Application Value of a Nomogram Integrating Contrast-enhanced CT Radiomics and Clinical Indicators in Evaluating Lymph Node Metastasis in Pediatric Peripheral Neuroblastoma

Application Value of a Nomogram Integrating Contrast-enhanced CT Radiomics and Clinical Indicators in Evaluating Lymph Node Metastasis in Pediatric Peripheral Neuroblastoma | 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 Application Value of a Nomogram Integrating Contrast-enhanced CT Radiomics and Clinical Indicators in Evaluating Lymph Node Metastasis in Pediatric Peripheral Neuroblastoma Wenbin Guo, Guangyong Yang, Zhengjun Dai, Jing Yang, Baoxin Qian, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8820742/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Objective To construct a machine learning model using contrast-enhanced CT radiomics and clinical indicators, aiming to improve the diagnostic accuracy for lymph node metastasis in children with peripheral neuroblastoma and provide practical evidence for clinical diagnosis and treatment. Methods A total of children with pathologically confirmed neuroblastoma were retrospectively enrolled between February 2014 and December 2024, and then randomly divided into a training set and a test set via random sampling. Radiomics features were extracted separately from CT images of the arterial phase and venous phase. In the training set, four radiomics models, one clinical model, and one combined model incorporating radiomics and clinical features were constructed respectively using the filtered radiomics features and clinical features. All models were validated against the pathological reference standard in the test set, and the area under the receiver operating characteristic curve of each model was calculated. The clinical utility of each model was evaluated using the decision curve analysis curve. The optimal model was visualized with a nomogram, and the diagnostic gain of the nomogram for evaluating lymph node metastasis in neuroblastoma children was quantified via human-machine comparison. Results A total of 225 children with neuroblastoma were enrolled in this study, (with a mean age of 2.23 ± 2.34 years and an age range of 0–13 years). All subjects were randomly divided into a training set ( n = 157) and a test set ( n = 68) at a ratio of 7:3. Compared with four radiomics models (Arterial phase, Venous phase, Delta-Absolute, Delta-Relative) and one clinical model (Ki-67), the nomogram integrating radiomics and clinical features (Arterial phase + Delta-Relative + Ki-67) exhibited superior diagnostic performance in evaluating lymph node metastasis in pediatric peripheral neuroblastoma. The AUC values of the nomogram reached 0.937 and 0.829 in the training set and validation set, respectively. In the human-machine comparison experiment, the diagnostic accuracy of radiologists for lymph node metastasis in neuroblastoma children was improved by 21% when assisted by the nomogram. Conclusion The nomogram combining contrast-enhanced CT radiomics and clinical indicators has significant diagnostic value in evaluating lymph node metastasis in pediatric patients with peripheral neuroblastoma. Moreover, it can substantially improve the diagnostic accuracy of radiologists with different levels of clinical experience. Children Neuroblastoma Lymph node metastasis Radiomics Clinical indicators Ki-67 Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1.Introduction Neuroblastoma(NB) is an embryonic tumor originating from the primitive neural crest cells of the sympathetic nervous system. This tumor can arise at any site within the sympathetic nervous system [ 1 ] . It encompasses a spectrum of tumors, such as malignant NB, mixed ganglioneuroblastoma, nodular ganglioneuroblastoma, and ganglioneuroma [ 2 ] . Peripheral neuroblastic tumors are the most common extracranial solid tumors in children [ 3 ] , with the majority of cases diagnosed before five years of age and a mean age at diagnosis of 17 months [ 4 ] . The clinical manifestations of NB are diverse, ranging from asymptomatic masses to aggressive primary tumors with extensive local invasion, widespread metastasis, or severe illness associated with these features [ 5 ] . Around 70% of NB patients already show signs of distant metastasis at the initial diagnosis, which is most frequently observed in bone marrow, followed by bones, lymph nodes, and the liver [ 6 ] . Although bone marrow metastasis is the most common in children with NB, determining the presence of lymph node metastasis is crucial for evaluating pre-treatment risk stratification, selecting surgical approaches, defining the target volume of postoperative radiotherapy, and assessing treatment outcomes in these patients [ 7 , 8 , 9 ] . In the imaging evaluation of NB, contrast-enhanced CT currently serves as a crucial diagnostic and assessment tool for pediatric peripheral NB. It enables preoperative tumor risk stratification, staging, and risk assessment based on the morphological features of lesions, intratumoral calcifications, and the involvement status of surrounding tissue structures. This facilitates determining the selection of radiotherapy and chemotherapy drugs, dosages, and treatment cycles, thereby achieving precise and personalized treatment as well as the evaluation of treatment outcomes [ 10 – 11 ] . However, due to the limitation of human visual resolution, the diagnostic accuracy of radiologists in detecting lymph node metastasis in NB patients based on conventional contrast-enhanced CT awaits further improvement. Radiomics based on contrast-enhanced CT has been widely applied in the diagnosis, grading, and prognostic evaluation of adult solid tumors. Models constructed using radiomics have also been extensively used to predict lymph node metastasis in adult tumors. These radiomics models have been explored to varying degrees in the prediction of lymph node metastasis across multiple systemic tumors in adults, including colorectal cancer, breast cancer, and endometrial cancer [ 12 , 13 , 14 ] , and have demonstrated favorable predictive benefits. However, owing to the unique characteristics of the pediatric population, such applications are far less common in children than in adults [ 15 ] . As of November 2025, the authors have not identified any relevant literature on contrast-enhanced CT-based radiomics models for predicting lymph node metastasis in pediatric patients with peripheral NB. Therefore, based on the above discussion, this study aimed to construct a contrast-enhanced CT-based radiomics model, a clinical model, and a nomogram model, to explore their application value in the diagnosis of lymph node metastasis in pediatric patients with peripheral NB, improve the diagnostic accuracy of radiologists, and provide more precise imaging evidence for prognostic evaluation and personalized treatment of NB patients. 2. Patients and Methods 2.1. Participants and Clinical Characteristics This retrospective study was approved by the Ethics Committee of the Children's Hospital Affiliated to Shandong University, and the requirement for informed consent was waived. We collected the imaging and clinical data of 358 pediatric patients diagnosed with NB via contrast-enhanced CT between February 2014 and December 2024. The inclusion criteria were as follows: (1) Newly diagnosed patients who had not received any tumor-related treatment; (2) Patients pathologically confirmed as NB via puncture or surgery; (3) Complete baseline data at initial diagnosis (including clinical data, laboratory test results, and relevant pathological data); (4) Complete pre-treatment CT imaging data, including both arterial and venous phases. The exclusion criteria were as follows: (1) Patients who had received relevant treatments such as radiotherapy, chemotherapy, or surgery prior to admission (n = 23); (2) Patients with tumors whose boundaries with surrounding tissue structures were extremely difficult to identify, or with multiple systemic masses that prevented complete delineation of the tumor region of interest (ROI) (n = 11); (3) Patients with imaging quality failing to meet the requirements for post-processing (n = 7). After screening, a total of 225 patients were enrolled in this study (112 males and 113 females; mean age, 2.23 ± 2.34 years; age range, 0–13 years), among whom 104 cases were pathologically confirmed to have lymph node metastasis. The clinical data of the patients included age and gender at initial diagnosis; laboratory examination indicators included serum ferritin (SF) and neuron-specific enolase (NSE); pathological data included the Ki-67 expression index and the presence of lymph node metastasis. The data were randomly divided into a training set and a test set at a ratio of 7:3 with a random seed of 305. Specifically, the training set comprised 157 cases, including 73 cases with lymph node metastasis and 84 cases without lymph node metastasis, while the test set included 68 cases, consisting of 31 cases with lymph node metastasis and 37 cases without lymph node metastasis.. The patient selection flowchart is shown in Fig. 1 . 2.2 CT Scanning Protocol All pediatric patients underwent contrast-enhanced CT scans without sedation. For those who were unable to cooperate, sedation was administered by certified physicians from the Sedation Center of our hospital.Scanning was performed using a GE Revolution GSI CT750 scanner. The scanning parameters were as follows: tube voltage, 80–120 kV; tube current, 50–200 mA with automatic modulation; slice thickness, 5.0 mm; reconstructed slice thickness for selected regions, 0.625 mm; pitch, 1.375:1. For contrast-enhanced scanning, an iodinated contrast agent was intravenously injected as a bolus via the antebrachial vein at a dose of 1–2 mL/kg of body weight and an injection rate of 1–3 mL/s. Two phases of enhanced scanning were performed at 25–35 s and 55–65 s after contrast administration, respectively, to acquire images of the arterial phase and venous phase. 3. Radiomics Analysis of CT Images 3.1. Lesion Segmentation We used RadCloud (Huiying Medical Technology Co., Ltd.) for the management of imaging data, clinical data, and subsequent radiomics statistical analysis. CT images were acquired in accordance with the standardized scanning protocol. Lesions were manually delineated by two radiologists who were blinded to the patients' clinical information: Observer 1 (J.C, with 6 years of experience in pediatric oncologic imaging) and Observer 2 (C.H.D, with 18 years of experience in pediatric oncologic imaging diagnosis). Tumor delineation was performed to preserve its integrity, including tumor calcification and necrosis areas, while avoiding the surrounding major blood vessels. (See Supplementary Material 1 for a schematic diagram). To assess the stability of the extracted radiomic features, we performed inter- and intra-observer reproducibility analyses. Thirty cases were randomly selected to perform two identical CT image annotation tasks.Observer 1 and Observer 2 independently segmented the entire tumor area on consecutive cross - sectional images to obtain ROI, and the acquired radiomic features were subjected to intraclass correlation coefficien ( ICC )analysis for inter - observer consistency. Observer 2 re - segmented the above 30 cases at an interval of one month, and the radiomic features derived from the first and second segmentations were used for intra - observer ICC analysis. Finally, Observer 2 completed the segmentation of all remaining cases. Subsequently, radiomic features were extracted from the ROIs of both the arterial phase and venous phase via the RadCloud platform, respectively. In addition, the two observers evaluated each case for the presence of lymph node metastasis based on imaging findings under the condition that they were only informed of the location of each lesion but remained completely unaware of the patients’ clinical and pathological information, and then provided their respective evaluation reports. 