Machine Learning based Radiomics from Multiparametric Magnetic Resonance Imaging for Predicting Lymph Node Metastasis in Cervical Cancer

Machine Learning based Radiomics from Multiparametric Magnetic Resonance Imaging for Predicting Lymph Node Metastasis in Cervical Cancer | 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 Machine Learning based Radiomics from Multiparametric Magnetic Resonance Imaging for Predicting Lymph Node Metastasis in Cervical Cancer Jing Liu, Mingxuan Zhu, Li Li, Lele Zang, Lan Luo, Fei Zhu, Huiqi Zhang, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4271155/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Construct and compare multiple machine-learning models to predict lymph node (LN) metastasis in cervical cancer, utilizing radiomic features extracted from preoperative multi-parametric magnetic resonance imaging (MRI). Methods This study retrospectively enrolled 407 patients with cervical cancer who were randomly divided into training cohort (n = 284) and validation cohort (n = 123). A total of 4065 radiomic features were extracted from the tumor regions of interest on contrast-enhanced T1-weighted imaging, T2-weighted imaging, and diffusion-weighted imaging for each patient. The Mann-Whitney U test, Spearman correlation analysis, and selection operator Cox regression analysis were employed for radiomic feature selection. The relationship between MRI radiomic features and LN status was analyzed by five machine-learning algorithms. Model performance was evaluated by measuring the area under the receiver-operating characteristic curve (AUC) and accuracy (ACC). Moreover, Kaplan–Meier analysis was used to validate the prognostic value of selected clinical and radiomics characteristics. Results LN metastasis was pathologically detected in 24.3% (99/407) of patients. Following three-step feature selection, 18 radiomic features were employed for model construction. The XGBoost model exhibited superior performance compared to other models, achieving an AUC, accuracy, sensitivity, specificity, and F1-score of 0.9268, 0.8969, 0.7419, 0.9891, and 0.8364, respectively, on the validation set. Additionally, Kaplan − Meier curves indicated a significant correlation between radiomic scores and progression-free survival in cervical cancer patients (p < 0.05). Conclusion Machine learning-based multi-parametric MRI radiomic analysis demonstrates a promising performance in the preoperative prediction of LN metastasis and clinical prognosis in cervical cancer. Cervical cancer Lymph node metastasis Machine learning Magnetic resonance imaging Figures Figure 1 Figure 2 Figure 3 Figure 4 1 Introduction Cervical cancer (CC) ranked as the fourth most frequently diagnosed cancer and the fourth leading cause of cancer-related deaths among women(Sung et al. 2021 ). Lymph node (LN) metastasis is a crucial factor in the prognosis of cervical cancer, influencing treatment decisions(Lapuz et al. 2016 ; Ikeda et al. 2021 ; Olawaiye et al. 2021 ). Despite the gold standard being pathological examination, its time delay hampers prompt treatment selection. Therefore, there is a crucial need for noninvasive and precise LN metastasis prediction to ensure accurate patient staging in CC and subsequently guide the selection of the most suitable treatment. In clinical practice, diverse imaging methods such as ultrasound (US), computed tomography (CT), and magnetic resonance imaging (MRI) are used to identify LN status in cancer patients(Petersen et al. 2020 ). MRI furnishes more intricate anatomical details for diagnosis and encompasses a richer array of texture information in its images than US and CT(L et al. 2020). Despite these advantages, imaging methods face constraints in evaluating LN metastasis in cervical cancer and are not recommended as the primary basis for determination, with limited sensitivity and accuracy(Huang and Fang 2018 ). Radiomics extracts and quantifies features from medical images to evaluate pixel-level characteristics that are imperceptible to the human eye.(Aerts 2016 ). Furthermore, the integration of machine-learning (ML) models with radiomics has demonstrated excellent outcomes in medical diagnosis, assisting clinicians in making improved decisions and predictions(Choy et al. 2018 ). According to reports(Yu et al. 2021 ; Pereira et al. 2021 ; Xu et al. 2022 ; Sheng et al. 2022 ; Zhang et al. 2023 ; Fan et al. 2023 ), radiomics and machine learning can be used for subtype classification, survival prediction, and assessing the impact of different diseases. Predicting LN metastasis using MRI features in CC poses a contemporary challenge. Several radiomics studies have been conducted to address this problem(Wang et al. 2019 ; Dong et al. 2020 ). Wang et al(Wang et al. 2019 ) constructed a machine learning model that constructed a Support Vector Machine (SVM) model based on T2WI and DWI images that predicted LN metastasis in early-stage cervical cancer. However, research on the integration of radiomics and machine learning for predicting LN metastasis and prognosis in cervical cancer is limited. Existing studies show a correlation between radiomic features in the primary tumor region and LN metastasis(Huang et al. 2016 ; Wu et al. 2018 ; Forghani et al. 2019 ). This study aims to investigate the diagnostic performance of machine-learning models based on multi-parametric 3D radiomics for preoperative cervical cancer LN metastasis and evaluate the role of radiomic features in predicting cervical cancer prognosis. 2 Material and methods 2.1 Patients The ethics committee approved this retrospective study, and no informative consent requirements were necessary. A total of 407 patients with histologically confirmed cervical cancer between February 2009 to June 2013 in Fujian Cancer Hospital were enrolled in the study. All patient characteristics had complete clinical information, imaging results, pathology diagnoses, and prognostic information. Patients were enrolled based on the following inclusion criteria: (1) 18 years or older; (2) performance status, Eastern Cooperative Oncology Group (ECOG) ≤ 2; (3) histologically confirmed squamous carcinoma, adenocarcinoma, or adenosquamous carcinoma of the cervix; (4) International Federation of Gynecology and Obstetrics (FIGO) stage (2009) from IB2 to IIB. The exclusion criteria were: (1) incomplete information or severe imaging artifacts of the MRI images; (2) lesion diameter < 5 mm on MRI images. According to the proportion of 7: 3, all cervical cancer patients were divided into the training cohort and the validation cohorts. 2.2 MRI image acquisition and segmentation Magnetic resonance (MR) images were obtained using a 1.5 Tesla MR system (GE, Fairfield City, USA). Conventional MR scanning utilized 8-channel phased-array surface coil pairs, while DWI scanning incorporated an external-body surface coil. MRI sequence scanning included Axial T1-weighted fast spin echo (FSE), axial T2-weighted FSE, axial DWI, Sagittal T2-weighted FSE, Axial CE three-dimensional spoiled gradient echo (3D SGRE) and Coronal CE 3D SGRE. DWI was performed before the intravenous injection of gadolinium. A 0.2 mmol/kg of body weight intravenous injection of gadopentetate dimeglumine (Magnevist, ScheringAG Germany) was administered for the post-gadolinium series. Preprocessing for image standardization included the entire gross tumor volume (GTV) cropping, cropping all images to a size of 32×256×256 based on the GTV position, and resampling anisotropic voxels into 1×1×1 mm³ using linear interpolation. One experienced radiologists (ten years of pelvic MRI reading expertise) manually outlined 3D tumor contours on axial slices using 3D Slicer software (version 4.11, https://www.slicer.org ). The segmentation results were confirmed by another gynecologic oncology expert (Q.X, with 20 years of clinical experience). Both radiologists remained blinded to the clinical and histopathological data throughout this process. Finally, images and image masks were exported as 3D files in the Neuroimaging Informatics Technology Initiative format. 2.3 Features extraction and selection The radiomics feature extraction was performed in Python (version 3.8.3) using a Pyradiomics package (version 3.0.1, https://github.com/Radiomics/pyradiomics)(van Griethuysen et al. 2017) . Quantitative radiomic features were extracted from the tumor regions of interest (ROI) in T1WI, T2WI, and DWI, including shape-based, first-order statistical, texture, and wavelet features. Before further analysis, all the extracted radiomics features underwent standardization into a normal distribution with z-scores, effectively mitigating the differences in the value scales of the data(Carré et al. 2020 ). In pursuit of selecting highly relevant and non-redundant features, a three-step method was applied. Firstly, the Mann–Whitney U test was conducted to select the features, retaining those with p < 0.05 as significantly different. Secondly, we sequentially used the Spearman correlation analysis to eliminate redundant radiomic features, with features exhibiting a Spearman correlation coefficient exceeding 0.9 being excluded. Finally, utilizing the Least Absolute Shrinkage and Selection Operator (LASSO) regression algorithm for further dimensionality reduction and optimized feature selection. 2.4 Development and Validation of machine learning–based models Based on the selected optimal feature subset, models were constructed using machine-learning algorithms such as logistic regression (LR), Naïve Bayes, support vector machine (SVM), AdaBoost, and XGBoost. LR learns a probabilistic model for binary classification by optimizing its parameters to align its predictions with the actual class labels as closely as possible. Naive Bayes is a classification algorithm based on Bayes' theorem. The objective of SVM is to identify an optimal hyperplane that separates samples of different classes while maximizing the margin between the two classes. AdaBoost is an ensemble learning algorithm that constructs a strong classifier by combining multiple weak classifiers. XGBoost is based on the framework of boosting trees, constructing multiple weak learners sequentially, with each learner correcting the errors of its predecessor, continuously optimizing predictive performance. The performance of each machine learning-based model was evaluated using receiver operating characteristic curve (ROC), decision curve analysis (DCA), and calibration curves. The performance of algorithms was also assessed in terms of accuracy, sensitivity, and specificity. Figure 1 illustrates the study design and workflow. 2.5 Statistical analysis Feature extraction, feature selection, and training and validation of machine learning models were conducted using Python (version 3.6, https://www.python.org ). To compare patient characteristics between the training and validation groups, T-test was applied for continuous variables, and Chi-square test (for groups with both n = 5) or Yates' corrected Chi-square test (for groups with both n = 5) was employed for categorical variables. A p-value greater than 0.05 indicated no significant difference between the groups(Hodneland et al. 2022 ; Lin et al. 2023 ). The reliability of the radiomics scoring risk stratification system was evaluated through Kaplan-Meier curves. All results with a p-value less than 0.05 were considered statistically significant. 