OL Oncology Letters 1792-1074 1792-1082 D.A. Spandidos 10.3892/ol.2024.14611 OL-28-4-14611 Articles Predicting central lymph node metastasis in papillary thyroid cancer: A nomogram based on clinical, ultrasound and contrast‑enhanced computed tomography characteristics ZhangQianru 1 * XuShangyan 1 * SongQi 2 MaYuanyuan 2 HuYan 1 YaoJiejie 1 ZhanWeiwei 1 Department of Ultrasound, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai 200025, P.R. China Department of Radiology, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai 200025, P.R. China Correspondence to: Professor Weiwei Zhan or Professor Jiejie Yao, Department of Ultrasound, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, 197 2nd Ruijin Road, Shanghai 200025, P.R. China, E-mail: jannyfulyao@126.com shanghairuijin@126.com

Contributed equally

 10 2024 05 08 2024 28 4 478 02022024 12072024 Copyright: © 2024 Zhang et al. 2024 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Central lymph node (CLN) status is considered to be an important risk factor in patients with papillary thyroid carcinoma (PTC). The aim of the present study was to identify risk factors associated with CLN metastasis (CLNM) for patients with PTC based on preoperative clinical, ultrasound (US) and contrast-enhanced computed tomography (CT) characteristics, and establish a prediction model for treatment plans. A total of 786 patients with a confirmed pathological diagnosis of PTC between January 2021 to December 2022 were included in the present retrospective study, with 550 patients included in the training group and 236 patients enrolled in the validation group (ratio of 7:3). Based on the preoperative clinical, US and contrast-enhanced CT features, univariate and multivariate logistic regression analyses were used to determine the independent predictive factors of CLNM, and a personalized nomogram was constructed. Calibration curve, receiver operating characteristic (ROC) curve and decision curve analyses were used to assess discrimination, calibration and clinical application of the prediction model. As a result, 38.9% (306/786) of patients with PTC and CLNM(−) status before surgery had confirmed CLNM using postoperative pathology. In multivariate analysis, a young age (≤45 years), the male sex, no presence of Hashimoto thyroiditis, isthmic location, microcalcification, inhomogeneous enhancement and capsule invasion were independent predictors of CLNM in patients with PTC. The nomogram integrating these 7 factors exhibited strong discrimination in both the training group [Area under the curve (AUC)=0.826] and the validation group (AUC=0.818). Furthermore, the area under the ROC curve for predicting CLNM based on clinical, US and contrast-enhanced CT features was higher than that without contrast-enhanced CT features (AUC=0.818 and AUC=0.712, respectively). In addition, the calibration curve was appropriately fitted and decision curve analysis confirmed the clinical utility of the nomogram. In conclusion, the present study developed a novel nomogram for preoperative prediction of CLNM, which could provide a basis for prophylactic central lymph node dissection in patients with PTC.

 CLNM PTC ultrasound contrast-enhanced CT prediction nomogram Funding: No funding was received. Introduction

Thyroid cancer, a malignancy ranking ninth worldwide in terms of incidence, can occur in people of any sex and any age (1). Papillary thyroid cancer (PTC) is the predominant type of thyroid cancer, making up ~80% of cases (2). Despite a more favorable overall prognosis compared with other forms of thyroid cancer (3), patients with PTC are more likely to have central lymph node metastasis (CLNM), and the incidence of CLNM is 30–80% (4,5), which is considered to be the most important risk factor of regional recurrence and poor survival (6). Therefore, to achieve the goal of radical tumor resection, surgeons will often perform therapeutic central lymph node dissection (CLND) in patients with PTC (7). However, it is still controversial as to whether CLN dissection should be performed in patients with PTC with clinically negative (cN0) CLNM. Certain cN0 patients have potential CLNM, therefore prophylactic CLND can lower the rate of postoperative regional recurrence rate and avoid a second surgery (8,9). However, prophylactic CLND is not particularly cost-effective, and the risk of recurrent laryngeal nerve injury, permanent hypoparathyroidism and other associated complications is greatly increased (10). Furthermore, ultrasound (US)-guided ablation is a safe, effective and minimally invasive substitute for surgical resection for patients with low-risk PTC without CLNM (11). Therefore, it is important to evaluate the central lymph nodes accurately and comprehensively before operation to avoid overtreatment and undertreatment, and to provide a more reasonable surgical plan for patients.

Given its non-invasiveness, non-radiation and high resolution, US is the preferred preoperative imaging technique for assessing thyroid nodules and cervical lymph nodes. It can clearly show the tumor size, location, shape, margin, composition, echogenicity, microcalcification and blood flow signal (12,13). However, owing to the influence of the anatomical structures of the central neck, the effect of US in detecting CLNM is not ideal (4,14). Meanwhile, US is also limited by the reliance on operator skills and the incapacity to visualize deep structures (15). Therefore, the American Thyroid Association guidelines recommend contrast-enhanced computed tomography (CT) as an adjunct to US to improve the accuracy of preoperative diagnosis (7). Contrast-enhanced CT effectively avoids the shortcomings of US. First, contrast-enhanced CT can provide comprehensive cross-sectional images of the thyroid gland and neighboring structures including the trachea, esophagus, blood vessels and lymph nodes (16,17). Second, due to the absence of gas and bone restrictions, contrast-enhanced CT may better visualize lymph node metastasis, capsule invasion and extrathyroidal extension (18–20).Therefore, the combination of US and contrast-enhanced CT diagnosis would be complementary, to make up for the deficiency of the single application of contrast-enhanced CT or US for diagnosing thyroid nodules, and improve the diagnostic specificity and sensitivity.

Currently, most studies predicting CLNM in patients with PTC have focused on clinical and US features, and the results are consistent (21–23). Furthermore, to the best of our knowledge, few studies have investigated the association between contrast-enhanced CT features and CLNM in patients with PTC. Therefore, the present study used contrast-enhanced CT features to determine the risk factors for CLNM in patients with cN0 PTC, aiming to identify key predictors and establish a new nomogram for predicting the risk of CLNM in patients with PTC to facilitate preoperative decision making for prophylactic CLND.

 Materials and methods <sec> <title>Patients selection

Owing to the retrospective study design, approval from the Ethics Committee of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China) was obtained and the requirement for informed consent was waived.

