Construction of a Predictive Model for T1-stage Low-risk Papillary Thyroid Carcinoma with Central Lymph Node Metastasis Using Ultrasound Radiomics Combined with Clinical Radiomics

Construction of a Predictive Model for T1-stage Low-risk Papillary Thyroid Carcinoma with Central Lymph Node Metastasis Using Ultrasound Radiomics Combined with Clinical Radiomics | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Construction of a Predictive Model for T1-stage Low-risk Papillary Thyroid Carcinoma with Central Lymph Node Metastasis Using Ultrasound Radiomics Combined with Clinical Radiomics Peng Zhao, Lulu Liang, Xian Wei, Duoping Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6733396/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Papillary thyroid carcinoma (PTC) is the most common tumor subtype of thyroid cancer and approximately 30–90% of patients with PTC exhibit occult central lymph node metastasis (CLNM) on postoperative pathological examination. This study aimed to explore the diagnostic efficacy and clinical application value of combining ultrasound radiomics features with clinical features to construct predictive models for patients with papillary thyroid carcinoma (PTC) and central lymph node metastasis (CLNM). Methods This study included the retrospective data from total 191 PTC patients hospitalized between June 2020 and June 2022 (training set: 134, validation set: 57). Additionally, 46 patients were included in the prospective validation set. Clinical features affecting CLNM in patients with PTC were identified using univariate and multivariate analyses. Logistic regression models were constructed based on clinical and radiomics features, individually and combined. The diagnostic efficacies of the three models were compared using receiver operating characteristic curves, and a nomogram was constructed for visualization. The clinical utility of the model was evaluated using decision curve analysis (DCA) and calibration curves. Results Male sex, unclear or irregular margins, and microcalcifications were independent risk factors in the clinical radiomics predictive model. The area under the curve (AUC) for the training, validation, and prospective validation sets was 0.740, 0.656, and 0.626, respectively. Twelve ultrasound radiomics features were selected to construct the radiomics model (AUC: 0.794, 0.720, and 0.766, respectively). The combined model demonstrated AUCs of 0.850, 0.750, and 0.786, for training, validation and prospective validation set respectively. The DCA and calibration curves indicated that the combined model had a better diagnostic efficiency and clinical utility. Conclusion This study presents a combined predictive model based on ultrasound radiomics and clinical features that can effectively predict preoperative CLNM in patients with PTC cN0 T1 stage, demonstrating its clinical applicability. papillary thyroid carcinoma central lymph node metastasis ultrasound radiomics predictive models Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. Introduction Papillary thyroid carcinoma (PTC) is the most common tumor subtype of thyroid cancer. Studies have revealed that approximately 30–90% of patients with PTC exhibit occult central lymph node metastasis (CLNM) on postoperative pathological examination [ 1 , 2 ]. Therefore, CLNM is closely associated with local tumor recurrence in patients with PTC after surgery [ 3 , 4 ]. Factors influencing lymph node metastasis in PTC include sex, age, number of cancer foci, tumor size, vascular invasion (VI), extrathyroidal extension (ETE), and the presence of BRAF V600E mutations [ 5 , 6 ]. Recent studies have identified a positive correlation between tumor size and CLNM in patients with PTC, especially in tumors with diameters > 2 cm [ 7 ]. Furthermore, Han et al. reported that a tumor diameter of ≥ 2 cm was an independent risk factor affecting the 5-year postoperative survival rate of patients postoperatively [ 8 ]. However, some tumors < 2 cm (T1 stage), including papillary thyroid microcarcinoma (PTMC), which has a tumor diameter < 1 cm, may also be highly invasive and prone to CLNM [ 9 ]. Therefore, the World Health Organization (WHO) emphasizes that low-risk PTC should not solely be defined by tumor size but should also consider tumor invasiveness and other imaging and clinical features [ 10 ]. However, the invasive nature of tumors is often fully assessed using only postoperative pathological reports. Although there is currently no consensus on whether routine prophylactic central compartment lymph node dissection (pCCND) should be performed during surgery for low-risk thyroid cancer, constructing a preoperative, non-invasive, and clinically applicable predictive model for patients with low-risk PTC would provide valuable guidance for clinical decision-making and reduce the occurrence of overtreatment. The development of nomograms based on clinical data has proven to be a simple, feasible, and reproducible approach, yielding promising predictive results [ 11 – 13 ]. However, an increasing number of patients with low-risk T1 stage PTC are being diagnosed with CLNM on fine needle aspiration biopsy (FNAB) during routine checkups [ 14 , 15 ]. Many existing predictive models for CLNM include both patients with stage T1 and T2, and fail to accurately reflect the distinct biological characteristics of T1 stage PTC. Research focusing specifically on CLNM in patients with T1 stage PTC remains limited. Therefore, this study aimed to explore the clinical application of a preoperative predictive model for CLNM in low-risk patients with T1 stage PTC. The model was constructed using a combination of ultrasound radiomics and clinical radiomics. Furthermore, we validated the accuracy of this model through prospective studies. 2. Methods 2.1. Patients and data collection This retrospective study collected data from patients with thyroid cancer who were hospitalized in the Department of Head and Neck Surgery at the Affiliated Cancer Hospital of Guangxi Medical University between June 2020 and June 2022, and between February 2023 and August 2023. All patients underwent routine two‑dimensional ultrasound. The inclusion criteria were as follows: 1) all patients underwent preoperative FNAB of the thyroid nodules, with BRAF V600E gene testing (fluorescence quantitative polymerase chain reaction); 2) postoperative pathology confirmed unilateral T1 stage PTC; 3) preoperative imaging did not indicate any obvious metastatic lymph nodes in the neck, or any suspicious lymph nodes identified were confirmed to be negative for metastasis through FNAB; 4) all included patients underwent hemithyroidectomy (HT) and pCCND; and 5) none of the patients received prior treatment for other cancers, including radiotherapy, chemotherapy, or any other treatments. The exclusion criteria were as follows: 1) pathology results that could not be identified as PTC or patients confirmed to have lateral neck lymph node metastasis preoperatively through FNAB; 2) incomplete or missing postoperative pathological reports; and 3) poor image quality preventing the extraction of radiomic features. A total of 237 patients were included in this study. Of these, 191 patients were treated between June 2020 and June 2022, and 46 patients were treated between February 2023 and August 2023. Using a random number method, the 191 patients were randomly divided into a training set (134 patients) and a validation set (57 patients) in a 7:3 ratio. The 46 patients from the later period were used as the prospective validation set. This study was approved by the Ethics Committee of the Affiliated Cancer Hospital of Guangxi Medical University (approval no. KY2025625). The requirement for individual consent for this retrospective analysis was waived. The details of the patient selection process and grouping are shown in Fig. 1 . 2.2. Thyroid and cervical ultrasound examination All patients underwent comprehensive preoperative thyroid and cervical ultrasound examinations performed by senior ultrasound specialists at our hospital. Ultrasound images were acquired using the GE Logiq E9 Ultrasound Diagnostic System, TOSHIBA Aplio 500 Ultrasound Diagnostic System, TOSHIBA Aplio 400 Ultrasound Diagnostic System, and CANON Aplio i800 Ultrasound Diagnostic System. The examinations were performed with the patient in the supine position, with the shoulders elevated and the head tilted downward. High-frequency 5–12 MHz ultrasound probes were used to scan the thyroid and cervical lymph nodes. The following thyroid nodule parameters were measured: maximum transverse diameter, maximum longitudinal diameter, height, nodule margin, presence of microcalcifications, texture, longitudinal-to-transverse (L/T) ratio, and presence of ETE. Thyroid nodules were classified in this study according to the Thyroid Imaging Reporting and Data System in China (C-TIRADS) [ 16 , 17 ]. Additionally, the central and lateral cervical neck compartments were carefully examined for enlarged or abnormally shaped lymph nodes [ 18 ]. 2.3. Surgical procedure All patients in the training, validation, and prospective validation sets underwent HT + pCCND under general anesthesia via open surgery or an endoscopic approach. All surgeries were performed by senior attending physicians at the Department of Head and Neck Surgery. The scope of lymph node dissection included the following boundaries: the upper boundary reached the level of the thyroid cartilage, the lower boundary extended to the superior border of the thymus or the level of the brachiocephalic trunk, the lateral boundary was the medial side of the carotid sheath on both sides, and the deep boundary was the deep layer of the cervical fascia and the anterior wall of the esophagus [ 18 – 20 ]. This area encompasses the Delphian, pretracheal, and tracheoesophageal groove lymph nodes. The lymph nodes behind the recurrent laryngeal nerve (LN-prRLN) were included for right-sided neck lymph node dissection. Lymph nodes were thoroughly removed during pCCND. 2.4. Extraction and selection of ultrasound radiomic features Ultrasound images were extracted from the electronic medical records of the hospital. The ITK-SNAP medical image segmentation tool identified the region of interest (ROI) [ 21 , 22 ]. ITK-SNAP has been widely used for the image segmentation of two-dimensional or three-dimensional imaging data, such as computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and positron emission tomography (PET)-CT [ 23 , 24 ], and has demonstrated good performance in segmenting thyroid ultrasound images [ 25 ]. Subsequently, the open-source Pyradiomics module in Python (v3.8.10, http://www.python.org ) was used to extract 851 radiomic features from the ultrasound images, including both the original and filtered (wavelet) features. The extracted radiomic features were categorized into several major groups: first-order statistics, shape features, gray-level co-occurrence matrix (GLCM), gray-level run-length matrix (GLRLM), gray-level size zone matrix (GLSZM), and neighboring gray-tone difference matrix (NGTDM). 2.5 Construction and evaluation of the radiomics, clinical, and combined models The radiomics, clinical, and combined models were constructed using logistic regression. The diagnostic performance of each model was evaluated using the receiver operating characteristic (ROC) curves. Decision curve analysis (DCA) and calibration curves were used to assess the clinical applicability and values of the three models. The goodness-of-fit test was conducted using the Hosmer–Lemeshow test. 2.6. Statistical analysis The univariate analysis was performed based on the data type. The chi-squared test or Fisher’s exact test was used for categorical data. The t-test was applied for continuous data, and the rank-sum test was used for ordinal data. In the univariate analysis, a p-value of < 0.05 was considered statistically significant and was included in the multivariate regression analysis. Multivariate regression analysis was performed using binary logistic regression. All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Nomograms, DCA, and calibration curves were generated using the R software (version 4.2.1; http://www.r-project.org/ ). All statistical tests were two-sided; a p-value < 0.05 was considered statistically significant. 