63 [0.51] | 0.69 [0.65] |
| NPV | 0.68 [0.71] | 0.77 [0.86] | 0.65 [0.80] | 0.79 [0.83] | 0.88 [0.90] | 0.82 [0.92] |
| Accuracy | 55% [51%] | 74% [78%] | 48% [49%] | 75% [74%] | 76% [70%] | 77% [82%] |
| LR+ | 1.263 [1.076] | 3.516 [4.017] | 1.066 [1.241] | 3.660 [2.905] | 2.940 [2.390] | 3.600 [4.250] |
| LR− | 0.761 [0.921] | 0.466 [0.347] | 0.861 [0.562] | 0.452 [0.470] | 0.224 [0.232] | 0.350 [0.187] |
LR−, negative likelihood ratio; LR+, positive likelihood ratio; NPV, negative predictive value; PPV, positive predictive value.
Risk of malignancy evaluation of the indeterminate thyroid nodules with available histology according to the risk score
The clinical, ultrasonographic, cytological and molecular data, the pre-surgical risk score and the histological diagnosis of the 65 patients submitted to surgery are reported in Table 2. In particular, 32/65 nodules had a score >4 and ≤ 6, 30/65 nodules had a score >6 and ≤8 and 3/65 nodules had a score >8. As expected, the ROM increased paralleling the score: in the category >4 and ≤ 6 (low suspicion) ROM was 10% for NIFTP benign and 12% for NIFTP malignant, in the category >6 ≤ 8, (intermediate suspicion) it was 47% for NIFTP benign and 60% for NITP malignant, and in the category >8 (high suspicion) it was 100% (Fig. 4).
Risk of malignancy (ROM) according to the thyroid risk score (TRS) and histological reports (NIFTP considered as malignant lesion). Values in square brackets report data considering NIFTPs as benign.
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Risk of malignancy (ROM) according to the thyroid risk score (TRS) and histological reports (NIFTP considered as malignant lesion). Values in square brackets report data considering NIFTPs as benign.
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Risk of malignancy (ROM) according to the thyroid risk score (TRS) and histological reports (NIFTP considered as malignant lesion). Values in square brackets report data considering NIFTPs as benign.
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Clinical, molecular and histological data, and risk score of the 65 patients with indeterminate cytology submitted to surgery. Molecular data were obtained by means of the PTC-MA assay (Nikiforova et al. 2018, Colombo et al. 2019).
| Patient ID | Gender/age at D (years) | Nodule size (mm) | EU-TIRADS | Cytology (Bethesda) | Molecular data | Score | Histological results |
|---|---|---|---|---|---|---|---|
| #1 | F/62 | 80 | 2 | III | WT | 5.5 | FA |
| #2 | F/69 | 9 | 2 | IV | WT | 5 | FA |
| #3 | M/55 | 50 | 3 | IV | TERTG228A | 8.5 | FTC pT2NX |
| #4 | F/19 | 40 | 5 | IV | BRAFV600E | 9.5 | CPTC pT3amN1a |
| #5 | F/40 | 10 | 4 | IV | BRAFV600E | 7.5 | CPTC pT1aNX |
| #6 | M/30 | 58 | 3 | III | PAX8/PPARγ | 6.5 | NIFTP |
| #7 | F/77 | 13 | 4 | III | TERTG228A | 6.5 | FTC pT1bNX |
| #8 | F/48 | 33 | 3 | IV | NRASQ61K | 7 | FVPTC pT2NX |