3 Results 3.1 Patient characteristics Table 1 outlines the clinical and demographic characteristics of the training cohort (n = 284) and the validation cohort (n = 123). The median age of patients in both the training and independent validation cohort was (47 ± 8) years. The number of patients with lymph node metastasis was 68 (23.9%) in the training cohort and 31 (25.2%) in the validation cohort. The median follow-up time for surviving patients was 103 months for the training cohort and 110 months for the validation cohort. The results indicate no significant statistical differences in clinical characteristics between the two cohorts of patients. Table 1 Clinical characteristics of patients in training and validation cohorts. Characteristics training (n = 284) validation(n = 123) P value age, mean ± sd 47.16 ± 8.304 47.25 ± 7.885 0.197 Macroscopic type, n (%) 0.343 NT 148(52.1%) 72(58.5%) CT 134(47.2%) 51(41.5%) PTC 2(0.7%) 0(0.0%) Tumor size (cm), n (%) 0.235 > 4 159(56.0%) 61(49.6%) ≤ 4 125(44.0%) 62(50.4%) Postoperative pathological, n (%) 0.393 SCC 265(93.3%) 110(89.4%) AC 17(6.0%) 12(9.8%) SAC 2(0.7%) 1(0.8%) Differentiation degree, n (%) 0.588 Low grade 64(22.5%) 25(20.3%) Middle grade 215(75.7%) 94(76.4%) High grade 5(1.8%) 4(3.3%) Depth Of Tumor Invasion, n (%) 0.619 No 22(7.7%) 8(6.5%) Near total 63(22.2%) 21(17.1%) Superficial 93(32.7%) 45(36.6%) Deep 106(37.3%) 49(39.8%) Corpus invasion, n (%) 0.991 No 261(91.9%) 113(91.9%) Yes 23(8.1%) 10(8.1%) Parametrial invasion, n (%) 0.38 No 272(95.8%) 120(97.6%) Yes 12(4.2%) 3(2.4%) Vaginal invasion, n (%) 0.271 No 263(92.6%) 118(95.9%) Yes 21(7.4%) 5(4.1%) LN, n (%) 0.786 No 216(76.1%) 92(74.8%) Yes 68(23.9%) 31(25.2%) NT, Nodular Type; CT, Cauliflower-like Type; PTC, Postoperative Pathological Type Change; SAC, adenosquamous carcinoma; AC, adenocarcinoma; SCC, squamous cell carcinoma; LN, lymph node. 3.2 Feature selection A total of 4650 radiomic features were initially extracted from the ROIs of contrast-enhanced T1-weighted imaging (ceT1WI), T2-weighted imaging (T2WI), and diffusion-weighted imaging (DWI). Features highly correlated with LN metastasis were selected using the Mann-Whitney U test (p < 0.05), reducing the features to 1356. Subsequent Spearman correlation testing retained 92 features with coefficients exceeding 0.9. The correlation of variables was assessed using a heatmap (Supplementary Figure S1 ). Finally, LASSO regression identified 18 radiomic features most correlated with lymph node metastasis, and Table 2 details these features along with their LASSO coefficients. Supplementary Figure S1 . The correlation between MRI radiomic feature. Correlation analysis was used to estimate the strength of the correlations with Spearman p. Table 2 Radiomics signature selection results with descriptions. Feature LASSO coefficient ceT1WI_original_glrlm_RunVariance 0.029388 ceT1WI _log-sigma-1-0-mm-3D_firstorder_90Percentile -0.00053 ceT1WI _squareroot_firstorder_10Percentile 0.184889 T2WI_original_shape_Elongation 0.086653 T2WI _original_shape_Flatness 0.045143 T2WI _original_glcm_Imc1 0.032342 T2WI _log-sigma-1-0-mm-3D_firstorder_Kurtosis 0.207755 T2WI _log-sigma-1-0-mm-3D_glcm_ClusterShade 0.003092 T2WI _log-sigma-3-0-mm-3D_firstorder_Skewness 0.072891 T2WI _wavelet-LLH_firstorder_Maximum -0.02597 T2WI _wavelet-HLL_firstorder_Minimum 0.266135 DWI_original_firstorder_90Percentile -0.00246 DWI_log-sigma-1-0-mm-3D_firstorder_Skewness 0.150593 DWI_wavelet-LLH_firstorder_Kurtosis 0.108992 DWI_wavelet-HLL_firstorder_Skewness -0.03973 DWI_wavelet-HHL_firstorder_Maximum 0.050664 DWI_exponential_glcm_Imc1 0.150122 DWI_squareroot_ngtdm_Coarseness -0.01102 3.3 Performance of prediction models Table 3 summarizes the predictive performance of all machine learning models. Notably, XGBoost exhibited superior performance, achieving a classification accuracy of 0.927 and an AUC of 0.897 for predicting the lymph node status of cervical cancer patients in the validation set. Its sensitivity, specificity, positive predictive value, negative predictive value, precision, and F1 score were 0.741, 0.989, 0.958, 0.919, 0.958, and 0.836, respectively. In comparison, other models showed varying levels of accuracy and AUC in the validation set: LR (0.667, 0.715), Naïve Bayes (0.724, 0.734), SVM (0.740, 0.791), AdaBoost (0.683, 0.737). These findings underscore the superior predictive performance of the XGBoost model in cervical cancer lymph node status prediction. Figure 2 visually depicts the prediction scores for each patient's lymph node status, highlighting XGBoost's effectiveness in distinguishing between LN-negative and LN-positive groups. For a clearer comparison of different models in predicting the lymph node status of cervical cancer, Fig. 3 A-B illustrate the AUCs for all models on both the training and validation cohorts. The DCA curve was used to evaluate the clinical values of these models (Fig. 3 C-D). Assuming no patients have LNM in cervical cancer, the solid black line (negative line) indicates that the net benefit is zero when no patient accepts therapy. Conversely, the solid grey line (positive line) represents the net benefits when all patients with LNM receive therapy. The XGBoost model demonstrated higher net benefits compared to the two extreme lines (negative line and positive line) in both cohorts. The noteworthy point is that XGBoost demonstrated significantly superior performance compared to the others across most threshold points. Calibration curves reflect the degree of consistency between the observed risk and predicted probabilities of a model. Figures 3 E-F confirm that the XGBoost model exhibits the optimal level of consistency. Table 3 Evaluation indicators of predictive performance of five models. Classifier ACC AUC SEN SPE PPV NPV PRE F1 AdaBoost Training 0.6620 0.7862 0.8824 0.5926 0.4054 0.9412 0.4054 0.5556 Validation 0.6829 0.7370 0.7419 0.6703 0.4259 0.8841 0.4259 0.5412 Naïve Bayes Training 0.5563 0.7044 0.8529 0.4630 0.3333 0.9091 0.3333 0.4793 Validation 0.7236 0.7342 0.7419 0.7174 0.4694 0.8919 0.4694 0.5750 LR Training 0.6901 0.7271 0.7206 0.6806 0.4153 0.8855 0.4153 0.5269 Validation 0.6667 0.7149 0.6774 0.6630 0.4038 0.8592 0.4038 0.5060 SVM Training 0.8486 0.8386 0.7059 0.8935 0.6761 0.9061 0.6761 0.6906 Validation 0.7398 0.7907 0.8065 0.7174 0.4902 0.9167 0.4902 0.6098 XGBoost Training 0.9296 0.9054 0.8088 0.9676 0.8871 0.9414 0.8871 0.8462 Validation 0.9268 0.8969 0.7419 0.9891 0.9583 0.9192 0.9583 0.8364 LR, Logistic Regression; SVM, Support Vector Machine; ACC, accuracy; AUC, area under ROC; SEN, sensitivity; SPE, specificity; PPV, positive predictive value; NPV, negative predictive value; PRE, precision. 3.4 Kaplan-Meier analysis Kaplan-Meier analysis further validates the prognostic value of the selected radiomic features. The Kaplan-Meier curves in Fig. 4 demonstrate a significant difference in survival probabilities between the two risk subgroups (P = 0.005). Patients with higher rad-score exhibit lower overall survival rates (HR 1.9, 95% CI 1.2-3.0, P = 0.006). Across the entire cohort, the stratification of rad-score is significantly associated with progression-free survival (PFS). 4 Discussion In this study, we developed and validated various radiomics-based machine learning diagnostic models for predicting lymph node metastasis in cervical cancer using multiparametric magnetic resonance imaging (MRI). Leveraging preoperative MRI with multiple parameters, we extracted 4065 imaging features, ultimately selecting 18 features and constructing multiple machine learning diagnostic models. The XGBoost model demonstrated superior performance in predicting lymph node metastasis in the validation set, with an accuracy of 0.9268 and an AUC of 0.8969, outperforming other models. Previous studies have indicated that radiomics analysis can extract quantitative image features from medical images, thereby improving tumor diagnosis, staging, and prognosis(Lambin et al. 2017 ; Fang et al. 2020 ; Yuan et al. 2021 ). Fang et al.'s research(Fang et al. 2020 ) demonstrated that radiomics scores derived from MRI as prognostic biomarkers for early-stage cervical cancer patients. Kan et al.(Kan et al. 2019 ) utilized T2WI and ceT1WI to extract radiomic features and employed an SVM model(AUC(95% CI), 0.754(0584–0.924)) to predict lymph node metastasis in early-stage cervical cancer. However, some studies only utilized features extracted from a single MRI image(Xu et al. 2019 ; Yamada et al. 2019 ), or performed radiomics analysis solely on 2D slices(Ytre-Hauge et al. 2018 ). In contrast, our approach involved using three common 3D MRI sequences in cervical cancer patients for feature extraction, allowing for the extraction of rich radiomic features and enhancing the model's reliability. Moreover, we systematically compared multiple machine learning models and identified the best-performing model, showing consistent and clinically excellent performance between the training and validation sets (AUC = 0.897 in the validation set), which suggests the excellence and reliability of our research methodology. Our research utilized a diverse set of machine learning algorithms, such as LR, Naïve Bayes, SVM, AdaBoost and XGBoost. XGBoost is an efficient implementation of the widely-used gradient boosting decision tree algorithm in the field of science. XGBoost is an efficient gradient boosting algorithm that iteratively trains decision trees, incorporates regularization and parallel processing, and constructs a powerful ensemble model suitable for regression and classification tasks. In various machine learning problems, especially in scenarios with a large number of features, high dimensions, and complex structures, XGBoost often exhibits excellent performance. In another study, based on another way of model construction, Sheng et al.(Sheng et al. 2022 ) found the XGBoost model demonstrated a better performance in Invasive ductal breast cancer molecular subtype prediction, especially in the triple-negative and non-triple-negative groups with an AUC of 0.903. Song(Song 2021 ) used the XGBoost algorithm to build a radiomics model for predicting axillary LNM in invasive ductal breast cancer with an AUC of 0.890. In our study, XGBoost model demonstrated superior prediction performance, achieving an AUC of 0.897, surpassing the SVM model which obtained an AUC of 0.791. Preceding investigations have elucidated the prospective efficacy of radiology in prognosticating the survival outcomes of individuals with cervical cancer, boasting superior accuracy than traditional clinical parameters(Liu et al. 2022 ; Zheng et al. 2022 ). Our study aimed to evaluate the prognostic significance of MRI-derived radiological characteristics in cervical cancer patients, with whom higher rad-score exhibit lower overall survival rates (HR 1.9, 95% CI 1.2-3.0, P = 0.006). The precise anticipation of disease progression via radiology holds pivotal importance in clinical settings, guiding the selection of optimal therapeutic strategies and thereby enhancing patient prognoses. Despite the positive results of this research, it is also fundamental to consider the limitations of this study. This study is retrospective and conducted in a single center, lacking external validation, which may introduce selection bias. To enhance the robustness of the models, a larger sample size from multiple centers is necessary. Furthermore, a manual segmentation approach is time-consuming and may impact the precision of feature extraction in certain cases. Previous research has indicated that utilizing an automatic segmentation method could streamline radiomics applications in clinical settings, promoting high levels of intra- and inter-observer reproducibility(Liu et al. 2019 ; Chen et al. 2020 ). 