A total of 6,275 patients who underwent thyroidectomy along with CLND and were histopathologically confirmed to have PTC at Ruijin Hospital from January 2021 to December 2022 were enrolled in the present study. The inclusion criteria were as follows: i) Treatment with primary thyroid surgery and CLND, and a BRAF V600E mutation test; ii) histopathologically confirmed PTC; iii) no signs of lymph node metastasis (cN0), conventional US and contrast-enhanced CT performed, and a medical history collection within 3 weeks before surgery; iv) preoperative thyroid function tests performed and no prior history of thyroxine treatment [including thyroid stimulating hormone (TSH), thyroglobulin (TG), TG antibody (TGAb) and thyroid peroxidase antibody (TPOAb)]; and v) only the largest nodule was included for patients with multiple nodules (with at least two of which confirmed as PTC). Exclusion criteria were as follows: i) Tumor size of <0.5 cm; ii) treatment of head and neck radiotherapy therapy; iii) presence of other malignant tumors, such as nasopharyngeal carcinoma and breast cancer; iv) incomplete or low quality medical records; and v) inconsistent imaging tumor lesions with the pathological results. Based on the aforementioned inclusion and exclusion criteria, the data from 786 patients with cN0 PTC were analyzed in the present study. The flowchart depicting the selection process is presented in Fig. 1.

 US assessment

All participants were evaluated using US equipment (MyLab™ 9, Esaote S.p.A; DC-8, Shenzhen Mindray Bio-Medical Electronics Co., Ltd.; and iU22, Philips Medical Systems B.V.) with 5–13 MHz linear probe. The patient was positioned in the supine position with the neck fully exposed. A total of two radiologists possessing 15 years of experience in thyroid US imaging evaluated the following sonographic features in consensus: Tumor location, size, orientation, margin, internal composition, echogenicity, microcalcification and blood flow signal. Representative US features are presented in Fig. 2. Any disagreements between the two radiologists were resolved by a third radiologist with 25 years of experience in thyroid sonography.

 BRAF V600E mutation testing

BRAF V600E mutation testing was performed and the results were reviewed by experienced technicians in the clinical laboratory of Ruijin Hospital. Genomic DNA was extracted from the thyroid tissue samples using QIAamp® DNA Micro Kit (Qiagen, Inc.; cat. no. 56304) according to the instructions. The extracted DNA was subjected to PCR amplification (reagents: Ampli Taq Gold™ 360 Premix of Applied Biosystems; Thermo Fisher Scientific, Inc.; thermocycling conditions: 95°C for 3 min, 95°C 15 sec, 58°C 30 sec, 72°C 1 min, 72°C 7 min, 35 cycles in total) and Sanger sequencing, and the sequencing data were interpreted using the low-frequency mutation analysis software Minor Variant Finder of Applied Biosystems (version 1.1; Thermo Fisher Scientific, Inc.). The sequencing primers used for BRAF V600E mutation testing are provided in Fig. S1. Sequencing traces for Sanger sequencing are shown in Fig. S2.

 Contrast-enhanced CT assessment

All patients underwent scanning using multidetector CT scanners (GE Discovery CT750 HD 64 Slice CT Scanner, Cytiva; uCT 760; Shanghai United Imaging Healthcare Co., Ltd; and Philips Brilliance iCT 256, Philips Medical Systems B.V.) to collect CT data. All patients provided written informed consent and underwent iodine allergy testing before the examination. The slice thickness was 3.0 or 2.5 mm. Contrast-enhanced scans were performed at 45–65 sec after intravenous injection of non-ionic iodine contrast agent (2.5 m/s). The scanning range was scanned from C7 up to the base of the posterior fossa. The CT findings of the following nodules were evaluated by two radiologists with extensive experience in thyroid CT imaging: i) Mean CT values of the lesions in the plain phase (UCT) and the venous phase (VCT). A circular region of interest was drawn at the maximum diameter of the lesion, excluding calcification, cystic components and artifacts, with the goal of covering >80% of the whole lesion area. ΔCT=VCT-UCT was used to evaluate the absolute enhanced CT value. The average value of two radiologists was used for further analysis; ii) homogeneity of enhancement was divided into homogeneity and inhomogeneity (Fig. 3A); iii) calcification (Fig. 3B); iv) capsule invasion (Fig. 3C); and v) tracheal deviation (Fig. 3D).

 Variable definition and evaluation

Data for the following characteristics were collected to construct a retrospective database: i) Basic features: Age (45 years old as the cut point in accordance with the 7th Union for International Cancer Control/American Joint Committee on Cancer tumor-node-metastasis staging system) (24), sex (male/female), body mass index [BMI; 20.92 kg/m2 as the cut point according to receiver operating characteristic (ROC) curve analysis], Hashimoto thyroiditis (HT; yes/no), BRAF V600E mutation (yes/no), TSH (reference, 0.35–4.94 µIU/ml), TG (reference, 3.5–77 ng/ml), TGAb (reference, <4.11) IU/ml) and TPOAb (reference, <5.61 IU/ml); ii) conventional US features: Tumor location (isthmus/non-isthmus), tumor number (unifocal/multifocal), tumor size (papillary thyroid microcarcinoma ≤1.0 cm and PTC >1.0 cm), tumor orientation (taller-than-wide/wider-than-tall), tumor margin (regular/irregular), internal composition (solid/non-solid), echogenicity (markedly hypoechoic/hypoechoic/isoechoic), capsule contact (yes/no; defined as thyroid nodule touching the thyroid boundary with or without capsule uplift), microcalcification (yes/no) and blood flow signal (poor/rich)' and iii) contrast-enhanced CT features: UCT, VCT, ΔCT, homogeneity of enhancement (homogeneity/inhomogeneity; defined as the degree of homogeneity of enhancement within the thyroid nodule), calcification (yes/no), capsule invasion (yes/no; defined as the maximum diameter of the nodule was located at the junction of the nodule and thyroid gland or at the lateral side of the thyroid gland, known as ‘cookie bite sign’) and tracheal deviation (yes/no). ROC curve analysis revealed that the UCT value was 65.35 Hu, the VCT value was 183.90 Hu, and ΔCT was 111.50 Hu as the cut-off point of CLNM in the population of the present study (data not presented).