| #9 | M/36 | 8 | 4 | IV | BRAFV600E | 8 | CPTC pT1amNX |
| #10 | M/67 | 15 | 4 | IV | BRAFV600E | 8 | CPTC pT1bNX |
| #11 | M/39 | 20 | 5 | III | BRAFV600E | 7.5 | CPTC pT1bNX |
| #12 | F/64 | 22 | 3 | IV | BRAFV600E | 8 | CPTC pT1bNX |
| #13 | M/62 | 10 | 5 | IV | BRAFV600E | 8.5 | CPTC pT1aNX |
| #14 | M/68 | 50 | 4 | IV | TERTG228A | 9 | FTC pT3mNX |
| #15 | M/59 | 36 | 4 | IV | WT | 7 | FA |
| #16 | M/44 | 36 | 4 | IV | WT | 7 | FA |
| #17 | M/50 | 28 | 3 | III | WT | 6 | NH |
| #18 | M/55 | 12 | 3 | IV | WT | 5 | FA |
| #19 | M/64 | 12 | 3 | IV | WT | 6 | FA |
| #20 | F/42 | 22 | 3 | IV | WT | 6.5 | FA |
| #21 | F/73 | 40 | 3 | III | WT | 6 | FA |
| #22 | F/40 | 22 | 3 | IV | WT | 6.5 | FA |
| #23 | F/50 | 4 | 4 | IV | WT | 6 | NH |
| #24 | F/52 | 28 | 4 | IV | WT | 7 | FTC pT2NX |
| #25 | M/54 | 20 | 4 | IV | WT | 7 | FTC pT1bNX |
| #26 | M/72 | 38 | 3 | III | WT | 6 | FA |
| #27 | F/52 | 7 | 3 | III | WT | 5 | NH |
| #28 | F/41 | 23 | 5 | IV | WT | 7.5 | SPTCpT3bN1b* |
| #29 | F/52 | 20 | 3 | IV | WT | 6 | FA |
| #30 | F/40 | 20 | 3 | III | WT | 5 | NH |
| #31 | F/62 | 10 | 3 | IV | WT | 4.5 | NH |
| #32 | F/31 | 35 | 4 | IV | HRASQ61K | 7.5 | NIFTP |
| #33 | M/49 | 20 | 3 | IV | WT | 6 | FA |
| #34 | M/36 | 11 | 3 | IV | WT | 6 | FA |
| #35 | F/72 | 30 | 3 | IV | WT | 6.5 | FA |
| #36 | F/69 | 45 | 3 | III | WT | 6 | FA |
| #37 | F/72 | 14 | 3 | IV | WT | 6 | FA |
| #38 | F/69 | 24 | 4 | III | WT | 6 | CPTC pT1amNX |
| #39 | F/76 | 60 | 3 | IV | HRASQ61R | 7.5 | NH |
| #40 | F/76 | 5 | 4 | IV | WT | 6 | FVPCT pT1aNX |
| #41 | F/62 | 38 | 3 | III | WT | 5.5 | FVPTC pT2NX |
| #42 | F/66 | 26 | 3 | IV | WT | 5.5 | NH |
| #43 | M/40 | 11 | 3 | IV | NRASQ61R | 6 | FA |
| #44 | F/66 | 18 | 3 | III | WT | 5 | FA |
| #45 | F/36 | 50 | 3 | IV | HRASQ61R | 7.5 | FA |
| #46 | F/65 | 38 | 3 | IV | HRASQ61R | 7 | NH |
| #47 | F/89 | 10 | 3 | IV | WT | 5.5 | FA |
| #48 | F/40 | 31 | 4 | IV | WT | 7 | FVPTC pT2NX |
| #49 | F/70 | 15 | 3 | IV | WT | 6.5 | NIFTP |
| #50 | M/71 | 50 | 4 | IV | WT | 7.5 | FA |
| #51 | F/70 | 13 | 3 | IV | WT | 6 | NH |
| #52 | F/77 | 16 | 3 | IV | TERTG228A/NRASQ61R | 7 | FA |
| #53 | F/59 | 20 | 3 | IV | WT | 6 | FA |
| #54 | F/58 | 14 | 3 | IV | WT | 6 | FA |
| #55 | F/44 | 20 | 3 | III | WT | 5 | FA |
| #56 | M/41 | 30 | 4 | III | WT | 6 | FA |
| #57 | F/63 | 16 | 5 | III | WT | 6 | FA |
| #58 | F/30 | 55 | 3 | III | NRASQ61R | 6.5 | FA |
| #59 | M/49 | 52 | 3 | III | WT | 6 | NIFTP |
| #60 | F/52 | 4 | 3 | IV | WT | 5.5 | FA |
| #61 | F/25 | 21 | 5 | IV | RET/PTC3 | 8 | CPTC pT3amN1b |
| #62 | F/51 | 50 | 4 | IV | WT | 7.5 | FA |
| #63 | F/32 | 14 | 5 | IV | WT | 7 | CPTC pT1aNX |
| #64 | F/52 | 44 | 3 | III | HRASQ61K | 6.5 | NIFTP |
| #65 | F/52 | 20 | 3 | III | EIF1AX | 6 | NH |
EU-TIRADS classification done according to Russ et al. (2017).
*No previous neck external irradiation associated thyroiditis.