3.2 Radiomic Feature Extraction This study strictly adhered to the latest recommended standards of the Image Biomarker Standardization Initiative (IBSI) [ 16 ] for the evaluation of radiomic features. All image preprocessing and feature extraction were performed using the RadCloud platform, which is built based on the IBSI-compliant PyRadiomics library (Version 3.1.0; https://pyradiomics.readthedocs.io/ ) and implemented in Python (Version 3.7.0; https://www.python.org ). Prior to feature extraction, all CT images and corresponding segmentation masks were resampled to isotropic voxels of 1×1×1 mm³ via B-spline interpolation; fixed bin width of 25 HU (Hounsfield units) was used for gray-level discretization to standardize intensity values [ 17 ] . Radiomic features were extracted from both original and filtered images. The image filtering methods applied included wavelet transform, square transform, square root transform, gradient transform, logarithm transform, exponential transform, as well as two-dimensional and three-dimensional local binary patterns (LBP-2D/3D). Additionally, Laplacian of Gaussian (LoG) filters with sigma values of 1.0, 2.0, and 3.0 were used to enhance multi-scale edge and texture information. Based on the ROIs in contrast-enhanced CT images, 1688 radiomic features were extracted for each patient at both the arterial phase and venous phase, including 324 First-Order Statistics features, 14 Three-Dimensional Shape features, 432 Gray Level Co-occurrence Matrix (GLCM) features, 288 Gray Level Run Length Matrix (GLRLM) features, 288 Gray Level Size Zone Matrix (GLSZM) features, 90 Neighboring Gray Tone Difference Matrix (NGTDM) features, and 252 Gray Level Dependence Matrix (GLDM) features. To quantify the magnitude of differences in radiomics features between the arterial phase and venous phase, this study introduced the Delta-Absolute and Delta-Relative, which reflect the absolute value fluctuation and relative change ratio of the features, respectively. The calculation formulas are as follows: $$\:\text{D}\text{e}\text{l}\text{t}\text{a}-\text{A}\text{b}\text{s}\text{o}\text{l}\text{u}\text{t}\text{e}=\left|{\text{A}\text{r}\text{t}\text{e}\text{r}\text{i}\text{a}\text{l}\:\text{p}\text{h}\text{a}\text{s}\text{e}}_{\text{R}\text{a}\text{d}\text{i}\text{o}\text{m}\text{i}\text{c}\text{s}}-\right.\left.{\text{V}\text{e}\text{n}\text{o}\text{u}\text{s}\:\text{p}\text{h}\text{a}\text{s}\text{e}}_{\text{R}\text{a}\text{d}\text{i}\text{o}\text{m}\text{i}\text{c}\text{s}}\right|$$ $$\:\text{D}\text{e}\text{l}\text{t}\text{a}-\text{R}\text{e}\text{l}\text{a}\text{t}\text{i}\text{v}\text{e}=\frac{\left|\left.{\text{A}\text{r}\text{t}\text{e}\text{r}\text{i}\text{a}\text{l}\:\text{p}\text{h}\text{a}\text{s}\text{e}}_{\text{R}\text{a}\text{d}\text{i}\text{o}\text{m}\text{i}\text{c}\text{s}}-{\text{V}\text{e}\text{n}\text{o}\text{u}\text{s}\:\text{p}\text{h}\text{a}\text{s}\text{e}}_{\text{R}\text{a}\text{d}\text{i}\text{o}\text{m}\text{i}\text{c}\text{s}}\right|\right.}{{\text{V}\text{e}\text{n}\text{o}\text{u}\text{s}\:\text{p}\text{h}\text{a}\text{s}\text{e}}_{\text{R}\text{a}\text{d}\text{i}\text{o}\text{m}\text{i}\text{c}\text{s}}}$$ Among these, \(\:{\text{A}\text{r}\text{t}\text{e}\text{r}\text{i}\text{a}\text{l}\:\text{p}\text{h}\text{a}\text{s}\text{e}}_{\text{R}\text{a}\text{d}\text{i}\text{o}\text{m}\text{i}\text{c}\text{s}}\) Represents the radiomics feature values in the arterial phase state, \(\:{\text{V}\text{e}\text{n}\text{o}\text{u}\text{s}\:\text{p}\text{h}\text{a}\text{s}\text{e}}_{\text{R}\text{a}\text{d}\text{i}\text{o}\text{m}\text{i}\text{c}\text{s}}\) Represents the radiomics feature values in the venous phase state. 3.3 Construction of Radiomics Model For the four sets of radiomic features(Arterial phase、Venous phase、Delta-Absolute、Delta-Relative) Z -score normalization was performed using the mean and standard deviation calculated from the training set, with the identical parameters applied independently to the test set.Feature selection for the training set was implemented in three steps:First, the intra- and inter-observer reproducibility of features was evaluated using ICC. Features with an ICC ≤ 0.75 were excluded to ensure robustness and methodological reliability.Second, the SelectKBest method combined with the analysis of variance (ANOVA) F -test was adopted to screen for features with statistically significant discriminative ability between the comparison groups ( P < 0.05).Third, the least absolute shrinkage and selection operator (LASSO) regression model was used to identify the optimal feature subset, where the regularization parameter λ was determined by the minimum error point via 5-fold cross-validation.The multivariate regression coefficients derived from LASSO regularization were used as feature weights. The final radiomics score (Rad-Score) for each subject was calculated as follows: When screening n features, let \(\:{X}_{i}\) denote the value of the i feature and denote the \(\:{\beta\:}_{i}\) coefficient corresponding to the i feature, which is expressed as follows: $$\:\text{R}\text{a}\text{d}-\text{S}\text{c}\text{o}\text{r}\text{e}={\sum\:}_{i=1}^{n}{\beta\:}_{i}{X}_{i}$$ In this study, five machine learning models, namely LR (Logistic Regression), SVM (Support Vector Machine), KNN (K-Nearest Neighbor), DT (Decision Tree) and GBDT (Gradient Boosting Decision Tree), were employed for radiomics-based model construction, and validation methods were adopted to improve the effectiveness of the models. 4. Implementation Protocol We used Radcloud for imaging data management, followed by subsequent statistical data analysis. The study workflow is illustrated in Fig. 2 . 5. Statistical Analysis Statistical analyses were performed using RadCloud and SPSS 26.0 software. A P -value < 0.05 was considered statistically significant. Univariate analysis with a cutoff of P < 0.05 was performed to evaluate the associations between the selected clinical features (including NSE, SF, sex, age, and Ki-67 expression index) and lymph node metastasis in the training set. Multivariate logistic regression analysis was subsequently performed on the factors with statistical significance ( P < 0.05) identified by univariate analysis. Features with statistically significant differences ( P < 0.05) in the multivariate analysis were regarded as independent risk factors associated with this task and were eligible for constructing the combined model. The odds ratio (OR) of each risk factor was calculated. For the selection of radiomics parameters, the ICC, SelectKBest algorithm, and LASSO method were sequentially applied for feature screening. In this study, to evaluate the predictive performance of the learning classifiers, four metrics were used to assess the classifier performance in both the training set and tset set: AUC, Accuracy,Sensitivity and Specificity. The Hosmer-Lemeshow test was performed to verify the goodness-of-fit of each model, with a P -value ≥ 0.05 indicating a good model fit. The diagnostic performance of each model was evaluated via ROC curve analysis. The clinical decision-making value of the predictive models was assessed using DCA curves. The gain in diagnostic accuracy brought by the models was evaluated through a human-machine comparison based on the area under the ROC curve. 6. Results 6.1 Clinical Characteristics Table 1 presents the clinical characteristics of NB patients in the training set (n = 157) and test set (n = 68), including NSE, SF, sex, age at initial diagnosis, and Ki-67 expression index. Univariate logistic regression analysis was performed to evaluate the contribution of individual variables to the outcome. The results demonstrated that the Ki-67 index ( P < 0.001) and NSE ( P = 0.001) were significantly correlated with lymph node metastasis in pediatric patients with NB. Subsequently, the Ki-67 index and NSE were included in the multivariate logistic regression analysis, which revealed that the Ki-67 index ( P = 0.018) was an independent predictive factor for lymph node metastasis in these patients. The results of univariate and multivariate analyses of clinical characteristics are summarized in Table 2 . Table 1 Clinical Characteristics of Patients Variables Total(n = 225) Training set(n = 157) Test set(n = 68) P Value lymph node metastasis,n(%) 1 No 121(54) 84(54) 37(54) Yes 104(46) 73(46) 31(46) Sex,n(%) 0.85 Male 112(50) 77(49) 35(51) Female 113(50) 80(51) 33(49) Age,Median(Q1,Q3) 2(0.83,4) 2(0.75,4) 1.71(0.98,3.94) 0.889 Ki67,Median(Q1,Q3) 0.4(0.1,0.6) 0.4(0.05,0.55) 0.4(0.15,0.66) 0.164 SF,Median(Q1,Q3) 71.9(33.78,214) 78.6(33.8,224) 66.12(30.85,153) 0.541 NSE,Median(Q1,Q3) 39.43(20.07,168.8) 39.43(19.47,163) 40.09(25.08,186.95) 0.287 Q1: 25th percentile; Q3: 75th percentile. SF: serum ferritin; NSE: neuron-specific enolase. Table 2 Clinical characteristics univariate-multivariate analysis Variables Univariate analysise OR(95%CI) P Value Multivariate analysis OR(95%CI) P Value Sex 0.837(0.446,1.569) 0.579 Age 0.975(0.857,1.104) 0.689 SF 1.002(1.000,1.004) 0.059 NSE 1.003(1.001,1.005) 0.001 1.002(1.000, 1.004) 0.073 Ki-67 24.643(6.598,102.164) < 0.001 8.988(1.532,59.356) 0.018 Numbers in parentheses are the 95% confidence interval. OR, odds ratio; CI, confidence interval. SF: serum ferritin; NSE: neuron-specific enolase 6.2 Radiomics Feature Analysis Radiomics feature selection was performed sequentially using the ICC, the SelectKBest method combined with ANOVA F -test, and the LASSO regression model. After screening, a total of 22 optimal features were selected for constructing the arterial phase radiomics model, 18 for the venous phase radiomics model, 28 for the Delta-Absolute radiomics model, and 30 for the Delta-Relative radiomics model (see Supplementary Material 2 for the optimal features and corresponding coefficients of the four radiomics models). Univariate analysis was conducted on the radiomics scores of the four groups, and the results showed that all four radiomics scores were significantly correlated with lymph node metastasis in pediatric patients with NB ( P < 0.001). Therefore, the arterial phase, venous phase, Delta-Absolute, and Delta-Relative features were used to construct the radiomics models. Subsequently, the radiomics scores of the four groups were included in the multivariate analysis, which ultimately revealed that the arterial phase ( P = 0.031) and Delta-Relative ( P < 0.001) were independent predictive factors for evaluating lymph node metastasis in pediatric patients with NB. The results of univariate and multivariate analyses of the four radiomics models are presented in Table 3 . Table 3 Univariate and Multivariate Analyses of Radiomics Features Variables Univariate analysis OR(95% CI) P Value Multivariate analysis OR(95% CI) P Value Arterial phase_Score 2.226(1.761, 2.931) < 0.001 2.316(1.110, 5.254) 0.031 Venous phase _Score 2.108(1.682, 2.748) < 0.001 0.554(0.256, 1.137) 0.114 Delta-Absolute-Score 3.097(2.221, 4.640) < 0.001 1.018(0.615, 1.758) 0.946 Delta-Relative_Score 4.571(2.944, 8.043) < 0.001 4.117(2.348, 8.293) < 0.001 Numbers in parentheses are the 95% confidence interval. OR, odds ratio; CI, confidence interval. 