3. Results 3.1. Baseline characteristics A total of 191 patients with T1 stage PTC between June 2020 and June 2022 were included. Of these, 48 were male and 143 were female. The patients were randomly divided into a training set (134 patients) and a validation set (57 patients) in a 7:3 ratio using random number generation. The training set included 33 male and 101 female patients, with a mean age of 42.55 ± 9.41 years. There were 15 male and 42 female patients in the validation set, with a mean age of 42.87 ± 10.86 years. From February 2023 to August 2023, an additional 46 patients with T1 stage PTC were enrolled as the prospective validation cohort, consisting of nine male and 37 female patients, with a mean age of 42.65 ± 10.16 years. The baseline characteristics and grouping details of the training, validation, and prospective validation sets are summarized in Table 1 . Table 1 Baseline data of patients in the training set, validation set and prospective validation set Characteristics training set validation set prospective validation set Total p (N = 134) (N = 57) (N = 46) (N = 237) Sex Male 33.0 (24.6%) 15.0 (26.3%) 9.00 (19.6%) 57.0 (24.1%) 0.875 Female 101 (75.4%) 42.0 (73.7%) 37.0 (80.4%) 180 (75.9%) Age ≤ 55 118 (88.1%) 48.0 (84.2%) 38.0 (82.6%) 204 (86.1%) 0.785 >55 16.0 (11.9%) 9.00 (15.8%) 8.00 (17.4%) 33.0 (13.9%) Family History No 120 (89.6%) 53.0 (93.0%) 42.0 (91.3%) 215 (90.7%) 0.901 Yes 14.0 (10.4%) 4.00 (7.0%) 4.00 (8.7%) 22.0 (9.3%) Hyperthyroidism No 129 (96.3%) 54.0 (94.7%) 42.0 (91.3%) 225 (94.9%) 0.623 Yes 5.00 (3.7%) 3.00 (5.3%) 4.00 (8.7%) 12.0 (5.1%) TSH <0.35 126 (94.0%) 50.0 (87.7%) 42.0 (91.3%) 218 (92.0%) 0.533 ≥ 0.35 8.00 (6.0%) 7.00 (12.3%) 4.00 (8.7%) 19.0 (8.0%) BRAF mutation No 35.0 (26.1%) 13.0 (22.8%) 11.0 (23.9%) 59.0 (24.9%) 0.967 Yes 99.0 (73.9%) 44.0 (77.2%) 35.0 (76.1%) 178 (75.1%) C-TIRRADS 3 13.0 (9.7%) 5.00 (8.8%) 4.00 (8.7%) 22.0 (9.3%) 0.917 4 91.0 (67.9%) 42.0 (73.7%) 35.0 (76.1%) 168 (70.9%) 5 18.0 (13.4%) 8.00 (14.0%) 6.00 (13.0%) 32.0 (13.5%) 6 12.0 (9.0%) 2.00 (3.5%) 1.00 (2.2%) 15.0 (6.3%) Echo Low 103 (76.9%) 49.0 (86.0%) 29.0 (63.0%) 181 (76.4%) 0.224 Medium 28.0 (20.9%) 8.00 (14.0%) 15.0 (32.6%) 51.0 (21.5%) High 3.00 (2.2%) 0 (0%) 2.00 (4.3%) 5.00 (2.1%) Margin Regular 39.0 (29.1%) 10.0 (17.5%) 16.0 (34.8%) 65.0 (27.4%) 0.237 Irregular 95.0 (70.9%) 47.0 (82.5%) 30.0 (65.2%) 172 (72.6%) Table 1 (Continued) Characteristics training set validation set prospective validation set Total p (N = 134) (N = 57) (N = 46) (N = 237) Aspect Ratio ≤ 1 75.0 (56.0%) 30.0 (52.6%) 31.0 (67.4%) 136 (57.4%) 0.472 >1 59.0 (44.0%) 27.0 (47.4%) 15.0 (32.6%) 101 (42.6%) No 69.0 (51.5%) 22.0 (38.6%) 18.0 (39.1%) 109 (46.0%) 0.289 Yes 65.0 (48.5%) 35.0 (61.4%) 28.0 (60.9%) 128 (54.0%) Solid No 13.0 (9.7%) 6.00 (10.5%) 5.00 (10.9%) 24.0 (10.1%) 0.996 Yes 121 (90.3%) 51.0 (89.5%) 41.0 (89.1%) 213 (89.9%) NLR Mean (SD) 1.95 (1.86) 1.68 (0.474) 2.16 (2.07) 1.93 (1.69) 0.551 Median [Min, Max] 1.75 [0.572, 21.9] 1.59 [0.841, 2.97] 1.59 [0.624, 11.0] 1.65 [0.572, 21.9] PLR Mean (SD) 132 (58.7) 122 (42.1) 138 (115) 130 (70.2) 0.676 Median [Min, Max] 121 [48.7, 584] 113 [2.15, 231] 115 [38.2, 822] 118 [2.15, 822] LMR Mean (SD) 5.71 (1.89) 6.03 (1.99) 4.99 (1.95) 5.65 (1.95) 0.055 Median [Min, Max] 5.33 [1.37, 11.2] 6.03 [1.80, 12.0] 4.61 [1.08, 10.5] 5.48 [1.08, 12.0] SII Mean (SD) 517 (761) 415(135) 516(493) 492(615) 0.755 Median [Min, Max] 415 [118, 8960] 399[7.89, 715] 403[96.0, 2980] 410[7.89, 8960] CLNM No 97.0 (72.4%) 44.0 (77.2%) 27.0 (58.7%) 168 (70.9%) 0.207 Yes 37.0 (27.6%) 13.0 (22.8%) 19.0 (41.3%) 69.0 (29.1%) 3.2. Construction of the clinical model in the training set The univariate analysis (Table 2 ) identified sex, indistinct or irregular margins, microcalcifications, and an L/T ratio > 1 as risk factors for CLNM in patients with T1 stage PTC. These four risk factors were subjected to multivariate analysis using binary logistic regression (Table 3 ), which revealed that sex, indistinct or irregular margins, and microcalcifications were independent risk factors for CLNM in T1 stage PTC. The diagnostic performance of the clinical model was evaluated using ROC curves for the training, validation, and prospective validation sets. The area under the curve (AUC) values for these sets were 0.740, 0.656, and 0.626, respectively. A nomogram was generated to visually represent the clinical model (Fig. 2 ). Table 2 Univariate Analysis of Factors Associated with CLNM in the Training set Characteristics B S.E Wald p ExpB Lower Upper Sex -0.916 0.425 4.65 0.031 0.400 0.174 0.920 Age -1.082 0.782 1.914 0.166 0.339 0.073 1.570 Family history 0.424 0.595 0.508 0.476 1.528 0.476 4.901 Hyperthyroidism -0.437 1.135 0.148 0.700 0.646 0.070 5.975 C-TIRADS 3 4.448 0.217 4 1.455 0.954 2.329 0.127 4.286 0.661 27.785 5 0.466 0.812 0.330 0.566 1.594 0.324 7.835 6 1.157 0.913 1.607 0.205 3.182 0.531 19.051 Echo Low 0.972 0.615 Medium 0.447 0.454 0.972 0.324 1.564 0.643 3.804 High -20.168 23205.422 0 0.999 0 0 . TSH(0 ≥ 0.35) 0.485 0.757 0.409 0.522 1.624 0.368 7.164 BRAF Mutation 0.130 0.447 0.085 0.770 1.139 0.475 2.733 Margin 1.538 0.570 7.282 0.007 4.657 1.524 14.236 Microcalcifications 1.257 0.415 9.174 0.002 3.515 1.558 7.929 Aspect Ratio >1 0.866 0.395 4.818 0.028 2.378 1.097 5.156 Solid 0.264 0.689 0.147 0.701 1.303 0.338 5.024 NLR 0.145 0.133 1.197 0.274 1.156 0.892 1.499 PLR 0 0.003 0.008 0.930 1 0.994 1.007 LMR 0.003 0.102 0.001 0.979 1.003 0.821 1.225 SII 0 0 1.056 0.304 1 1 1.001 Table 3 Multivariable Analysis of Factors Associated with CLNM in the Training set Characteristics B S.E Wald p ExpB Lower Upper Sex 1.085 0.473 5.260 0.022 2.959 1.171 7.477 Margin -1.379 0.615 5.025 0.025 0.252 0.075 0.841 Microcalcifications -0.976 0.446 4.785 0.029 0.377 0.157 0.903 Aspect Ratio >1 -0.572 0.426 1.805 0.179 0.564 0.245 1.300 Constant -0.255 0.349 0.533 0.465 0.775 3.3. Construction of the ultrasound radiomics model in the training set The 851 ultrasound radiomic features were first standardized using z-score normalization and then subjected to univariate analysis for feature selection, resulting in 201 features (Fig. 3 ). After performing five-fold cross-validation with the least absolute shrinkage and selection operator (LASSO) regression, 12 ultrasound radiomic features were selected (Table 4 ). The Rad score was calculated by summing the feature coefficients and a predictive model was built using the Rad score via a logistic regression algorithm. The AUC values for the training, validation, and prospective validation sets were 0.794, 0.720, and 0.766, respectively. Table 4 Ultrasonic radiomic features Radiomic features original_shape_Maximum 2D Diameter Slice original_shape_Maximum 3D Diameter original_firstorder_Robust Mean Absolute Deviation wavelet-HLH_glcm_Difference Entropy wavelet-HLH_glcm_Difference Variance wavelet-HLH_glrlm_Gray Level Non-Uniformity Normalized wavelet-HLH_glszm_Small Area Low Gray Level Emphasis wavelet-HHH_glcm_Cluster Prominence wavelet-HHH_glszm_High Gray Level Zone Emphasis wavelet-HHH_glszm_Low Gray Level Zone Emphasis wavelet-HHH_glszm_Zone Entropy wavelet-LLL_firstorder_Interquartile Range 3.4. Construction of the combined ultrasound radiomics–clinical model The Rad score derived from ultrasound radiomics features was combined with clinical characteristics, such as sex, margin irregularity, and microcalcifications, which were used to construct a combined predictive model via logistic regression. A nomogram for this combined predictive model was generated as a visual representation (Fig. 4 ). The AUC values for the combined models were 0.850, 0.750, and 0.786, respectively. Additionally, the goodness of fit of the combined model was assessed using the Hosmer–Lemeshow test. The results showed a p-value of 0.979, suggesting that the model fit the data well. 3.5. Comparison of the clinical, ultrasound radiomics, and combined models The clinical, ultrasound radiomics, and combined models were compared based on their diagnostic performance. As shown in Fig. 5 , which presents the ROC curve comparisons, the diagnostic performance of the models in the training, validation, and prospective validation sets were ranked. A comparison of the DCA and the calibration curve between these three models is shown in Fig. 6 and Fig. 7 . Overall, the combined predictive model demonstrated superior performance in predicting the occurrence of CLNM in T1 stage low-risk PTC. 4. Discussion Clinical prediction models are commonly visualized using nomograms that graphically represent the impact of various predictors on outcomes [ 26 ]. Many studies have focused on predicting CLNM in patients with PTC based on clinical features and logistic regression models. Wu et al. [ 27 ] developed a nomogram for predicting contralateral CLNM based on clinical and pathological data from 204 patients with unilateral PTC and achieved an AUC of 0.836. Chen et al. [ 28 ] constructed a similar nomogram for CLNM prediction using data from 691 patients with PTC, with an AUC of 0.717. Qiu et al. [ 29 ] also used clinical and ultrasound characteristics in a nomogram to predict CLNM in 119 patients with cN0 PTC, with AUCs of 0.703 and 0.656 in the training and validation sets, respectively. Although there are many studies on the prediction of CLNM in patients with PTC, few have specifically focused on patients with low-risk T1 PTC. Our study only included T1 stage patients and identified independent risk factors for CLNM after univariate and multivariate analyses, including sex, nodal margin irregularity, and microcalcifications. These findings are consistent with the previous literature [ 30 , 31 ]. The logistic regression model built using these three clinical features resulted in AUC values of 0.740 for the training set and 0.656 for the validation set, which was comparable to those of previous studies; however, our study stands out as we prospectively validated the clinical model which strengthens the reliability of our results. The AUC for the prospective validation set was 0.626. In summary, although the clinical model has a good predictive value for T1 PTC and CLNM, its diagnostic performance can be further improved by increasing the sample size and accounting for individual patient variability. Radiomics has gained attention owing to its ability to discover clinically relevant features in medical images that are often undetectable by the human eye. Unlike traditional qualitative imaging, radiomics allows quantitative analysis, making it a powerful tool for predicting clinical outcomes [ 32 , 33 ]. Ultrasound is the preferred imaging modality for thyroid nodules, and ultrasound radiomics have been widely studied for thyroid cancer. Several studies have demonstrated its ability to accurately predict CLNM and tumor invasiveness in patients with PTC [ 11 , 34 – 36 ]. In our study, the ultrasound radiomics model was constructed using 12 features (three original image features and nine wavelet-filtered features) selected through LASSO regression and cross-validation. The Rad score was computed by summing the products of each feature coefficient and its corresponding value. The logistic regression model built on the Rad score yielded AUC values of 0.794, 0.720, and 0.766 for the training, validation, and prospective validation sets, respectively, demonstrating good predictive ability. Furthermore, the accuracy was 0.754 and 0.596 for the training and validation sets, respectively; however, after prospective validation, the accuracy improved to 0.761, which matched the training set results. Compared with the clinical model, the ultrasound radiomics model showed superior diagnostic performance, which is consistent with previous studies [ 37 ]. This reinforces the ability of ultrasound radiomics features to reflect the biological characteristics of T1 low-risk PTC tumors. The constructed prediction model was validated to have good clinical value through DCA and calibration curves. As the need for more precise predictions grows, it has become clear that using a single-omics approach to construct prediction models often fails to achieve optimal results. With advancements in computational technology and artificial intelligence, integrated multi-omics models have become a mainstream trend. Similar integrated models have been reported for the prediction of CLNM in PTC. For instance, Zhang et al. [ 38 ] combined ultrasound radiomics and clinical data to predict occult central zone metastasis in cN0 PTC, achieving AUC values of 0.76 and 0.71 for the training and validation sets, respectively. Jiang et al. [ 39 ] developed a combined model using ultrasound radiomics and clinical data from 211 patients with PTC, resulting in AUC values of 0.82 and 0.81 for the training and validation sets. Furthermore, Haji et al. [ 40 ] found that models integrating radiomics and clinical data had a higher sensitivity and specificity than single-omics models, as shown in a meta-analysis of 25 studies. In our study, we combined three clinical features (sex, margin irregularity, and microcalcification) and the Rad score derived from ultrasound radiomics using logistic regression to build a combined prediction model. The AUCs for the combined model were 0.850 for the training set, 0.750 for the validation set, and 0.786 for the prospective validation set, indicating superior predictive performance compared with the clinical and ultrasound radiomics models. The model was visualized using nomograms. The goodness-of-fit test (Hosmer–Lemeshow) for the combined model showed no significant difference (p > 0.05), suggesting that the model fit well. The DCA and calibration curves further supported the clinical applicability of the combined model compared to the clinical and ultrasound radiomics models (Fig. 6 ). In this study, we visualized clinical and combined prediction models using nomograms and evaluated their diagnostic efficacy using ROC, DCA, and calibration curves. The results showed that the combined model outperformed the clinical and ultrasound radiomics models in all datasets (training, validation, and prospective validation), with AUC values of 0.850, 0.759, and 0.786, respectively. This combined model can