Predictive model and clinical application for lymph node metastasis in papillary thyroid microcarcinoma

Predictive model and clinical application for lymph node metastasis in papillary thyroid microcarcinoma | 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 Predictive model and clinical application for lymph node metastasis in papillary thyroid microcarcinoma Yuanhao Su, Tingkai Sun, Yongke Wu, Cheng Li, Yunhao Li, Xing Jin, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4560286/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 Purpose Papillary thyroid microcarcinomas (PTMC), small tumors under 10 mm, represent a major part of the increase in papillary thyroid cancer cases. The treatment plans for PTMC patients with lymph node metastasi should be different from those without lymph node metastasis. Therefore, accurately identifying patients with cervical lymph node metastasis is of great clinical significance. Methods We analyzed data from 256 patients diagnosed with PTMC, using age, gender, tumor size, lesion count, and ACR score as predictors. Outcomes were based on cervical lymph node pathology. Four machine learning models—Random Forest, Multivariate Logistic Regression, Support Vector Machine, and Xgboost—were tested for their predictive accuracy and clinical utility. We then created an online website for direct prediction and designed online platforms that allow other researchers to upload their data for model building and prediction. The website and platform design is based on "shiny" package. Results The Random Forest model proved optimal, achieving an AUC of 0.92. It showed high sensitivity (0.83) and specificity (0.90) at the best threshold of 0.46. The link to the website we built based on this model is as follows: http://yucemoxing.online:8082 . Additionally, the link to the online platforms that allows userss to upload their own data for model building and prediction is as follows: http://yucemoxing.online:8081 , http://yucemoxing.site:8089 , http://yucemoxing.online:8084 , http://yucemoxing.online:8085 , http://yucemoxing.online:8083 , http://yucemoxing.online:8088 , http://yucemoxing.online:8087 , http://yucemoxing.online:8086 Conclusions Machine learning tools can reliably predict cervical lymph node metastasis in PTMC patients. The developed websites offer valuable tools for clinical application, enhancing decision-making in treatment strategies. PTMC Predictive model Machine learning Lymph node metastasis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Introduction Thyroid cancer is the most common endocrine malignancy worldwide, with papillary thyroid carcinoma (PTC) being the predominant histological subtype, accounting for over 80% of cases[ 1 – 3 ] and the notable rise in thyroid cancer incidence is primarily attributed to increased diagnoses of the PTC subtype[ 4 ]. Studies have reported its potential to progress towards a more malignant state, significantly reducing life expectancy[ 5 , 6 ]. Papillary thyroid microcarcinoma (PTMC) is defined as a PTC measuring 1 centimeter or less in its longest dimension, with reported lower metastasis rates[ 7 – 10 ]. Nevertheless, studies have also shown instances of tumor progression in PTMC patients upon long-term observation[ 11 , 12 ]. Currently, treatment options for PTMC include thyroid lobectomy[ 10 ], as well as non-surgical methods such as active surveillance (AS)[ 13 , 14 ], and image-guided thermal ablation[ 15 – 17 ]. Recent studies suggest that non-surgical therapies are preferred to avoid overtreatment due to their less invasive nature and fewer complications[ 18 ]. However, for patients with lymph node metastasis, it is necessary to reconsider the treatment approach. Furthermore, if surgical treatment is chosen, the decision to perform prophylactic lymph node dissection should differ between PTMC patients with and without lymph node metastasis. Performing prophylactic lymph node dissection may lead to surgical complications such as recurrent laryngeal nerve injury and hypoparathyroidism[ 19 ]. Conversely, if prophylactic lymph node dissection is not performed, there is a risk of missing metastatic lymph nodes, which could result in tumor spread and poor prognosis. Therefore, accurately predicting lymph node metastasis in PTMC patients can significantly assist clinicians in making informed decisions. It is also well-known that selecting a treatment plan for PTMC patients is a complex process that also needs to consider patient preferences and the psychological anxiety of living with cancer[ 20 , 21 ], especially for patients with lymph node metastasis. These clinical issues highlight the importance of accurately predicting lymph node metastasis through imaging and clinical features and underscore the broad clinical application prospects of developing a practical prediction model for lymph node metastasis in patients with PTMC. Previous research has employed Multivariate Logistic Regression to construct prediction models for cervical lymph node metastasis in patients with PTMC[ 22 , 23 ], and studies have discussed the risk factors for lymph node metastasis in PTMC[ 24 , 25 ], identifying age[ 26 – 28 ], gender[ 29 – 31 ], tumor size[ 31 – 33 ], and multifocality[ 34 , 35 ] as significant predictors. Furthermore, studies have also suggested that ultrasound examination has a high specificity in evaluating abnormal lymph nodes[ 36 , 37 ]. The ACR (American College of Radiology) Thyroid Imaging Reporting and Data System (TI-RADS) is an internationally recognized method for evaluating thyroid nodule ultrasound results[ 38 ]. It provides a standardized system for describing thyroid ultrasound findings, including features such as size, shape, margins, echogenicity, etc., and assigns corresponding grades based on these features. Therefore, based on the considerations above, we included age, gender, maximum tumor diameter, multifocality or unifocality, and ACR score as risk factors in our study. These factors are all clinically feasible and cost-effective, enhancing the practicality of our clinical approach. Subsequently, we utilized four machine learning methods to construct predictive models (Multivariate Logistic Regression, Randon Forest, Support Vector Machine, Xgboost), considering the AUC value, sensitivity, and specificity at the optimal threshold, as well as clinical applicability, to select the best model. The AUC of our model exceeds 90%, ranking it among the top in similar studies. Moreover, we have developed a user-friendly prediction website that allows users to easily input relevant features to obtain prediction results, addressing the difficulty of applying prediction models practically in clinical settings. Additionally, we have created online platforms for researchers in this field, enabling medical professionals with access to patient imaging and clinical information to upload their collected data (following our provided template-supplementary material) to construct their prediction model easily. On online platforms, users can directly access the parameters of their constructed models, enabling them to select the most suitable prediction model and further facilitating research and application in this area. Figure 1 provides a intuitive illustration of our research process and significance. Materials and methods Patients inclusion and data collection The study included 256 patients, all from the General Surgery Department at the Second Affiliated Hospital of Xi'an Jiaotong University, with data collected from September 2021 to March 2024. The inclusion criteria were as follows: 1)Postoperative pathological confirmation of papillary thyroid carcinoma. 2)PTC with a maximum diameter of ≤ 1 cm. 3)Patients who had not received any other treatments related to thyroid disease before surgical treatment or radiofrequency ablation. 4)Ultrasound examination results retained complete images of each lesion, meeting the assessment requirements. The exclusion criteria were: 1) Patients with severe illnesses who cannot tolerate general anesthesia and surgery. 2) Patients with malignant tumors in other body parts and pathological detection of distant metastasis. 3)Patients with malignant tumors in other body parts and pathological detection of distant metastasis. 3) Patients with a history of exposure to radioactive substances in the head and neck area. 4) Patients' clinical data is incomplete and does not meet the requirements for collection and assessment. We collected data on the patients' gender, age, maximum tumor diameter, number of nodules (solitary or multiple), and ACR score (according to the 2017 ACR-TIRADS scoring criteria). The specific criteria for the ACR scoring can be found in the supplementary materials. The outcome variable was whether the patient was pathologically diagnosed with cervical lymph node metastasis. Gender was categorized as male or female, the number of lesions was classified as single or multiple, and the ACR score was dichotomized into ≤ 5 and > 5. The specific information about the patients included can also be found in the supplementary materials. Model construction We performed a statistical analysis using Multivariate Logistic Regression to examine the data collected from 256 patients and identify independent risk factors for lymph node metastasis (P < 0.05). Subsequently, we constructed models using all five factors as input indicators and separately using only the identified independent risk factors as predictive indicators. The outcome variable was the presence or absence of lymph node metastasis. Then, we employed Multivariate Logistic Regression, Randon Forest, Support Vector Machine, and Xgboost as the predictive models and evaluated their performance. The dataset was divided into two subsets: 70% designated as the training set for model training and 30% designated as the test set for prediction evaluation. The above process was conducted using R programming, and the specific R code is provided in the supplementary materials. Visualization of model performance To intuitively compare the performance of various models constructed, we utilized the "ggplot" R package to plot the ROC curves of each model and annotated the AUC values on the graphs. Additionally, to understand the proportion of weight that each predictive factor contributes within each model, we created bar charts for a clear visual representation. Moreover, to validate the stability of the best model, we conducted ten-fold cross-validation using the "mar" R package and the "ggplot" package to plot the resulting curves. 2.4 Selection of the optimal model The principles for conducting optimal model selection are as follows: 1. First, compare the AUC values, where a value closer to 1 indicates better model performance; 2. Second, the sensitivity at the optimal threshold should be compared. Sensitivity is a more important metric in clinical settings as it helps avoid missing lymph node metastasis. 