 Statistical analysis

Continuous data were transformed into categorical data using cut-off values established through ROC curve analysis for enhanced clinical comprehension. Data are presented as the frequency or mean ± standard deviation. UCT, VCT and ΔCT were analyzed using an independent samples t-test, and TSH and echogenicity were analyzed using Fisher's exact test. All other variables were analyzed using the χ2 test. Multivariate logistic regression analysis was used to determine independent factors. Based on the results of multivariate logistic regression analysis, a nomogram for predicting CLNM was developed and evaluated using ROC curves, calibration curves and decision curve analysis (DCA) curves. All statistical analyses were performed using SPSS version 27.0 (IBM Corp.) and R version 4.3.2 (The R Foundation) software. P<0.05 was considered to indicate a statistically significant difference.

 Results <sec> <title>Characteristics of patients

Patients were divided into the training group (n=550) and validation group (n=236). CLNM occurred in 39.1% (215/550) of the patients in the training group and 38.6% (91/236) of the patients in the validation group. In total, 38.9% of patients (306/786) had an CLNM(−) status before surgery, but had confirmed CLNM using postoperative pathology. As demonstrated in Table I, there was no significant difference between the two groups (P>0.05), which indicated their rationality as training and validation groups.

 Univariate analysis of CLNM

In the univariate analysis, CLNM was significantly associated with a younger age (≤45 years; P<0.001), the male sex (P<0.001), no HT (P=0.001), negative TGAb (P=0.004) and negative TPOAb (P=0.002). However, there were no significant differences for BMI, presence of BRAF V600E mutation, level of TSH or level of TG.

Among the US features, tumor location (isthmus; P<0.001), tumor size (>1.0 cm; P<0.001), presence of microcalcification (P<0.001) and capsule contact (yes; P<0.001) were significantly different between CLNM and non-CLNM groups. However, there was no significant difference for tumor number, tumor shape, tumor margin, internal composition, echogenicity or blood flow signal.

In terms of contrast-enhanced CT characteristics, there were significant differences for an inhomogeneous enhancement (P=0.002), presence of calcification (P=0.007) and capsule invasion (P<0.001), but there were no significant differences for UCT, VCT, ΔCT or tracheal deviation between the two groups (Table II).

 Multivariate logistic regression analysis of CLNM

The characteristics with statistical significance identified in the univariate analysis were further analyzed using multivariate logistic regression analysis. The results demonstrated that the following predictors were significantly independently associated with promoting CLNM in patients with PTC: Age of ≤45 years old [odds ratio (OR)=0.964; 95% confidence interval (CI), 0.945–0.982; P<0.001], male sex (OR=2.147; 95% CI, 1.332–3.459; P=0.002), no HT (OR=2.515; 95% CI, 1.208–5.239; P=0.014), isthmic tumor (OR=0.211; 95% CI, 0.067–0.669; P=0.008), presence of microcalcification (OR=0.589; 95% CI, 0.355–0.979; P=0.041), inhomogeneous enhancement (OR=2.711; 95% CI, 0.355–0.979; P=0.041). 95%CI 1.268–5.798, P=0.010) and capsule invasion (OR=6.463; 95% CI, 4.103–10.181; P<0.001; Table III).

 Development and validation of the individualized prediction nomogram

According to the results of multivariate logistic regression analysis, 7 variables including age, sex, presence of HT, tumor location, microcalcification, homogeneity of enhancement and capsule invasion were used in the development of a personalized prediction nomogram for predicting CLNM in patients with PTC (Fig. 4). According to the ROC curve, the area under the curve (AUC) was 0.826, the sensitivity was 0.824 and the specificity was 0.717 for the training group, whilst the AUC was 0.818, the sensitivity was 0.725 and the specificity was 0.781 for the validation group (Fig. 5A and B). In addition, the AUC for predicting CLNM without combined contrast-enhanced CT was 0.712, and the AUC of predicting CLNM increased to 0.818 when clinical and conventional US and contrast-enhanced CT features were combined (Fig. 6). This further demonstrates the advantage of the US combined CT model.

Furthermore, calibration curves depicting the CLNM risk nomogram in patients with PTC were generated to assess the effectiveness of the nomogram. The curves indicated a satisfactory agreement in both the training and validation groups, with mean absolute errors of 0.021 (Fig. 7A) and 0.023 (Fig. 7B), respectively.

 Clinical application

Finally, DCA was performed to evaluate the performance of the model in detecting CLNM for patients with PTC (Fig. 8). The DCA curve demonstrated that it would be beneficial to predict CLNM with the nomogram when the threshold probability ranges from 0.1 to 1.0.

 Discussion

Most PTCs show a slow and indolent growth pattern, and the overall prognosis is favorable, with a current 5-year survival rate of >90% (25). CLNM is occurs in 12–64% of patients with PTC, and exhibits a strong association with increased recurrence and poor overall survival (26). Therefore, precise preoperative prediction of CLNM can be advantageous for patients with PTC (cN0), and creating an effective prediction model would serve as a viable solution. Previous studies have reported that US features of PTC can help predict CLNM in patients, but few of them mentioned the role of CT in this (27–29). In the present study, the US and CT features of patients with PTC were reviewed and the value of US combined with contrast-enhanced CT for predicting CLNM was evaluated.

Many studies have reported that sex and age are independent risk factors for CLNM in PTC, among which the male sex and a younger age (≤45 years) have a greater risk of CLNM (30–32). This is consistent with the findings obtained in the present study. However, the multifocality and tumor size characteristics did not differ significantly between the two groups in the present study, which is not consistent with previous studies (22,31,33). The potential reasons contributing to this variation may be the different sample sizes and evaluation criteria used for characteristics.

HT is the most common autoimmune thyroid disease and 10–58% of patients with PTC have it (34). In most studies, HT has been regarded as a protective factor for CLNM in PTC (35,36), and Jara et al (37) also noted that HT correlated with less aggressive disease and a reduced incidence of lymph node metastasis. Moreover, the present study also demonstrated that patients with PTC but without coexistent HT were more prone to CLNM. However, Mao et al (38) and Liu et al (39) suggested that HT had no significant effect on the incidence of lymph node metastasis. Therefore, the effect of HT on CLNM in PTC is uncertain, and more clinical trials emphasizing the influence of HT on the progression of PTC are worth performing.