5 Conclusions In conclusion, machine learning is a step towards precision medicine in the field of gynecologic oncology. The preoperative assessment of the LNs is important for accurately staging patients with cervical cancer and for selecting the most suitable treatment. We extracted radiomic features from multi-parametric MRI and employed various machine learning models to predict lymph node metastasis in cervical cancer patients, achieving a commendable level of accuracy. Abbreviations MRI magnetic resonance imaging LN lymph node CC cervical cancer ceT1W contrast-enhanced T1-weighted T2W T2-weighted DWI diffusion-weighted imaging ROIs regions of interest AUC the area under the receiver-operating characteristic curve ACC accuracy PFS progression-free survival HR hazard ratio US ultrasound CT computed tomography ML machine learning SVM Support Vector Machine ECOG Eastern Cooperative Oncology Group FIGO International Federation of Gynecology and Obstetrics GTV gross tumor volume LASSO Least Absolute Shrinkage and Selection Operator LR Logistic Regression ROC receiver operating characteristic DCA decision curve analysis Declarations Statements & Declarations Funding The project was funded by the grants of Joint Funds for the National Clinical Key Specialty Construction Program (2021); Innovative Medicine Subject of Fujian Provincial Health Commission (2020CX0101); Natural Science Foundation of Fujian Province (2020J011126, 2023J011273); Fujian Provincial Clinical Research Center for Cancer Radiotherapy and Immunotherapy (2020Y2012); Startup Fund for scientific research, Fujian Medical University (Grant number：2020QH1233). Competing Interests The authors have no relevant financial or non-financial interests to disclose. Author Contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Jing Liu, Li Li and Lele Zang. The first draft of the manuscript was written by Mingxuan Zhu and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Data Availability The datasets generated during and analysed during the current study are available from the corresponding author on reasonable request. Ethics approval This is an observational study. The XYZ Research Ethics Committee has confirmed that no ethical approval is required. Consent to participate Informed consent was obtained from all individual participants included in the study. Consent to publish Not applicable. References Aerts HJWL (2016) The Potential of Radiomic-Based Phenotyping in Precision Medicine: A Review. JAMA Oncol 2(12):1636–1642. https://doi.org/10.1001/jamaoncol.2016.2631 Carré A, Klausner G, Edjlali M, Lerousseau M, Briend-Diop J, Sun R, Ammari S, Reuzé S, Alvarez Andres E, Estienne T, Niyoteka S, Battistella E, Vakalopoulou M, Dhermain F, Paragios N, Deutsch E, Oppenheim C, Pallud J, Robert C (2020) Standardization of brain MR images across machines and protocols: bridging the gap for MRI-based radiomics. Sci Rep 10(1):12340. https://doi.org/10.1038/s41598-020-69298-z Chen X, Men K, Chen B, Tang Y, Zhang T, Wang S, Li Y, Dai J (2020) CNN-Based Quality Assurance for Automatic Segmentation of Breast Cancer in Radiotherapy. Front Oncol 10:524. https://doi.org/10.3389/fonc.2020.00524 Choy G, Khalilzadeh O, Michalski M, Do S, Samir AE, Pianykh OS, Geis JR, Pandharipande PV, Brink JA, Dreyer KJ (2018) Current Applications and Future Impact of Machine Learning in Radiology. Radiology 288(2):318–328. https://doi.org/10.1148/radiol.2018171820 Dong T, Yang C, Cui B, Zhang T, Sun X, Song K, Wang L, Kong B, Yang X (2020) Development and Validation of a Deep Learning Radiomics Model Predicting Lymph Node Status in Operable Cervical Cancer. Front Oncol 10:464. https://doi.org/10.3389/fonc.2020.00464 Fan Y, Wang X, Dong Y, Cui E, Wang H, Sun X, Su J, Luo Y, Yu T, Jiang X (2023) Multiregional radiomics of brain metastasis can predict response to EGFR-TKI in metastatic NSCLC. Eur Radiol. https://doi.org/10.1007/s00330-023-09709-7 Fang J, Zhang B, Wang S, Jin Y, Wang F, Ding Y, Chen Q, Chen L, Li Y, Li M, Chen Z, Liu L, Liu Z, Tian J, Zhang S (2020) Association of MRI-derived radiomic biomarker with disease-free survival in patients with early-stage cervical cancer. Theranostics 10(5):2284–2292. https://doi.org/10.7150/thno.37429 Forghani R, Chatterjee A, Reinhold C, Pérez-Lara A, Romero-Sanchez G, Ueno Y, Bayat M, Alexander JWM, Kadi L, Chankowsky J, Seuntjens J, Forghani B (2019) Head and neck squamous cell carcinoma: prediction of cervical lymph node metastasis by dual-energy CT texture analysis with machine learning. Eur Radiol 29(11):6172–6181. https://doi.org/10.1007/s00330-019-06159-y Hodneland E, Kaliyugarasan S, Wagner-Larsen KS, Lura N, Andersen E, Bartsch H, Smit N, Halle MK, Krakstad C, Lundervold AS, Haldorsen IS (2022) Fully Automatic Whole-Volume Tumor Segmentation in Cervical Cancer. Cancers (Basel) 14(10):2372. https://doi.org/10.3390/cancers14102372 Huang B-X, Fang F (2018) Progress in the Study of Lymph Node Metastasis in Early-stage Cervical Cancer. Curr Med Sci 38(4):567–574. https://doi.org/10.1007/s11596-018-1915-0 Huang Y-Q, Liang C-H, He L, Tian J, Liang C-S, Chen X, Ma Z-L, Liu Z-Y (2016) Development and Validation of a Radiomics Nomogram for Preoperative Prediction of Lymph Node Metastasis in Colorectal Cancer. J Clin Oncol 34(18):2157–2164. https://doi.org/10.1200/JCO.2015.65.9128 Ikeda M, Shida M, Shigeta S, Nagase S, Takahashi F, Yamagami W, Katabuchi H, Yaegashi N, Aoki D, Mikami M (2021) The trend and outcome of postsurgical therapy for high-risk early-stage cervical cancer with lymph node metastasis in Japan: a report from the Japan Society of Gynecologic Oncology (JSGO) guidelines evaluation committee. J Gynecol Oncol 32(3):e44. https://doi.org/10.3802/jgo.2021.32.e44 Kan Y, Dong D, Zhang Y, Jiang W, Zhao N, Han L, Fang M, Zang Y, Hu C, Tian J, Li C, Luo Y (2019) Radiomic signature as a predictive factor for lymph node metastasis in early-stage cervical cancer. J Magn Reson Imaging 49(1):304–310. https://doi.org/10.1002/jmri.26209 L L, Q L, L T (2020) Diagnostic value and clinical significance of MRI and CT in detecting lymph node metastasis of early cervical cancer. Oncology letters 19(1). https://doi.org/10.3892/ol.2019.11180 Lambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, Sanduleanu S, Larue RTHM, Even AJG, Jochems A, van Wijk Y, Woodruff H, van Soest J, Lustberg T, Roelofs E, van Elmpt W, Dekker A, Mottaghy FM, Wildberger JE, Walsh S (2017) Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol 14(12):749–762. https://doi.org/10.1038/nrclinonc.2017.141 Lapuz C, Kondalsamy-Chennakesavan S, Bernshaw D, Khaw P, Narayan K (2016) Stage IB cervix cancer with nodal involvement treated with primary surgery or primary radiotherapy: Patterns of failure and outcomes in a contemporary population. J Med Imaging Radiat Oncol 60(2):274–282. https://doi.org/10.1111/1754-9485.12411 Lin Y-C, Lin G, Pandey S, Yeh C-H, Wang J-J, Lin C-Y, Ho T-Y, Ko S-F, Ng S-H (2023) Fully automated segmentation and radiomics feature extraction of hypopharyngeal cancer on MRI using deep learning. Eur Radiol 33(9):6548–6556. https://doi.org/10.1007/s00330-023-09827-2 Liu C, Gardner SJ, Wen N, Elshaikh MA, Siddiqui F, Movsas B, Chetty IJ (2019) Automatic Segmentation of the Prostate on CT Images Using Deep Neural Networks (DNN). Int J Radiat Oncol Biol Phys 104(4):924–932. https://doi.org/10.1016/j.ijrobp.2019.03.017 Liu S, Li R, Liu Q, Sun D, Yang H, Pan H, Wang L, Song S (2022) Radiomics model of 18F-FDG PET/CT imaging for predicting disease-free survival of early-stage uterine cervical squamous cancer. Cancer Biomark 33(2):249–259. https://doi.org/10.3233/CBM-210201 Olawaiye AB, Baker TP, Washington MK, Mutch DG (2021) The new (Version 9) American Joint Committee on Cancer tumor, node, metastasis staging for cervical cancer. CA Cancer J Clin 71(4):287–298. https://doi.org/10.3322/caac.21663 Pereira HM, Leite Duarte ME, Ribeiro Damasceno I, De Oliveira Moura Santos LA, Nogueira-Barbosa MH (2021) Machine learning-based CT radiomics features for the prediction of pulmonary metastasis in osteosarcoma. BJR 94(1124):20201391. https://doi.org/10.1259/bjr.20201391 Petersen LJ, Nielsen JB, Langkilde NC, Petersen A, Afshar-Oromieh A, De Souza NM, De Paepe K, Fisker RV, Arp DT, Carl J, Haberkorn U, Zacho HD (2020) 68Ga-PSMA PET/CT compared with MRI/CT and diffusion-weighted MRI for primary lymph node staging prior to definitive radiotherapy in prostate cancer: a prospective diagnostic test accuracy study. World J Urol 38(4):939–948. https://doi.org/10.1007/s00345-019-02846-z Sheng W, Xia S, Wang Y, Yan L, Ke S, Mellisa E, Gong F, Zheng Y, Tang T (2022) Invasive ductal breast cancer molecular subtype prediction by MRI radiomic and clinical features based on machine learning. Front Oncol 12:964605. https://doi.org/10.3389/fonc.2022.964605 Song B-I (2021) A machine learning-based radiomics model for the prediction of axillary lymph-node metastasis in breast cancer. Breast Cancer 28(3):664–671. https://doi.org/10.1007/s12282-020-01202-z Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 71(3):209–249. https://doi.org/10.3322/caac.21660 van Griethuysen JJM, Fedorov A, Parmar C, Hosny A, Aucoin N, Narayan V, Beets-Tan RGH, Fillion-Robin J-C, Pieper S, Aerts HJWL (2017) Computational Radiomics System to Decode the Radiographic Phenotype. Cancer Res 77(21):e104–e107. https://doi.org/10.1158/0008-5472.CAN-17-0339 Wang T, Gao T, Yang J, Yan X, Wang Y, Zhou X, Tian J, Huang L, Zhang M (2019) Preoperative prediction of pelvic lymph nodes metastasis in early-stage cervical cancer using radiomics nomogram developed based on T2-weighted MRI and diffusion-weighted imaging. Eur J Radiol 114:128–135. https://doi.org/10.1016/j.ejrad.2019.01.003 Wu S, Zheng J, Li Y, Wu Z, Shi S, Huang M, Yu H, Dong W, Huang J, Lin T (2018) Development and Validation of an MRI-Based Radiomics Signature for the Preoperative Prediction of Lymph Node Metastasis in Bladder Cancer. EBioMedicine 34:76–84. https://doi.org/10.1016/j.ebiom.2018.07.029 Xu H, Liu J, Chen Z, Wang C, Liu Y, Wang M, Zhou P, Luo H, Ren J (2022) Intratumoral and peritumoral radiomics based on dynamic contrast-enhanced MRI for preoperative prediction of intraductal component in invasive breast cancer. Eur Radiol 32(7):4845–4856. https://doi.org/10.1007/s00330-022-08539-3 Xu X, Li H, Wang S, Fang M, Zhong L, Fan W, Dong D, Tian J, Zhao X (2019) Multiplanar MRI-Based Predictive Model for Preoperative Assessment of Lymph Node Metastasis in Endometrial Cancer. Front Oncol 9:1007. https://doi.org/10.3389/fonc.2019.01007 Yamada I, Miyasaka N, Kobayashi D, Wakana K, Oshima N, Wakabayashi A, Sakamoto J, Saida Y, Tateishi U, Eishi Y (2019) Endometrial Carcinoma: Texture Analysis of Apparent Diffusion Coefficient Maps and Its Correlation with Histopathologic Findings and Prognosis. Radiol Imaging Cancer 1(2):e190054. https://doi.org/10.1148/rycan.2019190054 Ytre-Hauge S, Dybvik JA, Lundervold A, Salvesen ØO, Krakstad C, Fasmer KE, Werner HM, Ganeshan B, Høivik E, Bjørge L, Trovik J, Haldorsen IS (2018) Preoperative tumor texture analysis on MRI predicts high-risk disease and reduced survival in endometrial cancer. J Magn Reson Imaging 48(6):1637–1647. https://doi.org/10.1002/jmri.26184 Yu Y, He Z, Ouyang J, Tan Y, Chen Y, Gu Y, Mao L, Ren W, Wang J, Lin L, Wu Z, Liu J, Ou Q, Hu Q, Li A, Chen K, Li C, Lu N, Li X, Su F, Liu Q, Xie C, Yao H (2021) Magnetic resonance imaging radiomics predicts preoperative axillary lymph node metastasis to support surgical decisions and is associated with tumor microenvironment in invasive breast cancer: A machine learning, multicenter study. eBioMedicine 69. https://doi.org/10.1016/j.ebiom.2021.103460 Yuan Y, Ren J, Tao X (2021) Machine learning-based MRI texture analysis to predict occult lymph node metastasis in early-stage oral tongue squamous cell carcinoma. Eur Radiol 31(9):6429–6437. https://doi.org/10.1007/s00330-021-07731-1 Zhang Z, Wan X, Lei X, Wu Y, Zhang J, Ai Y, Yu B, Liu X, Jin J, Xie C, Jin X (2023) Intra- and peri-tumoral MRI radiomics features for preoperative lymph node metastasis prediction in early-stage cervical cancer. Insights Imaging 14(1):65. https://doi.org/10.1186/s13244-023-01405-w Zheng R-R, Cai M-T, Lan L, Huang XW, Yang YJ, Powell M, Lin F (2022) An MRI-based radiomics signature and clinical characteristics for survival prediction in early-stage cervical cancer. Br J Radiol 95(1129):20210838. https://doi.org/10.1259/bjr.20210838 Additional Declarations No competing interests reported. Supplementary Files 3Supplement.docx 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-4271155","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":291897807,"identity":"bae288d3-45f4-4c2c-ac46-aa7255fb92a2","order_by":0,"name":"Jing Liu","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":false,"prefix":"","firstName":"Jing","middleName":"","lastName":"Liu","suffix":""},{"id":291897809,"identity":"0fa43b94-56de-4ff6-a056-f0af6353c416","order_by":1,"name":"Mingxuan Zhu","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":false,"prefix":"","firstName":"Mingxuan","middleName":"","lastName":"Zhu","suffix":""},{"id":291897810,"identity":"8fa8f305-10f1-4886-8d8c-87374bce49bc","order_by":2,"name":"Li Li","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Li","suffix":""},{"id":291897813,"identity":"375a8e92-2113-43b7-9443-32f21380c165","order_by":3,"name":"Lele Zang","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":false,"prefix":"","firstName":"Lele","middleName":"","lastName":"Zang","suffix":""},{"id":291897814,"identity":"3ac57d78-3b33-419f-b57d-fb107547974c","order_by":4,"name":"Lan Luo","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":false,"prefix":"","firstName":"Lan","middleName":"","lastName":"Luo","suffix":""},{"id":291897815,"identity":"43b84aba-1c96-40c3-93ed-b60352917d2d","order_by":5,"name":"Fei Zhu","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":false,"prefix":"","firstName":"Fei","middleName":"","lastName":"Zhu","suffix":""},{"id":291897816,"identity":"88c4bfee-3931-4d46-89be-851830c7d027","order_by":6,"name":"Huiqi Zhang","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":false,"prefix":"","firstName":"Huiqi","middleName":"","lastName":"Zhang","suffix":""},{"id":291897817,"identity":"b69174c1-1039-4a4a-b17c-8f0520d13ec3","order_by":7,"name":"Qin Xu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1UlEQVRIiWNgGAWjYBACNv7+jw8kKmrk7I83HyBOC5/EAWMDizPHjBnOHEsgToscQ4KZRGUbc2LDjRwDIh3GcCBN4gYbmzHjjJyPN94w2MnpNhDSwtxw2HIGj4wcM8/bzZZzGJKNzQ4QtOVg420JCTZjNvbcbdI8DAcStxHWkswg/ceAObGHIecZsVrSmCQkEpgTZ3DksBGpReIMs4HEgWPGBjzHjC3nGBDhF/n+HsYHkv9q5AzYmx/eeFNhJ0dQCwqQ4CEyapC1kKpjFIyCUTAKRgQAAG0XP1KcUY3/AAAAAElFTkSuQmCC","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital (Fujian Branch of Fudan University Shanghai Cancer Center)","correspondingAuthor":true,"prefix":"","firstName":"Qin","middleName":"","lastName":"Xu","suffix":""}],"badges":[],"createdAt":"2024-04-15 17:27:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4271155/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4271155/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":55113930,"identity":"ab37d0d7-7f27-426a-95f2-20b36ae645c8","added_by":"auto","created_at":"2024-04-22 19:16:24","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":37925,"visible":true,"origin":"","legend":"\u003cp\u003eArtificial intelligence workflow and study flowchart.