3. Lastly, compare the specificity at the optimal threshold. Construction of website and online platforms To enhance the practicality of our model in clinical settings, we developed a user-friendly website that allows clinical practitioners to input several predictive indicators and quickly receive predictions about cervical lymph node metastasis. Additionally, to improve the model's generalization ability, we established an online platform that enables researchers in the field to input their collected data for model construction, selection, and prediction. The construction of the website and platform is based on R, utilizing the "shiny" R package. Detailed R code can be found in the supplementary materials. Statistical Methods For the statistical analysis of clinical data, variables such as gender, age (with a threshold of 45 years), the maximum diameter of the tumor (with a threshold of 5 cm), presence of single versus multiple nodules, and ACR score (with a threshold of 5) are all converted into binary variables. The chi-square test determines whether these indicators are associated with cervical lymph node metastasis in patients with PTMC. All statistical analyses are performed using IBM SPSS Statistics 26. Results Demographic and clinicopathological features In this study, 256 patients were enrolled, among which 81 were pathologically confirmed to have lymph node metastasis post-surgery, accounting for 31.6% of all patients. Significant differences were observed between the group with cervical lymph node metastasis (N1 group) and the group without cervical lymph node metastasis (N0 group) regarding maximum tumor diameter, number of lesions, and ultrasound imaging ACR score. However, age and gender did not show a statistical difference between the two groups according to the Chi-squared test analysis. Detailed information is described in Table 1 . Table 1 The demographic and clinicopathological characteristics of the patients included in the study Characteristics Level N N0 N1 P value Gender Male 56 34 22 0.194 Female 200 141 59 Age 5 73 36 37 0.000 *** ≤ 5 183 139 44 Number of lesions Single 121 102 19 0.000 *** Multiple 135 74 61 ACR score >5 90 31 59 0.000 *** ≤ 5 166 144 22 *** p < 0.001 The predictive factors were further analyzed using Multivariate Logistic Regression. Detailed statistical results can be found in Table 2 . It was also found that maximum tumor diameter, number of lesions, and ACR score are all independent risk factors for cervical lymph node metastasis in PTMC patients. Table 2 Statistical results of predictor variables from Multivariate Logistic Regression Characteristics Estimate Std.Error P value Gender 0.412 0.397 0.2995 Age -0.029 0.016 0.0718 Maximum tumor diameter(mm) 0.159 0.074 0.0307 * Number of lesions 0.916 0.356 0.0101 * ACR score 2.349 0.344 0.0000 *** * p < 0.05 ** p < 0.01 *** p < 0.001 Randon forest model Given the potential for complex and nonlinear relationships among predictive variables and outcome indicators, to construct the most suitable Randon Forest prediction model, we first use all five predictive factors (age, gender, maximum tumor diameter, number of lesions, and ACR score) to build Model 1. Then, we use independently verified risk factors (maximum tumor diameter, number of lesions, ACR score) to construct Model 2. The weights of each predictive factor in Model 1 and Model 2 are shown in Fig. 2 a and 2 b, respectively. From this, we can conclude that in Model 1, the maximum tumor diameter contributes the most to the model, while in Model 2, the ACR score has the most enormous contribution. The comparison of model parameters is as follows: Model 1 (AUC = 0.92, optimal threshold = 0.46, sensitivity = 0.83, specificity = 0.90), Model 2 (AUC = 0.89, optimal threshold = 0.28, sensitivity = 0.88, specificity = 0.83)(Table 3 ). Additionally, we have plotted the ROC curves for both models and annotated their AUC values (Fig. 3 a and 3 b). By comparing the above parameters and following the principles for selecting the optimal model, Model 1 is identified as the best model. The models mean that including all five predictive factors (age, gender, maximum tumor diameter, number of lesions, ACR score) in the model achieves better predictive performance. Table 3 Summary table of optimal thresholds, sensitivity, and specificity for four models Optimal Threshold Sensitivity Specificity AUC Randon Forest 0.46 1 0.83 1 0.90 1 0.92 1 0.28 2 0.88 2 0.83 2 0.89 2 SupportVector Machine 0.16 1 0.83 1 0.81 1 0.84 1 0.18 2 0.87 2 0.83 2 0.87 2 Logistic Regression 0.21 1 0.90 1 0.74 1 0.87 1 0.21 2 0.87 2 0.81 2 0.87 2 Xgboost 0.46 1 0.72 1 0.89 1 0.83 1 0.24 2 0.79 2 0.83 2 0.84 2 1 Model 1 2 Model 2 Support vector machine When constructing the Support Vector Machine model, we first use five predictive factors to build Model 1 and then three independent risk factors to construct Model 2. The weights of the variables in the models are shown in Fig. 2 c and 2 d. It can be observed that the number of lesions(nodes) is the factor contributing most significantly to the model. To visually compare the two models, we plotted the ROC curves of the models and annotated the AUC values (Fig. 1 c, 1 d). The specific parameters for Model 1 and Model 2 are as follows: Model 1 (AUC = 0.84, optimal threshold = 0.16, sensitivity = 0.83, specificity = 0.81); Model 2 (AUC = 0.87, optimal threshold = 0.18, sensitivity = 0.87, specificity = 0.83) (Table 3 ). According to the principles of selecting the optimal model, Model 2 is the superior Support Vector Machine model. Multivariate logistic regression Although our analysis of the collected data from all 256 patients suggests that age and gender are not independent risk factors, it is vital to consider domain-specific expertise when building models. Previous research has indicated that age and gender are independent risk factors for cervical lymph node metastasis in patients with PTMC[ 30 , 39 ]. Therefore, our study also constructed two models using multinomial Logistic regression: Model 1, which includes all five predictive factors, and Model 2, which provides only three independent risk factors. Based on the bar charts of the weights of various predictive factors (Fig. 4 a and 4 b), the ACR score is the most crucial variable in the Multivariate Logistic Regression model. The ROC curves for Model 1 and Model 2, along with their respective AUC values, can be seen in Fig. 5 a and Fig. 5 b. The specific parameters for the models are as follows: Model 1 (AUC = 0.87, optimal threshold = 0.21, sensitivity = 0.90, specificity = 0.74); Model 2 (AUC = 0.87, optimal threshold = 0.21, sensitivity = 0.87, specificity = 0.81) (Table 3 ). Both models have the same AUC value; however, Model 1 has higher sensitivity than Model 2, making it the better multivariate logistic model. Xgboost model Using the Xgboost model, a robust machine learning model capable of handling complex nonlinear relationships between factors, we constructed two models: Model 1 using all five predictive factors and Model 2 using only three independent risk factors. The maximum tumor diameter is the factor with the most excellent weight in both models (Fig. 4 c and 4 d). The ROC curves and AUC values for both models are also presented in Fig. 5 c and 5 d. The specific parameters are as follows: Model 1 (AUC = 0.83, optimal threshold = 0.46, sensitivity = 0.72, specificity = 0.89); Model 2 (AUC = 0.84, optimal threshold = 0.24, sensitivity = 0.79, specificity = 0.83) (Table 3 ). Model 2 is considered the better Xgboost model with a higher AUC value and sensitivity. Construction of the best model and predictive website Based on the information provided and by comparing the parameters of each model, as shown in Table 3 , we ultimately selected Randon Forest Model 1 as the best predictive model. To further validate the model's stability, we conducted a ten-fold cross-validation. The AUC values for the ten validations were as follows: 0.838, 0.910, 0.865, 0.873, 0.863, 0.872, 0.849, 0.825, 0.867, and 0.941, with an average AUC value of 0.870. A curve graph was also plotted to visually demonstrate this (Fig. 6 ). From the results, it can be observed that the model is highly stable, with the AUC values maintained above 0.80 in 10 validations. Based on this model, we have developed a website to allow clinicians to access the prediction results easily. The website's webpage is shown in Fig. 7 , and the link to the website is as follows: http://yucemoxing.online:8082 . Construction of the online platform To enhance the generalization ability of our model and allow it to train on more data, we have provided online open platforms that support users in uploading their data for model building, selection, and application. Since different data might yield varying performance outcomes with the models, we offer researchers ample choice to select the optimal model. We provide eight different model-building platforms, based respectively on two variations each of the Randon Forest model(Model1, Model2), Support Vector Machine model(Model 1, Model 2), multinomial Logistic Regression model(Model 1, Model 2), and the Xgboost model(Model 1, Model 2). The interfaces of the online platforms are shown in Fig. 8 and Fig. 9 , and the corresponding links to the platforms are as follows: http://yucemoxing.online:8081(Randon Forest Model 1); http://yucemoxing.site:8089 (Randon Forest Model 2); http://yucemoxing.online:8084( Support Vector Machine Model 1) ; http://yucemoxing.online:8085( Support Vector Machine Model 2) ; http://yucemoxing.online:8083( Multivariate Logistic Regression Model 1); http://yucemoxing.online:8088( Multivariate Logistic Regression Model 2); http://yucemoxing.online:8087(Xgboost Model 1); http://yucemoxing.online:8086(Xgboost Model 2) . Discussion Thyroid cancer is the most common endocrine malignancy[ 40 ], with papillary thyroid carcinoma being the most prevalent subtype[ 3 ]. In recent years, the detection rate of PTMC has dramatically contributed to the increase in new cases. For some low-risk microcarcinomas, many cases do not progress significantly, leading to the suggestion that such patients may forgo immediate surgical treatment in favor of active surveillance as a primary management strategy. And this approach reduces adverse reactions in patients and can significantly lower treatment costs[ 41 , 42 ]. However, there also are some patients with PTMC who may develop lymph node metastases, which can lead to disease progression and poorer prognosis, necessitating surgical intervention. The clinical challenge is that lymph node metastasis can only be confirmed postoperatively through pathological examination, and it is difficult to accurately predict preoperatively using standard imaging techniques. Thus, there is an urgent need in clinical practice for a highly usable and accurate predictive model to forecast lymph node metastasis in patients to determine the appropriate treatment plan. Additionally, the predictive factors used in constructing the model must be readily obtainable in clinical practice. Studies have shown that age [ 26 , 27 , 43 ], gender [ 29 – 31 ], tumor size [ 32 , 33 , 44 ], multifocal[ 26 , 45 , 46 ], and abnormal lymph nodes on ultrasound 36, 37 can predict lymph node metastasis in PTMC. Therefore, we primarily collected data on age, gender, maximum tumor diameter, multifocality or unifocality, ACR score, and the presence of cervical lymph node metastasis in patients with PTMC. Unlike similar studies, we did not rely solely on independent risk factors identified from our own data to build the model for two main reasons. First, any data from a single center can exhibit certain biases. Second, different models have unique characteristics and are suited for solving other problems; for instance, some models can uncover complex nonlinear relationships among various predictive factors. Therefore, in constructing our model, we thoroughly compared the effectiveness of using all five predictive factors versus using only the independent risk factors, thus avoiding the omission of important information. Many studies have attempted to create predictive models for the lymph node metastasis of papillary microcarcinoma of the thyroid[ 23 , 25 , 32 , 47 ]. These studies have significantly contributed to this field; however, some inconveniences may arise during practical application. One significant reason for this is that conventional prediction models, which often use scales[ 48 ], may increase the workload of clinicians and may not achieve the desired results in the fast-paced clinical environment. However, our predictive website can significantly enhance the usability of such models in clinical settings. Moreover, considering that predictive models require extensive sample data for training, and to eliminate difficulties researchers in this field might encounter while applying machine learning model codes, we provide online platforms that allow researchers to directly upload their data and construct their predictive models online, and make predictions