To the best of our knowledge, the relationship between PTC tumor location and lymph node metastasis is controversial. Certain studies reported that there is no significant association between tumor location and CLNM (22,40). However, Li et al (41) and Lyu et al (42) suggested that isthmic tumors are more prone to CLNM compared with lateral lobe tumors. The present study also demonstrated that tumor location in the isthmus was significantly associated with CLNM. The thyroid isthmus is typically situated anteriorly to the cartilaginous ring of the second to fourth trachea, where the gland thins to a thickness of only ~2 mm. Due to the specific location of the tumor, the isthmus tumor is adjacent to the trachea and thyroid capsule, so the incidence of extrathyroidal extension, CLNM and multifocality is higher than that of thyroid lobe tumor (43). In addition, the lymphatic drainage pattern of the thyroid isthmus differs from that of the thyroid lobe (41). The presence of the aforementioned features will increase the risk of CLNM in in isthmic tumors (42). Notably, although patients with isthmic tumors were demonstrated to have a higher risk of CLNM, the incidence of this feature was low, accounting for only 4.0% (22/550) of the total cases in the present study.

US is known to provide a better soft tissue resolution than CT, and microcalcifications seen by US are not necessarily shown on CT (44,45). Therefore, in the present study, microcalcifications were evaluated by US, whilst calcifications evaluated by CT were generally macrocalcifications. The present study confirmed that the presence of microcalcification was an independent predictor of CLNM in cN0 PTC and that macrocalcification was not significantly associated with CLNM. Previous studies have also reported an association between microcalcification and CLNM in PTC (46–48). Microcalcifications are characterized as punctate bright echoes with or without accompanying acoustic shadowing, mainly small psammoma bodies of 10–100 µm, arranged in concentric layers (49). Therefore, as microcalcification may be a predictive marker for CLNM (50), when microcalcification is found in thyroid nodules by preoperative examination, a more meticulous evaluation of the central cervical lymph nodes is warranted.

Angiogenesis is known to be associated with aggressive tumor growth and metastasis (51). Contrast-enhanced CT provides improved visualization of the tumor microvascular distribution (33). There were a large number of neovascularization in the thyroid tumor tissue, which appeared to be enhanced after enhancement. However, at the same time, this malignant growth will destroy a lot of tissue structures and blood vessels, so the degree of enhancement is lower than that of normal thyroid (52). Furthermore, heterogeneous vascular distribution can lead to inhomogeneous enhancement shown on contrast-enhanced CT images. The present study demonstrated that although the incidence of inhomogeneous enhancement was relatively high, it was also significantly associated with predicting CLNM.

Capsule invasion is generally regarded as being associated with CLNM. However, whether US or contrast-enhanced CT is superior in predicting capsule invasion is still controversial (19). Yang et al (23) reported that observation of the anterior thyroid capsule by US is influenced by US near-field artifacts, whilst observation of the lateral and posterior thyroid capsules is hindered by the presence of blood vessels and the trachea, which may not be distinctly depicted (8). In addition, considering the strong dependence of US on the operator (15), the present study included capsule contact on US images and capsule invasion on CT images, which were associated but not consistent. The present study demonstrated that capsule invasion assessed by CT is an independent risk factor for CLNM, and its mechanism may be linked to the abundant thyroid lymphatic network. If the tumor breaks through the capsule, it has the potential to readily induce lymph node metastasis in the central region (8,53).

Few studies have investigated the relationship between contrast-enhanced CT features and CLNM. For example, Peng et al (54) and Mou et al (55) collected data from preoperative CT images to predict CLNM in patients with cN0 PTC, but the studies only had small sample sizes. Moreover, Zhao et al (33) used a simple risk-scoring system to predict CLNM. To the best of our knowledge, the present study was the first with an adequate sample size to construct a nomogram combining US with contrast-enhanced CT for predicting CLNM.

Nonetheless, there were several limitations of the present study that should be acknowledged. First, the study design was retrospective, making it susceptible to inherent bias in patient recruitment and data collection. Second, the retrospective nature of the present study may limit the analysis of additional potential variables. Analyses were performed only on the basis of the characteristics of the primary tumor. Furthermore, the present study lacks external validation. Therefore, it is imperative to prioritize additional external validation cohorts from prospective studies to comprehensively assess the viability of the nomogram in the present study. In addition, for multifocal tumors, analysis was performed only on the largest tumor, and the features of the remaining tumors were unknown. Notably, the present study did not assess patients with PTCs that were <0.5 cm.

Despite the limitations, the present study presents certain highlights. Based on the aforementioned clinical, US and contrast-enhanced CT characteristics, the present study developed and validated a novel nomogram, which has an improved diagnostic performance in predicting CLNM than no combination of contrast-enhanced CT. Moreover, the nomogram serves as a user-friendly diagnostic tool for predicting CLNM. By adding the specific scores of each predictor, the corresponding CLNM probability for thyroid nodules can be obtained. Overall, this prediction model could make it possible to personalize the CLNM prediction of most patients with PTC and help surgeons make decisions on surgical options to maximize the benefits of patients.

In conclusion, the findings of the present study suggest that a young age, the male sex, no presence of HT, isthmic tumor, microcalcification, inhomogeneous enhancement and capsule invasion are significantly associated with CLNM in patients with cN0 PTC. Furthermore, the constructed nomogram has the potential to be used for preoperative risk assessment of CLNM, which can help surgeons better develop appropriate surgical plans, providing a novel approach to managing patients with cN0 PTC.

 Supplementary Material Supporting Data Acknowledgements

Not applicable.

 Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

 Authors' contributions

QZ, SX, JY and WZ conceived and designed the study. QZ, SX, QS and YM contributed to data collection and data analyses. QZ and SX, YH and YM performed the data interpretation. QZ, SX, YH and YM contributed to the statistical analysis. QZ, SX, YH and YM drafted the manuscript. QS, JY and WZ revised the manuscript critically for important intellectual content. QS, JY and WZ confirm the authenticity of all the raw data. QZ, SX, QS, YM, YH, JY and WZ discussed the results and contributed to the revision of the final manuscript. All authors read and approved the final version of the manuscript.