CPTC, conventional variant papillary thyroid carcinoma; D, diagnosis; FA, follicular adenoma; FCT, follicular thyroid carcinoma; FVPTC, follicular variant PTC; NH, nodular hyperplasia; NIFTP, noninvasive follicular thyroid neoplasm with papillary-like nuclear features; SPTC, diffuse sclerosing variant.
ROC curves for the identification of the best cut-off for the pre-surgical risk score and comparison with the performance of the single diagnostic tools
ROC plot analyses were used to find out the score threshold that able to pre-surgically differentiate between malignant (including papillary and follicular thyroid cancers) and benign thyroid disease (including hyperplastic nodules, follicular adenomas). Analyses were done considering NIFTP as either malignant or benign. The best threshold was >6.5 (AUC = 0.815, P < 0.001 for NIFTP malignant and AUC = 0.839, P < 0.0001 for NIFTP benign). Interactive dot Diagram confirmed that the score threshold > 6.5 showed a good reliability in the identification of benign and malignant nodules (SE = 72%, SP = 80%, positive likelihood ratio 4.2 for NIFTP malignant, and SE = 85%, SP = 80% for NIFTP benign) (Fig. 5). This threshold showed a better performance in the differential diagnosis between benign and malignant nodules than either the various parameters included in the risk score (nodule size, EU-TIRADS score, cytology, molecular data), or the cut-off arbitrarily identified at the beginning of this prospective study (≥6) (Table 2).
(A) ROC analysis to find the best threshold score to separate between benign and malignant nodules (including NIFTP); (B) Interactive dot Diagram: data of the benign and malignant nodules (including NIFTP) are displayed as dots on two vertical axes. Values in square brackets indicate data considering NIFTPs as benign. A horizontal line indicates the cut-off point with the best separation (minimal false negative and false positive results) between the two groups.
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
(A) ROC analysis to find the best threshold score to separate between benign and malignant nodules (including NIFTP); (B) Interactive dot Diagram: data of the benign and malignant nodules (including NIFTP) are displayed as dots on two vertical axes. Values in square brackets indicate data considering NIFTPs as benign. A horizontal line indicates the cut-off point with the best separation (minimal false negative and false positive results) between the two groups.
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
(A) ROC analysis to find the best threshold score to separate between benign and malignant nodules (including NIFTP); (B) Interactive dot Diagram: data of the benign and malignant nodules (including NIFTP) are displayed as dots on two vertical axes. Values in square brackets indicate data considering NIFTPs as benign. A horizontal line indicates the cut-off point with the best separation (minimal false negative and false positive results) between the two groups.
Citation: Endocrine-Related Cancer 28, 4; 10.1530/ERC-20-0511
Discussion
Thyroid nodules with indeterminate cytology are one of the most important challenges for the endocrinologist. The definite diagnosis of these nodules requires a histological evaluation after surgery, which represents an overtreatment and a considerable waste of resources for the national healthcare systems in most benign cases (60–80% of total indeterminate nodules). Ultrasound risk evaluation and molecular testing are useful tools to guide the clinical management of thyroid nodules with indeterminate cytology, though limited in the specificity and the costs, respectively. These parameters have been almost always considered as separate tools for the pre-surgical differential diagnosis of indeterminate nodules. To the best of our knowledge, this is the first report on a combined score including different parameters, previously validated as diagnostic tools, generated to increase the preoperative accuracy. In particular, we included in our ‘thyroid risk score’ (TRS), the EU-TIRADS scoring and the Bethesda classification, characterized by a high sensitivity (Bongiovanni et al. 2012, Kim et al. 2020). On the other hand, the PTC-MA platform was shown to be fully reliable, starting not only from tissues (Pesenti et al. 2018, Colombo et al. 2019) but also from FNA samples, and the molecular characterization of the nodules that was done by means of this cost-effective custom assay, showed a high specificity. Since the three former parameters have a lower accuracy on FTC cases, which often do not harbor specific suspicious features at ultrasound and for which a lower number of genetic alterations are known, we included the nodule size parameter, too, which has been found to correlate with the risk of malignancy mainly in this histotype (Kamran et al. 2013, Mehta et al. 2013, Nakamura et al. 2015). By the sum of the previously mentioned parameters, low-, intermediate-and high-risk TRS categories were defined (>4 and ≤ 6, >6 ≤ 8, and >8, respectively) for this rule-in test. We prospectively evaluated by the TRS 135 nodules, and 65 of them underwent surgery and histological evaluation. A progressively higher ROM was found according to the increase in the score, and ROC curves allowed to select the score >6.5 as the best threshold to pre-surgically differentiate between malignant and benign thyroid diseases, and did not change upon the inclusion of NIFTPs among benign or malignant lesions.