based on their models. Recognizing the diverse clinical needs of researchers for model parameters like sensitivity and specificity, we offer eight distinct machine-learning models on our platforms, each with detailed parameter information to help users select the most suitable model for their predictive requirements. Furthermore, it is essential to note that our study is single-center, and the model's excellent performance still needs to be further validated with multicenter data. During the patient inclusion process, some were excluded due to missing crucial information, meaning that the patients included in the study were not necessarily consecutive admissions over a period. This method could potentially affect the representativeness of the training data to some extent. However, we believe that our online platform can provide significant support for researchers in the field to conduct further in-depth studies and address these issues. Conclusions Predictive tools based on machine learning can accurately predict the probability of cervical lymph node metastasis in patients with PTMC. We have developed a highly accurate predictive model. Moreover, based on this model, we have created a user-friendly online website with practical clinical applications. We also provided online platforms for other researchers to upload their data, build and choose their models, and then predict using the selected model. We suggest that the outcomes of this study possess strong potential for future clinical applications. Declarations Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Ethics approval This study was approved by the The Ethics Committee of the Second Affiliated Hospital of Xi'an Jiaotong University(approval number: 2023318). Authors contributions YYJ and ZDW designed the study. TKS, YHS and YKW collected clinical data. YHS and TKS performed the statistical analysis. YHS wrote the manuscript, and YYJ and ZDW helps to revise it. All authors read and approved the final manuscript. Declaration of competing interest The authors declare no conflict of interest. References D W Chen, B H H Lang, D S A McLeod, K Newbold, M R Haymart, Thyroid cancer. Lancet 401, 1531–1544 (2023) W Chen, R Zheng, P D Baade, S Zhang, H Zeng, F Bray et al. Cancer statistics in China, 2015. CA Cancer J Clin 66, 115–132 (2016) U C Megwalu, P K Moon, Thyroid Cancer Incidence and Mortality Trends in the United States: 2000–2018. Thyroid 32, 560–570 (2022) H Lim, S S Devesa, J A Sosa, D Check, C M Kitahara, Trends in Thyroid Cancer Incidence and Mortality in the United States, 1974–2013. Jama 317, 1338–1348 (2017) H Luo, X Xia, G D Kim, Y Liu, Z Xue, L Zhang et al. Characterizing dedifferentiation of thyroid cancer by integrated analysis. Sci Adv 7, (2021) M Xing, B R Haugen, M Schlumberger, Progress in molecular-based management of differentiated thyroid cancer. Lancet 381, 1058–1069 (2013) E L Mazzaferri, Management of low-risk differentiated thyroid cancer. Endocr Pract 13, 498–512 (2007) I D Hay, Management of patients with low-risk papillary thyroid carcinoma. Endocr Pract 13, 521–533 (2007) J Tang, H B Liu, L Yu, X Meng, S X Leng, H Zhang, Clinical-pathological Characteristics and Prognostic Factors for Papillary Thyroid Microcarcinoma in the Elderly. J Cancer 9, 256–262 (2018) B R Haugen, E K Alexander, K C Bible, G M Doherty, S J Mandel, Y E Nikiforov et al. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid 26, 1–133 (2016) Y Ito, A Miyauchi, M Kihara, T Higashiyama, K Kobayashi, A Miya, Patient age is significantly related to the progression of papillary microcarcinoma of the thyroid under observation. Thyroid 24, 27–34 (2014) I Sugitani, K Toda, K Yamada, N Yamamoto, M Ikenaga, Y Fujimoto, Three distinctly different kinds of papillary thyroid microcarcinoma should be recognized: our treatment strategies and outcomes. World J Surg 34, 1222–1231 (2010) A Miyauchi, Y Ito, H Oda, Insights into the Management of Papillary Microcarcinoma of the Thyroid. Thyroid 28, 23–31 (2018) I Sugitani, Y Ito, D Takeuchi, H Nakayama, C Masaki, H Shindo et al. Indications and Strategy for Active Surveillance of Adult Low-Risk Papillary Thyroid Microcarcinoma: Consensus Statements from the Japan Association of Endocrine Surgery Task Force on Management for Papillary Thyroid Microcarcinoma. Thyroid 31, 183–192 (2021) G Mauri, L Hegedüs, S Bandula, R L Cazzato, A Czarniecka, O Dudeck et al. European Thyroid Association and Cardiovascular and Interventional Radiological Society of Europe 2021 Clinical Practice Guideline for the Use of Minimally Invasive Treatments in Malignant Thyroid Lesions. Eur Thyroid J 10, 185–197 (2021) J H Kim, J H Baek, H K Lim, H S Ahn, S M Baek, Y J Choi et al. 2017 Thyroid Radiofrequency Ablation Guideline: Korean Society of Thyroid Radiology. Korean J Radiol 19, 632–655 (2018) L Zhang, W Zhou, J Q Zhou, Q Shi, T Rago, G Gambelunghe et al. 2022 Expert consensus on the use of laser ablation for papillary thyroid microcarcinoma. Int J Hyperthermia 39, 1254–1263 (2022) L Yan, W Li, Y Zhu, X Li, Y Li, Y Li et al. Long-term comparison of Image-guided thermal ablation vs. lobectomy for solitary papillary thyroid microcarcinoma: a multicenter retrospective cohort study. Int J Surg, (2024) J L Roh, J Y Park, C I Park, Total thyroidectomy plus neck dissection in differentiated papillary thyroid carcinoma patients: pattern of nodal metastasis, morbidity, recurrence, and postoperative levels of serum parathyroid hormone. Ann Surg 245, 604–610 (2007) P Zhu, Q Zhang, Q Wu, G Shi, W Wang, H Xu et al. Barriers and Facilitators to the Choice of Active Surveillance for Low-Risk Papillary Thyroid Cancer in China: A Qualitative Study Examining Patient Perspectives. Thyroid 33, 826–834 (2023) L Davies, B R Roman, M Fukushima, Y Ito, A Miyauchi, Patient Experience of Thyroid Cancer Active Surveillance in Japan. JAMA Otolaryngol Head Neck Surg 145, 363–370 (2019) C Zhang, S Fu, H Liu, S Xue, Risk prediction for < 1 cm lateral lymph node metastasis in papillary thyroid microcarcinoma. Front Endocrinol (Lausanne) 14, 1235354 (2023) L Zhang, P Wang, K Li, S Xue, A novel nomogram for identifying high-risk patients among active surveillance candidates with papillary thyroid microcarcinoma. Front Endocrinol (Lausanne) 14, 1185327 (2023) X Wu, B L Li, C J Zheng, X D He, Predictive factors for central lymph node metastases in papillary thyroid microcarcinoma. World J Clin Cases 8, 1350–1360 (2020) Y Huang, Y Mao, L Xu, J Wen, G Chen, Exploring risk factors for cervical lymph node metastasis in papillary thyroid microcarcinoma: construction of a novel population-based predictive model. BMC Endocr Disord 22, 269 (2022) Y Luo, Y Zhao, K Chen, J Shen, J Shi, S Lu et al. Clinical analysis of cervical lymph node metastasis risk factors in patients with papillary thyroid microcarcinoma. J Endocrinol Invest 42, 227–236 (2019) Y Wang, Q Guan, J Xiang, Nomogram for predicting central lymph node metastasis in papillary thyroid microcarcinoma: A retrospective cohort study of 8668 patients. Int J Surg 55, 98–102 (2018) X Wu, B Li, C Zheng, X He, RISK FACTORS FOR CENTRAL LYMPH NODE METASTASES IN PATIENTS WITH PAPILLARY THYROID MICROCARCINOMA. Endocr Pract 24, 1057–1062 (2018) Q Zhang, Z Wang, X Meng, Q Y Duh, G Chen, Predictors for central lymph node metastases in CN0 papillary thyroid microcarcinoma (mPTC): A retrospective analysis of 1304 cases. Asian J Surg 42, 571–576 (2019) X Zheng, C Peng, M Gao, J Zhi, X Hou, J Zhao et al. Risk factors for cervical lymph node metastasis in papillary thyroid microcarcinoma: a study of 1,587 patients. Cancer Biol Med 16, 121–130 (2019) F Cheng, Y Chen, L Zhu, B Zhou, Y Xu, Y Chen et al. Risk Factors for Cervical Lymph Node Metastasis of Papillary Thyroid Microcarcinoma: A Single-Center Retrospective Study. Int J Endocrinol 2019, 8579828 (2019) W X Jin, D R Ye, Y H Sun, X F Zhou, O C Wang, X H Zhang et al. Prediction of central lymph node metastasis in papillary thyroid microcarcinoma according to clinicopathologic factors and thyroid nodule sonographic features: a case-control study. Cancer Manag Res 10, 3237–3243 (2018) N Qu, L Zhang, Q H Ji, J Y Chen, Y X Zhu, Y M Cao et al. Risk Factors for Central Compartment Lymph Node Metastasis in Papillary Thyroid Microcarcinoma: A Meta-Analysis. World J Surg 39, 2459–2470 (2015) W Zheng, K Wang, J Wu, W Wang, J Shang, Multifocality is associated with central neck lymph node metastases in papillary thyroid microcarcinoma. Cancer Manag Res 10, 1527–1533 (2018) Y L Zhou, E L Gao, W Zhang, H Yang, G L Guo, X H Zhang et al. Factors predictive of papillary thyroid micro-carcinoma with bilateral involvement and central lymph node metastasis: a retrospective study. World J Surg Oncol 10, 67 (2012) X P Huang, T T Ye, L Zhang, R F Liu, X J Lai, L Wang et al. Sonographic features of papillary thyroid microcarcinoma predicting high-volume central neck lymph node metastasis. Surg Oncol 27, 172–176 (2018) Y J Lee, D W Kim, H K Park, D H Kim, S J Jung, M Oh et al. Pre-operative ultrasound diagnosis of nodal metastasis in papillary thyroid carcinoma patients according to nodal compartment. Ultrasound Med Biol 41, 1294–1300 (2015) F N Tessler, W D Middleton, E G Grant, J K Hoang, L L Berland, S A Teefey et al. ACR Thyroid Imaging, Reporting and Data System (TI-RADS): White Paper of the ACR TI-RADS Committee. J Am Coll Radiol 14, 587–595 (2017) X Wang, J Tan, W Zheng, N Li, A retrospective study of the clinical features in papillary thyroid microcarcinoma depending on age. Nucl Med Commun 39, 713–719 (2018) H Sung, J Ferlay, R L Siegel, M Laversanne, I Soerjomataram, A Jemal et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 71, 209–249 (2021) A Miyauchi, Y Ito, Conservative Surveillance Management of Low-Risk Papillary Thyroid Microcarcinoma. Endocrinol Metab Clin North Am 48, 215–226 (2019) Y Ito, A Miyauchi, T Kudo, H Oda, M Yamamoto, H Sasai et al. Trends in the Implementation of Active Surveillance for Low-Risk Papillary Thyroid Microcarcinomas at Kuma Hospital: Gradual Increase and Heterogeneity in the Acceptance of This New Management Option. Thyroid 28, 488–495 (2018) X Yin, C Liu, Y Guo, X Li, N Shen, X Zhao et al. Influence of tumor extent on central lymph node metastasis in solitary papillary thyroid microcarcinomas: a retrospective study of 1092 patients. World J Surg Oncol 15, 133 (2017) X Yu, X Song, W Sun, S Zhao, J Zhao, Y G Wang, Independent Risk Factors Predicting Central Lymph Node Metastasis in Papillary Thyroid Microcarcinoma. Horm Metab Res 49, 201–207 (2017) C Y Gui, S L Qiu, Z H Peng, M Wang, Clinical and pathologic predictors of central lymph node metastasis in papillary thyroid microcarcinoma: a retrospective cohort study. J Endocrinol Invest 41, 403–409 (2018) Y Xu, L Xu, J Wang, Clinical predictors of lymph node metastasis and survival rate in papillary thyroid microcarcinoma: analysis of 3607 patients at a single institution. J Surg Res 221, 128–134 (2018) Y Wang, F Nie, G Wang, T Liu, T Dong, Y Sun, Value of Combining Clinical Factors, Conventional Ultrasound, and Contrast-Enhanced Ultrasound Features in Preoperative Prediction of Central Lymph Node Metastases of Different Sized Papillary Thyroid Carcinomas. Cancer Manag Res 13, 3403–3415 (2021) Y Akasu-Nagayoshi, T Hayashi, A Kawabata, N Shimizu, A Yamada, N Yokota et al. PHOSPHATE exporter XPR1/SLC53A1 is required for the tumorigenicity of epithelial ovarian cancer. Cancer Sci 113, 2034–2043 (2022) Additional Declarations No competing interests reported. Supplementary Files Modelcode.docx TableACRCriteria.docx datatemplate.xlsx patientsinformation.xlsx 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-4560286","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":318256282,"identity":"b4ac4034-0e27-4e3a-a0bf-f29c2c248fbc","order_by":0,"name":"Yuanhao Su","email":"","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Yuanhao","middleName":"","lastName":"Su","suffix":""},{"id":318256284,"identity":"f522552c-c5ea-4fe1-b383-8e33f47c267c","order_by":1,"name":"Tingkai