 Ethics approval and consent to participate

The present study was approved by the Ethics Committee of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China; approval no. 2023-129). The requirement for informed consent to participate was waived by the Ethics Committee as the present study is retrospective. All methods were performed in accordance with the Helsinki Declaration and local legislation and institutional requirements.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 Abbreviations PTC

papillary thyroid cancer

 CLNM

central lymph node metastasis

 CLND

central lymph node dissection

 US

ultrasound

 CT

computed tomography

 ROC

receiver operating characteristic

 DCA

decision curve analysis

 AUC

area under the curve

 TSH

thyroid stimulating hormone

 TG

thyroglobulin

 TGAb

TG antibody

 TPOAb

thyroid peroxidase antibody

 OR

odd ratio

 CI

confidence interval

 References SungHFerlayJSiegelRLLaversanneMSoerjomataramIJemalABrayFGlobal cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countriesCA Cancer J Clin71209249202110.3322/caac.2166033538338 ChenDWLangBHHMcLeodDSANewboldKHaymartMRThyroid cancerLancet40115311544202310.1016/S0140-6736(23)00020-X37023783 PapaleontiouMEvronJMEsfandiariNHReyes-GastelumDWardKCHamiltonASWordenFHaymartMRPatient report of recurrent and persistent thyroid cancerThyroid3012971305202010.1089/thy.2019.065232183609 DaiQLiuDTaoYDingCLiSZhaoCWangZTaoYTianJLengXNomograms based on preoperative multimodal ultrasound of papillary thyroid carcinoma for predicting central lymph node metastasisEur Radiol3245964608202210.1007/s00330-022-08565-135226156 Al AfifAWilliamsBARigbyMHBullockMJTaylorSMTritesJHartRDMultifocal papillary thyroid cancer increases the risk of central lymph node metastasisThyroid2510081012201510.1089/thy.2015.013026161997 HengYYangZLinJLiuQCaiWTaoLRisks of central lymph node metastasis in papillary thyroid carcinoma with or without multifocality in at least one lobe: A multi-center analysisOral Oncol134106185202210.1016/j.oraloncology.2022.10618536191477 HaugenBRAlexanderEKBibleKCDohertyGMMandelSJNikiforovYEPaciniFRandolphGWSawkaAMSchlumbergerM2015 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 cancerThyroid261133201610.1089/thy.2015.002026462967 GuangYHeWZhangWZhangHZhangYWanFClinical study of ultrasonographic risk factors for central lymph node metastasis of papillary thyroid carcinomaFront Endocrinol (Lausanne)12791970202110.3389/fendo.2021.79197034917039 MazzaferriELDohertyGMStewardDLThe pros and cons of prophylactic central compartment lymph node dissection for papillary thyroid carcinomaThyroid19683689200910.1089/thy.2009.157819583485 ShahaARCentral lymph node metastasis in papillary thyroid carcinomaWorld J Surg42630631201810.1007/s00268-017-4459-829305712 XueJTengDWangHEfficacy and safety of ultrasound-guided radiofrequency ablation for papillary thyroid microcarcinoma: A systematic review and meta-analysisInt J Hyperthermia3913001309202210.1080/02656736.2022.212910136195326 ZhouJYinLWeiXZhangSSongYLuoBLiJQianLCuiLChenW2020 Chinese guidelines for ultrasound malignancy risk stratification of thyroid nodules: The C-TIRADSEndocrine70256279202010.1007/s12020-020-02441-y32827126 AlexanderEKCibasESDiagnosis of thyroid nodulesLancet Diabetes Endocrinol10533539202210.1016/S2213-8587(22)00139-535752200 RohJLParkJYKimJMSongCJUse of preoperative ultrasonography as guidance for neck dissection in patients with papillary thyroid carcinomaJ Surg Oncol992831200910.1002/jso.2116418942072 HwangHSOrloffLAEfficacy of preoperative neck ultrasound in the detection of cervical lymph node metastasis from thyroid cancerLaryngoscope121487491201110.1002/lary.2122721344423 KimEParkJSSonKRKimJHJeonSJNaDGPreoperative diagnosis of cervical metastatic lymph nodes in papillary thyroid carcinoma: Comparison of ultrasound, computed tomography, and combined ultrasound with computed tomographyThyroid18411418200810.1089/thy.2007.026918358074 SuhCHBaekJHChoiYJLeeJHPerformance of CT in the preoperative diagnosis of cervical lymph node metastasis in patients with papillary thyroid cancer: A systematic review and meta-analysisAJNR Am J Neuroradiol38154161201710.3174/ajnr.A496727789450 TraylorKSComputed tomography and MR imaging of thyroid diseaseRadiol Clin North Am5810591070202010.1016/j.rcl.2020.07.00433040848 YooREKimJHHwangIKangKMYunTJChoiSHSohnCHParkSWAdded value of computed tomography to ultrasonography for assessing LN metastasis in preoperative patients with thyroid cancer: Node-by-node correlationCancers (Basel)121190202010.3390/cancers1205119032397148 JeonYHLeeJYYooRERhimJHLeeKHChoiKSHwangIKangKMKimJHValidation of ultrasound and computed tomography-based risk stratification system and biopsy criteria for cervical lymph nodes in preoperative patients with thyroid cancerKorean J Radiol24912923202310.3348/kjr.2023.021537634645 FengYMinYChenHXiangKWangXYinGConstruction and validation of a nomogram for predicting cervical lymph node metastasis in classic papillary thyroid carcinomaJ Endocrinol Invest4422032211202110.1007/s40618-021-01524-533586026 LiJSunPHuangTLiLHeSAiXXiaoHXueGPreoperative prediction of central lymph node metastasis in cN0T1/T2 papillary thyroid carcinoma: A nomogram based on clinical and ultrasound characteristicsEur J Surg Oncol4812721279202210.1016/j.ejso.2022.04.00135414404 YangZHengYLinJLuCYuDTaoLCaiWNomogram for predicting central lymph node metastasis in papillary thyroid cancer: A retrospective cohort study of two clinical centersCancer Res Treat5210101018202032599980 EdgeSBByrdDRComptonCCFritzAGGreeneFLTrottiAAJCC Cancer staging handbook: from the AJCC cancer staging manual7th editionSpringer-VerlagNew York, NY2010 Coca-PelazAShahJPHernandez-PreraJCGhosseinRARodrigoJPHartlDMOlsenKDShahaARZafereoMSuarezCPapillary thyroid cancer-aggressive variants and impact on management: A narrative reviewAdv Ther3731123128202010.1007/s12325-020-01391-132488657 LeeYMSungTYKimWBChungKWYoonJHHongSJRisk factors for recurrence in patients with papillary thyroid carcinoma undergoing modified radical neck dissectionBr J Surg10310201025201610.1002/bjs.1014427121346 FengJWHongLZWangFWuWXHuJLiuSYJiangYYeJA nomogram based on clinical and