accurately predict whether patients with low-risk T1 PTC will develop CLNM preoperatively, providing an objective basis for deciding whether prophylactic central neck dissection is needed, potentially preventing overtreatment, and offering valuable clinical guidance. This study has some limitations. First, this was a single-center retrospective study, and while prospective validation was conducted within the same center, the small sample size limited the generalizability of the results. Further multi-center studies are needed to externally validate these findings. Second, the images were obtained from four different ultrasound devices, which may have introduced some variability in image quality and parameters, potentially affecting the extraction of the ROI and feature interpretation. Although high-level ultrasound specialists performed the examinations, the inherent subjectivity of ultrasound reporting may have introduced a measurement bias. Finally, as with any radiomics study, feature extraction methods are manually defined, which could lead to biases in the tumor image representation. In conclusion, the combined model integrating ultrasound radiomics and clinical features offers a more accurate and clinically applicable tool for predicting CLNM in patients with T1-stage low-risk PTC. This model can assist in preoperative decision-making for pCCND and reduce the occurrence of overtreatment. Abbreviations Area under the curve (AUC) Central lymph node metastasis (CLNM) Computed tomography (CT) Decision curve analysis (DCA) Extrathyroidal extension (ETE) Fine needle aspiration biopsy (FNAB) Gray-level co-occurrence matrix (GLCM) Gray-level run-length matrix (GLRLM) Gray-level size zone matrix (GLSZM) Hemithyroidectomy (HT) Least absolute shrinkage and selection operator (LASSO) Longitudinal-to-transverse (L/T) Lymph nodes located behind the recurrent laryngeal nerve (LN-prRLN) Magnetic resonance imaging (MRI) Neighboring gray-tone difference matrix (NGTDM) Prophylactic central compartment lymph node dissection (pCCND) Papillary thyroid carcinoma (PTC) Papillary thyroid microcarcinoma (PTMC) Positron emission tomography (PET) Receiver operating characteristic (ROC) Region of interest (ROI) Thyroid Imaging Reporting and Data System in China (C-TIRADS) Vascular invasion (VI) World Health Organization (WHO) Declarations Acknowledgements We would like to thank Editage (www.editage.cn) for English language editing. Data Availability Statement The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Author Contributions PZ and LL selected the relevant studies, assessed the data, and wrote the manuscript. XW contributed to the methodological framework. DW revised the manuscript. All authors read and approved the final version of the manuscript. Funding This study received no funding. Ethical approval and consent to participate Ethical approval for this study was granted by the Medical Ethics Committee of Guangxi Medical University Cancer Hospital (Approval Number: KY2025625) in accordance with the Declaration of Helsinki. Informed consent was obtained from all participants. Conflicts of Interest The authors declare that they have no conflict of interest to disclose. References Zhu J ZJ, Li L, et al. Application of Machine Learning Algorithms to Predict Central Lymph Node Metastasis in T1-T2, Non-invasive, and Clinically Node Negative Papillary Thyroid Carcinoma [J]. Front Med (Lausanne). 2021, 9: 8:635771. Forest VI, Clark, J.R, Ebrahimi, A. et al. Central compartment dissection in thyroid papillary carcinoma [J]. Ann Surg. 2011, 253: 123-30. Ito Y KT, Kobayashi K, et al. Prognostic factors for recurrence of papillary thyroid carcinoma in the lymph nodes, lung, and bone: analysis of 5,768 patients with average 10-year follow-up [J]. World J Surg. 2012, 36(6): 1274-8. LI Hong-qiang, MA Run-sheng, TANG Yi-feng, et al. Discussion on the necessity of total thyroidectomy and central lymph node dissection in differentiated thyroid carcinoma with stage T1-2N0M0 [J]. Chinese Journal of Practical Surgery. 2021, 41(08): 892-4+9. Mao J ZQ, Zhang H, et al. Risk Factors for Lymph Node Metastasis in Papillary Thyroid Carcinoma: A Systematic Review and Meta-Analysis [J]. Front Endocrinol (Lausanne). 2020, 11: 265. Zheng H LV, Lu J, et al. Clinical Factors Predictive of Lymph Node Metastasis in Thyroid Cancer Patients: A Multivariate Analysis [J]. J Am Coll Surg. 2022, 234(4): 691-700. Huang JH. Predictive Factor of Large ‐ Volume Central Lymph Node Metastasis in Clinical N0 Papillary Thyroid Carcinoma Patients [D]. Shan Tou University. 2022. HAN Yuren，LI Limei，WANG Ｒui. Clinical Characteristics and Prognostic Factors of Thyroid Cancer Patients [J]. The Practical Journal of Cancer. 2022, 37(06): 1000-2. Nie X TZ, Ge M, et al. Risk factors analyses for lateral lymph node metastases in papillary thyroid carcinomas: a retrospective study of 356 patients [J]. Arch Endocrinol Metab. 2016, 60(5): 492-9. Baloch ZW AS, Barletta JA,et al. Overview of the 2022 WHO Classification of Thyroid Neoplasms [J]. Endocr Patho. 2022, 33(1): 27-63. Popović Krneta M ŠŠD, Mijatović Teodorović L, et al. Prediction of Cervical Lymph Node Metastasis in Clinically Node-Negative T1 and T2 Papillary Thyroid Carcinoma Using Supervised Machine Learning Approach [J]. J Clin Med. 2023, 24(11): 3641. Zhou SC LT, Zhou J, et al. An Ultrasound Radiomics Nomogram for Preoperative Prediction of Central Neck Lymph Node Metastasis in Papillary Thyroid Carcinoma [J]. Front Oncol. 2020, 10: 1591. Gong Y ZZ, Tang K, et al. Multimodal predictive factors of metastasis in lymph nodes posterior to the right recurrent laryngeal nerve in papillary thyroid carcinoma [J]. Front Endocrinol (Lausanne). 2023, 14: 1187825. Society of Tumor Ablation Therapy of the Chinese Anti‑Cancer Association, the Ablation Expert Committee of the Chinese Society of Clinical Oncology (CSCO), Chinese Medical Doctor Association College of Interventionalists Tumor Ablation Committee, et al. Expert consensus on thermal ablation of papillary thyroid cancer (2024 edition) [J]. Chin J Intern Med. 2024, 63(04): 355-64. Haddad RI BL, Ball D, et al. Thyroid Carcinoma, Version 2.2022, NCCN Clinical Practice Guidelines in Oncology [J]. J Natl Compr Canc Netw. 2022, 20(8): 925-51. Tessler FN MW, Grant EG. Thyroid Imaging Reporting and Data System (TI-RADS): A User's Guide [J]. Radiology. 2018, 287(1): 29-36. Kim DH KS, Basurrah MA, et al. Diagnostic Performance of Six Ultrasound Risk Stratification Systems for Thyroid Nodules: A Systematic Review and Network Meta-Analysis [J]. AJR Am J Roentgenol. 2023, 220(6): 791-803. Chinese Society of Endocrinology; Thyroid and Metabolism Surgery Group of the Chinese Society of Surgery; China Anti-Cancer Association, et al. Guidelines for the Diagnosis and Management of Thyroid Nodules and Differentiated Thyroid Cancer (Second Edition) [J]. Int J Endocrinol Metab. 2023, 39(03): 181-226. ROBBINS K T SAＲ, MEDINA J E, et al. Consensusstatement on the classification and terminology of neck dissection [J]. Arch Otolaryngol Head Neck Surg. 2008, 134(5): 536-8. Group ATASW EA, Otolaryngology-Head A, et al. Consensus statement on the terminology and classification of central neck dissection for thyroid cancer [J]. Thyroid. 2009, 19(11): 1153-8. Paul A. Yushkevich JP, et al. User-guided 3D active contour segmentation of anatomical structures: Significantly improved efficiency and reliability [J]. Neuroimage. 2006, 31(3): 1116-28. Yushkevich PA YG, Gerig G. ITK-SNAP: An interactive tool for semi-automatic segmentation of multi-modality biomedical images [J]. Annu Int Conf IEEE Eng Med Biol Soc. 2016: 3342-5. Street JS PA, Toma AK. Predicting vasospasm risk using first presentation aneurysmal subarachnoid hemorrhage volume: A semi-automated CT image segmentation analysis using ITK-SNAP [J]. PLoS One. 2023, 18(6): e0286485. Lu S RY, Lu C, et al. Radiomics features from whole thyroid gland tissue for prediction of cervical lymph node metastasis in the patients with papillary thyroid carcinoma [J]. J Cancer Res Clin Oncol. 2023, 149(14): 13005-16. Li F PD, He Y, et al. Using ultrasound features and radiomics analysis to predict lymph node metastasis in patients with thyroid cancer [J]. BMC Surg. 2020, 20(1): 315. Balachandran VP GM, Smith JJ, et al. . Nomograms in oncology: more than meets the eye [J]. Lancet Oncol. 2015, 16(4): e173-18. Wu F HK, Huang X, et al. Nomogram model based on preoperative clinical characteristics of unilateral papillary thyroid carcinoma to predict contralateral medium-volume central lymph node metastasis [J]. Front Endocrinol (Lausanne). 2024, 14(1271). Chen F JS, Yao F, et al. A nomogram based on clinicopathological and ultrasound characteristics to predict central neck lymph node metastases in papillary thyroid cancer [J]. Front Endocrinol (Lausanne). 2024, 14: 1267494. Qiu P GQ, Pan K, et al. Development of a nomogram for prediction of central lymph node metastasis of papillary thyroid microcarcinoma [J]. BMC Cancer. 2024, 24(1): 235. Feng JW LS, Qi GF,et al. Development and Validation of Clinical-Radiomics Nomogram for Preoperative Prediction of Central Lymph Node Metastasis in Papillary Thyroid Carcinoma [J]. Acad Radiol. 2024, S1076-6332(23): 00682-7. Lu J LJ, Chen Y, et al. Risk factor analysis and prediction model for papillary thyroid carcinoma with lymph node metastasis [J]. Front Endocrinol (Lausanne). 2023, 14: 1287593. Lambin P R-VE, Leijenaar R, et al. . Radiomics: Extracting more information from medical images using advanced feature analysis [J]. Eur J Cancer. 2012, 48(4): 441-46. Avanzo M WL, Stancanello J, et al. Machine and deep learning methods for radiomics [J]. Med Phys. 2020, 47(5): :e185-e202. Huang J LZ, Zhong Q, Fang J, et al. Developing and validating a multivariable machine learning model for the preoperative prediction of lateral lymph node metastasis of papillary thyroid cancer [J]. Gland Surg. 2023, 12(1): 101-9. Lu WW ZD, Ni XJ. A Review of the Role of Ultrasound Radiomics and Its Application and Limitations in the Investigation of Thyroid Disease. Med Sci Monit [J]. Med Sci Monit. 2022, 28: e937738. Wang X AE, Ren Y, et al. A Radiomic Nomogram for the Ultrasound-Based Evaluation of Extrathyroidal Extension in Papillary Thyroid Carcinoma [J]. Front Oncol 2021, 11: 625646. Fan F LF, Wang Y, et al. Integration of ultrasound-based radiomics with clinical features for predicting cervical lymph node metastasis in postoperative patients with differentiated thyroid carcinoma [J]. Endocrine. 2023. Zhang M LS, Yang L, et al. A nomogram based on ultrasound radiomics for predicting the invasiveness of cN0 single papillary thyroid microcarcinoma [J]. Gland Surg. 2023, 12(12): 1735-45 Jiang L ZZ, Guo S, et al. Clinical-Radiomics Nomogram Based on Contrast-Enhanced Ultrasound for Preoperative Prediction of Cervical Lymph Node Metastasis in Papillary Thyroid Carcinoma [J]. Cancers (Basel). 2023, 15(5): 1613. HajiEsmailPoor Z KZ, Tabnak P. Radiomics diagnostic performance in predicting lymph node metastasis of papillary thyroid carcinoma: A systematic review and meta-analysis [J]. Eur J Radiol. 2023, 168: 111129. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6733396","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":478214410,"identity":"d424ffb4-ccdf-4fea-a924-132eb1ccdfec","order_by":0,"name":"Peng Zhao","email":"","orcid":"","institution":"Guangxi Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Peng","middleName":"","lastName":"Zhao","suffix":""},{"id":478214411,"identity":"5d7ee3e6-2874-4d7a-affb-6e9679881ea8","order_by":1,"name":"Lulu Liang","email":"","orcid":"","institution":"The Second Nanning People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Lulu","middleName":"","lastName":"Liang","suffix":""},{"id":478214412,"identity":"f1c0f071-ec6d-477d-9e5f-9ec7e77c10ed","order_by":2,"name":"Xian Wei","email":"","orcid":"","institution":"Guangxi Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xian","middleName":"","lastName":"Wei","suffix":""},{"id":478214413,"identity":"aad0e164-fb20-48d3-bee6-6682adbd1bd3","order_by":3,"name":"Duoping Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYLACHiBmY2BgfJBQUUOaFmaDB2eOkaAFpEvyYQszYdUGx88efvG2zSaPT7r9WkViAxsDf3t3An4tZ/LSLOe2pRWzyZwpu5G4Q4ZB4szZDXi1mB3IMTPmbTuc2CaRk3Yj8Qwbg4FELgEt598gtBQktjEToeVGjvFjiJb0YwxEabG/8caMcc45oF8kcpglEs4c4yHoF8n+HOMPb8ps8uRnpD/8+KOiRo6/vRe/FiBgkwASCcDYMQDxeAgpBwHmDxAt7A+IUT0KRsEoGAUjEAAAlNhK74F0QysAAAAASUVORK5CYII=","orcid":"","institution":"Guangxi Medical University Cancer Hospital","correspondingAuthor":true,"prefix":"","firstName":"Duoping","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2025-05-23 13:38:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6733396/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6733396/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85827116,"identity":"db8d8125-4835-466e-b4e5-68447f22a1bf","added_by":"auto","created_at":"2025-07-02 07:22:09","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1577422,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart of case selection and technical route in this study\u003c/p\u003e","description":"","filename":"figure01.