As already reported for 7- and 5-genes panels, the PTC-MA assay showed a lower sensitivity than Thyroseq and GEC but a higher specificity (Muzza et al. 2020). BRAFV600E mutation and RET/PTC3 fusion were always associated with PTC, whereas the other genetic alterations were found in either malignant or benign or NIFTP cases. Interestingly, TERTG228A mutation, mostly known as a feature of aggressive cancer behavior (Muzza et al. 2015), was found in combination with the NRASQ61R alteration in a follicular adenoma with solid pattern and oxyphilous cell component, consistent with literature data (Wang et al. 2014, Topf et al. 2018) reporting the presence of TERT mutations in benign thyroid disease, too.
On the other hand, the Bethesda classification was highly sensitive but poorly specific, while the EU-TIRADS classification in our expert hands was highly accurate. The EU-TIRADS classification and the molecular analysis showed a lower sensitivity but similar specificity and accuracy than the TRS score. Nevertheless, the combination of different parameters allowed to reduce the number of false negatives inherent in each classification system, and the TRS showed a higher performance if compared with cytological, ultrasound or molecular evaluations taken separately. Thus, we support the importance to fully evaluate all the features of indeterminate nodules, especially considering that ultrasound is included in the standard diagnostic workup and does not imply additional time and costs.
In the present series, the Bethesda III and IV categories rates were 2.8 and 3.3 % respectively, fairly lower if compared to some published studies but comparable to another very large Italian series (Rago et al. 2014). Indeed, the prevalence of cytologically indeterminate thyroid nodules is highly variable among the series reported in literature (reviewed in Bongiovanni et al. 2012) probably due to the different ethnicities, the different selection of the nodules to be submitted to cytologic examination, and the possible inter-individual variability in the pathologic evaluation. Still, the TRS score is predicted to be equally reliable in series with higher prevalence of indeterminate nodules.
A proper role of the TRS can also be predicted in the evaluation of large nodules, routed to surgery in most cases. Indeed, in the era of mini-invasive procedures, such as thermal ablation (Papini et al. 2020), the genetic definition of a nodule, regardless of its size, is crucial. A genetically ‘negative’ nodule, especially in older patients and in cases with co-morbidities, could be successfully treated by these techniques.
The use of a custom PTC-MA assay which is not commercially available is a limitation of this study. Nevertheless, though 14 point mutations and 11 fusions were tested , only seven genetic alterations were found (in the HRAS, NRAS, BRAF, TERT, EIF1AX, PAX8, RET genes), which are likely to be the more frequently represented in the indeterminate samples. Thus, we speculate that the exclusive evaluation of these genes, which can be easily and cost-effectively set up in many laboratories, could not significantly decrease the accuracy of the scoring system.
Another drawback of this study is the number of follicular cancers included, which is likely too low to definitely confirm the reliability of TRS for this histotype. Nevertheless, the rate of FTCs included (8%) is even higher than that reported in large validation case series (Nikiforova et al. 2018, Patel et al. 2018, Steward et al. 2019). Still, our five FTCs all had a TR score ≥6.5, thus demonstrating a good performance in the preoperative diagnosis even for this histotype.
The next step of our prospective study will include the validation of our TRS in a new larger series. Moreover, we plan to implement the molecular panel by the inclusion of less frequent molecular alterations in order to improve specificity, still maintaining costs in a range suitable for a wide use.