\u003c/p\u003e","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4271155/v1/f8fbde58b7174c886bc8b929.png"},{"id":55113932,"identity":"d95b0083-b92d-4f52-a008-f9efaa08fb60","added_by":"auto","created_at":"2024-04-22 19:16:24","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":16869,"visible":true,"origin":"","legend":"\u003cp\u003eComparisons of predict scores of each patient for predicting the LN metastasis status in the training cohort. AdaBoost (A); Logistic Regression (B); Naïve Bayes (C); SVM(D); and XGBoost (E). Red bars represent patients who developed LNM. Blue bars represent that LNM was negative.\u003c/p\u003e","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-4271155/v1/9c7fb0aeaeec0933a45afa57.png"},{"id":55113931,"identity":"675f570f-9f3e-4520-bff4-048d5001fae6","added_by":"auto","created_at":"2024-04-22 19:16:24","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":57525,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic (ROC) curves, Calibration curves, and decision curves of machine-learning models. The ROC curves in the training cohort (A) and the validation cohort (B); calibration curves in the training cohort (C) and the validation cohort (D); decision curve in the training cohort (E) and the validation cohort (F).\u003c/p\u003e","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-4271155/v1/982cb09bf5f96aae1170910d.png"},{"id":55113933,"identity":"ec0aa20b-9c46-49dd-9937-db24ff2c6515","added_by":"auto","created_at":"2024-04-22 19:16:24","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":16028,"visible":true,"origin":"","legend":"\u003cp\u003eKaplan-Meier curves of survival probability between the low and high radiomics score (RS) in all patients.\u003c/p\u003e","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-4271155/v1/c943c6d835c217913dc4386e.png"},{"id":64727477,"identity":"1dc89d03-1a63-4ddf-8adf-ace379a7ed41","added_by":"auto","created_at":"2024-09-18 06:00:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":824310,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4271155/v1/7a951b5d-8a10-40af-ac8d-2e1c504b57a6.pdf"},{"id":55113934,"identity":"cf166086-317f-46b1-8c20-38b8a5e423e1","added_by":"auto","created_at":"2024-04-22 19:16:24","extension":"docx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":415788,"visible":true,"origin":"","legend":"","description":"","filename":"3Supplement.docx","url":"https://assets-eu.researchsquare.com/files/rs-4271155/v1/883308da34d3ce993286b175.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine Learning based Radiomics from Multiparametric Magnetic Resonance Imaging for Predicting Lymph Node Metastasis in Cervical Cancer","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eCervical cancer (CC) ranked as the fourth most frequently diagnosed cancer and the fourth leading cause of cancer-related deaths among women(Sung et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Lymph node (LN) metastasis is a crucial factor in the prognosis of cervical cancer, influencing treatment decisions(Lapuz et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Ikeda et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Olawaiye et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Despite the gold standard being pathological examination, its time delay hampers prompt treatment selection. Therefore, there is a crucial need for noninvasive and precise LN metastasis prediction to ensure accurate patient staging in CC and subsequently guide the selection of the most suitable treatment.\u003c/p\u003e \u003cp\u003eIn clinical practice, diverse imaging methods such as ultrasound (US), computed tomography (CT), and magnetic resonance imaging (MRI) are used to identify LN status in cancer patients(Petersen et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). MRI furnishes more intricate anatomical details for diagnosis and encompasses a richer array of texture information in its images than US and CT(L et al. 2020). Despite these advantages, imaging methods face constraints in evaluating LN metastasis in cervical cancer and are not recommended as the primary basis for determination, with limited sensitivity and accuracy(Huang and Fang \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRadiomics extracts and quantifies features from medical images to evaluate pixel-level characteristics that are imperceptible to the human eye.(Aerts \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Furthermore, the integration of machine-learning (ML) models with radiomics has demonstrated excellent outcomes in medical diagnosis, assisting clinicians in making improved decisions and predictions(Choy et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). According to reports(Yu et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Pereira et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Xu et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Sheng et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zhang et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Fan et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), radiomics and machine learning can be used for subtype classification, survival prediction, and assessing the impact of different diseases. Predicting LN metastasis using MRI features in CC poses a contemporary challenge. Several radiomics studies have been conducted to address this problem(Wang et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Dong et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Wang et al(Wang et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) constructed a machine learning model that constructed a Support Vector Machine (SVM) model based on T2WI and DWI images that predicted LN metastasis in early-stage cervical cancer. However, research on the integration of radiomics and machine learning for predicting LN metastasis and prognosis in cervical cancer is limited. Existing studies show a correlation between radiomic features in the primary tumor region and LN metastasis(Huang et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Wu et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Forghani et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis study aims to investigate the diagnostic performance of machine-learning models based on multi-parametric 3D radiomics for preoperative cervical cancer LN metastasis and evaluate the role of radiomic features in predicting cervical cancer prognosis.\u003c/p\u003e"},{"header":"2 Material and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Patients\u003c/h2\u003e \u003cp\u003e The ethics committee approved this retrospective study, and no informative consent requirements were necessary. A total of 407 patients with histologically confirmed cervical cancer between February 2009 to June 2013 in Fujian Cancer Hospital were enrolled in the study. All patient characteristics had complete clinical information, imaging results, pathology diagnoses, and prognostic information. Patients were enrolled based on the following inclusion criteria: (1) 18 years or older; (2) performance status, Eastern Cooperative Oncology Group (ECOG)\u0026thinsp;\u0026le;\u0026thinsp;2; (3) histologically confirmed squamous carcinoma, adenocarcinoma, or adenosquamous carcinoma of the cervix; (4) International Federation of Gynecology and Obstetrics (FIGO) stage (2009) from IB2 to IIB. The exclusion criteria were: (1) incomplete information or severe imaging artifacts of the MRI images; (2) lesion diameter\u0026thinsp;\u0026lt;\u0026thinsp;5 mm on MRI images. According to the proportion of 7: 3, all cervical cancer patients were divided into the training cohort and the validation cohorts.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 MRI image acquisition and segmentation\u003c/h2\u003e \u003cp\u003eMagnetic resonance (MR) images were obtained using a 1.5 Tesla MR system (GE, Fairfield City, USA). Conventional MR scanning utilized 8-channel phased-array surface coil pairs, while DWI scanning incorporated an external-body surface coil. MRI sequence scanning included Axial T1-weighted fast spin echo (FSE), axial T2-weighted FSE, axial DWI, Sagittal T2-weighted FSE, Axial CE three-dimensional spoiled gradient echo (3D SGRE) and Coronal CE 3D SGRE. DWI was performed before the intravenous injection of gadolinium. A 0.2 mmol/kg of body weight intravenous injection of gadopentetate dimeglumine (Magnevist, ScheringAG Germany) was administered for the post-gadolinium series. Preprocessing for image standardization included the entire gross tumor volume (GTV) cropping, cropping all images to a size of 32\u0026times;256\u0026times;256 based on the GTV position, and resampling anisotropic voxels into 1\u0026times;1\u0026times;1 mm\u0026sup3; using linear interpolation.\u003c/p\u003e \u003cp\u003eOne experienced radiologists (ten years of pelvic MRI reading expertise) manually outlined 3D tumor contours on axial slices using 3D Slicer software (version 4.11, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.slicer.org\u003c/span\u003e\u003cspan address=\"https://www.slicer.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The segmentation results were confirmed by another gynecologic oncology expert (Q.X, with 20 years of clinical experience). Both radiologists remained blinded to the clinical and histopathological data throughout this process. Finally, images and image masks were exported as 3D files in the Neuroimaging Informatics Technology Initiative format.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Features extraction and selection\u003c/h2\u003e \u003cp\u003eThe radiomics feature extraction was performed in Python (version 3.8.3) using a Pyradiomics package (version 3.0.1, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/Radiomics/pyradiomics)(van Griethuysen et al. 2017)\u003c/span\u003e\u003cspan address=\"https://github.com/Radiomics/pyradiomics)(van Griethuysen et al. 2017)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Quantitative radiomic features were extracted from the tumor regions of interest (ROI) in T1WI, T2WI, and DWI, including shape-based, first-order statistical, texture, and wavelet features.\u003c/p\u003e \u003cp\u003eBefore further analysis, all the extracted radiomics features underwent standardization into a normal distribution with z-scores, effectively mitigating the differences in the value scales of the data(Carr\u0026eacute; et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In pursuit of selecting highly relevant and non-redundant features, a three-step method was applied. Firstly, the Mann\u0026ndash;Whitney U test was conducted to select the features, retaining those with p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 as significantly different. Secondly, we sequentially used the Spearman correlation analysis to eliminate redundant radiomic features, with features exhibiting a Spearman correlation coefficient exceeding 0.9 being excluded. Finally, utilizing the Least Absolute Shrinkage and Selection Operator (LASSO) regression algorithm for further dimensionality reduction and optimized feature selection.