ultrasound characteristics to predict central lymph node metastasis of papillary thyroid carcinomaFront Endocrinol (Lausanne)12666315202110.3389/fendo.2021.66631533995284 GaoXLuoWHeLChengJYangLPredictors and a prediction model for central cervical lymph node metastasis in papillary thyroid carcinoma (cN0)Front Endocrinol (Lausanne)12789310202110.3389/fendo.2021.78931035154002 LuoQWGaoSLvXLiSJWangBFHanQQWangYPGuanQLGongTA novel tool for predicting the risk of central lymph node metastasis in patients with papillary thyroid microcarcinoma: A retrospective cohort studyBMC Cancer22606202210.1186/s12885-022-09768-x35655253 GuiCYQiuSLPengZHWangMClinical and pathologic predictors of central lymph node metastasis in papillary thyroid microcarcinoma: A retrospective cohort studyJ Endocrinol Invest41403409201810.1007/s40618-017-0759-y28884301 LiXZhangHZhouYChengRRisk factors for central lymph node metastasis in the cervical region in papillary thyroid carcinoma: A retrospective studyWorld J Surg Oncol19138202110.1186/s12957-021-02247-w33941214 WangZChangQZhangHDuGLiSLiuYSunHYinDA clinical predictive model of central lymph node metastases in papillary thyroid carcinomaFront Endocrinol (Lausanne)13856278202210.3389/fendo.2022.85627835784530 ZhaoSYueWWangHYaoJPengCLiuXXuDCombined conventional ultrasound and contrast-enhanced computed tomography for cervical lymph node metastasis prediction in papillary thyroid carcinomaJ Ultrasound Med42385398202310.1002/jum.1602435634760 LeeJHKimYChoiJWKimYSThe association between papillary thyroid carcinoma and histologically proven Hashimoto's thyroiditis: A meta-analysisEur J Endocrinol168343349201310.1530/EJE-12-090323211578 WangYZhengJHuXChangQQiaoYYaoXZhouXA retrospective study of papillary thyroid carcinoma: Hashimoto's thyroiditis as a protective biomarker for lymph node metastasisEur J Surg Oncol49560567202310.1016/j.ejso.2022.11.01436404253 IssaPPOmarMButiYIssaCPChabotBCarnabatuCJMunshiRHusseinMAboueishaMShamaMHashimoto's thyroiditis minimizes lymph node metastasis in BRAF mutant papillary thyroid carcinomasBiomedicines102051202210.3390/biomedicines1008205136009596 JaraSMCarsonKAPaiSIAgrawalNRichmonJDPrescottJDDackiwAZeigerMABishopJATufanoRPThe relationship between chronic lymphocytic thyroiditis and central neck lymph node metastasis in North American patients with papillary thyroid carcinomaSurgery15412721280201310.1016/j.surg.2013.07.02124238047 MaoJZhangQZhangHZhengKWangRWangGRisk factors for lymph node metastasis in papillary thyroid carcinoma: A systematic review and meta-analysisFront Endocrinol (Lausanne)11265202010.3389/fendo.2020.0026532477264 LiuYLvHZhangSShiBSunYThe impact of coexistent Hashimoto's thyroiditis on central compartment lymph node metastasis in papillary thyroid carcinomaFront Endocrinol (Lausanne)12772071202110.3389/fendo.2021.77207134867817 SunJJiangQWangXLiuWWangXNomogram for preoperative estimation of cervical lymph node metastasis risk in papillary thyroid microcarcinomaFront Endocrinol (Lausanne)12613974202110.3389/fendo.2021.61397433868164 LiYGaoXGuoTLiuJDevelopment and validation of nomograms for predicting the risk of central lymph node metastasis of solitary papillary thyroid carcinoma of the isthmusJ Cancer Res Clin Oncol1491485314868202310.1007/s00432-023-05146-737598343 LyuYSPyoJSChoWJKimSYKimJHClinicopathological significance of papillary thyroid carcinoma located in the isthmus: A meta-analysisWorld J Surg4527592768202110.1007/s00268-021-06178-134117511 ZhuFLiFXieXWuYWangWInvestigating the impact of tumor location and size on the risk of recurrence for papillary thyroid carcinoma in the isthmusCancer Med121329013299202310.1002/cam4.602337132252 SchmitzGDencksSUltrasound imagingRecent Results Cancer Res216135154202010.1007/978-3-030-42618-7_432594386 Bin SaeedanMAljohaniIMKhushaimAOBukhariSQElnaasSTThyroid computed tomography imaging: Pictorial review of variable pathologiesInsights Imaging7601617201610.1007/s13244-016-0506-527271508 TianXSongQXieFRenLZhangYTangJZhangYJinZZhuYZhangMLuoYPapillary thyroid carcinoma: An ultrasound-based nomogram improves the prediction of lymph node metastases in the central compartmentEur Radiol3058815893202010.1007/s00330-020-06906-632588211 PyoJSKangGKimDHParkCKimJHSohnJHThe prognostic relevance of psammoma bodies and ultrasonographic intratumoral calcifications in papillary thyroid carcinomaWorld J Surg3723302335201310.1007/s00268-013-2107-523716027 FerreiraLBLimaRTBastosACSDFSilvaAMTavaresCPestanaARiosEEloyCSobrinho-SimõesMGimbaERPSoaresPOPNa overexpression is associated with matrix calcification in thyroid cancer cell linesInt J Mol Sci192990201810.3390/ijms1910299030274371 LiXZhouWZhanWClinical and ultrasonographic features of medullary thyroid microcarcinomas compared with papillary thyroid microcarcinomas: A retrospective analysisBMC Med Imaging2049202010.1186/s12880-020-00444-932410587 LiuJJiaXGuYChenXGuanLYanJZhaiHZhouNDongYZhanWThyroid parenchyma microcalcifications on ultrasound for predicting lymph node metastasis in papillary thyroid carcinoma: A prospective multicenter study in ChinaFront Oncol11609075202110.3389/fonc.2021.60907533747925 LuganoRRamachandranMDimbergATumor angiogenesis: Causes, consequences, challenges and opportunitiesCell Mol Life Sci7717451770202010.1007/s00018-019-03351-731690961 ZhangFQiaoYZhangHValue of CT features in the diagnosis of papillary thyroid tumors in incidental thyroid nodulesInt J Endocrinol20209342317202010.1155/2020/934231733123196 ParkJPRohJLLeeJHBaekJHGongGChoKJChoiSHNamSYKimSYRisk factors for central neck lymph node metastasis of clinically noninvasive, node-negative papillary thyroid microcarcinomaAm J Surg208412418201410.1016/j.amjsurg.2013.10.03224602323 PengYZhangZTWangTTWangYLiCHZuoMJLinHSGongLGPrediction of central lymph node metastasis in cN0 papillary thyroid carcinoma by CT radiomicsAcad Radiol3014001407202310.1016/j.acra.2022.09.00236220726 MouYHanXLiJYuPWangCSongZWangXZhangMZhangHMaoNSongXDevelopment and validation of a computed tomography-based radiomics nomogram for the preoperative prediction of central lymph node metastasis in papillary thyroid microcarcinomaAcad Radiol3118051817202310.1016/j.acra.2023.11.03038071100