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/d79317300614a12300d66929.jpg"},{"id":85828437,"identity":"a7723f76-ba3a-4f5b-b9b0-75973dee403f","added_by":"auto","created_at":"2025-07-02 07:30:09","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1265772,"visible":true,"origin":"","legend":"\u003cp\u003eColumnar plots for predicting central lymph node metastasis for T1 PTC\u003c/p\u003e","description":"","filename":"figure02.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/9d98fe5fd14977e0ddf3e45d.jpg"},{"id":85826378,"identity":"7be7cbfa-8127-4975-95b1-8a694ee1af89","added_by":"auto","created_at":"2025-07-02 07:14:09","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":982813,"visible":true,"origin":"","legend":"\u003cp\u003eLASSO regression feature screening\u003c/p\u003e","description":"","filename":"figure03.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/0c6eaf4c7efec224362f3f7f.jpg"},{"id":85827118,"identity":"afb0426b-3949-4a7f-8da1-faf41a80df02","added_by":"auto","created_at":"2025-07-02 07:22:09","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":973487,"visible":true,"origin":"","legend":"\u003cp\u003eCombined prediction model diagram\u003c/p\u003e","description":"","filename":"figure04.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/d75377ee99c10adda23ba268.jpg"},{"id":85826382,"identity":"4cc4d39a-c904-4cb3-9ba8-87727ed5ede7","added_by":"auto","created_at":"2025-07-02 07:14:10","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":873801,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of ROC curves among the three models\u003c/p\u003e","description":"","filename":"figure05.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/e579bce331ea550627149655.jpg"},{"id":85826380,"identity":"71b11b1c-396f-40df-b2a8-14b0c9297959","added_by":"auto","created_at":"2025-07-02 07:14:09","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":925754,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of DCA curves among the three models\u003c/p\u003e","description":"","filename":"figure06.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/9a9a2cd13c57ede2adfb9369.jpg"},{"id":85827119,"identity":"f7057274-e6cd-4102-b9bb-9b24efd96949","added_by":"auto","created_at":"2025-07-02 07:22:10","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":909395,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of calibration curves among the three models\u003c/p\u003e","description":"","filename":"figure07.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/560dfb5be4f7992f2c361b29.jpg"},{"id":99789234,"identity":"53a5356a-0a50-4725-ba79-87104e46b6e8","added_by":"auto","created_at":"2026-01-08 12:49:09","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":8789141,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6733396/v1/8dc1c481-d686-4aa4-a320-27ffd9722db5.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eConstruction of a Predictive Model for T1-stage Low-risk Papillary Thyroid Carcinoma with Central Lymph Node Metastasis Using Ultrasound Radiomics Combined with Clinical Radiomics\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003ePapillary thyroid carcinoma (PTC) is the most common tumor subtype of thyroid cancer. Studies have revealed that approximately 30\u0026ndash;90% of patients with PTC exhibit occult central lymph node metastasis (CLNM) on postoperative pathological examination [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Therefore, CLNM is closely associated with local tumor recurrence in patients with PTC after surgery [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Factors influencing lymph node metastasis in PTC include sex, age, number of cancer foci, tumor size, vascular invasion (VI), extrathyroidal extension (ETE), and the presence of \u003cem\u003eBRAF V600E\u003c/em\u003e mutations [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRecent studies have identified a positive correlation between tumor size and CLNM in patients with PTC, especially in tumors with diameters\u0026thinsp;\u0026gt;\u0026thinsp;2 cm [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Furthermore, Han et al. reported that a tumor diameter of \u0026ge;\u0026thinsp;2 cm was an independent risk factor affecting the 5-year postoperative survival rate of patients postoperatively [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. However, some tumors\u0026thinsp;\u0026lt;\u0026thinsp;2 cm (T1 stage), including papillary thyroid microcarcinoma (PTMC), which has a tumor diameter\u0026thinsp;\u0026lt;\u0026thinsp;1 cm, may also be highly invasive and prone to CLNM [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Therefore, the World Health Organization (WHO) emphasizes that low-risk PTC should not solely be defined by tumor size but should also consider tumor invasiveness and other imaging and clinical features [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. However, the invasive nature of tumors is often fully assessed using only postoperative pathological reports. Although there is currently no consensus on whether routine prophylactic central compartment lymph node dissection (pCCND) should be performed during surgery for low-risk thyroid cancer, constructing a preoperative, non-invasive, and clinically applicable predictive model for patients with low-risk PTC would provide valuable guidance for clinical decision-making and reduce the occurrence of overtreatment.\u003c/p\u003e \u003cp\u003eThe development of nomograms based on clinical data has proven to be a simple, feasible, and reproducible approach, yielding promising predictive results [\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. However, an increasing number of patients with low-risk T1 stage PTC are being diagnosed with CLNM on fine needle aspiration biopsy (FNAB) during routine checkups [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Many existing predictive models for CLNM include both patients with stage T1 and T2, and fail to accurately reflect the distinct biological characteristics of T1 stage PTC. Research focusing specifically on CLNM in patients with T1 stage PTC remains limited.\u003c/p\u003e \u003cp\u003eTherefore, this study aimed to explore the clinical application of a preoperative predictive model for CLNM in low-risk patients with T1 stage PTC. The model was constructed using a combination of ultrasound radiomics and clinical radiomics. Furthermore, we validated the accuracy of this model through prospective studies.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1. Patients and data collection\u003c/h2\u003e \u003cp\u003eThis retrospective study collected data from patients with thyroid cancer who were hospitalized in the Department of Head and Neck Surgery at the Affiliated Cancer Hospital of Guangxi Medical University between June 2020 and June 2022, and between February 2023 and August 2023. All patients underwent routine two‑dimensional ultrasound. The inclusion criteria were as follows: 1) all patients underwent preoperative FNAB of the thyroid nodules, with \u003cem\u003eBRAF V600E\u003c/em\u003e gene testing (fluorescence quantitative polymerase chain reaction); 2) postoperative pathology confirmed unilateral T1 stage PTC; 3) preoperative imaging did not indicate any obvious metastatic lymph nodes in the neck, or any suspicious lymph nodes identified were confirmed to be negative for metastasis through FNAB; 4) all included patients underwent hemithyroidectomy (HT) and pCCND; and 5) none of the patients received prior treatment for other cancers, including radiotherapy, chemotherapy, or any other treatments. The exclusion criteria were as follows: 1) pathology results that could not be identified as PTC or patients confirmed to have lateral neck lymph node metastasis preoperatively through FNAB; 2) incomplete or missing postoperative pathological reports; and 3) poor image quality preventing the extraction of radiomic features.\u003c/p\u003e \u003cp\u003eA total of 237 patients were included in this study. Of these, 191 patients were treated between June 2020 and June 2022, and 46 patients were treated between February 2023 and August 2023. Using a random number method, the 191 patients were randomly divided into a training set (134 patients) and a validation set (57 patients) in a 7:3 ratio. The 46 patients from the later period were used as the prospective validation set.\u003c/p\u003e \u003cp\u003eThis study was approved by the Ethics Committee of the Affiliated Cancer Hospital of Guangxi Medical University (approval no. KY2025625). The requirement for individual consent for this retrospective analysis was waived. The details of the patient selection process and grouping are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2. Thyroid and cervical ultrasound examination\u003c/h2\u003e \u003cp\u003eAll patients underwent comprehensive preoperative thyroid and cervical ultrasound examinations performed by senior ultrasound specialists at our hospital. Ultrasound images were acquired using the GE Logiq E9 Ultrasound Diagnostic System, TOSHIBA Aplio 500 Ultrasound Diagnostic System, TOSHIBA Aplio 400 Ultrasound Diagnostic System, and CANON Aplio i800 Ultrasound Diagnostic System.\u003c/p\u003e \u003cp\u003eThe examinations were performed with the patient in the supine position, with the shoulders elevated and the head tilted downward. High-frequency 5\u0026ndash;12 MHz ultrasound probes were used to scan the thyroid and cervical lymph nodes. The following thyroid nodule parameters were measured: maximum transverse diameter, maximum longitudinal diameter, height, nodule margin, presence of microcalcifications, texture, longitudinal-to-transverse (L/T) ratio, and presence of ETE. Thyroid nodules were classified in this study according to the Thyroid Imaging Reporting and Data System in China (C-TIRADS) [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Additionally, the central and lateral cervical neck compartments were carefully examined for enlarged or abnormally shaped lymph nodes [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3. Surgical procedure\u003c/h2\u003e \u003cp\u003eAll patients in the training, validation, and prospective validation sets underwent HT\u0026thinsp;+\u0026thinsp;pCCND under general anesthesia via open surgery or an endoscopic approach. All surgeries were performed by senior attending physicians at the Department of Head and Neck Surgery.\u003c/p\u003e \u003cp\u003eThe scope of lymph node dissection included the following boundaries: the upper boundary reached the level of the thyroid cartilage, the lower boundary extended to the superior border of the thymus or the level of the brachiocephalic trunk, the lateral boundary was the medial side of the carotid sheath on both sides, and the deep boundary was the deep layer of the cervical fascia and the anterior wall of the esophagus [\u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. This area encompasses the Delphian, pretracheal, and tracheoesophageal groove lymph nodes. The lymph nodes behind the recurrent laryngeal nerve (LN-prRLN) were included for right-sided neck lymph node dissection. Lymph nodes were thoroughly removed during pCCND.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4. Extraction and selection of ultrasound radiomic features\u003c/h2\u003e \u003cp\u003eUltrasound images were extracted from the electronic medical records of the hospital. The ITK-SNAP medical image segmentation tool identified the region of interest (ROI) [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. ITK-SNAP has been widely used for the image segmentation of two-dimensional or three-dimensional imaging data, such as computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and positron emission tomography (PET)-CT [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e], and has demonstrated good performance in segmenting thyroid ultrasound images [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSubsequently, the open-source Pyradiomics module in Python (v3.8.10, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.python.org\u003c/span\u003e\u003cspan address=\"http://www.python.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was used to extract 851 radiomic features from the ultrasound images, including both the original and filtered (wavelet) features. The extracted radiomic features were categorized into several major groups: first-order statistics, shape features, gray-level co-occurrence matrix (GLCM), gray-level run-length matrix (GLRLM), gray-level size zone matrix (GLSZM), and neighboring gray-tone difference matrix (NGTDM).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Construction and evaluation of the radiomics, clinical, and combined models\u003c/h2\u003e \u003cp\u003eThe radiomics, clinical, and combined models were constructed using logistic regression. The diagnostic performance of each model was evaluated using the receiver operating characteristic (ROC) curves. Decision curve analysis (DCA) and calibration curves were used to assess the clinical applicability and values of the three models. The goodness-of-fit test was conducted using the Hosmer\u0026ndash;Lemeshow test.