6.3 Model Construction and Nomogram Visualization of the Optimal Model After univariate and multivariate analyses of the clinical and radiomics data as described above, four radiomics models were constructed using the arterial phase, venous phase, Delta-Absolute, and Delta-Relative features, respectively; a clinical model was established using the Ki-67 index; and a combined model was built by incorporating the arterial phase, Delta-Relative, and Ki-67 index. Machine learning models based on clinical and radiomics features were developed using algorithms including LR and SVM. Among these models, the combined LR model integrating the arterial phase, Delta-Relative, and Ki-67 index exhibited the optimal overall performance in predicting lymph node metastasis in pediatric patients with NB. In the training and test sets, the AUC values of the LR model were 0.937 and 0.829, respectively; the accuracy rates were 85.4% and 76.5%, respectively; the sensitivities were 87.7% and 77.4%, respectively; and the specificities were 83.3% and 75.7%, respectively, all demonstrating excellent predictive performance. Therefore, this model was visualized using a nomogram, as shown in Fig. 5 g. 6.4 Validation of Predictive Model Efficacy and Human-Machine Comparison The results of the Hosmer-Lemeshow test demonstrated good goodness-of-fit for all models ( P ≥ 0.05 for each model; see Supplementary Material 3). Among all the constructed models, the combined model incorporating the arterial phase radiomics score, Delta-Relative radiomics score, and Ki-67 index exhibited the most favorable performance in predicting the risk of lymph node metastasis in patients with NB, outperforming the four individual radiomics models and the single clinical model based on the Ki-67 index alone. In the training and test sets, this combined model achieved an AUC of 0.937 and 0.829, an accuracy of 85.4% and 76.5%, a sensitivity of 87.7% and 77.4%, and a specificity of 83.3% and 75.7%, respectively, for predicting lymph node metastasis in NB. In the human-machine comparison, Observer 1 achieved an AUC of 0.537 without nomogram assistance and 0.744 with nomogram assistance for diagnosing lymph node metastasis in NB, accompanied by corresponding accuracy of 54.4% vs. 75%, sensitivity of 45.2% vs. 67.7%, and specificity of 62.2% vs. 81.1%. Similarly, Observer 2 attained an AUC of 0.593 (without nomogram assistance) and 0.806 (with nomogram assistance) for the same diagnostic task, with respective accuracy of 60.3% vs. 80.9%, sensitivity of 48.4% vs. 77.4%, and specificity of 70.3% vs. 83.8%.With the aid of the nomogram, the diagnostic accuracy of the two radiologists for lymph node metastasis in pediatric patients with NB was improved by 21%. Table 4 presents the AUC (with confidence intervals), accuracy, sensitivity, specificity of each model in the training and test sets, as well as the results of the human-machine comparison. The ROC curves of all models in the training and test sets and the human-machine comparison results in the test set are illustrated in Fig. 3 . Table 4 Performances of radiomics,clinical model, nomogram and Human-Machine comparison Cohort Model AUC(95% CI) Accuracy Sensitivity Specificity Training set Arterial phase 0.844(0.800- 0.888) 0.771 0.726 0.810 Venous phase 0.831(0.790–0.873) 0.739 0.685 0.786 Delta-Absolute 0.868(0.821–0.914) 0.777 0.753 0.798 Delta-Relative 0.934(0.878–0.990) 0.841 0.836 0.845 Clinical (Ki-67) 0.725(0.700-.0750) 0.688 0.603 0.762 Nomogram 0.937(0.880–0.994) 0.854 0.877 0.833 Test set Arterial phase 0.779(0.721–0.836) 0.735 0.710 0.757 Venous phase 0.793(0.737–0.848) 0.750 0.774 0.730 Delta-Absolute 0.773(0.708–0.838) 0.691 0.742 0.649 Delta-Relative 0.744(0.669–0.819) 0.691 0.710 0.676 Clinical (Ki-67) 0.726(0.686–0.766) 0.662 0.613 0.703 Nomogram 0.829(0.755–0.904) 0.765 0.774 0.757 Human-Machine comparison Radiologist 1 NA 0.544 0.452 0.622 Radiologist1 + Nomogram NA 0.750 0.677 0.811 Radiologist 2 NA 0.603 0.484 0.703 Radiologist2 + Nomogram NA 0.809 0.774 0.838 6.5 Clinical Efficacy Evaluation of Predictive Models and Clinical Application of the Nomogram The DCA of the individual radiomics models, clinical model, and nomogram model in the training and test sets are presented in Fig. 4 . Across all threshold probabilities, the nomogram model yielded the highest net benefit in predicting lymph node metastasis in patients with NB, compared with the individual radiomics models and clinical model. Figure 5 demonstrates the favorable performance of the nomogram in predicting the risk scores for lymph node metastasis in NB. 7. Discussion NB exhibits complex and diverse biological behaviors and clinical characteristics, leading to substantial variability in clinical treatment strategies and prognostic assessments. Therefore, accurate pre-treatment evaluation of lymph node metastasis in NB is crucial for guiding clinical assessments and personalized treatment decisions [ 9 ] . As an emerging field in molecular imaging, radiomics facilitates the extraction and analysis of feature information associated with the pathophysiological basis of lesions by constructing radiomics models. This approach enables comprehensive, objective, and quantitative assessment of the spatial and temporal heterogeneity of lesions [ 18 ] . Previous radiomics studies on pediatric NB have been relatively limited due to the particularities of the pediatric population and constraints associated with pathological sample collection. Ilker et al. developed a CT-based clinical radiomics model to predict and differentiate between pediatric Wilms' tumor and adrenal NB [ 19 ] . Wang, Liu et al. separately employed contrast-enhanced CT-based radiomics and computational pathology analyses to predict the risk classification (INPC) and pathological prognostic classification of NB [ 20 – 21 ] . However, radiomics research focusing on lymph node metastasis in NB has not been reported to date. In the present study, we retrospectively collected NB cases over the past decade, expanding the research scope of pediatric NB to the entire peripheral body region to achieve full coverage of lesion locations in individual cases. This approach ensured maximum diversity of the study sample. More importantly, we established a nomogram model integrating contrast-enhanced CT radiomics with validated clinical and laboratory data. This nomogram model further improved the detection rate and diagnostic accuracy of imaging for lymph node metastasis in pediatric body NB. All six predictive models constructed in this study for lymph node metastasis in NB exhibited excellent clinical goodness-of-fit and demonstrated high diagnostic performance in both the training and test sets. The DCA also revealed substantial clinical decision-making benefits of these predictive models. Particularly, the nomogram model incorporating arterial phase radiomics score, Delta-Relative radiomics score, and Ki-67 index showed the highest predictive value, with AUC values of 0.937 and 0.829 in the training and test sets, respectively. In the human-machine comparison, the diagnostic accuracy of junior radiologists for lymph node metastasis in NB increased from 54% to 75% with the assistance of the nomogram, while that of senior radiologists improved from 60% to 81%. Both groups achieved an accuracy improvement of 21 percentage points. Previously, on conventional CT images, subtle morphological and enhancement pattern changes of partially metastatic lymph nodes in NB were difficult to identify visually, which resulted in a high rate of missed diagnoses of lymph node metastasis in pediatric patients with NB. The nomogram-assisted diagnostic approach thus represents a substantial improvement in diagnostic accuracy for radiologists and provides significant support for the clinical evaluation and management of pediatric NB. In addition to constructing radiomics models using radiomic features, this study also incorporated screened clinical indicators to establish clinical models for evaluating lymph node metastasis status in NB patients. These non-radiomic clinical indicators included basic patient information (age at diagnosis, gender), tumor markers (NSE), SF, and Ki-67 index. Basic patient information, as non-quantitative indicators, is not elaborated here. Tumor markers are laboratory indicators reflecting the malignant potential of tumor cells, which can indicate tumor development, progression, and prognosis. The combined application of CT, MRI, and NSE has shown high diagnostic value and specificity in pediatric patients with NB [ 22 ] . SF is a ubiquitous intracellular protein that stores iron and releases it in a controlled manner, serving as a buffer against iron deficiency and iron overload [ 23 ]. Evans, Veronica, et al. proposed risk stratification for NB patients based on SF levels, age, and disease stage [ 24 – 25 ] ; however, its concentration can also increase significantly in acute inflammation [ 26 ] , resulting in poor specificity.Ki-67 is a biological indicator for evaluating tumor proliferative activity and also reflects tumor invasiveness [ 27 ] . In malignant tumors, the expression level of Ki-67 is correlated with patient prognosis, with higher Ki-67 expression levels associated with poorer patient survival rates [ 28 ] . Univariate analysis in this study demonstrated that there were no statistically significant differences in the age at initial diagnosis, gender, or SF levels of patients for the diagnosis of lymph node metastasis in NB( P ≥ 0.05). Meanwhile, multivariate analysis indicated that NSE levels also showed no statistically significant difference in the diagnosis of lymph node metastasis in NB( P ≥ 0.05). The absence of statistical differences in age and gender for diagnosing lymph node metastasis in NB is an objective fact and thus requires no further elaboration. Although both SF and NSE levels were elevated to varying degrees in NB patients, these two laboratory parameters exhibited no statistically significant differences in the diagnosis of lymph node metastasis in NB ( P ≥ 0.05). We hypothesize that this may be attributed to the incomplete correlation between the levels of these two biochemical indicators and the invasiveness of tumors to the surrounding and distant lymph nodes; in other words, these two indicators cannot fully reflect the biological behaviors of the tumor. Numerous previous and recent studies have employed radiomics to predict the Ki-67 expression index in various tumors, including glioma, gastrointestinal stromal tumor, and breast cancer [ 29 – 31 ] . In the present study, the Ki-67 expression index, as a biological indicator for evaluating tumor proliferative activity, was ultimately included in the construction of the clinical and nomogram predictive models after screening. This indicator has also been incorporated into the development of tumor predictive models in many previous studies. Qiu [ 32 ] , Zhang [ 33 ] et al. constructed a combined model based on radiomics combined with biochemical indicators such as Ki-67 and Her, which effectively predicted lymph node metastasis in breast cancer. Cozzi [ 34 ] et al. confirmed that contrast-enhanced CT radiomics combined with indicators including Ki-67 could serve as an effective tool for predicting the subtypes and invasiveness of pulmonary neuroendocrine neoplasms. Chen [ 35 ] et al. established a model based on radiomics analysis of preoperative contrast-enhanced CT images combined with clinical indicators such as Ki-67 to predict vascular invasion (VI) in gastric cancer patients, thereby optimizing the clinical treatment of these patients. The aforementioned studies indicate that radiomics combined with Ki-67 has considerable research potential and clinical utility in evaluating tumor invasiveness, heterogeneity, and intratumoral and peritumoral microenvironmental characteristics, providing new insights and approaches for future clinical diagnosis and treatment.In summary, this is the rationale and original intention for our further investigation into tumor biological behaviors using radiomics information combined with clinical indicators, which further confirms the clinical practical value of the nomogram predictive model integrating radiomics and clinical indicators. This study has several limitations. First, it was a single-center, single-region retrospective analysis without external validation or multi-regional research, which represents a critical concern regarding the general applicability of radiomics models. In the future, we will further expand the sample size and conduct multi-center, multi-regional collaborations to improve the broad applicability of the predictive model. In addition, the methods used to determine Ki-67 expression in this study included preoperative core needle biopsy (CNB) and postoperative immunohistochemical analysis of surgical specimens, and these two approaches may yield inconsistent results [ 36 ] . A previous study on breast cancer demonstrated that the discrepancy between the Ki-67 index measured by preoperative biopsy and that determined by postoperative pathological examination could be as high as 10–40% [ 37 ] , which may be attributed to tumor heterogeneity and the limited tissue volume obtained during biopsy procedures. Third, to ensure the integrity of the data samples, children with NB who had received prior treatment were excluded from the study cohort. Therefore, the diagnostic accuracy of the predictive model constructed in this study for evaluating lymph node metastasis in pediatric patients with NB after treatment requires further investigation and validation. 