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Development and Validation of machine learning\u0026ndash;based models\u003c/h2\u003e \u003cp\u003eBased on the selected optimal feature subset, models were constructed using machine-learning algorithms such as logistic regression (LR), Na\u0026iuml;ve Bayes, support vector machine (SVM), AdaBoost, and XGBoost. LR learns a probabilistic model for binary classification by optimizing its parameters to align its predictions with the actual class labels as closely as possible. Naive Bayes is a classification algorithm based on Bayes' theorem. The objective of SVM is to identify an optimal hyperplane that separates samples of different classes while maximizing the margin between the two classes. AdaBoost is an ensemble learning algorithm that constructs a strong classifier by combining multiple weak classifiers. XGBoost is based on the framework of boosting trees, constructing multiple weak learners sequentially, with each learner correcting the errors of its predecessor, continuously optimizing predictive performance. The performance of each machine learning-based model was evaluated using receiver operating characteristic curve (ROC), decision curve analysis (DCA), and calibration curves. The performance of algorithms was also assessed in terms of accuracy, sensitivity, and specificity. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrates the study design and workflow.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Statistical analysis\u003c/h2\u003e \u003cp\u003eFeature extraction, feature selection, and training and validation of machine learning models were conducted using Python (version 3.6, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.python.org\u003c/span\u003e\u003cspan address=\"https://www.python.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). To compare patient characteristics between the training and validation groups, T-test was applied for continuous variables, and Chi-square test (for groups with both n\u0026thinsp;=\u0026thinsp;5) or Yates' corrected Chi-square test (for groups with both n\u0026thinsp;=\u0026thinsp;5) was employed for categorical variables. A p-value greater than 0.05 indicated no significant difference between the groups(Hodneland et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Lin et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The reliability of the radiomics scoring risk stratification system was evaluated through Kaplan-Meier curves. All results with a p-value less than 0.05 were considered statistically significant.\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Patient characteristics\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e outlines the clinical and demographic characteristics of the training cohort (n\u0026thinsp;=\u0026thinsp;284) and the validation cohort (n\u0026thinsp;=\u0026thinsp;123). The median age of patients in both the training and independent validation cohort was (47\u0026thinsp;\u0026plusmn;\u0026thinsp;8) years. The number of patients with lymph node metastasis was 68 (23.9%) in the training cohort and 31 (25.2%) in the validation cohort. The median follow-up time for surviving patients was 103 months for the training cohort and 110 months for the validation cohort. The results indicate no significant statistical differences in clinical characteristics between the two cohorts of patients.\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\u003eClinical characteristics of patients in training and validation cohorts.\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=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"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\u003etraining (n\u0026thinsp;=\u0026thinsp;284)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003evalidation(n\u0026thinsp;=\u0026thinsp;123)\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\u003eage, mean\u0026thinsp;\u0026plusmn;\u0026thinsp;sd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e47.16\u0026thinsp;\u0026plusmn;\u0026thinsp;8.304\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47.25\u0026thinsp;\u0026plusmn;\u0026thinsp;7.885\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.197\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMacroscopic type, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.343\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e148(52.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e72(58.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e134(47.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51(41.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePTC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2(0.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0(0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTumor size (cm), n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.235\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e159(56.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e61(49.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e125(44.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e62(50.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePostoperative pathological, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.393\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSCC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e265(93.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e110(89.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17(6.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12(9.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSAC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2(0.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1(0.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDifferentiation degree, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.588\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow grade\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64(22.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25(20.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMiddle grade\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e215(75.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e94(76.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHigh grade\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5(1.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4(3.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDepth Of Tumor Invasion, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.619\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22(7.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8(6.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNear total\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e63(22.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21(17.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSuperficial\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e93(32.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45(36.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e106(37.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49(39.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCorpus invasion, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.991\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e261(91.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e113(91.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23(8.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10(8.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eParametrial invasion, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e272(95.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e120(97.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12(4.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3(2.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVaginal invasion, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.271\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e263(92.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e118(95.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21(7.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5(4.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLN, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.786\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e216(76.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e92(74.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e68(23.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31(25.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eNT, Nodular Type; CT, Cauliflower-like Type; PTC, Postoperative Pathological Type Change; SAC, adenosquamous carcinoma; AC, adenocarcinoma; SCC, squamous cell carcinoma; LN, lymph node.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Feature selection\u003c/h2\u003e \u003cp\u003eA total of 4650 radiomic features were initially extracted from the ROIs of contrast-enhanced T1-weighted imaging (ceT1WI), T2-weighted imaging (T2WI), and diffusion-weighted imaging (DWI). Features highly correlated with LN metastasis were selected using the Mann-Whitney U test (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), reducing the features to 1356. Subsequent Spearman correlation testing retained 92 features with coefficients exceeding 0.9. The correlation of variables was assessed using a heatmap (Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Finally, LASSO regression identified 18 radiomic features most correlated with lymph node metastasis, and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e details these features along with their LASSO coefficients.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eSupplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e.\u003c/b\u003e The correlation between MRI radiomic feature. Correlation analysis was used to estimate the strength of the correlations with Spearman p.\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\u003eRadiomics signature selection results with descriptions.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFeature\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLASSO coefficient\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eceT1WI_original_glrlm_RunVariance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.029388\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eceT1WI _log-sigma-1-0-mm-3D_firstorder_90Percentile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.00053\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eceT1WI _squareroot_firstorder_10Percentile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.184889\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI_original_shape_Elongation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.086653\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI _original_shape_Flatness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.045143\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI _original_glcm_Imc1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.032342\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI _log-sigma-1-0-mm-3D_firstorder_Kurtosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.207755\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI _log-sigma-1-0-mm-3D_glcm_ClusterShade\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.003092\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI _log-sigma-3-0-mm-3D_firstorder_Skewness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.072891\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI _wavelet-LLH_firstorder_Maximum\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.02597\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2WI _wavelet-HLL_firstorder_Minimum\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.266135\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI_original_firstorder_90Percentile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.00246\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI_log-sigma-1-0-mm-3D_firstorder_Skewness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.150593\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI_wavelet-LLH_firstorder_Kurtosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.108992\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI_wavelet-HLL_firstorder_Skewness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.03973\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI_wavelet-HHL_firstorder_Maximum\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.050664\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI_exponential_glcm_Imc1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.150122\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI_squareroot_ngtdm_Coarseness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.01102\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Performance of prediction models\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e summarizes the predictive performance of all machine learning models. Notably, XGBoost exhibited superior performance, achieving a classification accuracy of 0.927 and an AUC of 0.897 for predicting the lymph node status of cervical cancer patients in the validation set. Its sensitivity, specificity, positive predictive value, negative predictive value, precision, and F1 score were 0.741, 0.989, 0.958, 0.919, 0.958, and 0.836, respectively. In comparison, other models showed varying levels of accuracy and AUC in the validation set: LR (0.667, 0.715), Na\u0026iuml;ve Bayes (0.724, 0.734), SVM (0.740, 0.791), AdaBoost (0.683, 0.737). These findings underscore the superior predictive performance of the XGBoost model in cervical cancer lymph node status prediction. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e visually depicts the prediction scores for each patient's lymph node status, highlighting XGBoost's effectiveness in distinguishing between LN-negative and LN-positive groups.\u003c/p\u003e \u003cp\u003eFor a clearer comparison of different models in predicting the lymph node status of cervical cancer, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-B illustrate the AUCs for all models on both the training and validation cohorts. The DCA curve was used to evaluate the clinical values of these models (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC-D). Assuming no patients have LNM in cervical cancer, the solid black line (negative line) indicates that the net benefit is zero when no patient accepts therapy. Conversely, the solid grey line (positive line) represents the net benefits when all patients with LNM receive therapy. The XGBoost model demonstrated higher net benefits compared to the two extreme lines (negative line and positive line) in both cohorts. The noteworthy point is that XGBoost demonstrated significantly superior performance compared to the others across most threshold points. Calibration curves reflect the degree of consistency between the observed risk and predicted probabilities of a model. Figures\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE-F confirm that the XGBoost model exhibits the optimal level of consistency.\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\u003eEvaluation indicators of predictive performance of five models.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"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 \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eClassifier\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eACC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSEN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSPE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003ePRE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAdaBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTraining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.6620\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7862\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.8824\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.5926\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.4054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.9412\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.4054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.5556\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\u003eValidation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.6829\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7370\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.7419\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.6703\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.4259\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.8841\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.4259\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.5412\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNa\u0026iuml;ve Bayes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTraining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.5563\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7044\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.8529\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.4630\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.3333\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.9091\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.3333\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.4793\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\u003eValidation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.7236\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7342\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.7419\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.7174\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.4694\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.8919\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.4694\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.5750\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTraining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.6901\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7271\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.7206\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.6806\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.4153\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.8855\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.4153\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.5269\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\u003eValidation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.6667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7149\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.6774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.6630\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.4038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.8592\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.4038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.5060\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTraining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.8486\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.8386\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.7059\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.8935\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.6761\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.9061\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.6761\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.6906\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\u003eValidation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.7398\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7907\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.8065\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.7174\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.4902\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.9167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.4902\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.6098\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=\"left\" colname=\"c2\"\u003e \u003cp\u003eTraining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.9296\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.9054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.8088\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.9676\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.8871\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.9414\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.8871\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.8462\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\u003eValidation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.9268\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.8969\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.7419\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.9891\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.9583\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.9192\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.9583\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.8364\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eLR, Logistic Regression; SVM, Support Vector Machine; ACC, accuracy; AUC, area under ROC; SEN, sensitivity; SPE, specificity; PPV, positive predictive value; NPV, negative predictive value; PRE, precision.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Kaplan-Meier analysis\u003c/h2\u003e \u003cp\u003eKaplan-Meier analysis further validates the prognostic value of the selected radiomic features. The Kaplan-Meier curves in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e demonstrate a significant difference in survival probabilities between the two risk subgroups (P\u0026thinsp;=\u0026thinsp;0.005). Patients with higher rad-score exhibit lower overall survival rates (HR 1.9, 95% CI 1.2-3.0, P\u0026thinsp;=\u0026thinsp;0.006). Across the entire cohort, the stratification of rad-score is significantly associated with progression-free survival (PFS).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eIn this study, we developed and validated various radiomics-based machine learning diagnostic models for predicting lymph node metastasis in cervical cancer using multiparametric magnetic resonance imaging (MRI). Leveraging preoperative MRI with multiple parameters, we extracted 4065 imaging features, ultimately selecting 18 features and constructing multiple machine learning diagnostic models. The XGBoost model demonstrated superior performance in predicting lymph node metastasis in the validation set, with an accuracy of 0.9268 and an AUC of 0.8969, outperforming other models.\u003c/p\u003e \u003cp\u003ePrevious studies have indicated that radiomics analysis can extract quantitative image features from medical images, thereby improving tumor diagnosis, staging, and prognosis(Lambin et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fang et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Yuan et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Fang et al.'s research(Fang et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) demonstrated that radiomics scores derived from MRI as prognostic biomarkers for early-stage cervical cancer patients. Kan et al.(Kan et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) utilized T2WI and ceT1WI to extract radiomic features and employed an SVM model(AUC(95% CI), 0.754(0584\u0026ndash;0.924)) to predict lymph node metastasis in early-stage cervical cancer. However, some studies only utilized features extracted from a single MRI image(Xu et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Yamada et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), or performed radiomics analysis solely on 2D slices(Ytre-Hauge et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). In contrast, our approach involved using three common 3D MRI sequences in cervical cancer patients for feature extraction, allowing for the extraction of rich radiomic features and enhancing the model's reliability. Moreover, we systematically compared multiple machine learning models and identified the best-performing model, showing consistent and clinically excellent performance between the training and validation sets (AUC\u0026thinsp;=\u0026thinsp;0.897 in the validation set), which suggests the excellence and reliability of our research methodology.\u003c/p\u003e \u003cp\u003eOur research utilized a diverse set of machine learning algorithms, such as LR, Na\u0026iuml;ve Bayes, SVM, AdaBoost and XGBoost. XGBoost is an efficient implementation of the widely-used gradient boosting decision tree algorithm in the field of science. XGBoost is an efficient gradient boosting algorithm that iteratively trains decision trees, incorporates regularization and parallel processing, and constructs a powerful ensemble model suitable for regression and classification tasks. In various machine learning problems, especially in scenarios with a large number of features, high dimensions, and complex structures, XGBoost often exhibits excellent performance. In another study, based on another way of model construction, Sheng et al.