Sun","email":"","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Tingkai","middleName":"","lastName":"Sun","suffix":""},{"id":318256285,"identity":"64922549-0c0e-45f0-ab9a-db0ec851faf0","order_by":2,"name":"Yongke Wu","email":"","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Yongke","middleName":"","lastName":"Wu","suffix":""},{"id":318256286,"identity":"de893c7d-534a-44de-add3-22ff7c749016","order_by":3,"name":"Cheng Li","email":"","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Cheng","middleName":"","lastName":"Li","suffix":""},{"id":318256287,"identity":"7350a2fb-d26b-44ef-8b12-232d98ced9ec","order_by":4,"name":"Yunhao Li","email":"","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Yunhao","middleName":"","lastName":"Li","suffix":""},{"id":318256289,"identity":"515b7e65-ede2-48bf-8dd8-5edd40d23d61","order_by":5,"name":"Xing Jin","email":"","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Xing","middleName":"","lastName":"Jin","suffix":""},{"id":318256290,"identity":"46602106-4dcd-479f-8a6a-982feabab9ef","order_by":6,"name":"Yuanyuan Ji","email":"","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Yuanyuan","middleName":"","lastName":"Ji","suffix":""},{"id":318256292,"identity":"a69ae228-c79a-42dc-a597-2c4739fa18a7","order_by":7,"name":"Zhidong Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAoUlEQVRIiWNgGAWjYLACngoJOXkStZyxMDZsIEkLb1tFIsMBYlUbHD9j/OHtPIkExgbmh49uEKNFsifHwHDuNok8dgY2Y+McYrTwM+RuSObdJlHM2MDDJk2UFjb+txsO886RSGw4QKwWfoncjc28DaRokZzx/jPjnGMSxobNxPrF4Hxa8oc3NXVy8uzNDx8TpQUBmElTPgpGwSgYBaMAHwAASLUsnwnQSQMAAAAASUVORK5CYII=","orcid":"","institution":"Xi'an Jiaotong University","correspondingAuthor":true,"prefix":"","firstName":"Zhidong","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2024-06-10 23:08:45","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4560286/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4560286/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":59526097,"identity":"b2df2cbc-2b3b-4ba0-aeaa-966ca6d842e8","added_by":"auto","created_at":"2024-07-02 21:01:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1899632,"visible":true,"origin":"","legend":"\u003cp\u003eAn overview of the main research process and its clinical application in predicting cervical lymph node metastasis in patients with papillary thyroid microcarcinoma.\u003c/p\u003e","description":"","filename":"Fig.1.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/4c0cc1c2ef4c2d62a115f908.png"},{"id":59524469,"identity":"6b75e23f-5c49-40c2-bb76-c6785dc0e617","added_by":"auto","created_at":"2024-07-02 20:45:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2791236,"visible":true,"origin":"","legend":"\u003cp\u003eBar chart of factor weights for four machine learning models: a. Randon Forest model 1; b. Randon Forest Model 2; c. Support Vector Machine Model 1; d. Support Vector Machine Model 2.\u003c/p\u003e","description":"","filename":"Fig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/332514fe7796d8647c68b285.png"},{"id":59525745,"identity":"a5eb0f63-64cd-41d7-982e-67ef7deb7ca4","added_by":"auto","created_at":"2024-07-02 20:53:42","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":248424,"visible":true,"origin":"","legend":"\u003cp\u003eThe ROC curves and AUC values of Logistic Regression and Logistic Regression models: a. The ROC curve and AUC value of Randon Forest model 1. b. The ROC curve and AUC value of Randon Forest model 2. c. The ROC curve and AUC value of Support Vector Machine model 1. d. The ROC curve and AUC value of Support Vector Machine model 2.\u003c/p\u003e","description":"","filename":"Fig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/e0c4c3d73fb1252b35187e94.png"},{"id":59524470,"identity":"0f4e0ee7-5ec3-48d6-974d-599892ea81ed","added_by":"auto","created_at":"2024-07-02 20:45:42","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1291071,"visible":true,"origin":"","legend":"\u003cp\u003eBar chart of factor weights for four types of machine learning models: a. Multivariate Logistic Regression Model 1; b. Multivariate Logistic Regression Model 2; c. Xgboost Model 1; d. Xgboost Model 2.\u003c/p\u003e","description":"","filename":"Fig.4.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/8477b637bfc24457e21bd3b8.png"},{"id":59524467,"identity":"e28a9d7e-b1c9-4005-9eb5-87b193d08a5e","added_by":"auto","created_at":"2024-07-02 20:45:42","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1805931,"visible":true,"origin":"","legend":"\u003cp\u003eThe ROC curves and AUC values of Multivariate Logistic Regression and Xgboost models: a. The ROC curve and AUC value of Multivariate Logistic Regression model 1. b. The ROC curve and AUC value of Multivariate Logistic Regression model 2. c. The ROC curve and AUC value of Xgboost model 1. d. The ROC curve and AUC value of Xgboost model 2.\u003c/p\u003e","description":"","filename":"Fig.5.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/3cc3be0408d06fa36f4a0acb.png"},{"id":59526098,"identity":"e9152eae-8729-46b0-898c-8075c59c3e25","added_by":"auto","created_at":"2024-07-02 21:01:43","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1027459,"visible":true,"origin":"","legend":"\u003cp\u003eThe AUC value curve of Randon Forest Model 1 with ten-fold cross-validation.\u003c/p\u003e","description":"","filename":"Fig.6.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/0fdebb43d0b6dc88f407f300.png"},{"id":59525747,"identity":"548e3beb-4333-4362-b20b-43e5e2b0f887","added_by":"auto","created_at":"2024-07-02 20:53:43","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1853412,"visible":true,"origin":"","legend":"\u003cp\u003eWebsite interface of the prediction model based on the Randon Forest Model 1.\u003c/p\u003e","description":"","filename":"Fig.7.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/08cfc086439050406fac868b.png"},{"id":59524476,"identity":"9631bcb6-1c4f-49ca-a02c-5c3b5170ec83","added_by":"auto","created_at":"2024-07-02 20:45:43","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":5927242,"visible":true,"origin":"","legend":"\u003cp\u003eInterfaces of online platforms: a. Platform built on Randon Forest Model 1; b. Platform built on Randon Forest Model 2; c. Platform based on Support Vector Machine Model 1; d. Platform based on Support Vector Machine Model 2.\u003c/p\u003e","description":"","filename":"Fig.8.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/9144818e3fbe63757f6393c4.png"},{"id":59524477,"identity":"4b888569-a13b-4dda-a3bb-f28210fbddb1","added_by":"auto","created_at":"2024-07-02 20:45:43","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":5913481,"visible":true,"origin":"","legend":"\u003cp\u003eInterfaces of online platforms: a. Platform built on Multivariate Logistic Regression Model 1; b. Platform built on Multivariate Logistic Regression Model 2; c. Platform based on Xgboost Model 1; d. Platform based on Xgboost Model 2.\u003c/p\u003e","description":"","filename":"Fig.9.png","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/e297c7b49bb8ff86047dfa79.png"},{"id":60404221,"identity":"dee03d98-bb76-4cab-abc1-359305d5dfff","added_by":"auto","created_at":"2024-07-16 11:39:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":19992018,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/69dbc6a8-619a-4adc-bf61-27715308f5b6.pdf"},{"id":59524475,"identity":"d04a3a1b-88aa-4a30-b21e-9676cc5c9226","added_by":"auto","created_at":"2024-07-02 20:45:43","extension":"docx","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":20093,"visible":true,"origin":"","legend":"","description":"","filename":"Modelcode.docx","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/b6be7d0f8b0af0a38214f3e6.docx"},{"id":59524474,"identity":"c9738e25-633c-4d6e-a970-28cbb01ab4f9","added_by":"auto","created_at":"2024-07-02 20:45:43","extension":"docx","order_by":12,"title":"","display":"","copyAsset":false,"role":"supplement","size":12292,"visible":true,"origin":"","legend":"","description":"","filename":"TableACRCriteria.docx","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/f7300c0df1d4f72f73ccabd4.docx"},{"id":59524471,"identity":"d84b5df6-710a-4125-a477-87206d1def4e","added_by":"auto","created_at":"2024-07-02 20:45:43","extension":"xlsx","order_by":13,"title":"","display":"","copyAsset":false,"role":"supplement","size":291694,"visible":true,"origin":"","legend":"","description":"","filename":"datatemplate.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/6795033cf8641cb8416316a4.xlsx"},{"id":59524479,"identity":"a022640e-b3c3-45ee-9ff7-5e339ac225d8","added_by":"auto","created_at":"2024-07-02 20:45:44","extension":"xlsx","order_by":14,"title":"","display":"","copyAsset":false,"role":"supplement","size":517135,"visible":true,"origin":"","legend":"","description":"","filename":"patientsinformation.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4560286/v1/c4467742be629d2e6c2731bb.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Predictive model and clinical application for lymph node metastasis in papillary thyroid microcarcinoma","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThyroid cancer is the most common endocrine malignancy worldwide, with papillary thyroid carcinoma (PTC) being the predominant histological subtype, accounting for over 80% of cases[\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] and the notable rise in thyroid cancer incidence is primarily attributed to increased diagnoses of the PTC subtype[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Studies have reported its potential to progress towards a more malignant state, significantly reducing life expectancy[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Papillary thyroid microcarcinoma (PTMC) is defined as a PTC measuring 1 centimeter or less in its longest dimension, with reported lower metastasis rates[\u003cspan additionalcitationids=\"CR8 CR9\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Nevertheless, studies have also shown instances of tumor progression in PTMC patients upon long-term observation[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eCurrently, treatment options for PTMC include thyroid lobectomy[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], as well as non-surgical methods such as active surveillance (AS)[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e], and image-guided thermal ablation[\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Recent studies suggest that non-surgical therapies are preferred to avoid overtreatment due to their less invasive nature and fewer complications[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. However, for patients with lymph node metastasis, it is necessary to reconsider the treatment approach. Furthermore, if surgical treatment is chosen, the decision to perform prophylactic lymph node dissection should differ between PTMC patients with and without lymph node metastasis. Performing prophylactic lymph node dissection may lead to surgical complications such as recurrent laryngeal nerve injury and hypoparathyroidism[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Conversely, if prophylactic lymph node dissection is not performed, there is a risk of missing metastatic lymph nodes, which could result in tumor spread and poor prognosis. Therefore, accurately predicting lymph node metastasis in PTMC patients can significantly assist clinicians in making informed decisions. It is also well-known that selecting a treatment plan for PTMC patients is a complex process that also needs to consider patient preferences and the psychological anxiety of living with cancer[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], especially for patients with lymph node metastasis. These clinical issues highlight the importance of accurately predicting lymph node metastasis through imaging and clinical features and underscore the broad clinical application prospects of developing a practical prediction model for lymph node metastasis in patients with PTMC.