8. Conclusion In conclusion, the nomogram integrating contrast-enhanced CT radiomics and clinical indicators has high practical value in the diagnostic assessment of lymph node metastasis in pediatric patients with peripheral NB. It can significantly improve the diagnostic accuracy of radiologists and provide clinically applicable evidence for the clinical evaluation and formulation of personalized treatment regimens for these patients. Declarations Ethics approval and consent to participate This retrospective study complied with the Declaration of Helsinki and was approved by the Ethics Committee of the Children's Hospital Affiliated to Shandong University (Children's Hospital of Jinan) (No.SDFE-IRB/T-2026014). Written informed consent was waived by the ethics committee of the Children's Hospital Affiliated to Shandong University. Consent for publication Not applicable. Competing interests The authors declare that they have no conflicts of interest. Funding No funding. Author Contribution WB (1): Conceptualization, Methodology, Investigation, Writing - Original Draft.GY (2): Data Curation.ZD (3): Investigation, Data Collection, Validation.JY (4): Data Curation.BQ (5): Visualization.JZ (1)*: Conceptualization, Supervision. Corresponding author.GF (2)*: Writing - Review & Editing. Corresponding author.All authors have read and approved the final manuscript. Acknowledgements Not applicable. Data Availability The data are not publicly available because they contain information that could compromise research participant privacy or consent but are available from the corresponding author upon reasonable request. References Maris JM. Recent advances in neuroblastoma. N Engl J Med. 2010;362:2202–11. Shimada H, DeLellis RA, Marx A. Neuroblastic tumours of the adrenal gland. In: Lloyd RV, Osamura RY, Kloppel ¨ G, Rosai J, editors. WHO Classification of Tumours of Endocrine Organs. 4th ed. Lyon: IARC; 2017. pp. 196–203. Van KJ, Chung DH. Neuroblastoma: tumor biology and its implications for staging and treatment[J]. Child (Basel). 2019;6(1):12–8. Goodman MTGJ, Smith MA, Olshan AF. Sympathetic nervous system tumors. Cancer incidence and survival among children and adolescents: United States SEER Program, 1975–1995. Bethesda, MD; 1999. Kholodenko IV, Kalinovsky DV, Doronin SM, Deyev II, Kholodenko RV. Neuroblastoma Origin and Therapeutic Targets for Immunotherapy. J Immunol Res 2018 (2018) 7394268. Louis CU, Shohet JM. Neuroblastoma: molecular pathogenesis and therapy. Annu Rev Med. 2015;66:49–63. Santiago T, Polanco AC, Fuentes-Alabi S, et al. Multinational retrospective central pathology review of neuroblastoma: lessons learned to establish a regional pathology referral center in resource-limited settings. Arch Pathol Lab Med. 2021;145:214–21. La Quaglia MP. Surgical management of neuroblastoma. Semin Pediatr Surg. 2001;10(3):132–9. Park JR, Bagatell R, Cohn SL, et al. Revisions to the International Neuroblastoma Response Criteria: A Consensus Statement From the National Cancer Institute Clinical Trials Planning Meeting. J Clin Oncol. 2017;35(22):2580–7. Bagatell R, Park JR, Acharya S, Neuroblastoma, et al. Version 2.2024, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2024;22(6):413–33. Voss SD. Staging and following common pediatric malignancies: MRI versus CT versus functional imaging. Pediatr Radiol. 2018;48(9):1324–36. Abbaspour E, Karimzadhagh S, Monsef A, Joukar F, Mansour-Ghanaei F, Hassanipour S. Application of radiomics for preoperative prediction of lymph node metastasis in colorectal cancer: a systematic review and meta-analysis. Int J Surg. 2024;110(6):3795–813. Wang Q, Lin Y, Ding C, et al. Multi-modality radiomics model predicts axillary lymph node metastasis of breast cancer using MRI and mammography. Eur Radiol. 2024;34(9):6121–31. Li S, Wang Y, Sun Y, et al. Both intra- and peri-tumoral radiomics signatures can be used to predict lymphatic vascular space invasion and lymphatic metastasis positive status from endometrial cancer MR imaging. Abdom Radiol (NY). 2024;49(11):4140–50. Song J, Yin Y, Wang H, Chang Z, Liu Z, Cui L. A review of original articles published in the emerging field of radiomics. Eur J Radiol. 2020;127:108991. Zwanenburg A, Vallières M, Abdalah MA, et al. The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping. Radiology. 2020;295(2):328–38. Shin J, Seo N, Baek S-E et al. MRI Radiomics Model Predicts Pathologic Complete Response of Rectal Cancer Following Chemoradiotherapy. Radiology.2022;303(2):351–358. Fornacon-Wood I, Faivre-Finn C, O'Connor JPB, Price GJ. Radiomics as a personalized medicine tool in lung cancer: Separating the hope from the hype. Lung Cancer. 2020;146:197–208. Koska IO, Ozcan HN, Tan AA, et al. Radiomics in differential diagnosis of Wilms tumor and neuroblastoma with adrenal location in children. Eur Radiol. 2024;34(8):5016–27. Wang H, Xie M, Chen X, et al. Radiomics analysis of contrast-enhanced computed tomography in predicting the International Neuroblastoma Pathology Classification in neuroblastoma. Insights Imaging. 2023;14(1):106. Published 2023 Jun 14. Liu Y, Jia Y, Hou C, et al. Pathological prognosis classification of patients with neuroblastoma using computational pathology analysis. Comput Biol Med. 2022;149:105980. Hao J, Sang J, Xu X, Bao A. Diagnostic value of CT and MRI combined with serum LDH, NSE, CEA, and MYCN in pediatric neuroblastoma. World J Surg Oncol. 2023;21(1):251. Wang W, Knovich MA, Coffman LG, Torti FM, Torti SV. Serum ferritin:past, present and future. Biochim Biophys Acta. 2010;1800(8):760–9. Evans AE, D’angio GJ, Propert K, Anderson J, Hann HL. Prognostic factors in neuroblastoma. Cancer. 1987;59:1853–9. Moroz V, Machin D, Hero B, et al. The prognostic strength of serum LDH and serum ferritin in children with neuroblastoma: A report from the International Neuroblastoma Risk Group (INRG) project. Pediatr Blood Cancer. 2020;67(8):e28359. Ong DS, Wang L, Zhu Y, Ho B, Ding JL. The response of ferritin to LPS and acute phase of Pseudomonas infection. J Endotoxin Res. 2005;11(5):267–80. Bi S, Li J, Wang T, Man F, Zhang P, Hou F, Wang H, Hao D. Multi-parametric MRI-based radiomics signature for preoperative prediction of Ki-67 proliferation status in sinonasal malignancies: a two-centre study. Eur Radiol. 2022;32(10):6933–42. Maeshima AM, Taniguchi H, Hori Y, Ida H, Hosoba R, Makita S, Fukuhara S, Munakata W, Suzuki T, Maruyama D, et al. Diagnostic utility and prognostic significance of the Ki-67 labeling index in diffuse large B-cell lymphoma transformed from follicular lymphoma: a study of 76 patients. Pathol Int. 2021;71(10):674–81. Ni J, Zhang H, Yang Q, et al. Machine-Learning and Radiomics-Based Preoperative Prediction of Ki-67 Expression in Glioma Using MRI Data. Acad Radiol. 2024;31(8):3397–405. Cai W, Guo K, Chen Y, Shi Y, Chen J. Sub-regional CT Radiomics for the Prediction of Ki-67 Proliferation Index in Gastrointestinal Stromal Tumors: A Multi-center Study. Acad Radiol. 2024;31(12):4974–84. Hu B, Xu Y, Gong H, Tang L, Li H. Radiomics Analysis of Intratumoral and Various Peritumoral Regions From Automated Breast Volume Scanning for Accurate Ki-67 Prediction in Breast Cancer Using Machine Learning. Acad Radiol. 2025;32(2):651–63. Qiu X, Fu Y, Ye Y, Wang Z, Cao C. A Nomogram Based on Molecular Biomarkers and Radiomics to Predict Lymph Node Metastasis in Breast Cancer. Front Oncol. 2022;12:790076. Published 2022 Mar 15. Zhang D, Shen M, Zhang L, He X, Huang X. Establishment of an interpretable MRI radiomics-based machine learning model capable of predicting axillary lymph node metastasis in invasive breast cancer. Sci Rep. 2025;15(1):26030. Cozzi D, Bicci E, Cavigli E, et al. Radiomics in pulmonary neuroendocrine tumours (NETs). Radiol Med. 2022;127(6):609–15. Chen Z, Zhang G, Liu Y, Zhu K. Radiomics analysis in predicting vascular invasion in gastric cancer based on enhanced CT: a preliminary study. BMC Cancer. 2024;24(1):1020. Dowsett M, Nielsen TO, A’Hern R, et al. Assessment of Ki67 in breast cancer: recommendations from the international Ki67 in breast cancer working group. J Natl Cancer Inst. 2011;103(22):1656–64. Rossi C, Fraticelli S, Fanizza M, et al. Concordance of immunohistochemistry for predictive and prognostic factors in breast cancer between biopsy and surgical excision: a single-centre experience and review of the literature. Breast Cancer Res Treat. 2023;198(3):573–82. Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial1.docx Supplementarymaterial2.docx Supplementarymaterial3.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 04 May, 2026 Reviews received at journal 10 Apr, 2026 Reviewers agreed at journal 10 Mar, 2026 Reviewers invited by journal 02 Mar, 2026 Editor assigned by journal 11 Feb, 2026 Submission checks completed at journal 11 Feb, 2026 First submitted to journal 08 Feb, 2026 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-8820742","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":599896493,"identity":"1be99612-be0b-4da8-bbac-6c4c37b1d304","order_by":0,"name":"Wenbin Guo","email":"","orcid":"","institution":"Children's Hospital Affiliated to Shandong University(Jinan Children’s Hospital)","correspondingAuthor":false,"prefix":"","firstName":"Wenbin","middleName":"","lastName":"Guo","suffix":""},{"id":599896494,"identity":"b8f542a4-4168-4dca-82b6-9b2706481312","order_by":1,"name":"Guangyong Yang","email":"","orcid":"","institution":"Children's Hospital Affiliated to Shandong University(Jinan Children’s Hospital)","correspondingAuthor":false,"prefix":"","firstName":"Guangyong","middleName":"","lastName":"Yang","suffix":""},{"id":599896495,"identity":"930f85a8-a5f8-4e72-b077-54524e1623b4","order_by":2,"name":"Zhengjun Dai","email":"","orcid":"","institution":"Huiying Medical Technology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Zhengjun","middleName":"","lastName":"Dai","suffix":""},{"id":599896496,"identity":"120e1a78-926c-4884-b020-5227179a9c7d","order_by":3,"name":"Jing Yang","email":"","orcid":"","institution":"Huiying Medical Technology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Jing","middleName":"","lastName":"Yang","suffix":""},{"id":599896497,"identity":"b92c7d42-d33a-48e3-8d3f-61e2c8b2f731","order_by":4,"name":"Baoxin Qian","email":"","orcid":"","institution":"Huiying Medical Technology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Baoxin","middleName":"","lastName":"Qian","suffix":""},{"id":599896498,"identity":"05fb2e55-1ee5-438e-bb55-f5aca3a58278","order_by":5,"name":"Jianshe Zhao","email":"","orcid":"","institution":"Children's Hospital Affiliated to Shandong University(Jinan Children’s Hospital)","correspondingAuthor":false,"prefix":"","firstName":"Jianshe","middleName":"","lastName":"Zhao","suffix":""},{"id":599896499,"identity":"4c1bb9e0-aaa1-4bd8-8a83-54b27c26e785","order_by":6,"name":"Guangfeng Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA40lEQVRIie3RsQrCMBCA4UggLqldU5S6OR8InXyYBKFTLYJLBkFBSQfrrm/h6CpCXeKeUfEF6uamdVbaujnkn++D5A4hm+0PI93jKecwmBGMDxcup9WkxUiIchlyt6mGcNFZNfEZ7TU2+si9VPe86xLXeFg7JTdH4RhMRKSYE+QmK15OOudT31FkAmaUGbHvIKbPu3KCYt52FG3sTBwaoQkCNqoiERSEFSQKxkLhGoRFgbfRILapDlA9QrMh5JL330tmXGe08i/dZFFcEJ7++5T3h5z6brIuJx/R38ZtNpvN9rUXb+5L0pT4VkYAAAAASUVORK5CYII=","orcid":"","institution":"Children's Hospital Affiliated to Shandong University(Jinan Children’s Hospital)","correspondingAuthor":true,"prefix":"","firstName":"Guangfeng","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2026-02-08 10:08:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8820742/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8820742/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104181984,"identity":"8289b52f-288d-4482-a1ff-ec4b1c1d34d8","added_by":"auto","created_at":"2026-03-08 17:33:08","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":54194,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart for selecting patients\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/4b86d5813d01726f7559ded2.png"},{"id":104181986,"identity":"0af083be-3cd9-478d-a601-98fa1771a8e9","added_by":"auto","created_at":"2026-03-08 17:33:08","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":478280,"visible":true,"origin":"","legend":"\u003cp\u003eWorkflow of necessary steps in current study. LASSO least absolute shrinkage and selection operator, ROC receiver operating character\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/e7f604993b25d49783800f67.png"},{"id":104404406,"identity":"57237c10-aced-420c-a731-bb8c5b15abc5","added_by":"auto","created_at":"2026-03-11 12:20:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":158444,"visible":true,"origin":"","legend":"\u003cp\u003eROC Curves of the Training set and Test set.To predict the risk of lymph node metastasis, the ROC curves of the clinical model, arterial phase model, venous phase model, Delta-Absolute model, Delta-Relative model, and nomogram were compared in the Training set (a) and Test set (b). Furthermore, a human-machine comparison was conducted in the Test set.