In conclusion, we generated a new rule-in test, the TRS, which combines a cost-effective platform for the genetic characterization of indeterminate nodules with other diagnostic parameters, such as cytology, ultrasonography and nodule size evaluation, in order to increase the accuracy in the differential diagnosis of indeterminate nodules. The identification of a TRS > 6.5, highly suggestive for malignancy, would allow a more reliable identification of patients for whom a surgical treatment is definitely needed.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/ERC-20-0511.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
Partially funded by Ricerca Corrente Istituto Auxologico Italiano IRCCS (PTC-array, 05C825_2018) and by the Italian Ministry of University and Research (PRIN 2017 – 2017YTWKWH)
References
Alexander EK, Kennedy GC, Baloch ZW, Cibas ES, Chudova D, Diggans J, Friedman L, Kloos RT, LiVolsi VA, Mandel SJet al. 2012 Preoperative diagnosis of benign thyroid nodules with indeterminate cytology. New England Journal of Medicine 367 705–715. (https://doi.org/10.1056/NEJMoa1203208)
Amin MB, Edge S, Greene F, Byrd DR, Brookland RK, Washington MK, Gershenwald JE, Compton CC, Hess KR & Sullivan DC et al.2017 AJCC Cancer Staging Manual, 8th ed. New York: Springer.
Bongiovanni M, Spitale A, Faquin WC, Mazzucchelli L & Baloch ZW 2012 The Bethesda system for reporting thyroid cytopathology: a meta-analysis. Acta Cytologica 56 333–339. (https://doi.org/10.1159/000339959)
Ciampi R & Nikiforov YE 2005 Alterations of the BRAF gene in thyroid tumors. Endocrine Pathology 16 163–172. (https://doi.org/10.1385/ep:16:3:163)
Cibas ES & Ali SZ 2017 The 2017 Bethesda system for reporting thyroid cytopathology. Thyroid 27 1341–1346. (https://doi.org/10.1089/thy.2017.0500)
Colombo C, Muzza M, Proverbio MC, Tosi D, Soranna D, Pesenti C, Rossi S, Cirello V, De Leo S & Fusco N et al.2019 Impact of mutation density and heterogeneity on papillary thyroid cancer clinical features and remission probability. Thyroid 29 237–251. (https://doi.org/10.1089/thy.2018.0339)
Cozzolino A, Pozza C, Pofi R, Sbardella E, Faggiano A, Isidori AM, Giannetta E, Pernazza A, Rullo E & Ascoli V et al.2020 Predictors of malignancy in high-risk indeterminate (TIR3B) cytopathology thyroid nodules. Journal of Endocrinological Investigation 43 1115–1123. (https://doi.org/10.1007/s40618-020-01200-0)
Duick DS, Klopper JP, Diggans JC, Friedman L, Kennedy GC, Lanman RB & McIver B 2012 The impact of benign gene expression classifier test results on the endocrinologist-patient decision to operate on patients with thyroid nodules with indeterminate fine-needle aspiration cytopathology. Thyroid 22 996–1001. (https://doi.org/10.1089/thy.2012.0180)
Fugazzola L, Mannavola D, Cirello V, Vannucchi G, Muzza M, Vicentini L & Beck-Peccoz P 2004 BRAF mutations in an Italian cohort of thyroid cancers. Clinical Endocrinology 61 239–243. (https://doi.org/10.1111/j.1365-2265.2004.02089.x)
Gharib H, Papini E, Garber JR, Duick DS, Harrell RM, Hegedüs L, Paschke R, Valcavi R, Vitti PAACE/ACE/AME Task Force on Thyroid Nodules 2016 American Association of Clinical Endocrinologists, American College of Endocrinology, and Associazione Medici Endocrinology Medical Guidelines for clinical practice for the diagnosis and management of thyroid nodules-2016 Update. Endocrine Practice 22 622–639. (https://doi.org/10.4158/EP161208.GL)
Haugen BR, Alexander EK, Bible KC, Doherty GM, Mandel SJ, Nikiforov YE, Pacini F, Randolph GW, Sawka AM & Schlumberger M et al.2016 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 26 1–133. (https://doi.org/10.1089/thy.2015.0020)
He H, Li W, Yan P, Bundschuh R, Killian JA, Labanowska J, Brock P, Shen R, Heerema NA & de la Chapelle A 2018 Identification of a recurrent LMO7-BRAF fusion in papillary thyroid carcinoma. Thyroid 28 748–754. (https://doi.org/10.1089/thy.2017.0258)