\u003c/p\u003e \u003cp\u003ePrevious research has employed Multivariate Logistic Regression to construct prediction models for cervical lymph node metastasis in patients with PTMC[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e], and studies have discussed the risk factors for lymph node metastasis in PTMC[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], identifying age[\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], gender[\u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e], tumor size[\u003cspan additionalcitationids=\"CR32\" citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e], and multifocality[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e] as significant predictors. Furthermore, studies have also suggested that ultrasound examination has a high specificity in evaluating abnormal lymph nodes[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. The ACR (American College of Radiology) Thyroid Imaging Reporting and Data System (TI-RADS) is an internationally recognized method for evaluating thyroid nodule ultrasound results[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. It provides a standardized system for describing thyroid ultrasound findings, including features such as size, shape, margins, echogenicity, etc., and assigns corresponding grades based on these features. Therefore, based on the considerations above, we included age, gender, maximum tumor diameter, multifocality or unifocality, and ACR score as risk factors in our study. These factors are all clinically feasible and cost-effective, enhancing the practicality of our clinical approach. Subsequently, we utilized four machine learning methods to construct predictive models (Multivariate Logistic Regression, Randon Forest, Support Vector Machine, Xgboost), considering the AUC value, sensitivity, and specificity at the optimal threshold, as well as clinical applicability, to select the best model. The AUC of our model exceeds 90%, ranking it among the top in similar studies. Moreover, we have developed a user-friendly prediction website that allows users to easily input relevant features to obtain prediction results, addressing the difficulty of applying prediction models practically in clinical settings. Additionally, we have created online platforms for researchers in this field, enabling medical professionals with access to patient imaging and clinical information to upload their collected data (following our provided template-supplementary material) to construct their prediction model easily. On online platforms, users can directly access the parameters of their constructed models, enabling them to select the most suitable prediction model and further facilitating research and application in this area. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e provides a intuitive illustration of our research process and significance.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatients inclusion and data collection\u003c/h2\u003e \u003cp\u003eThe study included 256 patients, all from the General Surgery Department at the Second Affiliated Hospital of Xi'an Jiaotong University, with data collected from September 2021 to March 2024. The inclusion criteria were as follows:\u003c/p\u003e \u003cp\u003e1)Postoperative pathological confirmation of papillary thyroid carcinoma.\u003c/p\u003e \u003cp\u003e2)PTC with a maximum diameter of \u0026le;\u0026thinsp;1 cm.\u003c/p\u003e \u003cp\u003e3)Patients who had not received any other treatments related to thyroid disease before surgical treatment or radiofrequency ablation.\u003c/p\u003e \u003cp\u003e4)Ultrasound examination results retained complete images of each lesion, meeting the assessment requirements.\u003c/p\u003e \u003cp\u003eThe exclusion criteria were:\u003c/p\u003e \u003cp\u003e1) Patients with severe illnesses who cannot tolerate general anesthesia and surgery.\u003c/p\u003e \u003cp\u003e2) Patients with malignant tumors in other body parts and pathological detection of distant metastasis.\u003c/p\u003e \u003cp\u003e3)Patients with malignant tumors in other body parts and pathological detection of distant metastasis.\u003c/p\u003e \u003cp\u003e3) Patients with a history of exposure to radioactive substances in the head and neck area.\u003c/p\u003e \u003cp\u003e4) Patients' clinical data is incomplete and does not meet the requirements for collection and assessment.\u003c/p\u003e \u003cp\u003eWe collected data on the patients' gender, age, maximum tumor diameter, number of nodules (solitary or multiple), and ACR score (according to the 2017 ACR-TIRADS scoring criteria). The specific criteria for the ACR scoring can be found in the supplementary materials. The outcome variable was whether the patient was pathologically diagnosed with cervical lymph node metastasis. Gender was categorized as male or female, the number of lesions was classified as single or multiple, and the ACR score was dichotomized into \u0026le;\u0026thinsp;5 and \u0026gt;\u0026thinsp;5. The specific information about the patients included can also be found in the supplementary materials.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eModel construction\u003c/h2\u003e \u003cp\u003eWe performed a statistical analysis using Multivariate Logistic Regression to examine the data collected from 256 patients and identify independent risk factors for lymph node metastasis (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Subsequently, we constructed models using all five factors as input indicators and separately using only the identified independent risk factors as predictive indicators. The outcome variable was the presence or absence of lymph node metastasis. Then, we employed Multivariate Logistic Regression, Randon Forest, Support Vector Machine, and Xgboost as the predictive models and evaluated their performance. The dataset was divided into two subsets: 70% designated as the training set for model training and 30% designated as the test set for prediction evaluation. The above process was conducted using R programming, and the specific R code is provided in the supplementary materials.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eVisualization of model performance\u003c/h2\u003e \u003cp\u003eTo intuitively compare the performance of various models constructed, we utilized the \"ggplot\" R package to plot the ROC curves of each model and annotated the AUC values on the graphs. Additionally, to understand the proportion of weight that each predictive factor contributes within each model, we created bar charts for a clear visual representation. Moreover, to validate the stability of the best model, we conducted ten-fold cross-validation using the \"mar\" R package and the \"ggplot\" package to plot the resulting curves.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Selection of the optimal model\u003c/h2\u003e \u003cp\u003eThe principles for conducting optimal model selection are as follows: 1. First, compare the AUC values, where a value closer to 1 indicates better model performance; 2. Second, the sensitivity at the optimal threshold should be compared. Sensitivity is a more important metric in clinical settings as it helps avoid missing lymph node metastasis. 3. Lastly, compare the specificity at the optimal threshold.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eConstruction of website and online platforms\u003c/h2\u003e \u003cp\u003eTo enhance the practicality of our model in clinical settings, we developed a user-friendly website that allows clinical practitioners to input several predictive indicators and quickly receive predictions about cervical lymph node metastasis. Additionally, to improve the model's generalization ability, we established an online platform that enables researchers in the field to input their collected data for model construction, selection, and prediction. The construction of the website and platform is based on R, utilizing the \"shiny\" R package. Detailed R code can be found in the supplementary materials.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Methods\u003c/h2\u003e \u003cp\u003eFor the statistical analysis of clinical data, variables such as gender, age (with a threshold of 45 years), the maximum diameter of the tumor (with a threshold of 5 cm), presence of single versus multiple nodules, and ACR score (with a threshold of 5) are all converted into binary variables. The chi-square test determines whether these indicators are associated with cervical lymph node metastasis in patients with PTMC. All statistical analyses are performed using IBM SPSS Statistics 26.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eDemographic and clinicopathological features\u003c/h2\u003e \u003cp\u003eIn this study, 256 patients were enrolled, among which 81 were pathologically confirmed to have lymph node metastasis post-surgery, accounting for 31.6% of all patients. Significant differences were observed between the group with cervical lymph node metastasis (N1 group) and the group without cervical lymph node metastasis (N0 group) regarding maximum tumor diameter, number of lesions, and ultrasound imaging ACR score. However, age and gender did not show a statistical difference between the two groups according to the Chi-squared test analysis. Detailed information is described in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\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\u003eThe demographic and clinicopathological characteristics of the patients included in the study\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=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" 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\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLevel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN0\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.194\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e200\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e141\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e59\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 \u003cp\u003e\u0026lt;\u0026thinsp;45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e109\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.225\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e147\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e105\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e42\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\u003eMaximum tumor diameter(mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.000\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e183\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e139\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e44\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\u003eNumber of lesions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSingle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e121\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e102\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.000\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMultiple\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e135\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e61\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\u003eACR score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.000\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e166\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e144\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e22\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 \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e*** \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe predictive factors were further analyzed using Multivariate Logistic Regression. Detailed statistical results can be found in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. It was also found that maximum tumor diameter, number of lesions, and ACR score are all independent risk factors for cervical lymph node metastasis in PTMC patients.