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e2.6. Statistical analysis\u003c/b\u003e\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe univariate analysis was performed based on the data type. The chi-squared test or Fisher\u0026rsquo;s exact test was used for categorical data. The t-test was applied for continuous data, and the rank-sum test was used for ordinal data. In the univariate analysis, a p-value of \u0026lt;\u0026thinsp;0.05 was considered statistically significant and was included in the multivariate regression analysis. Multivariate regression analysis was performed using binary logistic regression. All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Nomograms, DCA, and calibration curves were generated using the R software (version 4.2.1; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.r-project.org/\u003c/span\u003e\u003cspan address=\"http://www.r-project.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). All statistical tests were two-sided; a p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Baseline characteristics\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eA total of 191 patients with T1 stage PTC between June 2020 and June 2022 were included. Of these, 48 were male and 143 were female. The patients were randomly divided into a training set (134 patients) and a validation set (57 patients) in a 7:3 ratio using random number generation. The training set included 33 male and 101 female patients, with a mean age of 42.55\u0026thinsp;\u0026plusmn;\u0026thinsp;9.41 years. There were 15 male and 42 female patients in the validation set, with a mean age of 42.87\u0026thinsp;\u0026plusmn;\u0026thinsp;10.86 years. From February 2023 to August 2023, an additional 46 patients with T1 stage PTC were enrolled as the prospective validation cohort, consisting of nine male and 37 female patients, with a mean age of 42.65\u0026thinsp;\u0026plusmn;\u0026thinsp;10.16 years.\u003c/p\u003e \u003cp\u003eThe baseline characteristics and grouping details of the training, validation, and prospective validation sets are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline data of patients in the training set, validation set and prospective validation set\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003etraining set\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003evalidation set\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eprospective validation set\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;134)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;57)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;46)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;237)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e33.0 (24.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15.0 (26.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e9.00 (19.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e57.0 (24.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.875\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e101 (75.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42.0 (73.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e37.0 (80.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e180 (75.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e118 (88.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e48.0 (84.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e38.0 (82.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e204 (86.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.785\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e16.0 (11.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.00 (15.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e8.00 (17.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e33.0 (13.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily History\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e120 (89.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e53.0 (93.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e42.0 (91.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e215 (90.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.901\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14.0 (10.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.00 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.00 (8.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e22.0 (9.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHyperthyroidism\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e129 (96.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54.0 (94.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e42.0 (91.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e225 (94.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.623\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e5.00 (3.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.00 (5.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.00 (8.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e12.0 (5.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTSH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;0.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e126 (94.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.0 (87.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e42.0 (91.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e218 (92.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.533\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;0.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e8.00 (6.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7.00 (12.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.00 (8.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e19.0 (8.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBRAF mutation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e35.0 (26.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.0 (22.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e11.0 (23.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e59.0 (24.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.967\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e99.0 (73.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e44.0 (77.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e35.0 (76.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e178 (75.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eC-TIRRADS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e13.0 (9.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.00 (8.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.00 (8.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e22.0 (9.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.917\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e91.0 (67.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42.0 (73.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e35.0 (76.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e168 (70.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e18.0 (13.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8.00 (14.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.00 (13.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e32.0 (13.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12.0 (9.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.00 (3.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.00 (2.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e15.0 (6.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEcho\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e103 (76.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49.0 (86.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e29.0 (63.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e181 (76.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.224\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e28.0 (20.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8.00 (14.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e15.0 (32.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e51.0 (21.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3.00 (2.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.00 (4.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.00 (2.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMargin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRegular\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e39.0 (29.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.0 (17.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e16.0 (34.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e65.0 (27.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.237\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIrregular\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e95.0 (70.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47.0 (82.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e30.0 (65.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e172 (72.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e(Continued)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003etraining set\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003evalidation set\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003eprospective validation set\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" morerows=\"1\" nameend=\"c10\" namest=\"c9\" rowspan=\"2\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;134)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;57)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;46)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;237)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAspect Ratio\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e75.0 (56.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e30.0 (52.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e31.0 (67.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e136 (57.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e0.472\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e59.0 (44.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e27.0 (47.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e15.0 (32.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e101 (42.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e69.0 (51.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e22.0 (38.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e18.0 (39.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e109 (46.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e0.289\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e65.0 (48.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e35.0 (61.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e28.0 (60.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e128 (54.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSolid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.0 (9.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e6.00 (10.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e5.00 (10.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e24.0 (10.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e0.996\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e121 (90.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e51.0 (89.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e41.0 (89.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e213 (89.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean (SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.95 (1.86)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e1.68 (0.474)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.16 (2.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e1.93 (1.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e0.551\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Min, Max]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.75 [0.572, 21.9]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e1.59 [0.841, 2.97]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e1.59 [0.624, 11.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e1.65 [0.572, 21.9]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean (SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e132 (58.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e122 (42.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e138 (115)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e130 (70.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e0.676\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Min, Max]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e121 [48.7, 584]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e113 [2.15, 231]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e115 [38.2, 822]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e118 [2.15, 822]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLMR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean (SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.71 (1.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e6.03 (1.99)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e4.99 (1.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e5.65 (1.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e0.055\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Min, Max]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.33 [1.37, 11.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e6.03 [1.80, 12.