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/ffdc70620b70027273efac57.png"},{"id":104181991,"identity":"3eff74d5-2962-4fc5-bd08-22bfd2f3ffc3","added_by":"auto","created_at":"2026-03-08 17:33:09","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":170005,"visible":true,"origin":"","legend":"\u003cp\u003eDCA Curves of the Training Set(a) and Test Set(b).\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/b7716cdda496f02612477830.png"},{"id":104181990,"identity":"1a54e9b4-a47a-4ce4-a545-5787b33d0af7","added_by":"auto","created_at":"2026-03-08 17:33:09","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":453715,"visible":true,"origin":"","legend":"\u003cp\u003eClinical Application Example of the Nomogram.This figure illustrates the application of the nomogram in predicting lymph node metastasis in NB. The nomogram enables risk assessment by assigning corresponding scores to each predictor: the total score is obtained by summing the scores of all variables, and the corresponding risk of lymph node metastasis is then determined based on this total score. Contrast-enhanced CT images in the arterial phase (a) and venous phase (b) of a 6-year-old girl with NB revealed a large irregular retroperitoneal mass with heterogeneous enhancement; immunohistochemistry confirmed the presence of lymph node metastasis (Figure c, magnification ×100). Contrast-enhanced CT images in the arterial phase (d) and venous phase (e) of a 1-year-old boy with NB showed an irregular hypodense retroperitoneal mass on the right side with heterogeneous enhancement; immunohistochemistry confirmed the absence of lymph node metastasis (Figure f, magnification ×200). The nomogram (g) demonstrates the risk assessment results for the above two patients: the patient with lymph node metastasis (red solid arrow) had a total nomogram score of 103.8, corresponding to a lymph node metastasis risk of \u0026gt;97%; the patient without lymph node metastasis (blue solid arrow) had a total nomogram score of 55.8, corresponding to a lymph node metastasis risk of \u0026lt;12%.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/0ba3b24a4ea73165574770b1.png"},{"id":104408880,"identity":"71c0cb4d-10af-41ce-b311-16230a22b54c","added_by":"auto","created_at":"2026-03-11 12:43:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2514995,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/9b142655-b805-4f0d-b278-bea91eca5ac7.pdf"},{"id":104403986,"identity":"aca27f5b-c536-41c5-b47f-03e558f5b022","added_by":"auto","created_at":"2026-03-11 12:19:32","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":218516,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial1.docx","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/94304b0c2bef22d411dc318e.docx"},{"id":104181985,"identity":"0b027cd6-fba4-40e6-81a4-8de897d0811c","added_by":"auto","created_at":"2026-03-08 17:33:08","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":25890,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial2.docx","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/f3a67fcb978f8f1ae754f3d5.docx"},{"id":104403716,"identity":"b8f1dd1a-4d5f-4601-b7b8-b10273a67e62","added_by":"auto","created_at":"2026-03-11 12:18:53","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":18074,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial3.docx","url":"https://assets-eu.researchsquare.com/files/rs-8820742/v1/628f063e2629c08358afe36c.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Application Value of a Nomogram Integrating Contrast-enhanced CT Radiomics and Clinical Indicators in Evaluating Lymph Node Metastasis in Pediatric Peripheral Neuroblastoma","fulltext":[{"header":"1.Introduction","content":"\u003cp\u003eNeuroblastoma(NB) is an embryonic tumor originating from the primitive neural crest cells of the sympathetic nervous system. This tumor can arise at any site within the sympathetic nervous system \u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. It encompasses a spectrum of tumors, such as malignant NB, mixed ganglioneuroblastoma, nodular ganglioneuroblastoma, and ganglioneuroma \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. Peripheral neuroblastic tumors are the most common extracranial solid tumors in children \u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e, with the majority of cases diagnosed before five years of age and a mean age at diagnosis of 17 months \u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. The clinical manifestations of NB are diverse, ranging from asymptomatic masses to aggressive primary tumors with extensive local invasion, widespread metastasis, or severe illness associated with these features \u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e. Around 70% of NB patients already show signs of distant metastasis at the initial diagnosis, which is most frequently observed in bone marrow, followed by bones, lymph nodes, and the liver \u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/sup\u003e. Although bone marrow metastasis is the most common in children with NB, determining the presence of lymph node metastasis is crucial for evaluating pre-treatment risk stratification, selecting surgical approaches, defining the target volume of postoperative radiotherapy, and assessing treatment outcomes in these patients \u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn the imaging evaluation of NB, contrast-enhanced CT currently serves as a crucial diagnostic and assessment tool for pediatric peripheral NB. It enables preoperative tumor risk stratification, staging, and risk assessment based on the morphological features of lesions, intratumoral calcifications, and the involvement status of surrounding tissue structures. This facilitates determining the selection of radiotherapy and chemotherapy drugs, dosages, and treatment cycles, thereby achieving precise and personalized treatment as well as the evaluation of treatment outcomes \u003csup\u003e[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e. However, due to the limitation of human visual resolution, the diagnostic accuracy of radiologists in detecting lymph node metastasis in NB patients based on conventional contrast-enhanced CT awaits further improvement.\u003c/p\u003e \u003cp\u003eRadiomics based on contrast-enhanced CT has been widely applied in the diagnosis, grading, and prognostic evaluation of adult solid tumors. Models constructed using radiomics have also been extensively used to predict lymph node metastasis in adult tumors. These radiomics models have been explored to varying degrees in the prediction of lymph node metastasis across multiple systemic tumors in adults, including colorectal cancer, breast cancer, and endometrial cancer \u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/sup\u003e, and have demonstrated favorable predictive benefits. However, owing to the unique characteristics of the pediatric population, such applications are far less common in children than in adults \u003csup\u003e[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e. As of November 2025, the authors have not identified any relevant literature on contrast-enhanced CT-based radiomics models for predicting lymph node metastasis in pediatric patients with peripheral NB.\u003c/p\u003e \u003cp\u003eTherefore, based on the above discussion, this study aimed to construct a contrast-enhanced CT-based radiomics model, a clinical model, and a nomogram model, to explore their application value in the diagnosis of lymph node metastasis in pediatric patients with peripheral NB, improve the diagnostic accuracy of radiologists, and provide more precise imaging evidence for prognostic evaluation and personalized treatment of NB patients.\u003c/p\u003e"},{"header":"2. Patients and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1. Participants and Clinical Characteristics\u003c/h2\u003e \u003cp\u003eThis retrospective study was approved by the Ethics Committee of the Children's Hospital Affiliated to Shandong University, and the requirement for informed consent was waived. We collected the imaging and clinical data of 358 pediatric patients diagnosed with NB via contrast-enhanced CT between February 2014 and December 2024. The inclusion criteria were as follows: (1) Newly diagnosed patients who had not received any tumor-related treatment; (2) Patients pathologically confirmed as NB via puncture or surgery; (3) Complete baseline data at initial diagnosis (including clinical data, laboratory test results, and relevant pathological data); (4) Complete pre-treatment CT imaging data, including both arterial and venous phases. The exclusion criteria were as follows: (1) Patients who had received relevant treatments such as radiotherapy, chemotherapy, or surgery prior to admission (n\u0026thinsp;=\u0026thinsp;23); (2) Patients with tumors whose boundaries with surrounding tissue structures were extremely difficult to identify, or with multiple systemic masses that prevented complete delineation of the tumor region of interest (ROI) (n\u0026thinsp;=\u0026thinsp;11); (3) Patients with imaging quality failing to meet the requirements for post-processing (n\u0026thinsp;=\u0026thinsp;7).\u003c/p\u003e \u003cp\u003eAfter screening, a total of 225 patients were enrolled in this study (112 males and 113 females; mean age, 2.23\u0026thinsp;\u0026plusmn;\u0026thinsp;2.34 years; age range, 0\u0026ndash;13 years), among whom 104 cases were pathologically confirmed to have lymph node metastasis. The clinical data of the patients included age and gender at initial diagnosis; laboratory examination indicators included serum ferritin (SF) and neuron-specific enolase (NSE); pathological data included the Ki-67 expression index and the presence of lymph node metastasis. The data were randomly divided into a training set and a test set at a ratio of 7:3 with a random seed of 305. Specifically, the training set comprised 157 cases, including 73 cases with lymph node metastasis and 84 cases without lymph node metastasis, while the test set included 68 cases, consisting of 31 cases with lymph node metastasis and 37 cases without lymph node metastasis.. The patient selection flowchart is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 CT Scanning Protocol\u003c/h2\u003e \u003cp\u003eAll pediatric patients underwent contrast-enhanced CT scans without sedation. For those who were unable to cooperate, sedation was administered by certified physicians from the Sedation Center of our hospital.Scanning was performed using a GE Revolution GSI CT750 scanner. The scanning parameters were as follows: tube voltage, 80\u0026ndash;120 kV; tube current, 50\u0026ndash;200 mA with automatic modulation; slice thickness, 5.0 mm; reconstructed slice thickness for selected regions, 0.625 mm; pitch, 1.375:1. For contrast-enhanced scanning, an iodinated contrast agent was intravenously injected as a bolus via the antebrachial vein at a dose of 1\u0026ndash;2 mL/kg of body weight and an injection rate of 1\u0026ndash;3 mL/s. Two phases of enhanced scanning were performed at 25\u0026ndash;35 s and 55\u0026ndash;65 s after contrast administration, respectively, to acquire images of the arterial phase and venous phase.