\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 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eStatistical results of predictor variables from Multivariate Logistic Regression\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\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\u003eEstimate\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStd.Error\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.412\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.397\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.2995\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=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0718\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMaximum tumor diameter(mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.159\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.074\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0307\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber of lesions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.916\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.356\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0101\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eACR score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.349\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.344\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0000\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e* \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 ** \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01 *** \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eRandon forest model\u003c/h2\u003e \u003cp\u003eGiven the potential for complex and nonlinear relationships among predictive variables and outcome indicators, to construct the most suitable Randon Forest prediction model, we first use all five predictive factors (age, gender, maximum tumor diameter, number of lesions, and ACR score) to build Model 1. Then, we use independently verified risk factors (maximum tumor diameter, number of lesions, ACR score) to construct Model 2. The weights of each predictive factor in Model 1 and Model 2 are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea and \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb, respectively. From this, we can conclude that in Model 1, the maximum tumor diameter contributes the most to the model, while in Model 2, the ACR score has the most enormous contribution. The comparison of model parameters is as follows: Model 1 (AUC\u0026thinsp;=\u0026thinsp;0.92, optimal threshold\u0026thinsp;=\u0026thinsp;0.46, sensitivity\u0026thinsp;=\u0026thinsp;0.83, specificity\u0026thinsp;=\u0026thinsp;0.90), Model 2 (AUC\u0026thinsp;=\u0026thinsp;0.89, optimal threshold\u0026thinsp;=\u0026thinsp;0.28, sensitivity\u0026thinsp;=\u0026thinsp;0.88, specificity\u0026thinsp;=\u0026thinsp;0.83)(Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Additionally, we have plotted the ROC curves for both models and annotated their AUC values (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). By comparing the above parameters and following the principles for selecting the optimal model, Model 1 is identified as the best model. The models mean that including all five predictive factors (age, gender, maximum tumor diameter, number of lesions, ACR score) in the model achieves better predictive performance.\u003c/p\u003e \u003cp\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 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSummary table of optimal thresholds, sensitivity, and specificity for four models\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\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=\"char\" char=\".\" 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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOptimal Threshold\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRandon Forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.46\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.83\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.90\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.92\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.28\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.88\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.83\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.89\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSupportVector Machine\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.16\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.83\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.81\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.84\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.18\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.87\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.83\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.87\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLogistic Regression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.21\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.90\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.74\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.87\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.21\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.87\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.81\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.87\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eXgboost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.46\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.72\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.89\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.83\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.24\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.79\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.83\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.84\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003csup\u003e1\u003c/sup\u003e Model 1 \u003csup\u003e2\u003c/sup\u003eModel 2\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\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\u003eSupport vector machine\u003c/h2\u003e \u003cp\u003eWhen constructing the Support Vector Machine model, we first use five predictive factors to build Model 1 and then three independent risk factors to construct Model 2. The weights of the variables in the models are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec and \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed. It can be observed that the number of lesions(nodes) is the factor contributing most significantly to the model. To visually compare the two models, we plotted the ROC curves of the models and annotated the AUC values (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec, \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed). The specific parameters for Model 1 and Model 2 are as follows: Model 1 (AUC\u0026thinsp;=\u0026thinsp;0.84, optimal threshold\u0026thinsp;=\u0026thinsp;0.16, sensitivity\u0026thinsp;=\u0026thinsp;0.83, specificity\u0026thinsp;=\u0026thinsp;0.81); Model 2 (AUC\u0026thinsp;=\u0026thinsp;0.87, optimal threshold\u0026thinsp;=\u0026thinsp;0.18, sensitivity\u0026thinsp;=\u0026thinsp;0.87, specificity\u0026thinsp;=\u0026thinsp;0.83) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). According to the principles of selecting the optimal model, Model 2 is the superior Support Vector Machine model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMultivariate logistic regression\u003c/h2\u003e \u003cp\u003eAlthough our analysis of the collected data from all 256 patients suggests that age and gender are not independent risk factors, it is vital to consider domain-specific expertise when building models. Previous research has indicated that age and gender are independent risk factors for cervical lymph node metastasis in patients with PTMC[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. Therefore, our study also constructed two models using multinomial Logistic regression: Model 1, which includes all five predictive factors, and Model 2, which provides only three independent risk factors. Based on the bar charts of the weights of various predictive factors (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), the ACR score is the most crucial variable in the Multivariate Logistic Regression model. The ROC curves for Model 1 and Model 2, along with their respective AUC values, can be seen in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea and Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb. The specific parameters for the models are as follows: Model 1 (AUC\u0026thinsp;=\u0026thinsp;0.87, optimal threshold\u0026thinsp;=\u0026thinsp;0.21, sensitivity\u0026thinsp;=\u0026thinsp;0.90, specificity\u0026thinsp;=\u0026thinsp;0.74); Model 2 (AUC\u0026thinsp;=\u0026thinsp;0.87, optimal threshold\u0026thinsp;=\u0026thinsp;0.21, sensitivity\u0026thinsp;=\u0026thinsp;0.87, specificity\u0026thinsp;=\u0026thinsp;0.81) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Both models have the same AUC value; however, Model 1 has higher sensitivity than Model 2, making it the better multivariate logistic model.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eXgboost model\u003c/h2\u003e \u003cp\u003eUsing the Xgboost model, a robust machine learning model capable of handling complex nonlinear relationships between factors, we constructed two models: Model 1 using all five predictive factors and Model 2 using only three independent risk factors. The maximum tumor diameter is the factor with the most excellent weight in both models (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed). The ROC curves and AUC values for both models are also presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed. The specific parameters are as follows: Model 1 (AUC\u0026thinsp;=\u0026thinsp;0.83, optimal threshold\u0026thinsp;=\u0026thinsp;0.46, sensitivity\u0026thinsp;=\u0026thinsp;0.72, specificity\u0026thinsp;=\u0026thinsp;0.89); Model 2 (AUC\u0026thinsp;=\u0026thinsp;0.84, optimal threshold\u0026thinsp;=\u0026thinsp;0.24, sensitivity\u0026thinsp;=\u0026thinsp;0.79, specificity\u0026thinsp;=\u0026thinsp;0.83) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Model 2 is considered the better Xgboost model with a higher AUC value and sensitivity.