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e4.61 [1.08, 10.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e5.48 [1.08, 12.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSII\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean (SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e517 (761)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e415(135)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e516(493)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e492(615)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.755\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Min, Max]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e415 [118, 8960]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e399[7.89, 715]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e403[96.0, 2980]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e410[7.89, 8960]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCLNM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97.0 (72.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e44.0 (77.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e27.0 (58.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e168 (70.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e0.207\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e37.0 (27.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e13.0 (22.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e19.0 (41.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e69.0 (29.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Construction of the clinical model in the training set\u003c/h2\u003e \u003cp\u003eThe univariate analysis (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e2\u003c/span\u003e) identified sex, indistinct or irregular margins, microcalcifications, and an L/T ratio\u0026thinsp;\u0026gt;\u0026thinsp;1 as risk factors for CLNM in patients with T1 stage PTC. These four risk factors were subjected to multivariate analysis using binary logistic regression (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e3\u003c/span\u003e), which revealed that sex, indistinct or irregular margins, and microcalcifications were independent risk factors for CLNM in T1 stage PTC. The diagnostic performance of the clinical model was evaluated using ROC curves for the training, validation, and prospective validation sets. The area under the curve (AUC) values for these sets were 0.740, 0.656, and 0.626, respectively. A nomogram was generated to visually represent the clinical model (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUnivariate Analysis of Factors Associated with CLNM in the Training set\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eS.E\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWald\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eExpB\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eLower\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eUpper\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.916\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.425\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.031\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.400\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.174\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.920\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-1.082\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.782\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.914\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.166\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.339\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.073\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.570\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily history\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.424\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.595\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.508\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.476\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.528\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.476\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4.901\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHyperthyroidism\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.437\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.135\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.148\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.700\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.646\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.070\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.975\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eC-TIRADS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.448\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.217\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.455\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.954\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.329\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.127\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4.286\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.661\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e27.785\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.466\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.812\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.330\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.566\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.594\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.324\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e7.835\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.157\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.913\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.607\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.205\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.182\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.531\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e19.051\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEcho\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.972\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.615\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.447\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.454\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.972\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.324\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.564\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.643\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.804\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-20.168\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23205.422\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.999\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTSH(0\u0026thinsp;\u0026ge;\u0026thinsp;0.35)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.485\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.757\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.409\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.522\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.624\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.368\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e7.164\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBRAF Mutation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.130\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.447\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.085\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.770\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.139\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.475\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.733\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMargin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.538\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.570\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.282\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4.657\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.524\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e14.236\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMicrocalcifications\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.257\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.415\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e9.174\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.515\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.558\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e7.929\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAspect Ratio \u0026gt;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.866\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.395\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.818\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.028\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2.378\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.097\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.156\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSolid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.264\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.689\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.147\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.701\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.303\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.338\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.024\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.145\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.133\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.197\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.274\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.156\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.892\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.499\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.930\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.994\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.007\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLMR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.102\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.979\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.821\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.225\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSII\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.056\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.304\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMultivariable Analysis of Factors Associated with CLNM in the Training set\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eS.E\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWald\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eExpB\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eLower\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eUpper\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.085\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.473\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.260\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2.959\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.171\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e7.477\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMargin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-1.379\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.615\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.252\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.075\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.841\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMicrocalcifications\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.976\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.446\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.785\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.377\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.157\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAspect Ratio \u0026gt;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.572\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.426\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.805\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.179\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.564\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.245\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.300\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConstant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.255\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.349\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.533\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.465\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.775\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.3. Construction of the ultrasound radiomics model in the training set\u003c/h2\u003e \u003cp\u003eThe 851 ultrasound radiomic features were first standardized using z-score normalization and then subjected to univariate analysis for feature selection, resulting in 201 features (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). After performing five-fold cross-validation with the least absolute shrinkage and selection operator (LASSO) regression, 12 ultrasound radiomic features were selected (Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The Rad score was calculated by summing the feature coefficients and a predictive model was built using the Rad score via a logistic regression algorithm. The AUC values for the training, validation, and prospective validation sets were 0.794, 0.720, and 0.766, respectively.