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Radiomics Analysis of CT Images","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Lesion Segmentation\u003c/h2\u003e \u003cp\u003eWe used RadCloud (Huiying Medical Technology Co., Ltd.) for the management of imaging data, clinical data, and subsequent radiomics statistical analysis. CT images were acquired in accordance with the standardized scanning protocol. Lesions were manually delineated by two radiologists who were blinded to the patients' clinical information: Observer 1 (J.C, with 6 years of experience in pediatric oncologic imaging) and Observer 2 (C.H.D, with 18 years of experience in pediatric oncologic imaging diagnosis). Tumor delineation was performed to preserve its integrity, including tumor calcification and necrosis areas, while avoiding the surrounding major blood vessels. (See Supplementary Material 1 for a schematic diagram).\u003c/p\u003e \u003cp\u003eTo assess the stability of the extracted radiomic features, we performed inter- and intra-observer reproducibility analyses. Thirty cases were randomly selected to perform two identical CT image annotation tasks.Observer 1 and Observer 2 independently segmented the entire tumor area on consecutive cross - sectional images to obtain ROI, and the acquired radiomic features were subjected to intraclass correlation coefficien\u003cb\u003e(\u003c/b\u003eICC )analysis for inter - observer consistency. Observer 2 re - segmented the above 30 cases at an interval of one month, and the radiomic features derived from the first and second segmentations were used for intra - observer ICC analysis. Finally, Observer 2 completed the segmentation of all remaining cases. Subsequently, radiomic features were extracted from the ROIs of both the arterial phase and venous phase via the RadCloud platform, respectively.\u003c/p\u003e \u003cp\u003eIn addition, the two observers evaluated each case for the presence of lymph node metastasis based on imaging findings under the condition that they were only informed of the location of each lesion but remained completely unaware of the patients\u0026rsquo; clinical and pathological information, and then provided their respective evaluation reports.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Radiomic Feature Extraction\u003c/h2\u003e \u003cp\u003eThis study strictly adhered to the latest recommended standards of the Image Biomarker Standardization Initiative (IBSI) \u003csup\u003e[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e for the evaluation of radiomic features. All image preprocessing and feature extraction were performed using the RadCloud platform, which is built based on the IBSI-compliant PyRadiomics library (Version 3.1.0; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pyradiomics.readthedocs.io/\u003c/span\u003e\u003cspan address=\"https://pyradiomics.readthedocs.io/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and implemented in Python (Version 3.7.0; \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). Prior to feature extraction, all CT images and corresponding segmentation masks were resampled to isotropic voxels of 1\u0026times;1\u0026times;1 mm\u0026sup3; via B-spline interpolation; fixed bin width of 25 HU (Hounsfield units) was used for gray-level discretization to standardize intensity values \u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eRadiomic features were extracted from both original and filtered images. The image filtering methods applied included wavelet transform, square transform, square root transform, gradient transform, logarithm transform, exponential transform, as well as two-dimensional and three-dimensional local binary patterns (LBP-2D/3D). Additionally, Laplacian of Gaussian (LoG) filters with sigma values of 1.0, 2.0, and 3.0 were used to enhance multi-scale edge and texture information. Based on the ROIs in contrast-enhanced CT images, 1688 radiomic features were extracted for each patient at both the arterial phase and venous phase, including 324 First-Order Statistics features, 14 Three-Dimensional Shape features, 432 Gray Level Co-occurrence Matrix (GLCM) features, 288 Gray Level Run Length Matrix (GLRLM) features, 288 Gray Level Size Zone Matrix (GLSZM) features, 90 Neighboring Gray Tone Difference Matrix (NGTDM) features, and 252 Gray Level Dependence Matrix (GLDM) features.\u003c/p\u003e \u003cp\u003eTo quantify the magnitude of differences in radiomics features between the arterial phase and venous phase, this study introduced the Delta-Absolute and Delta-Relative, which reflect the absolute value fluctuation and relative change ratio of the features, respectively. The calculation formulas are as follows:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:\\text{D}\\text{e}\\text{l}\\text{t}\\text{a}-\\text{A}\\text{b}\\text{s}\\text{o}\\text{l}\\text{u}\\text{t}\\text{e}=\\left|{\\text{A}\\text{r}\\text{t}\\text{e}\\text{r}\\text{i}\\text{a}\\text{l}\\:\\text{p}\\text{h}\\text{a}\\text{s}\\text{e}}_{\\text{R}\\text{a}\\text{d}\\text{i}\\text{o}\\text{m}\\text{i}\\text{c}\\text{s}}-\\right.\\left.{\\text{V}\\text{e}\\text{n}\\text{o}\\text{u}\\text{s}\\:\\text{p}\\text{h}\\text{a}\\text{s}\\text{e}}_{\\text{R}\\text{a}\\text{d}\\text{i}\\text{o}\\text{m}\\text{i}\\text{c}\\text{s}}\\right|$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:\\text{D}\\text{e}\\text{l}\\text{t}\\text{a}-\\text{R}\\text{e}\\text{l}\\text{a}\\text{t}\\text{i}\\text{v}\\text{e}=\\frac{\\left|\\left.{\\text{A}\\text{r}\\text{t}\\text{e}\\text{r}\\text{i}\\text{a}\\text{l}\\:\\text{p}\\text{h}\\text{a}\\text{s}\\text{e}}_{\\text{R}\\text{a}\\text{d}\\text{i}\\text{o}\\text{m}\\text{i}\\text{c}\\text{s}}-{\\text{V}\\text{e}\\text{n}\\text{o}\\text{u}\\text{s}\\:\\text{p}\\text{h}\\text{a}\\text{s}\\text{e}}_{\\text{R}\\text{a}\\text{d}\\text{i}\\text{o}\\text{m}\\text{i}\\text{c}\\text{s}}\\right|\\right.}{{\\text{V}\\text{e}\\text{n}\\text{o}\\text{u}\\text{s}\\:\\text{p}\\text{h}\\text{a}\\text{s}\\text{e}}_{\\text{R}\\text{a}\\text{d}\\text{i}\\text{o}\\text{m}\\text{i}\\text{c}\\text{s}}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eAmong these,\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{A}\\text{r}\\text{t}\\text{e}\\text{r}\\text{i}\\text{a}\\text{l}\\:\\text{p}\\text{h}\\text{a}\\text{s}\\text{e}}_{\\text{R}\\text{a}\\text{d}\\text{i}\\text{o}\\text{m}\\text{i}\\text{c}\\text{s}}\\)\u003c/span\u003e\u003c/span\u003e Represents the radiomics feature values in the arterial phase state, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{V}\\text{e}\\text{n}\\text{o}\\text{u}\\text{s}\\:\\text{p}\\text{h}\\text{a}\\text{s}\\text{e}}_{\\text{R}\\text{a}\\text{d}\\text{i}\\text{o}\\text{m}\\text{i}\\text{c}\\text{s}}\\)\u003c/span\u003e\u003c/span\u003e Represents the radiomics feature values in the venous phase state.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Construction of Radiomics Model\u003c/h2\u003e \u003cp\u003eFor the four sets of radiomic features(Arterial phase、Venous phase、Delta-Absolute、Delta-Relative)\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eZ\u003c/span\u003e-score normalization was performed using the mean and standard deviation calculated from the training set, with the identical parameters applied independently to the test set.Feature selection for the training set was implemented in three steps:First, the intra- and inter-observer reproducibility of features was evaluated using ICC. Features with an ICC\u0026thinsp;\u0026le;\u0026thinsp;0.75 were excluded to ensure robustness and methodological reliability.Second, the SelectKBest method combined with the analysis of variance (ANOVA) \u003cem\u003eF\u003c/em\u003e-test was adopted to screen for features with statistically significant discriminative ability between the comparison groups (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05).Third, the least absolute shrinkage and selection operator (LASSO) regression model was used to identify the optimal feature subset, where the regularization parameter λ was determined by the minimum error point via 5-fold cross-validation.The multivariate regression coefficients derived from LASSO regularization were used as feature weights. The final radiomics score (Rad-Score) for each subject was calculated as follows: When screening n features, let \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{X}_{i}\\)\u003c/span\u003e\u003c/span\u003e denote the value of the \u003cem\u003ei\u003c/em\u003e feature and denote the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{i}\\)\u003c/span\u003e\u003c/span\u003ecoefficient corresponding to the \u003cem\u003ei\u003c/em\u003e feature, which is expressed as follows:\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:\\text{R}\\text{a}\\text{d}-\\text{S}\\text{c}\\text{o}\\text{r}\\text{e}={\\sum\\:}_{i=1}^{n}{\\beta\\:}_{i}{X}_{i}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eIn this study, five machine learning models, namely LR (Logistic Regression), SVM (Support Vector Machine), KNN (K-Nearest Neighbor), DT (Decision Tree) and GBDT (Gradient Boosting Decision Tree), were employed for radiomics-based model construction, and validation methods were adopted to improve the effectiveness of the models.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Implementation Protocol","content":"\u003cp\u003eWe used Radcloud for imaging data management, followed by subsequent statistical data analysis. The study workflow is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"5. Statistical Analysis","content":"\u003cp\u003eStatistical analyses were performed using RadCloud and SPSS 26.0 software. A \u003cem\u003eP\u003c/em\u003e-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant. Univariate analysis with a cutoff of \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was performed to evaluate the associations between the selected clinical features (including NSE, SF, sex, age, and Ki-67 expression index) and lymph node metastasis in the training set. Multivariate logistic regression analysis was subsequently performed on the factors with statistical significance (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) identified by univariate analysis. Features with statistically significant differences (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in the multivariate analysis were regarded as independent risk factors associated with this task and were eligible for constructing the combined model. The odds ratio (OR) of each risk factor was calculated. For the selection of radiomics parameters, the ICC, SelectKBest algorithm, and LASSO method were sequentially applied for feature screening. In this study, to evaluate the predictive performance of the learning classifiers, four metrics were used to assess the classifier performance in both the training set and tset set: AUC, Accuracy,Sensitivity and Specificity. The Hosmer-Lemeshow test was performed to verify the goodness-of-fit of each model, with a \u003cem\u003eP\u003c/em\u003e-value\u0026thinsp;\u0026ge;\u0026thinsp;0.05 indicating a good model fit. The diagnostic performance of each model was evaluated via ROC curve analysis. The clinical decision-making value of the predictive models was assessed using DCA curves. The gain in diagnostic accuracy brought by the models was evaluated through a human-machine comparison based on the area under the ROC curve.