Ho AS, Sarti EE, Jain KS, Wang H, Nixon IJ, Shaha AR, Shah JP, Kraus DH, Ghossein R & Fish SA et al.2014 Malignancy rate in thyroid nodules classified as Bethesda category III (AUS/FLUS). Thyroid 24 832–839. (https://doi.org/10.1089/thy.2013.0317)
Ianni F, Campanella P, Rota CA, Prete A, Castellino L, Pontecorvi A & Corsello SM 2016 A meta-analysis-derived proposal for a clinical, ultrasonographic, and cytological scoring system to evaluate thyroid nodules: the ‘CUT’ score. Endocrine 52 313–321. (https://doi.org/10.1007/s12020-015-0785-5)
Kamran SC, Marqusee E, Kim MI, Frates MC, Ritner J, Peters H, Benson CB, Doubilet PM, Cibas ES & Barletta J et al.2013 Thyroid nodule size and prediction of cancer. Journal of Clinical Endocrinology and Metabolism 98 564–570. (https://doi.org/10.1210/jc.2012-2968)
Kim PH, Suh CH, Baek JH, Chung SR, Choi YJ & Lee JH 2020 Diagnostic performance of four ultrasound risk stratification systems: a systematic review and meta-analysis. Thyroid 30 1159–1168. (https://doi.org/10.1089/thy.2019.0812)
Li H, Robinson KA, Anton B, Saldanha IJ & Ladenson PW 2011 Cost-effectiveness of a novel molecular test for cytologically indeterminate thyroid nodules. Journal of Clinical Endocrinology and Metabolism 96 E1719–E1726. (https://doi.org/10.1210/jc.2011-0459)
McIver B, Castro MR, Morris JC, Bernet V, Smallridge R, Henry M, Kosok L & Reddi H 2014 An independent study of a gene expression classifier (Afirma) in the evaluation of cytologically indeterminate thyroid nodules. Journal of Clinical Endocrinology and Metabolism 99 4069–4077. (https://doi.org/10.1210/jc.2013-3584)
Mehta RS, Carty SE, Ohori NP, Hodak SP, Coyne C, LeBeau SO, Tublin ME, Stang MT, Johnson JT & McCoy KL et al.2013 Nodule size is an independent predictor of malignancy in mutation-negative nodules with follicular lesion of undetermined significance cytology. Surgery 154 730–736; discussion 736. (https://doi.org/10.1016/j.surg.2013.05.015)
Muzza M, Colombo C, Rossi S, Tosi D, Cirello V, Perrino M, De Leo S, Magnani E, Pignatti E & Vigo B et al.2015 Telomerase in differentiated thyroid cancer: promoter mutations, expression and localization. Molecular and Cellular Endocrinology 399 288–295. (https://doi.org/10.1016/j.mce.2014.10.019)
Muzza M, Colombo C, Pogliaghi G, Karapanou O & Fugazzola L 2020 Molecular markers for the classification of cytologically indeterminate thyroid nodules. Journal of Endocrinological Investigation 43 703–716. (https://doi.org/10.1007/s40618-019-01164-w)
Nakamura H, Hirokawa M, Ota H, Kihara M, Miya A & Miyauchi A 2015 Is an increase in thyroid nodule volume a risk factor for malignancy? Thyroid 25 804–811. (https://doi.org/10.1089/thy.2014.0567)
Nardi F, Basolo F, Crescenzi A, Fadda G, Frasoldati A, Orlandi F, Palombini L, Papini E, Zini M & Pontecorvi A et al.2014 Italian consensus for the classification and reporting of thyroid cytology. Journal of Endocrinological Investigation 37 593–599. (https://doi.org/10.1007/s40618-014-0062-0)
Nicholson KJ, Roberts MS, McCoy KL, Carty SE & Yip L 2019 Molecular testing versus diagnostic lobectomy in Bethesda III/IV thyroid nodules: a cost-effectiveness analysis. Thyroid 29 1237–1243. (https://doi.org/10.1089/thy.2018.0779)
Nikiforov YE 2017 Role of molecular markers in thyroid nodule management: then and now. Endocrine Practice 23 979–988. (https://doi.org/10.4158/EP171805.RA)
Nikiforov YE, Carty SE, Chiosea SI, Coyne C, Duvvuri U, Ferris RL, Gooding WE, Hodak SP, LeBeau SO & Ohori NP et al.2014 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 120 3627–3634. (https://doi.org/10.1002/cncr.29038)