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUltrasonic radiomic features\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"1\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiomic features\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eoriginal_shape_Maximum 2D Diameter Slice\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eoriginal_shape_Maximum 3D Diameter\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eoriginal_firstorder_Robust Mean Absolute Deviation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HLH_glcm_Difference Entropy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HLH_glcm_Difference Variance\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HLH_glrlm_Gray Level Non-Uniformity Normalized\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HLH_glszm_Small Area Low Gray Level Emphasis\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HHH_glcm_Cluster Prominence\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HHH_glszm_High Gray Level Zone Emphasis\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HHH_glszm_Low Gray Level Zone Emphasis\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-HHH_glszm_Zone Entropy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ewavelet-LLL_firstorder_Interquartile Range\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.4. Construction of the combined ultrasound radiomics\u0026ndash;clinical model\u003c/h2\u003e \u003cp\u003eThe Rad score derived from ultrasound radiomics features was combined with clinical characteristics, such as sex, margin irregularity, and microcalcifications, which were used to construct a combined predictive model via logistic regression. A nomogram for this combined predictive model was generated as a visual representation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe AUC values for the combined models were 0.850, 0.750, and 0.786, respectively. Additionally, the goodness of fit of the combined model was assessed using the Hosmer\u0026ndash;Lemeshow test. The results showed a p-value of 0.979, suggesting that the model fit the data well.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.5. Comparison of the clinical, ultrasound radiomics, and combined models\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe clinical, ultrasound radiomics, and combined models were compared based on their diagnostic performance. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, which presents the ROC curve comparisons, the diagnostic performance of the models in the training, validation, and prospective validation sets were ranked. A comparison of the DCA and the calibration curve between these three models is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. Overall, the combined predictive model demonstrated superior performance in predicting the occurrence of CLNM in T1 stage low-risk PTC.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eClinical prediction models are commonly visualized using nomograms that graphically represent the impact of various predictors on outcomes [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Many studies have focused on predicting CLNM in patients with PTC based on clinical features and logistic regression models. Wu et al. [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] developed a nomogram for predicting contralateral CLNM based on clinical and pathological data from 204 patients with unilateral PTC and achieved an AUC of 0.836. Chen et al. [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] constructed a similar nomogram for CLNM prediction using data from 691 patients with PTC, with an AUC of 0.717. Qiu et al. [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e] also used clinical and ultrasound characteristics in a nomogram to predict CLNM in 119 patients with cN0 PTC, with AUCs of 0.703 and 0.656 in the training and validation sets, respectively. Although there are many studies on the prediction of CLNM in patients with PTC, few have specifically focused on patients with low-risk T1 PTC.\u003c/p\u003e \u003cp\u003eOur study only included T1 stage patients and identified independent risk factors for CLNM after univariate and multivariate analyses, including sex, nodal margin irregularity, and microcalcifications. These findings are consistent with the previous literature [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. The logistic regression model built using these three clinical features resulted in AUC values of 0.740 for the training set and 0.656 for the validation set, which was comparable to those of previous studies; however, our study stands out as we prospectively validated the clinical model which strengthens the reliability of our results. The AUC for the prospective validation set was 0.626. In summary, although the clinical model has a good predictive value for T1 PTC and CLNM, its diagnostic performance can be further improved by increasing the sample size and accounting for individual patient variability.\u003c/p\u003e \u003cp\u003eRadiomics has gained attention owing to its ability to discover clinically relevant features in medical images that are often undetectable by the human eye. Unlike traditional qualitative imaging, radiomics allows quantitative analysis, making it a powerful tool for predicting clinical outcomes [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Ultrasound is the preferred imaging modality for thyroid nodules, and ultrasound radiomics have been widely studied for thyroid cancer. Several studies have demonstrated its ability to accurately predict CLNM and tumor invasiveness in patients with PTC [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan additionalcitationids=\"CR35\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn our study, the ultrasound radiomics model was constructed using 12 features (three original image features and nine wavelet-filtered features) selected through LASSO regression and cross-validation. The Rad score was computed by summing the products of each feature coefficient and its corresponding value. The logistic regression model built on the Rad score yielded AUC values of 0.794, 0.720, and 0.766 for the training, validation, and prospective validation sets, respectively, demonstrating good predictive ability. Furthermore, the accuracy was 0.754 and 0.596 for the training and validation sets, respectively; however, after prospective validation, the accuracy improved to 0.761, which matched the training set results. Compared with the clinical model, the ultrasound radiomics model showed superior diagnostic performance, which is consistent with previous studies [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. This reinforces the ability of ultrasound radiomics features to reflect the biological characteristics of T1 low-risk PTC tumors. The constructed prediction model was validated to have good clinical value through DCA and calibration curves.\u003c/p\u003e \u003cp\u003eAs the need for more precise predictions grows, it has become clear that using a single-omics approach to construct prediction models often fails to achieve optimal results. With advancements in computational technology and artificial intelligence, integrated multi-omics models have become a mainstream trend. Similar integrated models have been reported for the prediction of CLNM in PTC. For instance, Zhang et al. [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e] combined ultrasound radiomics and clinical data to predict occult central zone metastasis in cN0 PTC, achieving AUC values of 0.76 and 0.71 for the training and validation sets, respectively. Jiang et al. [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e] developed a combined model using ultrasound radiomics and clinical data from 211 patients with PTC, resulting in AUC values of 0.82 and 0.81 for the training and validation sets. Furthermore, Haji et al. [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e] found that models integrating radiomics and clinical data had a higher sensitivity and specificity than single-omics models, as shown in a meta-analysis of 25 studies.\u003c/p\u003e \u003cp\u003eIn our study, we combined three clinical features (sex, margin irregularity, and microcalcification) and the Rad score derived from ultrasound radiomics using logistic regression to build a combined prediction model. The AUCs for the combined model were 0.850 for the training set, 0.750 for the validation set, and 0.786 for the prospective validation set, indicating superior predictive performance compared with the clinical and ultrasound radiomics models. The model was visualized using nomograms. The goodness-of-fit test (Hosmer\u0026ndash;Lemeshow) for the combined model showed no significant difference (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05), suggesting that the model fit well. The DCA and calibration curves further supported the clinical applicability of the combined model compared to the clinical and ultrasound radiomics models (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn this study, we visualized clinical and combined prediction models using nomograms and evaluated their diagnostic efficacy using ROC, DCA, and calibration curves. The results showed that the combined model outperformed the clinical and ultrasound radiomics models in all datasets (training, validation, and prospective validation), with AUC values of 0.850, 0.759, and 0.786, respectively. This combined model can accurately predict whether patients with low-risk T1 PTC will develop CLNM preoperatively, providing an objective basis for deciding whether prophylactic central neck dissection is needed, potentially preventing overtreatment, and offering valuable clinical guidance.\u003c/p\u003e \u003cp\u003eThis study has some limitations. First, this was a single-center retrospective study, and while prospective validation was conducted within the same center, the small sample size limited the generalizability of the results. Further multi-center studies are needed to externally validate these findings. Second, the images were obtained from four different ultrasound devices, which may have introduced some variability in image quality and parameters, potentially affecting the extraction of the ROI and feature interpretation. Although high-level ultrasound specialists performed the examinations, the inherent subjectivity of ultrasound reporting may have introduced a measurement bias. Finally, as with any radiomics study, feature extraction methods are manually defined, which could lead to biases in the tumor image representation.\u003c/p\u003e \u003cp\u003eIn conclusion, the combined model integrating ultrasound radiomics and clinical features offers a more accurate and clinically applicable tool for predicting CLNM in patients with T1-stage low-risk PTC. This model can assist in preoperative decision-making for pCCND and reduce the occurrence of overtreatment.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eArea under the curve (AUC)\u003c/p\u003e\n\u003cp\u003eCentral lymph node metastasis (CLNM)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eComputed tomography (CT)\u003c/p\u003e\n\u003cp\u003eDecision curve analysis (DCA)\u003c/p\u003e\n\u003cp\u003eExtrathyroidal extension (ETE)\u003c/p\u003e\n\u003cp\u003eFine needle aspiration biopsy (FNAB)\u003c/p\u003e\n\u003cp\u003eGray-level co-occurrence matrix (GLCM)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGray-level run-length matrix (GLRLM)\u003c/p\u003e\n\u003cp\u003eGray-level size zone matrix (GLSZM)\u003c/p\u003e\n\u003cp\u003eHemithyroidectomy (HT)\u003c/p\u003e\n\u003cp\u003eLeast absolute shrinkage and selection operator (LASSO)\u003c/p\u003e\n\u003cp\u003eLongitudinal-to-transverse (L/T)\u003c/p\u003e\n\u003cp\u003eLymph nodes located behind the recurrent laryngeal nerve (LN-prRLN)\u003c/p\u003e\n\u003cp\u003eMagnetic resonance imaging (MRI)\u003c/p\u003e\n\u003cp\u003eNeighboring gray-tone difference matrix (NGTDM)\u003c/p\u003e\n\u003cp\u003eProphylactic central compartment lymph node dissection (pCCND)\u003c/p\u003e\n\u003cp\u003ePapillary thyroid carcinoma (PTC)\u003c/p\u003e\n\u003cp\u003ePapillary thyroid microcarcinoma (PTMC)\u003c/p\u003e\n\u003cp\u003ePositron emission tomography (PET)\u003c/p\u003e\n\u003cp\u003eReceiver operating characteristic (ROC)\u003c/p\u003e\n\u003cp\u003eRegion of interest (ROI)\u003c/p\u003e\n\u003cp\u003eThyroid Imaging Reporting and Data System in China (C-TIRADS)\u003c/p\u003e\n\u003cp\u003eVascular invasion (VI)\u003c/p\u003e\n\u003cp\u003eWorld Health Organization (WHO)\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank Editage (www.editage.cn) for English language editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePZ and LL selected the relevant studies, assessed the data, and wrote the manuscript. \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; XW contributed to the methodological framework. DW revised the manuscript. All authors read and approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study received no funding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthical approval for this study was granted by the Medical Ethics Committee of Guangxi Medical University Cancer Hospital (Approval Number: KY2025625) in accordance with the Declaration of Helsinki. Informed consent was obtained from all participants.