(Sheng et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) found the XGBoost model demonstrated a better performance in Invasive ductal breast cancer molecular subtype prediction, especially in the triple-negative and non-triple-negative groups with an AUC of 0.903. Song(Song \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) used the XGBoost algorithm to build a radiomics model for predicting axillary LNM in invasive ductal breast cancer with an AUC of 0.890. In our study, XGBoost model demonstrated superior prediction performance, achieving an AUC of 0.897, surpassing the SVM model which obtained an AUC of 0.791.\u003c/p\u003e \u003cp\u003ePreceding investigations have elucidated the prospective efficacy of radiology in prognosticating the survival outcomes of individuals with cervical cancer, boasting superior accuracy than traditional clinical parameters(Liu et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zheng et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Our study aimed to evaluate the prognostic significance of MRI-derived radiological characteristics in cervical cancer patients, with whom higher rad-score exhibit lower overall survival rates (HR 1.9, 95% CI 1.2-3.0, P\u0026thinsp;=\u0026thinsp;0.006). The precise anticipation of disease progression via radiology holds pivotal importance in clinical settings, guiding the selection of optimal therapeutic strategies and thereby enhancing patient prognoses.\u003c/p\u003e \u003cp\u003eDespite the positive results of this research, it is also fundamental to consider the limitations of this study. This study is retrospective and conducted in a single center, lacking external validation, which may introduce selection bias. To enhance the robustness of the models, a larger sample size from multiple centers is necessary. Furthermore, a manual segmentation approach is time-consuming and may impact the precision of feature extraction in certain cases. Previous research has indicated that utilizing an automatic segmentation method could streamline radiomics applications in clinical settings, promoting high levels of intra- and inter-observer reproducibility(Liu et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Chen et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e"},{"header":"5 Conclusions","content":"\u003cp\u003eIn conclusion, machine learning is a step towards precision medicine in the field of gynecologic oncology. The preoperative assessment of the LNs is important for accurately staging patients with cervical cancer and for selecting the most suitable treatment. We extracted radiomic features from multi-parametric MRI and employed various machine learning models to predict lymph node metastasis in cervical cancer patients, achieving a commendable level of accuracy.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eMRI \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;magnetic resonance imaging\u003c/p\u003e\n\u003cp\u003eLN \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;lymph node\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCC \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;cervical cancer\u003c/p\u003e\n\u003cp\u003eceT1W \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;contrast-enhanced\u0026nbsp;T1-weighted\u003c/p\u003e\n\u003cp\u003eT2W \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; T2-weighted\u003c/p\u003e\n\u003cp\u003eDWI \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; diffusion-weighted imaging\u003c/p\u003e\n\u003cp\u003eROIs \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; regions of interest\u003c/p\u003e\n\u003cp\u003eAUC \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; the area under the receiver-operating characteristic curve\u003c/p\u003e\n\u003cp\u003eACC \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; accuracy\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePFS \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;progression-free survival\u003c/p\u003e\n\u003cp\u003eHR \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;hazard ratio\u003c/p\u003e\n\u003cp\u003eUS \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;ultrasound\u003c/p\u003e\n\u003cp\u003eCT \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;computed tomography\u003c/p\u003e\n\u003cp\u003eML \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; machine learning\u003c/p\u003e\n\u003cp\u003eSVM \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Support Vector Machine\u003c/p\u003e\n\u003cp\u003eECOG \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Eastern Cooperative Oncology Group\u003c/p\u003e\n\u003cp\u003eFIGO \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;International Federation of Gynecology and Obstetrics\u003c/p\u003e\n\u003cp\u003eGTV \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; gross tumor volume\u003c/p\u003e\n\u003cp\u003eLASSO \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Least Absolute Shrinkage and Selection Operator\u003c/p\u003e\n\u003cp\u003eLR \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Logistic Regression\u003c/p\u003e\n\u003cp\u003eROC \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;receiver operating characteristic\u003c/p\u003e\n\u003cp\u003eDCA \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;decision curve analysis\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eStatements \u0026amp; Declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe project was funded by the grants of Joint Funds for the National Clinical Key Specialty Construction Program (2021); Innovative Medicine Subject of Fujian Provincial Health Commission (2020CX0101); Natural Science Foundation of Fujian Province (2020J011126, 2023J011273); Fujian Provincial Clinical Research Center for Cancer Radiotherapy and Immunotherapy (2020Y2012); Startup Fund for scientific research, Fujian Medical University (Grant number：2020QH1233).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Jing Liu, Li Li and Lele Zang. The first draft of the manuscript was written by Mingxuan Zhu and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated during and 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\u003e\u003c/p\u003e\n\u003cp\u003eThis is an observational study. The XYZ Research Ethics Committee has confirmed that no ethical approval is required.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInformed consent was obtained from all individual participants included in the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publish\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAerts HJWL (2016) The Potential of Radiomic-Based Phenotyping in Precision Medicine: A Review. JAMA Oncol 2(12):1636\u0026ndash;1642. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jamaoncol.2016.2631\u003c/span\u003e\u003cspan address=\"10.1001/jamaoncol.2016.2631\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarr\u0026eacute; A, Klausner G, Edjlali M, Lerousseau M, Briend-Diop J, Sun R, Ammari S, Reuz\u0026eacute; S, Alvarez Andres E, Estienne T, Niyoteka S, Battistella E, Vakalopoulou M, Dhermain F, Paragios N, Deutsch E, Oppenheim C, Pallud J, Robert C (2020) Standardization of brain MR images across machines and protocols: bridging the gap for MRI-based radiomics. Sci Rep 10(1):12340. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-020-69298-z\u003c/span\u003e\u003cspan address=\"10.1038/s41598-020-69298-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen X, Men K, Chen B, Tang Y, Zhang T, Wang S, Li Y, Dai J (2020) CNN-Based Quality Assurance for Automatic Segmentation of Breast Cancer in Radiotherapy. Front Oncol 10:524. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fonc.2020.00524\u003c/span\u003e\u003cspan address=\"10.3389/fonc.2020.00524\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoy G, Khalilzadeh O, Michalski M, Do S, Samir AE, Pianykh OS, Geis JR, Pandharipande PV, Brink JA, Dreyer KJ (2018) Current Applications and Future Impact of Machine Learning in Radiology. Radiology 288(2):318\u0026ndash;328. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1148/radiol.2018171820\u003c/span\u003e\u003cspan address=\"10.1148/radiol.2018171820\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDong T, Yang C, Cui B, Zhang T, Sun X, Song K, Wang L, Kong B, Yang X (2020) Development and Validation of a Deep Learning Radiomics Model Predicting Lymph Node Status in Operable Cervical Cancer. Front Oncol 10:464. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fonc.2020.00464\u003c/span\u003e\u003cspan address=\"10.3389/fonc.2020.00464\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFan Y, Wang X, Dong Y, Cui E, Wang H, Sun X, Su J, Luo Y, Yu T, Jiang X (2023) Multiregional radiomics of brain metastasis can predict response to EGFR-TKI in metastatic NSCLC. Eur Radiol. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-023-09709-7\u003c/span\u003e\u003cspan address=\"10.1007/s00330-023-09709-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFang J, Zhang B, Wang S, Jin Y, Wang F, Ding Y, Chen Q, Chen L, Li Y, Li M, Chen Z, Liu L, Liu Z, Tian J, Zhang S (2020) Association of MRI-derived radiomic biomarker with disease-free survival in patients with early-stage cervical cancer. Theranostics 10(5):2284\u0026ndash;2292. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7150/thno.37429\u003c/span\u003e\u003cspan address=\"10.7150/thno.37429\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eForghani R, Chatterjee A, Reinhold C, P\u0026eacute;rez-Lara A, Romero-Sanchez G, Ueno Y, Bayat M, Alexander JWM, Kadi L, Chankowsky J, Seuntjens J, Forghani B (2019) Head and neck squamous cell carcinoma: prediction of cervical lymph node metastasis by dual-energy CT texture analysis with machine learning. Eur Radiol 29(11):6172\u0026ndash;6181. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-019-06159-y\u003c/span\u003e\u003cspan address=\"10.1007/s00330-019-06159-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHodneland E, Kaliyugarasan S, Wagner-Larsen KS, Lura N, Andersen E, Bartsch H, Smit N, Halle MK, Krakstad C, Lundervold AS, Haldorsen IS (2022) Fully Automatic Whole-Volume Tumor Segmentation in Cervical Cancer. Cancers (Basel) 14(10):2372. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/cancers14102372\u003c/span\u003e\u003cspan address=\"10.3390/cancers14102372\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang B-X, Fang F (2018) Progress in the Study of Lymph Node Metastasis in Early-stage Cervical Cancer. Curr Med Sci 38(4):567\u0026ndash;574. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11596-018-1915-0\u003c/span\u003e\u003cspan address=\"10.1007/s11596-018-1915-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang Y-Q, Liang C-H, He L, Tian J, Liang C-S, Chen X, Ma Z-L, Liu Z-Y (2016) Development and Validation of a Radiomics Nomogram for Preoperative Prediction of Lymph Node Metastasis in Colorectal Cancer. J Clin Oncol 34(18):2157\u0026ndash;2164. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1200/JCO.2015.65.9128\u003c/span\u003e\u003cspan address=\"10.1200/JCO.2015.65.9128\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIkeda M, Shida M, Shigeta S, Nagase S, Takahashi F, Yamagami W, Katabuchi H, Yaegashi N, Aoki D, Mikami M (2021) The trend and outcome of postsurgical therapy for high-risk early-stage cervical cancer with lymph node metastasis in Japan: a report from the Japan Society of Gynecologic Oncology (JSGO) guidelines evaluation committee. J Gynecol Oncol 32(3):e44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3802/jgo.2021.32.e44\u003c/span\u003e\u003cspan address=\"10.3802/jgo.2021.32.e44\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKan Y, Dong D, Zhang Y, Jiang W, Zhao N, Han L, Fang M, Zang Y, Hu C, Tian J, Li C, Luo Y (2019) Radiomic signature as a predictive factor for lymph node metastasis in early-stage cervical cancer. J Magn Reson Imaging 49(1):304\u0026ndash;310. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jmri.26209\u003c/span\u003e\u003cspan address=\"10.1002/jmri.26209\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL L, Q L, L T (2020) Diagnostic value and clinical significance of MRI and CT in detecting lymph node metastasis of early cervical cancer. Oncology letters 19(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3892/ol.2019.11180\u003c/span\u003e\u003cspan address=\"10.3892/ol.2019.11180\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, Sanduleanu S, Larue RTHM, Even AJG, Jochems A, van Wijk Y, Woodruff H, van Soest J, Lustberg T, Roelofs E, van Elmpt W, Dekker A, Mottaghy FM, Wildberger JE, Walsh S (2017) Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol 14(12):749\u0026ndash;762. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nrclinonc.2017.141\u003c/span\u003e\u003cspan address=\"10.1038/nrclinonc.2017.141\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLapuz C, Kondalsamy-Chennakesavan S, Bernshaw D, Khaw P, Narayan K (2016) Stage IB cervix cancer with nodal involvement treated with primary surgery or primary radiotherapy: Patterns of failure and outcomes in a contemporary population. J Med Imaging Radiat Oncol 60(2):274\u0026ndash;282. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/1754-9485.12411\u003c/span\u003e\u003cspan address=\"10.1111/1754-9485.12411\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin Y-C, Lin G, Pandey S, Yeh C-H, Wang J-J, Lin C-Y, Ho T-Y, Ko S-F, Ng S-H (2023) Fully automated segmentation and radiomics feature extraction of hypopharyngeal cancer on MRI using deep learning. Eur Radiol 33(9):6548\u0026ndash;6556. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-023-09827-2\u003c/span\u003e\u003cspan address=\"10.1007/s00330-023-09827-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu C, Gardner SJ, Wen N, Elshaikh MA, Siddiqui F, Movsas B, Chetty IJ (2019) Automatic Segmentation of the Prostate on CT Images Using Deep Neural Networks (DNN). Int J Radiat Oncol Biol Phys 104(4):924\u0026ndash;932. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ijrobp.2019.03.017\u003c/span\u003e\u003cspan address=\"10.1016/j.ijrobp.2019.03.017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu S, Li R, Liu Q, Sun D, Yang H, Pan H, Wang L, Song S (2022) Radiomics model of 18F-FDG PET/CT imaging for predicting disease-free survival of early-stage uterine cervical squamous cancer. Cancer Biomark 33(2):249\u0026ndash;259. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3233/CBM-210201\u003c/span\u003e\u003cspan address=\"10.3233/CBM-210201\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOlawaiye AB, Baker TP, Washington MK, Mutch DG (2021) The new (Version 9) American Joint Committee on Cancer tumor, node, metastasis staging for cervical cancer. CA Cancer J Clin 71(4):287\u0026ndash;298. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3322/caac.21663\u003c/span\u003e\u003cspan address=\"10.3322/caac.21663\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePereira HM, Leite Duarte ME, Ribeiro Damasceno I, De Oliveira Moura Santos LA, Nogueira-Barbosa MH (2021) Machine learning-based CT radiomics features for the prediction of pulmonary metastasis in osteosarcoma. BJR 94(1124):20201391. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1259/bjr.20201391\u003c/span\u003e\u003cspan address=\"10.1259/bjr.20201391\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetersen LJ, Nielsen JB, Langkilde NC, Petersen A, Afshar-Oromieh A, De Souza NM, De Paepe K, Fisker RV, Arp DT, Carl J, Haberkorn U, Zacho HD (2020) 68Ga-PSMA PET/CT compared with MRI/CT and diffusion-weighted MRI for primary lymph node staging prior to definitive radiotherapy in prostate cancer: a prospective diagnostic test accuracy study. World J Urol 38(4):939\u0026ndash;948. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00345-019-02846-z\u003c/span\u003e\u003cspan address=\"10.1007/s00345-019-02846-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSheng W, Xia S, Wang Y, Yan L, Ke S, Mellisa E, Gong F, Zheng Y, Tang T (2022) Invasive ductal breast cancer molecular subtype prediction by MRI radiomic and clinical features based on machine learning. Front Oncol 12:964605. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fonc.2022.964605\u003c/span\u003e\u003cspan address=\"10.3389/fonc.2022.964605\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSong B-I (2021) A machine learning-based radiomics model for the prediction of axillary lymph-node metastasis in breast cancer. Breast Cancer 28(3):664\u0026ndash;671. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s12282-020-01202-z\u003c/span\u003e\u003cspan address=\"10.1007/s12282-020-01202-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 71(3):209\u0026ndash;249. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3322/caac.21660\u003c/span\u003e\u003cspan address=\"10.3322/caac.21660\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Griethuysen JJM, Fedorov A, Parmar C, Hosny A, Aucoin N, Narayan V, Beets-Tan RGH, Fillion-Robin J-C, Pieper S, Aerts HJWL (2017) Computational Radiomics System to Decode the Radiographic Phenotype. Cancer Res 77(21):e104\u0026ndash;e107. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1158/0008-5472.CAN-17-0339\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-17-0339\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang T, Gao T, Yang J, Yan X, Wang Y, Zhou X, Tian J, Huang L, Zhang M (2019) Preoperative prediction of pelvic lymph nodes metastasis in early-stage cervical cancer using radiomics nomogram developed based on T2-weighted MRI and diffusion-weighted imaging. Eur J Radiol 114:128\u0026ndash;135. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejrad.2019.01.003\u003c/span\u003e\u003cspan address=\"10.1016/j.ejrad.2019.01.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu S, Zheng J, Li Y, Wu Z, Shi S, Huang M, Yu H, Dong W, Huang J, Lin T (2018) Development and Validation of an MRI-Based Radiomics Signature for the Preoperative Prediction of Lymph Node Metastasis in Bladder Cancer. EBioMedicine 34:76\u0026ndash;84. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ebiom.2018.07.029\u003c/span\u003e\u003cspan address=\"10.1016/j.ebiom.2018.07.029\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu H, Liu J, Chen Z, Wang C, Liu Y, Wang M, Zhou P, Luo H, Ren J (2022) Intratumoral and peritumoral radiomics based on dynamic contrast-enhanced MRI for preoperative prediction of intraductal component in invasive breast cancer. Eur Radiol 32(7):4845\u0026ndash;4856. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-022-08539-3\u003c/span\u003e\u003cspan address=\"10.1007/s00330-022-08539-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu X, Li H, Wang S, Fang M, Zhong L, Fan W, Dong D, Tian J, Zhao X (2019) Multiplanar MRI-Based Predictive Model for Preoperative Assessment of Lymph Node Metastasis in Endometrial Cancer. Front Oncol 9:1007. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fonc.2019.01007\u003c/span\u003e\u003cspan address=\"10.3389/fonc.2019.01007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYamada I, Miyasaka N, Kobayashi D, Wakana K, Oshima N, Wakabayashi A, Sakamoto J, Saida Y, Tateishi U, Eishi Y (2019) Endometrial Carcinoma: Texture Analysis of Apparent Diffusion Coefficient Maps and Its Correlation with Histopathologic Findings and Prognosis. Radiol Imaging Cancer 1(2):e190054. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1148/rycan.2019190054\u003c/span\u003e\u003cspan address=\"10.1148/rycan.2019190054\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYtre-Hauge S, Dybvik JA, Lundervold A, Salvesen \u0026Oslash;O, Krakstad C, Fasmer KE, Werner HM, Ganeshan B, H\u0026oslash;ivik E, Bj\u0026oslash;rge L, Trovik J, Haldorsen IS (2018) Preoperative tumor texture analysis on MRI predicts high-risk disease and reduced survival in endometrial cancer. J Magn Reson Imaging 48(6):1637\u0026ndash;1647. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jmri.26184\u003c/span\u003e\u003cspan address=\"10.1002/jmri.26184\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu Y, He Z, Ouyang J, Tan Y, Chen Y, Gu Y, Mao L, Ren W, Wang J, Lin L, Wu Z, Liu J, Ou Q, Hu Q, Li A, Chen K, Li C, Lu N, Li X, Su F, Liu Q, Xie C, Yao H (2021) Magnetic resonance imaging radiomics predicts preoperative axillary lymph node metastasis to support surgical decisions and is associated with tumor microenvironment in invasive breast cancer: A machine learning, multicenter study. eBioMedicine 69. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ebiom.2021.103460\u003c/span\u003e\u003cspan address=\"10.1016/j.ebiom.2021.103460\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan Y, Ren J, Tao X (2021) Machine learning-based MRI texture analysis to predict occult lymph node metastasis in early-stage oral tongue squamous cell carcinoma. Eur Radiol 31(9):6429\u0026ndash;6437. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-021-07731-1\u003c/span\u003e\u003cspan address=\"10.1007/s00330-021-07731-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Z, Wan X, Lei X, Wu Y, Zhang J, Ai Y, Yu B, Liu X, Jin J, Xie C, Jin X (2023) Intra- and peri-tumoral MRI radiomics features for preoperative lymph node metastasis prediction in early-stage cervical cancer. Insights Imaging 14(1):65. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13244-023-01405-w\u003c/span\u003e\u003cspan address=\"10.1186/s13244-023-01405-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng R-R, Cai M-T, Lan L, Huang XW, Yang YJ, Powell M, Lin F (2022) An MRI-based radiomics signature and clinical characteristics for survival prediction in early-stage cervical cancer. Br J Radiol 95(1129):20210838. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1259/bjr.20210838\u003c/span\u003e\u003cspan address=\"10.1259/bjr.20210838\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\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":"Cervical cancer, Lymph node metastasis, Machine learning, Magnetic resonance imaging","lastPublishedDoi":"10.21203/rs.3.rs-4271155/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4271155/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eConstruct and compare multiple machine-learning models to predict lymph node (LN) metastasis in cervical cancer, utilizing radiomic features extracted from preoperative multi-parametric magnetic resonance imaging (MRI).\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThis study retrospectively enrolled 407 patients with cervical cancer who were randomly divided into training cohort (n\u0026thinsp;=\u0026thinsp;284) and validation cohort (n\u0026thinsp;=\u0026thinsp;123). A total of 4065 radiomic features were extracted from the tumor regions of interest on contrast-enhanced T1-weighted imaging, T2-weighted imaging, and diffusion-weighted imaging for each patient. The Mann-Whitney U test, Spearman correlation analysis, and selection operator Cox regression analysis were employed for radiomic feature selection. The relationship between MRI radiomic features and LN status was analyzed by five machine-learning algorithms. Model performance was evaluated by measuring the area under the receiver-operating characteristic curve (AUC) and accuracy (ACC). Moreover, Kaplan\u0026ndash;Meier analysis was used to validate the prognostic value of selected clinical and radiomics characteristics.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eLN metastasis was pathologically detected in 24.3% (99/407) of patients. Following three-step feature selection, 18 radiomic features were employed for model construction. The XGBoost model exhibited superior performance compared to other models, achieving an AUC, accuracy, sensitivity, specificity, and F1-score of 0.9268, 0.8969, 0.7419, 0.9891, and 0.8364, respectively, on the validation set. Additionally, Kaplan\u0026thinsp;\u0026minus;\u0026thinsp;Meier curves indicated a significant correlation between radiomic scores and progression-free survival in cervical cancer patients (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eMachine learning-based multi-parametric MRI radiomic analysis demonstrates a promising performance in the preoperative prediction of LN metastasis and clinical prognosis in cervical cancer.\u003c/p\u003e","manuscriptTitle":"Machine Learning based Radiomics from Multiparametric Magnetic Resonance Imaging for Predicting Lymph Node Metastasis in Cervical Cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-04-22 19:16:19","doi":"10.21203/rs.3.rs-4271155/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":"2d29b54d-df0c-4d97-864b-73aaac409610","owner":[],"postedDate":"April 22nd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-09-18T05:36:06+00:00","versionOfRecord":[],"versionCreatedAt":"2024-04-22 19:16:19","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4271155","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4271155","identity":"rs-4271155","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}