Flowchart of the present study. PTC, papillary thyroid cancer; CLND, central lymph node dissection; US, ultrasound; CT, computed tomography; cN0, clinically negative.

Representative ultrasound images. (A) Tumor location (isthmus). (B) Microcalcification. (C) Echogenicity (markedly hypoechoic). (D) Capsule contact.

Representative contrast-enhanced computed tomography images. (A) Inhomogeneous enhancement (red arrow). (B) Calcification (red arrow). (C) Capsule invasion (red arrow). (D) Tracheal deviation (red arrow).

Nomogram for predicting central lymph node metastasis. US, ultrasound; CT, computed tomography.

Receiver operating characteristic curves. (A) Training group. (B) Validation group. AUC, area under the curve.

ROC curves. The area under the ROC curve for predicting central lymph node metastasis based on clinical, US and contrast-enhanced CT features was higher than that of not combining contrast-enhanced CT features. ROC, receiver operating characteristic; US, ultrasound; CT, computed tomography; AUC, area under the curve.

Calibration curves. (A) Training group. (B) Validation group. CLNM, central lymph node metastasis.

Decision curve analysis for the nomogram.

Characteristics of all patients in the training and validation group.

A, Clinical characteristics
Characteristic Training group (n=550) Validation group (n=236) P-value
Age (%) 0.620
  ≤45 years 401 (72.91) 168 (71.19)
  >45 years 149 (27.09) 68 (28.81)
Sex (%) 0.399
  Male 138 (25.09) 66 (27.97)
  Female 412 (74.91) 170 (72.03)
BMI, kg/m2 23.74±3.57 23.59±3.59 0.596
With HT (%) 0.150
  Yes 100 (18.18) 33 (13.98)
  No 450 (81.82) 203 (86.02)
BRAF V600E mutation (%) 0.863
  Yes 448 (81.45) 191 (80.93)
  No 102 (18.55) 45 (19.07)
TSH (%) 0.191
  Low 9 (1.64) 3 (1.27)
  Normal 536 (97.45) 227 (96.19)
  High 5 (0.91) 6 (2.54)
TG (%) 0.105
  Low 95 (17.27) 29 (12.29)
  Normal 425 (77.27) 188 (79.66)
  High 30 (5.45) 19 (8.05)
TGAb (%) 0.125
  Negative 351 (63.82) 164 (69.49)
  Positive 199 (36.18) 72 (30.51)
TPOAb (%) 0.658
  Negative 428 (77.82) 187 (79.24)
  Positive 122 (22.18) 49 (20.76)
B, US characteristics
Characteristic Training group (n=550) Validation group (n=236) P-value
Tumor location (%) 0.682
  Isthmus 22 (4.00) 8 (3.39)
  Non-isthmus 528 (96.00) 228 (96.61)
Tumor number (%) 0.472
  Unifocal 412 (74.91) 171 (72.46)
  Multifocal 138 (25.09) 65 (27.54)
  Tumor size (%) 0.589
  ≤1.0 mm 395 (71.82) 165 (69.92)
  >1.0 mm 155 (28.18) 71 (30.08)
Tumor shape (%) 0.768
  Taller-than-wide 278 (50.55) 122 (51.69)
  Wider-than-tall 272 (49.45) 114 (48.31)
Margin (%) 0.500
  Irregular 511 (92.91) 216 (91.53)
  Regular 39 (7.09) 20 (8.47)
Composition (%) 0.943
  Solid 525 (95.45) 225 (95.34)
  Non-solid 25 (4.55) 11 (4.66)
Echogenicity (%) 0.140
  Markedly hypoechoic 19 (3.45) 8 (3.39)
  Hypoechoic 527 (95.82) 222 (94.07)
  Isoechoic 4 (0.73) 6 (2.54)
Capsule contact (%) 0.557
  Yes 405 (73.64) 169 (71.61)
  No 145 (26.36) 67 (28.39)
Microcalcification (%) 0.238
  Yes 407 (74.00) 184 (77.97)
  No 143 (26.00) 52 (22.03)
Blood flow signal (%) 0.159
  Rich 89 (16.18) 48 (20.34)
  poor 461 (83.82) 188 (79.66)
C, CT characteristics
Characteristic Training group (n=550) Validation group (n=236) P-value
UCT, Hu 61.11±17.51 62.92±16.84 0.174
VCT, Hu 130.69±35.10 131.64±34.84 0.726
ΔCT, Hu 69.58±28.95 67.72±29.71 0.419
Homogeneity of enhancement (%) 0.219
  Inhomogeneity 488 (88.73) 202 (85.59)
  Homogeneity 62 (11.27) 34 (14.41)
Calcification (%) 0.091
  Yes 91 (16.55) 51 (21.61)
  No 459 (83.45) 185 (78.39)
Capsule invasion (%) 0.382
  Yes 254 (46.18) 117(49.58)
  No 296 (53.82) 119 (50.42)
Tracheal deviation (%) 0.852
  Yes 15 (2.73) 7 (2.97)
  No 535 (97.27) 229 (97.03)

UCT, mean CT values in plain phase; VCT, mean CT values in venous phase; ΔCT=VCT-UCT. Data are presented as n or mean ± standard deviation. UCT, VCT and ΔCT were analyzed using the t test, TSH and echogenicity were analyzed using Fisher's exact test, and all other variables were analyzed using the χ2 test. BMI, body mass index; HT, Hashimoto thyroiditis; TSH, thyroid stimulating hormone; TG, thyroglobulin; TGAb, TG antibody; TPOAb, thyroid peroxidase antibody; US, ultrasound; CT, computed tomography.