Nikiforova MN, Mercurio S, Wald AI, Barbi de Moura M, Callenberg K, Santana-Santos L, Gooding WE, Yip L, Ferris RL & Nikiforov YE 2018 Analytical performance of the ThyroSeq v3 genomic classifier for cancer diagnosis in thyroid nodules. Cancer 124 1682–1690. (https://doi.org/10.1002/cncr.31245)
Papini E, Monpeyssen H, Frasoldati A & Hegedüs L 2020 2020 European Thyroid Association clinical practice guideline for the use of image-guided ablation in benign thyroid nodules. European Thyroid Journal 9 172–185. (https://doi.org/10.1159/000508484)
Patel KN, Angell TE, Babiarz J, Barth NM, Blevins T, Duh QY, Ghossein RA, Harrell RM, Huang J & Kennedy GC et al.2018 Performance of a genomic sequencing classifier for the preoperative diagnosis of cytologically indeterminate thyroid nodules. JAMA Surgery 153 817–824. (https://doi.org/10.1001/jamasurg.2018.1153)
Pesenti C, Muzza M, Colombo C, Proverbio MC, Farè C, Ferrero S, Miozzo M, Fugazzola L & Tabano S 2018 MassARRAY-based simultaneous detection of hotspot somatic mutations and recurrent fusion genes in papillary thyroid carcinoma: the PTC-MA assay. Endocrine 61 36–41. (https://doi.org/10.1007/s12020-017-1483-2)
Placzkowski KA, Reddi HV, Grebe SK, Eberhardt NL & McIver B 2008 The role of the PAX8/PPARgamma fusion oncogene in thyroid cancer. PPAR Research 2008 672829. (https://doi.org/10.1155/2008/672829)
Rago T, Scutari M, Latrofa F, Loiacono V, Piaggi P, Marchetti I, Romani R, Basolo F, Miccoli P & Tonacchera M et al.2014 The large majority of 1520 patients with indeterminate thyroid nodule at cytology have a favorable outcome, and a clinical risk score has a high negative predictive value for a more cumbersome cancer disease. Journal of Clinical Endocrinology and Metabolism 99 3700–3707. (https://doi.org/10.1210/jc.2013-4401)
Russ G, Bonnema SJ, Erdogan MF, Durante C, Ngu R & Leenhardt L 2017 European Thyroid Association guidelines for ultrasound malignancy risk stratification of thyroid nodules in adults: the EU-TIRADS. European Thyroid Journal 6 225–237. (https://doi.org/10.1159/000478927)
Sciacchitano S, Lavra L, Ulivieri A, Magi F, De Francesco GP, Bellotti C, Salehi LB, Trovato M, Drago C & Bartolazzi A 2017 Comparative analysis of diagnostic performance, feasibility and cost of different test-methods for thyroid nodules with indeterminate cytology. Oncotarget 8 49421–49442. (https://doi.org/10.18632/oncotarget.17220)
Seethala RR, Baloch ZW, Barletta JA, Khanafshar E, Mete O, Sadow PM, LiVolsi VA, Nikiforov YE, Tallini G & Thompson LD 2018 Noninvasive follicular thyroid neoplasm with papillary-like nuclear features: a review for pathologists. Modern Pathology 31 39–55. (https://doi.org/10.1038/modpathol.2017.130)
Steward DL, Carty SE, Sippel RS, Yang SP, Sosa JA, Sipos JA, Figge JJ, Mandel S, Haugen BR & Burman KD et al.2019 Performance of a multigene genomic classifier in thyroid nodules with indeterminate cytology: a prospective blinded multicenter study. JAMA Oncology 5 204–212. (https://doi.org/10.1001/jamaoncol.2018.4616)
Topf MC, Wang ZX, Tuluc M & Pribitkin EA 2018 Tert, HRAS, and EIF1AX mutations in a patient with follicular adenoma. Thyroid 28 815–817. (https://doi.org/10.1089/thy.2017.0504)
Wang N, Liu T, Sofiadis A, Juhlin CC, Zedenius J, Höög A, Larsson C & Xu D 2014 Tert promoter mutation as an early genetic event activating telomerase in follicular thyroid adenoma (FTA) and atypical FTA. Cancer 120 2965–2979. (https://doi.org/10.1002/cncr.28800)
Xing M 2019 Genetic-guided risk assessment and management of thyroid cancer. Endocrinology and Metabolism Clinics of North America 48 109–124. (https://doi.org/10.1016/j.ecl.2018.11.007)
Yoon JH, Lee HS, Kim EK, Moon HJ, Park VY & Kwak JY 2020 Pattern-based vs. score-based guidelines using ultrasound features have different strengths in risk stratification of thyroid nodules. European Radiology 30 3793–3802. (https://doi.org/10.1007/s00330-020-06722-y)