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eConstruction of the best model and predictive website\u003c/h2\u003e \u003cp\u003eBased on the information provided and by comparing the parameters of each model, as shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, we ultimately selected Randon Forest Model 1 as the best predictive model. To further validate the model's stability, we conducted a ten-fold cross-validation. The AUC values for the ten validations were as follows: 0.838, 0.910, 0.865, 0.873, 0.863, 0.872, 0.849, 0.825, 0.867, and 0.941, with an average AUC value of 0.870. A curve graph was also plotted to visually demonstrate this (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). From the results, it can be observed that the model is highly stable, with the AUC values maintained above 0.80 in 10 validations. Based on this model, we have developed a website to allow clinicians to access the prediction results easily. The website's webpage is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e, and the link to the website is as follows: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8082\u003c/span\u003e\u003cspan address=\"http://yucemoxing.online:8082\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eConstruction of the online platform\u003c/h2\u003e \u003cp\u003eTo enhance the generalization ability of our model and allow it to train on more data, we have provided online open platforms that support users in uploading their data for model building, selection, and application. Since different data might yield varying performance outcomes with the models, we offer researchers ample choice to select the optimal model. We provide eight different model-building platforms, based respectively on two variations each of the Randon Forest model(Model1, Model2), Support Vector Machine model(Model 1, Model 2), multinomial Logistic Regression model(Model 1, Model 2), and the Xgboost model(Model 1, Model 2). The interfaces of the online platforms are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e, and the corresponding links to the platforms are as follows:\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8081(Randon\u003c/span\u003e \u003cspan address=\"http://yucemoxing.online:8081(Randon\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e Forest Model 1);\u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.site:8089\u003c/span\u003e \u003cspan address=\"http://yucemoxing.site:8089\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e (Randon Forest Model 2);\u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8084(\u003c/span\u003e \u003cspan address=\"http://yucemoxing.online:8084(\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e Support Vector Machine Model 1) ;\u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8085(\u003c/span\u003e \u003cspan address=\"http://yucemoxing.online:8085(\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e Support Vector Machine Model 2) ;\u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8083(\u003c/span\u003e \u003cspan address=\"http://yucemoxing.online:8083(\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e Multivariate Logistic Regression Model 1);\u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8088(\u003c/span\u003e \u003cspan address=\"http://yucemoxing.online:8088(\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e Multivariate Logistic Regression Model 2);\u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8087(Xgboost\u003c/span\u003e \u003cspan address=\"http://yucemoxing.online:8087(Xgboost\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e Model 1);\u003c/p\u003e \u003cp\u003e \u003cspan class=\"ExternalRef\"\u003e \u003cspan class=\"RefSource\"\u003ehttp://yucemoxing.online:8086(Xgboost\u003c/span\u003e \u003cspan address=\"http://yucemoxing.online:8086(Xgboost\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e \u003c/span\u003e Model 2) .\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThyroid cancer is the most common endocrine malignancy[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e], with papillary thyroid carcinoma being the most prevalent subtype[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. In recent years, the detection rate of PTMC has dramatically contributed to the increase in new cases. For some low-risk microcarcinomas, many cases do not progress significantly, leading to the suggestion that such patients may forgo immediate surgical treatment in favor of active surveillance as a primary management strategy. And this approach reduces adverse reactions in patients and can significantly lower treatment costs[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. However, there also are some patients with PTMC who may develop lymph node metastases, which can lead to disease progression and poorer prognosis, necessitating surgical intervention. The clinical challenge is that lymph node metastasis can only be confirmed postoperatively through pathological examination, and it is difficult to accurately predict preoperatively using standard imaging techniques. Thus, there is an urgent need in clinical practice for a highly usable and accurate predictive model to forecast lymph node metastasis in patients to determine the appropriate treatment plan.\u003c/p\u003e \u003cp\u003eAdditionally, the predictive factors used in constructing the model must be readily obtainable in clinical practice. Studies have shown that age [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e], gender [\u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e], tumor size [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e], multifocal[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e], and abnormal lymph nodes on ultrasound\u003csup\u003e36, 37\u003c/sup\u003e can predict lymph node metastasis in PTMC. Therefore, we primarily collected data on age, gender, maximum tumor diameter, multifocality or unifocality, ACR score, and the presence of cervical lymph node metastasis in patients with PTMC. Unlike similar studies, we did not rely solely on independent risk factors identified from our own data to build the model for two main reasons. First, any data from a single center can exhibit certain biases. Second, different models have unique characteristics and are suited for solving other problems; for instance, some models can uncover complex nonlinear relationships among various predictive factors. Therefore, in constructing our model, we thoroughly compared the effectiveness of using all five predictive factors versus using only the independent risk factors, thus avoiding the omission of important information.\u003c/p\u003e \u003cp\u003eMany studies have attempted to create predictive models for the lymph node metastasis of papillary microcarcinoma of the thyroid[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. These studies have significantly contributed to this field; however, some inconveniences may arise during practical application. One significant reason for this is that conventional prediction models, which often use scales[\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e], may increase the workload of clinicians and may not achieve the desired results in the fast-paced clinical environment. However, our predictive website can significantly enhance the usability of such models in clinical settings. Moreover, considering that predictive models require extensive sample data for training, and to eliminate difficulties researchers in this field might encounter while applying machine learning model codes, we provide online platforms that allow researchers to directly upload their data and construct their predictive models online, and make predictions based on their models. Recognizing the diverse clinical needs of researchers for model parameters like sensitivity and specificity, we offer eight distinct machine-learning models on our platforms, each with detailed parameter information to help users select the most suitable model for their predictive requirements.\u003c/p\u003e \u003cp\u003eFurthermore, it is essential to note that our study is single-center, and the model's excellent performance still needs to be further validated with multicenter data. During the patient inclusion process, some were excluded due to missing crucial information, meaning that the patients included in the study were not necessarily consecutive admissions over a period. This method could potentially affect the representativeness of the training data to some extent. However, we believe that our online platform can provide significant support for researchers in the field to conduct further in-depth studies and address these issues.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003ePredictive tools based on machine learning can accurately predict the probability of cervical lymph node metastasis in patients with PTMC. We have developed a highly accurate predictive model. Moreover, based on this model, we have created a user-friendly online website with practical clinical applications. We also provided online platforms for other researchers to upload their data, build and choose their models, and then predict using the selected model. We suggest that the outcomes of this study possess strong potential for future clinical applications.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability \u003c/strong\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\u003eEthics approval \u003c/strong\u003eThis study was approved by the The Ethics Committee of the Second Affiliated Hospital of Xi'an Jiaotong University(approval number: 2023318).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors contributions \u003c/strong\u003eYYJ and ZDW designed the study. TKS, YHS and YKW collected clinical data. YHS and TKS performed the statistical analysis. YHS wrote the manuscript, and YYJ and ZDW helps to revise it. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of competing interest \u003c/strong\u003eThe authors declare no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eD W Chen, B H H Lang, D S A McLeod, K Newbold, M R Haymart, Thyroid cancer. Lancet 401, 1531\u0026ndash;1544 (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW Chen, R Zheng, P D Baade, S Zhang, H Zeng, F Bray et al. Cancer statistics in China, 2015. CA Cancer J Clin 66, 115\u0026ndash;132 (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eU C Megwalu, P K Moon, Thyroid Cancer Incidence and Mortality Trends in the United States: 2000\u0026ndash;2018. Thyroid 32, 560\u0026ndash;570 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH Lim, S S Devesa, J A Sosa, D Check, C M Kitahara, Trends in Thyroid Cancer Incidence and Mortality in the United States, 1974\u0026ndash;2013. Jama 317, 1338\u0026ndash;1348 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH Luo, X Xia, G D Kim, Y Liu, Z Xue, L Zhang et al. Characterizing dedifferentiation of thyroid cancer by integrated analysis. Sci Adv 7, (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM Xing, B R Haugen, M Schlumberger, Progress in molecular-based management of differentiated thyroid cancer. Lancet 381, 1058\u0026ndash;1069 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eE L Mazzaferri, Management of low-risk differentiated thyroid cancer. Endocr Pract 13, 498\u0026ndash;512 (2007)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eI D Hay, Management of patients with low-risk papillary thyroid carcinoma. Endocr Pract 13, 521\u0026ndash;533 (2007)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ Tang, H B Liu, L Yu, X Meng, S X Leng, H Zhang, Clinical-pathological Characteristics and Prognostic Factors for Papillary Thyroid Microcarcinoma in the Elderly. J Cancer 9, 256\u0026ndash;262 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eB R Haugen, E K Alexander, K C Bible, G M Doherty, S J Mandel, Y E Nikiforov et al. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid 26, 1\u0026ndash;133 (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Ito, A Miyauchi, M Kihara, T Higashiyama, K Kobayashi, A Miya, Patient age is significantly related to the progression of papillary microcarcinoma of the thyroid under observation. Thyroid 24, 27\u0026ndash;34 (2014)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eI Sugitani, K Toda, K Yamada, N Yamamoto, M Ikenaga, Y Fujimoto, Three distinctly different kinds of papillary thyroid microcarcinoma should be recognized: our treatment strategies and outcomes. World J Surg 34, 1222\u0026ndash;1231 (2010)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA Miyauchi, Y Ito, H Oda, Insights into the Management of Papillary Microcarcinoma of the Thyroid. Thyroid 28, 23\u0026ndash;31 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eI Sugitani, Y Ito, D Takeuchi, H Nakayama, C Masaki, H Shindo et al. Indications and Strategy for Active Surveillance of Adult Low-Risk Papillary Thyroid Microcarcinoma: Consensus Statements from the Japan Association of Endocrine Surgery Task Force on Management for Papillary Thyroid Microcarcinoma. Thyroid 31, 183\u0026ndash;192 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG Mauri, L Heged\u0026uuml;s, S Bandula, R L Cazzato, A Czarniecka, O Dudeck et al. European Thyroid Association and Cardiovascular and Interventional Radiological Society of Europe 2021 Clinical Practice Guideline for the Use of Minimally Invasive Treatments in Malignant Thyroid Lesions. Eur Thyroid J 10, 185\u0026ndash;197 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ H Kim, J H Baek, H K Lim, H S Ahn, S M Baek, Y J Choi et al. 2017 Thyroid Radiofrequency Ablation Guideline: Korean Society of Thyroid Radiology. Korean J Radiol 19, 632\u0026ndash;655 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL Zhang, W Zhou, J Q Zhou, Q Shi, T Rago, G Gambelunghe et al. 2022 Expert consensus on the use of laser ablation for papillary thyroid microcarcinoma. Int J Hyperthermia 39, 1254\u0026ndash;1263 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL Yan, W Li, Y Zhu, X Li, Y Li, Y Li et al. Long-term comparison of Image-guided thermal ablation vs. lobectomy for solitary papillary thyroid microcarcinoma: a multicenter retrospective cohort study. Int J Surg, (2024)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ L Roh, J Y Park, C I Park, Total thyroidectomy plus neck dissection in differentiated papillary thyroid carcinoma patients: pattern of nodal metastasis, morbidity, recurrence, and postoperative levels of serum parathyroid hormone. Ann Surg 245, 604\u0026ndash;610 (2007)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP Zhu, Q Zhang, Q Wu, G Shi, W Wang, H Xu et al. Barriers and Facilitators to the Choice of Active Surveillance for Low-Risk Papillary Thyroid Cancer in China: A Qualitative Study Examining Patient Perspectives. Thyroid 33, 826\u0026ndash;834 (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL Davies, B R Roman, M Fukushima, Y Ito, A Miyauchi, Patient Experience of Thyroid Cancer Active Surveillance in Japan. JAMA Otolaryngol Head Neck Surg 145, 363\u0026ndash;370 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC Zhang, S Fu, H Liu, S Xue, Risk prediction for \u0026lt;\u0026thinsp;1 cm lateral lymph node metastasis in papillary thyroid microcarcinoma. Front Endocrinol (Lausanne) 14, 1235354 (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL Zhang, P Wang, K Li, S Xue, A novel nomogram for identifying high-risk patients among active surveillance candidates with papillary thyroid microcarcinoma. Front Endocrinol (Lausanne) 14, 1185327 (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX Wu, B L Li, C J Zheng, X D He, Predictive factors for central lymph node metastases in papillary thyroid microcarcinoma. World J Clin Cases 8, 1350\u0026ndash;1360 (2020)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Huang, Y Mao, L Xu, J Wen, G Chen, Exploring risk factors for cervical lymph node metastasis in papillary thyroid microcarcinoma: construction of a novel population-based predictive model. BMC Endocr Disord 22, 269 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Luo, Y Zhao, K Chen, J Shen, J Shi, S Lu et al. Clinical analysis of cervical lymph node metastasis risk factors in patients with papillary thyroid microcarcinoma. J Endocrinol Invest 42, 227\u0026ndash;236 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Wang, Q Guan, J Xiang, Nomogram for predicting central lymph node metastasis in papillary thyroid microcarcinoma: A retrospective cohort study of 8668 patients. Int J Surg 55, 98\u0026ndash;102 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX Wu, B Li, C Zheng, X He, RISK FACTORS FOR CENTRAL LYMPH NODE METASTASES IN PATIENTS WITH PAPILLARY THYROID MICROCARCINOMA. Endocr Pract 24, 1057\u0026ndash;1062 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQ Zhang, Z Wang, X Meng, Q Y Duh, G Chen, Predictors for central lymph node metastases in CN0 papillary thyroid microcarcinoma (mPTC): A retrospective analysis of 1304 cases. Asian J Surg 42, 571\u0026ndash;576 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX Zheng, C Peng, M Gao, J Zhi, X Hou, J Zhao et al. Risk factors for cervical lymph node metastasis in papillary thyroid microcarcinoma: a study of 1,587 patients. Cancer Biol Med 16, 121\u0026ndash;130 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF Cheng, Y Chen, L Zhu, B Zhou, Y Xu, Y Chen et al. Risk Factors for Cervical Lymph Node Metastasis of Papillary Thyroid Microcarcinoma: A Single-Center Retrospective Study. Int J Endocrinol 2019, 8579828 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW X Jin, D R Ye, Y H Sun, X F Zhou, O C Wang, X H Zhang et al. Prediction of central lymph node metastasis in papillary thyroid microcarcinoma according to clinicopathologic factors and thyroid nodule sonographic features: a case-control study. Cancer Manag Res 10, 3237\u0026ndash;3243 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eN Qu, L Zhang, Q H Ji, J Y Chen, Y X Zhu, Y M Cao et al. Risk Factors for Central Compartment Lymph Node Metastasis in Papillary Thyroid Microcarcinoma: A Meta-Analysis. World J Surg 39, 2459\u0026ndash;2470 (2015)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW Zheng, K Wang, J Wu, W Wang, J Shang, Multifocality is associated with central neck lymph node metastases in papillary thyroid microcarcinoma. Cancer Manag Res 10, 1527\u0026ndash;1533 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY L Zhou, E L Gao, W Zhang, H Yang, G L Guo, X H Zhang et al. Factors predictive of papillary thyroid micro-carcinoma with bilateral involvement and central lymph node metastasis: a retrospective study. World J Surg Oncol 10, 67 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX P Huang, T T Ye, L Zhang, R F Liu, X J Lai, L Wang et al. Sonographic features of papillary thyroid microcarcinoma predicting high-volume central neck lymph node metastasis. Surg Oncol 27, 172\u0026ndash;176 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY J Lee, D W Kim, H K Park, D H Kim, S J Jung, M Oh et al. Pre-operative ultrasound diagnosis of nodal metastasis in papillary thyroid carcinoma patients according to nodal compartment. Ultrasound Med Biol 41, 1294\u0026ndash;1300 (2015)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF N Tessler, W D Middleton, E G Grant, J K Hoang, L L Berland, S A Teefey et al. ACR Thyroid Imaging, Reporting and Data System (TI-RADS): White Paper of the ACR TI-RADS Committee. J Am Coll Radiol 14, 587\u0026ndash;595 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX Wang, J Tan, W Zheng, N Li, A retrospective study of the clinical features in papillary thyroid microcarcinoma depending on age. Nucl Med Commun 39, 713\u0026ndash;719 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH Sung, J Ferlay, R L Siegel, M Laversanne, I Soerjomataram, A Jemal et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 71, 209\u0026ndash;249 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA Miyauchi, Y Ito, Conservative Surveillance Management of Low-Risk Papillary Thyroid Microcarcinoma. Endocrinol Metab Clin North Am 48, 215\u0026ndash;226 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Ito, A Miyauchi, T Kudo, H Oda, M Yamamoto, H Sasai et al. Trends in the Implementation of Active Surveillance for Low-Risk Papillary Thyroid Microcarcinomas at Kuma Hospital: Gradual Increase and Heterogeneity in the Acceptance of This New Management Option. Thyroid 28, 488\u0026ndash;495 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX Yin, C Liu, Y Guo, X Li, N Shen, X Zhao et al. Influence of tumor extent on central lymph node metastasis in solitary papillary thyroid microcarcinomas: a retrospective study of 1092 patients. World J Surg Oncol 15, 133 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX Yu, X Song, W Sun, S Zhao, J Zhao, Y G Wang, Independent Risk Factors Predicting Central Lymph Node Metastasis in Papillary Thyroid Microcarcinoma. Horm Metab Res 49, 201\u0026ndash;207 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC Y Gui, S L Qiu, Z H Peng, M Wang, Clinical and pathologic predictors of central lymph node metastasis in papillary thyroid microcarcinoma: a retrospective cohort study. J Endocrinol Invest 41, 403\u0026ndash;409 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Xu, L Xu, J Wang, Clinical predictors of lymph node metastasis and survival rate in papillary thyroid microcarcinoma: analysis of 3607 patients at a single institution. J Surg Res 221, 128\u0026ndash;134 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Wang, F Nie, G Wang, T Liu, T Dong, Y Sun, Value of Combining Clinical Factors, Conventional Ultrasound, and Contrast-Enhanced Ultrasound Features in Preoperative Prediction of Central Lymph Node Metastases of Different Sized Papillary Thyroid Carcinomas. Cancer Manag Res 13, 3403\u0026ndash;3415 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY Akasu-Nagayoshi, T Hayashi, A Kawabata, N Shimizu, A Yamada, N Yokota et al. PHOSPHATE exporter XPR1/SLC53A1 is required for the tumorigenicity of epithelial ovarian cancer. Cancer Sci 113, 2034\u0026ndash;2043 (2022)\u003c/span\u003e\u003c/li\u003e\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":"PTMC, Predictive model, Machine learning, Lymph node metastasis","lastPublishedDoi":"10.21203/rs.3.rs-4560286/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4560286/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePapillary thyroid microcarcinomas (PTMC), small tumors under 10 mm, represent a major part of the increase in papillary thyroid cancer cases. The treatment plans for PTMC patients with lymph node metastasi should be different from those without lymph node metastasis. Therefore, accurately identifying patients with cervical lymph node metastasis is of great clinical significance.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe analyzed data from 256 patients diagnosed with PTMC, using age, gender, tumor size, lesion count, and ACR score as predictors. Outcomes were based on cervical lymph node pathology. Four machine learning models—Random Forest, Multivariate Logistic Regression, Support Vector Machine, and Xgboost—were tested for their predictive accuracy and clinical utility. We then created an online website for direct prediction and designed online platforms that allow other researchers to upload their data for model building and prediction. The website and platform design is based on \"shiny\" package.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Random Forest model proved optimal, achieving an AUC of 0.92. It showed high sensitivity (0.83) and specificity (0.90) at the best threshold of 0.46. The link to the website we built based on this model is as follows: \u003ca href=\"http://yucemoxing.online:8082/\"\u003ehttp://yucemoxing.online:8082\u003c/a\u003e. Additionally, the link to the online platforms that allows userss to upload their own data for model building and prediction is as follows: \u003ca href=\"http://yucemoxing.online:8081/\"\u003ehttp://yucemoxing.online:8081\u003c/a\u003e,\u003ca href=\"http://yucemoxing.site:8089/\"\u003ehttp://yucemoxing.site:8089\u003c/a\u003e,\u003ca href=\"http://yucemoxing.online:8084/\"\u003ehttp://yucemoxing.online:8084\u003c/a\u003e,\u003ca href=\"http://yucemoxing.online:8085/\"\u003ehttp://yucemoxing.online:8085\u003c/a\u003e,\u003ca href=\"http://yucemoxing.online:8083/\"\u003ehttp://yucemoxing.online:8083\u003c/a\u003e,\u003ca href=\"http://yucemoxing.online:8088/\"\u003ehttp://yucemoxing.online:8088\u003c/a\u003e, \u003ca href=\"http://yucemoxing.online:8087/\"\u003ehttp://yucemoxing.online:8087\u003c/a\u003e, http://yucemoxing.online:8086\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMachine learning tools can reliably predict cervical lymph node metastasis in PTMC patients. The developed websites offer valuable tools for clinical application, enhancing decision-making in treatment strategies.\u003c/p\u003e","manuscriptTitle":"Predictive model and clinical application for lymph node metastasis in papillary thyroid microcarcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-02 20:45:37","doi":"10.21203/rs.3.rs-4560286/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":"f607a996-e229-4fe6-80cd-79ccd5092923","owner":[],"postedDate":"July 2nd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-07-16T11:31:32+00:00","versionOfRecord":[],"versionCreatedAt":"2024-07-02 20:45:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4560286","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4560286","identity":"rs-4560286","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}