Univariate analysis of characteristics in the training group.

A, Clinical characteristics
Characteristic CLNM(+) group (n=215) CLNM(−) group (n=335) P-value
Age (%) <0.001a
  ≤45 years 174 (80.93) 227 (67.76)
  >45 years 41 (19.07) 108 (32.24)
Sex (%) <0.001a
  Male 77 (35.81) 61 (18.21)
  Female 138 (64.19) 274 (81.79)
BMI (%) 0.168
  ≤20.92 kg/m2 43 (20.00) 84 (25.07)
  >20.92 kg/m2 172 (80.00) 251(74.93)
With HT (%) 0.001a
  Yes 25 (11.63) 75 (22.39)
  No 190 (88.37) 260 (77.61)
BRAF V600E mutation (%) 0.187
  Yes 181 (84.19) 267 (79.70)
  No 34 (15.81) 68 (20.30)
TSH (%) 0.757
  Low 3 (1.40) 6 (1.79)
  Normal 211 (98.14) 325 (97.01)
  High 1 (0.47) 4 (1.19)
TG (%) 0.107
  Low 28 (13.02) 67 (20.00)
  Normal 175 (81.40) 250 (74.63)
  High 12 (5.58) 18 (5.37)
TGAb (%) 0.004a
  Negative 153 (71.16) 198 (59.10)
  Positive 62 (28.84) 137 (40.90)
TPOAb (%) 0.002a
  Negative 182 (84.65) 246 (73.43)
  Positive 33 (15.35) 89 (26.57)
B, US characteristics
Characteristic CLNM(+) group (n=215) CLNM(−) group (n=335) P-value
Tumor location (%) <0.001a
  Isthmus 16 (7.44) 6 (1.79)
  Non-isthmus 199 (92.56) 329 (98.21)
Tumor number (%) 0.538
  Unifocal 158 (73.49) 254 (75.82)
  Multifocal 57 (26.51) 81 (24.18)
Tumor size (%) <0.001a
  ≤1.0 mm 128 (59.53) 267 (79.70)
  >1.0 mm 87 (40.47) 68 (20.30)
Tumor shape (%) 0.414
  Taller-than-wide 104 (48.37) 174 (51.94)
  Wider-than-tall 111 (51.63) 161 (48.06)
Margin (%) 0.148
  Irregular 204 (94.88) 307 (91.64)
  Regular 11 (5.12) 28 (8.36)
Composition (%) 0.457
  Solid 207 (96.28) 318 (94.93)
  Non-solid 8 (3.72) 17 (5.07)
Echogenicity (%) 0.263
  Markedly hypoechoic 4 (1.86) 15 (4.48)
  Hypoechoic 209 (97.21) 318 (94.93)
  Isoechoic 2 (0.93) 2 (0.60)
Capsule contact (%) <0.001a
  Yes 179 (83.26) 226 (67.46)
  No 36 (16.74) 109 (32.54)
Microcalcification (%) <0.001a
  Yes 178 (82.79) 229 (68.36)
  No 37 (17.21) 106 (31.64)
Blood flow signal (%) 0.508
  Rich 32 (14.88) 57 (17.01)
  Poor 183 (85.12) 278 (82.99)
C, CT characteristics
Characteristics CLNM(+) group (n=215) CLNM(−) group (n=335) P-value
UCT (%) 0.155
  ≤65.35 Hu 135 (62.79) 230 (68.66)
  >65.35 Hu 80 (37.21) 105 (31.34)
VCT (%) 0.208
  ≤183.90 Hu 200 (93.02) 320 (95.52)
  >183.90 Hu 15 (6.98) 15 (4.48)
ΔCT (%) 0.173
  ≤111.50 Hu 194 (90.23) 313 (93.43)
  >111.50 Hu 21 (9.77) 22 (6.57)
Homogeneity of enhancement (%) 0.002a
  Inhomogeneity 202 (93.95) 286 (85.37)
  Homogeneity 13 (6.05) 49 (14.63)
Calcification (%) 0.007a
  Yes 47 (21.86) 44 (13.13)
  No 168 (78.14) 291 (86.87)
Capsule invasion (%) <0.001a
  Yes 155 (72.09) 99 (29.55)
  No 60 (27.91) 236 (70.45)
Tracheal deviation (%) 0.643
  Yes 5 (2.33) 10 (2.99)
  No 210 (97.67) 325 (97.01)

UCT, mean CT values in plain phase; VCT, mean CT values in venous phase; ΔCT=VCT-UCT.

P<0.05. BMI, body mass index; HT, Hashimoto thyroiditis; TSH, thyroid stimulating hormone; TG, thyroglobulin; TGAb, TG antibody; TPOAb, thyroid peroxidase antibody; US, ultrasound; CT, computed tomography; CLNM, central lymph node metastasis.

Multivariate analysis of characteristics in the training group.

Variable B coefficient OR 95% CI P-value
Age
  ≤45 years −0.037 0.964 0.945–0.982 <0.001a
  >45 years
Sex
  Male 0.764 2.147 1.332–3.459 0.002a
  Female
With HT
  Yes 0.922 2.515 1.208–5.239 0.014a
  No
TGAb
  Negative 0.196 1.216 0.681–2.173 0.509
  Positive
TPOAb
  Negative −0.629 0.533 0.277–1.025 0.059
  Positive
US-tumor location
  Isthmus −1.554 0.211 0.067–0.669 0.008a
  Non-isthmus
US-tumor size
  ≤1.0 mm −0.264 0.768 0.466–1.265 0.300
  >1.0 mm
US-capsule contact
  Yes −0.418 0.659 0.399–1.088 0.103
  No
US-microcalcification
  Yes −0.529 0.589 0.355–0.979 0.041a
  No
CT-homogeneity of enhancement
  Inhomogeneity 0.997 2.711 1.268–5.798 0.010a
  Homogeneity
CT-capsule invasion
  Yes 1.866 6.463 4.103–10.181 <0.001a
  No
CT-calcification
  Yes −0.450 0.638 0.368–1.106 0.110
  No

P<0.05. HT, Hashimoto thyroiditis; TSH, thyroid stimulating hormone; TGAb, TG antibody; TPOAb, thyroid peroxidase antibody; US, ultrasound; CT, computed tomography; OR, odds ratio; CI, confidence interval.