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no conflict of interest to disclose.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eZhu J ZJ, Li L, et al. Application of Machine Learning Algorithms to Predict Central Lymph Node Metastasis in T1-T2, Non-invasive, and Clinically Node Negative Papillary Thyroid Carcinoma [J]. Front Med (Lausanne). 2021, 9: 8:635771.\u003c/li\u003e\n\u003cli\u003eForest VI, Clark, J.R, Ebrahimi, A. et al. Central compartment dissection in thyroid papillary carcinoma [J]. Ann Surg. 2011, 253: 123-30.\u003c/li\u003e\n\u003cli\u003eIto Y KT, Kobayashi K, et al. Prognostic factors for recurrence of papillary thyroid carcinoma in the lymph nodes, lung, and bone: analysis of 5,768 patients with average 10-year follow-up [J]. World J Surg. 2012, 36(6): 1274-8.\u003c/li\u003e\n\u003cli\u003eLI Hong-qiang, MA Run-sheng, TANG Yi-feng, et al. Discussion on the necessity of total thyroidectomy and central lymph node dissection in differentiated thyroid carcinoma with stage T1-2N0M0 [J]. Chinese Journal of Practical Surgery. 2021, 41(08): 892-4+9.\u003c/li\u003e\n\u003cli\u003eMao J ZQ, Zhang H, et al. Risk Factors for Lymph Node Metastasis in Papillary Thyroid Carcinoma: A Systematic Review and Meta-Analysis [J]. Front Endocrinol (Lausanne). 2020, 11: 265.\u003c/li\u003e\n\u003cli\u003eZheng H LV, Lu J, et al. Clinical Factors Predictive of Lymph Node Metastasis in Thyroid Cancer Patients: A Multivariate Analysis [J]. J Am Coll Surg. 2022, 234(4): 691-700.\u003c/li\u003e\n\u003cli\u003eHuang JH. Predictive Factor of Large ‐ Volume Central Lymph Node Metastasis in Clinical N0 Papillary Thyroid Carcinoma Patients [D]. Shan Tou University. 2022.\u003c/li\u003e\n\u003cli\u003eHAN Yuren，LI Limei，WANG Ｒui. Clinical Characteristics and Prognostic Factors of Thyroid Cancer Patients [J]. The Practical Journal of Cancer. 2022, 37(06): 1000-2.\u003c/li\u003e\n\u003cli\u003eNie X TZ, Ge M, et al. Risk factors analyses for lateral lymph node metastases in papillary thyroid carcinomas: a retrospective study of 356 patients [J]. Arch Endocrinol Metab. 2016, 60(5): 492-9.\u003c/li\u003e\n\u003cli\u003eBaloch ZW AS, Barletta JA,et al. Overview of the 2022 WHO Classification of Thyroid Neoplasms [J]. Endocr Patho. 2022, 33(1): 27-63.\u003c/li\u003e\n\u003cli\u003ePopović Krneta M \u0026Scaron;\u0026Scaron;D, Mijatović Teodorović L, et al. Prediction of Cervical Lymph Node Metastasis in Clinically Node-Negative T1 and T2 Papillary Thyroid Carcinoma Using Supervised Machine Learning Approach [J]. J Clin Med. 2023, 24(11): 3641.\u003c/li\u003e\n\u003cli\u003eZhou SC LT, Zhou J, et al. An Ultrasound Radiomics Nomogram for Preoperative Prediction of Central Neck Lymph Node Metastasis in Papillary Thyroid Carcinoma [J]. Front Oncol. 2020, 10: 1591.\u003c/li\u003e\n\u003cli\u003eGong Y ZZ, Tang K, et al. Multimodal predictive factors of metastasis in lymph nodes posterior to the right recurrent laryngeal nerve in papillary thyroid carcinoma [J]. Front Endocrinol (Lausanne). 2023, 14: 1187825.\u003c/li\u003e\n\u003cli\u003eSociety of Tumor Ablation Therapy of the Chinese Anti‑Cancer Association, the Ablation Expert Committee of the Chinese Society of Clinical Oncology (CSCO), Chinese Medical Doctor Association College of Interventionalists Tumor Ablation Committee, et al. Expert consensus on thermal ablation of papillary thyroid cancer (2024 edition) [J]. Chin J Intern Med. 2024, 63(04): 355-64.\u003c/li\u003e\n\u003cli\u003eHaddad RI BL, Ball D, et al. Thyroid Carcinoma, Version 2.2022, NCCN Clinical Practice Guidelines in Oncology [J]. J Natl Compr Canc Netw. 2022, 20(8): 925-51.\u003c/li\u003e\n\u003cli\u003eTessler FN MW, Grant EG. Thyroid Imaging Reporting and Data System (TI-RADS): A User\u0026apos;s Guide [J]. Radiology. 2018, 287(1): 29-36.\u003c/li\u003e\n\u003cli\u003eKim DH KS, Basurrah MA, et al. Diagnostic Performance of Six Ultrasound Risk Stratification Systems for Thyroid Nodules: A Systematic Review and Network Meta-Analysis [J]. AJR Am J Roentgenol. 2023, 220(6): 791-803.\u003c/li\u003e\n\u003cli\u003eChinese Society of Endocrinology; Thyroid and Metabolism Surgery Group of the Chinese Society of Surgery; China Anti-Cancer Association, et al. Guidelines for the Diagnosis and Management of Thyroid Nodules and Differentiated Thyroid Cancer (Second Edition) [J]. Int J Endocrinol Metab. 2023, 39(03): 181-226.\u003c/li\u003e\n\u003cli\u003eROBBINS K T SAＲ, MEDINA J E, et al. Consensusstatement on the classification and terminology of neck dissection [J]. Arch Otolaryngol Head Neck Surg. 2008, 134(5): 536-8.\u003c/li\u003e\n\u003cli\u003eGroup ATASW EA, Otolaryngology-Head A, et al. Consensus statement on the terminology and classification of central neck dissection for thyroid cancer [J]. Thyroid. 2009, 19(11): 1153-8.\u003c/li\u003e\n\u003cli\u003ePaul A. Yushkevich JP, et al. User-guided 3D active contour segmentation of anatomical structures: Significantly improved efficiency and reliability [J]. Neuroimage. 2006, 31(3): 1116-28.\u003c/li\u003e\n\u003cli\u003eYushkevich PA YG, Gerig G. ITK-SNAP: An interactive tool for semi-automatic segmentation of multi-modality biomedical images [J]. Annu Int Conf IEEE Eng Med Biol Soc. 2016: 3342-5.\u003c/li\u003e\n\u003cli\u003eStreet JS PA, Toma AK. Predicting vasospasm risk using first presentation aneurysmal subarachnoid hemorrhage volume: A semi-automated CT image segmentation analysis using ITK-SNAP [J]. PLoS One. 2023, 18(6): e0286485.\u003c/li\u003e\n\u003cli\u003eLu S RY, Lu C, et al. Radiomics features from whole thyroid gland tissue for prediction of cervical lymph node metastasis in the patients with papillary thyroid carcinoma [J]. J Cancer Res Clin Oncol. 2023, 149(14): 13005-16.\u003c/li\u003e\n\u003cli\u003eLi F PD, He Y, et al. Using ultrasound features and radiomics analysis to predict lymph node metastasis in patients with thyroid cancer [J]. BMC Surg. 2020, 20(1): 315.\u003c/li\u003e\n\u003cli\u003eBalachandran VP GM, Smith JJ, et al. . Nomograms in oncology: more than meets the eye [J]. Lancet Oncol. 2015, 16(4): e173-18.\u003c/li\u003e\n\u003cli\u003eWu F HK, Huang X, et al. Nomogram model based on preoperative clinical characteristics of unilateral papillary thyroid carcinoma to predict contralateral medium-volume central lymph node metastasis [J]. Front Endocrinol (Lausanne). 2024, 14(1271).\u003c/li\u003e\n\u003cli\u003eChen F JS, Yao F, et al. A nomogram based on clinicopathological and ultrasound characteristics to predict central neck lymph node metastases in papillary thyroid cancer [J]. Front Endocrinol (Lausanne). 2024, 14: 1267494.\u003c/li\u003e\n\u003cli\u003eQiu P GQ, Pan K, et al. Development of a nomogram for prediction of central lymph node metastasis of papillary thyroid microcarcinoma [J]. BMC Cancer. 2024, 24(1): 235.\u003c/li\u003e\n\u003cli\u003eFeng JW LS, Qi GF,et al. Development and Validation of Clinical-Radiomics Nomogram for Preoperative Prediction of Central Lymph Node Metastasis in Papillary Thyroid Carcinoma [J]. Acad Radiol. 2024, S1076-6332(23): 00682-7.\u003c/li\u003e\n\u003cli\u003eLu J LJ, Chen Y, et al. Risk factor analysis and prediction model for papillary thyroid carcinoma with lymph node metastasis [J]. Front Endocrinol (Lausanne). 2023, 14: 1287593.\u003c/li\u003e\n\u003cli\u003eLambin P R-VE, Leijenaar R, et al. . Radiomics: Extracting more information from medical images using advanced feature analysis [J]. Eur J Cancer. 2012, 48(4): 441-46.\u003c/li\u003e\n\u003cli\u003eAvanzo M WL, Stancanello J, et al. Machine and deep learning methods for radiomics [J]. Med Phys. 2020, 47(5): :e185-e202.\u003c/li\u003e\n\u003cli\u003eHuang J LZ, Zhong Q, Fang J, et al. Developing and validating a multivariable machine learning model for the preoperative prediction of lateral lymph node metastasis of papillary thyroid cancer [J]. Gland Surg. 2023, 12(1): 101-9.\u003c/li\u003e\n\u003cli\u003eLu WW ZD, Ni XJ. A Review of the Role of Ultrasound Radiomics and Its Application and Limitations in the Investigation of Thyroid Disease. Med Sci Monit [J]. Med Sci Monit. 2022, 28: e937738.\u003c/li\u003e\n\u003cli\u003eWang X AE, Ren Y, et al. A Radiomic Nomogram for the Ultrasound-Based Evaluation of Extrathyroidal Extension in Papillary Thyroid Carcinoma [J]. Front Oncol 2021, 11: 625646.\u003c/li\u003e\n\u003cli\u003eFan F LF, Wang Y, et al. Integration of ultrasound-based radiomics with clinical features for predicting cervical lymph node metastasis in postoperative patients with differentiated thyroid carcinoma [J]. Endocrine. 2023.\u003c/li\u003e\n\u003cli\u003eZhang M LS, Yang L, et al. A nomogram based on ultrasound radiomics for predicting the invasiveness of cN0 single papillary thyroid microcarcinoma [J]. Gland Surg. 2023, 12(12): 1735-45 \u003c/li\u003e\n\u003cli\u003eJiang L ZZ, Guo S, et al. Clinical-Radiomics Nomogram Based on Contrast-Enhanced Ultrasound for Preoperative Prediction of Cervical Lymph Node Metastasis in Papillary Thyroid Carcinoma [J]. Cancers (Basel). 2023, 15(5): 1613.\u003c/li\u003e\n\u003cli\u003eHajiEsmailPoor Z KZ, Tabnak P. Radiomics diagnostic performance in predicting lymph node metastasis of papillary thyroid carcinoma: A systematic review and meta-analysis [J]. Eur J Radiol. 2023, 168: 111129.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"papillary thyroid carcinoma, central lymph node metastasis, ultrasound radiomics, predictive models","lastPublishedDoi":"10.21203/rs.3.rs-6733396/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6733396/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003ePapillary thyroid carcinoma (PTC) is the most common tumor subtype of thyroid cancer and approximately 30\u0026ndash;90% of patients with PTC exhibit occult central lymph node metastasis (CLNM) on postoperative pathological examination. This study aimed to explore the diagnostic efficacy and clinical application value of combining ultrasound radiomics features with clinical features to construct predictive models for patients with papillary thyroid carcinoma (PTC) and central lymph node metastasis (CLNM).\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThis study included the retrospective data from total 191 PTC patients hospitalized between June 2020 and June 2022 (training set: 134, validation set: 57). Additionally, 46 patients were included in the prospective validation set. Clinical features affecting CLNM in patients with PTC were identified using univariate and multivariate analyses. Logistic regression models were constructed based on clinical and radiomics features, individually and combined. The diagnostic efficacies of the three models were compared using receiver operating characteristic curves, and a nomogram was constructed for visualization. The clinical utility of the model was evaluated using decision curve analysis (DCA) and calibration curves.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eMale sex, unclear or irregular margins, and microcalcifications were independent risk factors in the clinical radiomics predictive model. The area under the curve (AUC) for the training, validation, and prospective validation sets was 0.740, 0.656, and 0.626, respectively. Twelve ultrasound radiomics features were selected to construct the radiomics model (AUC: 0.794, 0.720, and 0.766, respectively). The combined model demonstrated AUCs of 0.850, 0.750, and 0.786, for training, validation and prospective validation set respectively. The DCA and calibration curves indicated that the combined model had a better diagnostic efficiency and clinical utility.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThis study presents a combined predictive model based on ultrasound radiomics and clinical features that can effectively predict preoperative CLNM in patients with PTC cN0 T1 stage, demonstrating its clinical applicability.\u003c/p\u003e","manuscriptTitle":"Construction of a Predictive Model for T1-stage Low-risk Papillary Thyroid Carcinoma with Central Lymph Node Metastasis Using Ultrasound Radiomics Combined with Clinical Radiomics","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-02 07:14:05","doi":"10.21203/rs.3.rs-6733396/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"54ee16a0-8430-4398-a2d5-4c8637f4b53c","owner":[],"postedDate":"July 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-02T13:09:31+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-02 07:14:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6733396","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6733396","identity":"rs-6733396","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}