\u003c/p\u003e"},{"header":"6. Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e6.1 Clinical Characteristics\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents the clinical characteristics of NB patients in the training set (n\u0026thinsp;=\u0026thinsp;157) and test set (n\u0026thinsp;=\u0026thinsp;68), including NSE, SF, sex, age at initial diagnosis, and Ki-67 expression index. Univariate logistic regression analysis was performed to evaluate the contribution of individual variables to the outcome. The results demonstrated that the Ki-67 index (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and NSE (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001) were significantly correlated with lymph node metastasis in pediatric patients with NB. Subsequently, the Ki-67 index and NSE were included in the multivariate logistic regression analysis, which revealed that the Ki-67 index (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.018) was an independent predictive factor for lymph node metastasis in these patients. The results of univariate and multivariate analyses of clinical characteristics are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eClinical Characteristics of Patients\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTotal(n\u0026thinsp;=\u0026thinsp;225)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTraining set(n\u0026thinsp;=\u0026thinsp;157)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTest set(n\u0026thinsp;=\u0026thinsp;68)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e Value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003elymph node metastasis,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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\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\u003e121(54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e84(54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e37(54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\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\u003e104(46)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e73(46)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e31(46)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex,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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e112(50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e77(49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e35(51)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e113(50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e80(51)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e33(49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge,Median(Q1,Q3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2(0.83,4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2(0.75,4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.71(0.98,3.94)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.889\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKi67,Median(Q1,Q3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.4(0.1,0.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.4(0.05,0.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.4(0.15,0.66)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.164\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSF,Median(Q1,Q3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e71.9(33.78,214)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e78.6(33.8,224)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e66.12(30.85,153)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.541\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNSE,Median(Q1,Q3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e39.43(20.07,168.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39.43(19.47,163)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e40.09(25.08,186.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.287\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\u003eQ1: 25th percentile; Q3: 75th percentile. SF: serum ferritin; NSE: neuron-specific enolase.\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\u003eClinical characteristics univariate-multivariate analysis\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eUnivariate analysise\u003c/p\u003e \u003cp\u003eOR(95%CI) \u003cem\u003eP\u003c/em\u003e Value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eMultivariate analysis\u003c/p\u003e \u003cp\u003eOR(95%CI) \u003cem\u003eP\u003c/em\u003e Value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.837(0.446,1.569)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.579\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.975(0.857,1.104)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.689\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.002(1.000,1.004)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.059\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNSE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.003(1.001,1.005)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.002(1.000, 1.004)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.073\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKi-67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e24.643(6.598,102.164)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e8.988(1.532,59.356)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.018\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eNumbers in parentheses are the 95% confidence interval. OR, odds ratio; CI, confidence interval. SF: serum ferritin; NSE: neuron-specific enolase\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e6.2 Radiomics Feature Analysis\u003c/h2\u003e \u003cp\u003eRadiomics feature selection was performed sequentially using the ICC, the SelectKBest method combined with ANOVA \u003cem\u003eF\u003c/em\u003e-test, and the LASSO regression model. After screening, a total of 22 optimal features were selected for constructing the arterial phase radiomics model, 18 for the venous phase radiomics model, 28 for the Delta-Absolute radiomics model, and 30 for the Delta-Relative radiomics model (see Supplementary Material 2 for the optimal features and corresponding coefficients of the four radiomics models). Univariate analysis was conducted on the radiomics scores of the four groups, and the results showed that all four radiomics scores were significantly correlated with lymph node metastasis in pediatric patients with NB (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Therefore, the arterial phase, venous phase, Delta-Absolute, and Delta-Relative features were used to construct the radiomics models. Subsequently, the radiomics scores of the four groups were included in the multivariate analysis, which ultimately revealed that the arterial phase (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.031) and Delta-Relative (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) were independent predictive factors for evaluating lymph node metastasis in pediatric patients with NB. The results of univariate and multivariate analyses of the four radiomics models are presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUnivariate and Multivariate Analyses of Radiomics Features\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnivariate analysis OR(95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e Value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMultivariate analysis OR(95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e Value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArterial phase_Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.226(1.761, 2.931)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.316(1.110, 5.254)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.031\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVenous phase _Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.108(1.682, 2.748)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.554(0.256, 1.137)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.114\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDelta-Absolute-Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3.097(2.221, 4.640)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.018(0.615, 1.758)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.946\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDelta-Relative_Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.571(2.944, 8.043)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.117(2.348, 8.293)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eNumbers in parentheses are the 95% confidence interval. OR, odds ratio; CI, confidence interval.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e6.3 Model Construction and Nomogram Visualization of the Optimal Model\u003c/h2\u003e \u003cp\u003eAfter univariate and multivariate analyses of the clinical and radiomics data as described above, four radiomics models were constructed using the arterial phase, venous phase, Delta-Absolute, and Delta-Relative features, respectively; a clinical model was established using the Ki-67 index; and a combined model was built by incorporating the arterial phase, Delta-Relative, and Ki-67 index. Machine learning models based on clinical and radiomics features were developed using algorithms including LR and SVM. Among these models, the combined LR model integrating the arterial phase, Delta-Relative, and Ki-67 index exhibited the optimal overall performance in predicting lymph node metastasis in pediatric patients with NB. In the training and test sets, the AUC values of the LR model were 0.937 and 0.829, respectively; the accuracy rates were 85.4% and 76.5%, respectively; the sensitivities were 87.7% and 77.4%, respectively; and the specificities were 83.3% and 75.7%, respectively, all demonstrating excellent predictive performance. Therefore, this model was visualized using a nomogram, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eg.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e6.4 Validation of Predictive Model Efficacy and Human-Machine Comparison\u003c/h2\u003e \u003cp\u003eThe results of the Hosmer-Lemeshow test demonstrated good goodness-of-fit for all models (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026ge;\u0026thinsp;0.05 for each model; see Supplementary Material 3). Among all the constructed models, the combined model incorporating the arterial phase radiomics score, Delta-Relative radiomics score, and Ki-67 index exhibited the most favorable performance in predicting the risk of lymph node metastasis in patients with NB, outperforming the four individual radiomics models and the single clinical model based on the Ki-67 index alone. In the training and test sets, this combined model achieved an AUC of 0.937 and 0.829, an accuracy of 85.4% and 76.5%, a sensitivity of 87.7% and 77.4%, and a specificity of 83.3% and 75.7%, respectively, for predicting lymph node metastasis in NB.\u003c/p\u003e \u003cp\u003eIn the human-machine comparison, Observer 1 achieved an AUC of 0.537 without nomogram assistance and 0.744 with nomogram assistance for diagnosing lymph node metastasis in NB, accompanied by corresponding accuracy of 54.4% vs. 75%, sensitivity of 45.2% vs. 67.7%, and specificity of 62.2% vs. 81.1%. Similarly, Observer 2 attained an AUC of 0.593 (without nomogram assistance) and 0.806 (with nomogram assistance) for the same diagnostic task, with respective accuracy of 60.3% vs. 80.9%, sensitivity of 48.4% vs. 77.4%, and specificity of 70.3% vs. 83.8%.With the aid of the nomogram, the diagnostic accuracy of the two radiologists for lymph node metastasis in pediatric patients with NB was improved by 21%. Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e presents the AUC (with confidence intervals), accuracy, sensitivity, specificity of each model in the training and test sets, as well as the results of the human-machine comparison. The ROC curves of all models in the training and test sets and the human-machine comparison results in the test set are illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformances of radiomics,clinical model, nomogram and Human-Machine comparison\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAUC(95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"5\" rowspan=\"6\"\u003e \u003cp\u003eTraining set\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArterial phase\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.844(0.800- 0.888)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.771\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.726\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.810\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVenous phase\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.831(0.790\u0026ndash;0.873)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.739\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.685\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.786\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDelta-Absolute\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.868(0.821\u0026ndash;0.914)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.777\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.753\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.798\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDelta-Relative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.934(0.878\u0026ndash;0.990)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.841\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.836\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.845\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClinical (Ki-67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.725(0.700-.0750)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.688\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.603\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.762\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNomogram\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.937(0.880\u0026ndash;0.994)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.854\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.877\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"5\" rowspan=\"6\"\u003e \u003cp\u003eTest set\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArterial phase\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.779(0.721\u0026ndash;0.836)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.735\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.710\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.757\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVenous phase\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.793(0.737\u0026ndash;0.848)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.750\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.730\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDelta-Absolute\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.773(0.708\u0026ndash;0.838)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.691\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.742\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.649\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDelta-Relative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.744(0.669\u0026ndash;0.819)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.691\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.710\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.676\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClinical (Ki-67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.726(0.686\u0026ndash;0.766)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.662\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.613\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.703\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNomogram\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.829(0.755\u0026ndash;0.904)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.765\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.757\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eHuman-Machine comparison\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRadiologist 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.544\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.452\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.622\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRadiologist1\u0026thinsp;+\u0026thinsp;Nomogram\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.750\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.677\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.811\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRadiologist 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.603\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.484\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.703\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRadiologist2\u0026thinsp;+\u0026thinsp;Nomogram\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.809\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.838\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e6.5 Clinical Efficacy Evaluation of Predictive Models and Clinical Application of the Nomogram\u003c/h2\u003e \u003cp\u003eThe DCA of the individual radiomics models, clinical model, and nomogram model in the training and test sets are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Across all threshold probabilities, the nomogram model yielded the highest net benefit in predicting lymph node metastasis in patients with NB, compared with the individual radiomics models and clinical model. Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e demonstrates the favorable performance of the nomogram in predicting the risk scores for lymph node metastasis in NB.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"7. Discussion","content":"\u003cp\u003eNB exhibits complex and diverse biological behaviors and clinical characteristics, leading to substantial variability in clinical treatment strategies and prognostic assessments. Therefore, accurate pre-treatment evaluation of lymph node metastasis in NB is crucial for guiding clinical assessments and personalized treatment decisions \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. As an emerging field in molecular imaging, radiomics facilitates the extraction and analysis of feature information associated with the pathophysiological basis of lesions by constructing radiomics models. This approach enables comprehensive, objective, and quantitative assessment of the spatial and temporal heterogeneity of lesions \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003ePrevious radiomics studies on pediatric NB have been relatively limited due to the particularities of the pediatric population and constraints associated with pathological sample collection. Ilker et al. developed a CT-based clinical radiomics model to predict and differentiate between pediatric Wilms' tumor and adrenal NB \u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e. Wang, Liu et al. separately employed contrast-enhanced CT-based radiomics and computational pathology analyses to predict the risk classification (INPC) and pathological prognostic classification of NB \u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e. However, radiomics research focusing on lymph node metastasis in NB has not been reported to date. In the present study, we retrospectively collected NB cases over the past decade, expanding the research scope of pediatric NB to the entire peripheral body region to achieve full coverage of lesion locations in individual cases. This approach ensured maximum diversity of the study sample. More importantly, we established a nomogram model integrating contrast-enhanced CT radiomics with validated clinical and laboratory data. This nomogram model further improved the detection rate and diagnostic accuracy of imaging for lymph node metastasis in pediatric body NB.\u003c/p\u003e \u003cp\u003eAll six predictive models constructed in this study for lymph node metastasis in NB exhibited excellent clinical goodness-of-fit and demonstrated high diagnostic performance in both the training and test sets. The DCA also revealed substantial clinical decision-making benefits of these predictive models. Particularly, the nomogram model incorporating arterial phase radiomics score, Delta-Relative radiomics score, and Ki-67 index showed the highest predictive value, with AUC values of 0.937 and 0.829 in the training and test sets, respectively. In the human-machine comparison, the diagnostic accuracy of junior radiologists for lymph node metastasis in NB increased from 54% to 75% with the assistance of the nomogram, while that of senior radiologists improved from 60% to 81%. Both groups achieved an accuracy improvement of 21 percentage points. Previously, on conventional CT images, subtle morphological and enhancement pattern changes of partially metastatic lymph nodes in NB were difficult to identify visually, which resulted in a high rate of missed diagnoses of lymph node metastasis in pediatric patients with NB. The nomogram-assisted diagnostic approach thus represents a substantial improvement in diagnostic accuracy for radiologists and provides significant support for the clinical evaluation and management of pediatric NB.\u003c/p\u003e \u003cp\u003eIn addition to constructing radiomics models using radiomic features, this study also incorporated screened clinical indicators to establish clinical models for evaluating lymph node metastasis status in NB patients. These non-radiomic clinical indicators included basic patient information (age at diagnosis, gender), tumor markers (NSE), SF, and Ki-67 index. Basic patient information, as non-quantitative indicators, is not elaborated here. Tumor markers are laboratory indicators reflecting the malignant potential of tumor cells, which can indicate tumor development, progression, and prognosis. The combined application of CT, MRI, and NSE has shown high diagnostic value and specificity in pediatric patients with NB \u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e. SF is a ubiquitous intracellular protein that stores iron and releases it in a controlled manner, serving as a buffer against iron deficiency and iron overload \u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e]. Evans, Veronica, et al. proposed risk stratification for NB patients based on SF levels, age, and disease stage \u003csup\u003e[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e; however, its concentration can also increase significantly in acute inflammation \u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/sup\u003e, resulting in poor specificity.Ki-67 is a biological indicator for evaluating tumor proliferative activity and also reflects tumor invasiveness \u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/sup\u003e. In malignant tumors, the expression level of Ki-67 is correlated with patient prognosis, with higher Ki-67 expression levels associated with poorer patient survival rates \u003csup\u003e[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eUnivariate analysis in this study demonstrated that there were no statistically significant differences in the age at initial diagnosis, gender, or SF levels of patients for the diagnosis of lymph node metastasis in NB(\u003cem\u003eP\u0026thinsp;\u0026ge;\u003c/em\u003e\u0026thinsp;0.05). Meanwhile, multivariate analysis indicated that NSE levels also showed no statistically significant difference in the diagnosis of lymph node metastasis in NB(\u003cem\u003eP\u0026thinsp;\u0026ge;\u003c/em\u003e\u0026thinsp;0.05). The absence of statistical differences in age and gender for diagnosing lymph node metastasis in NB is an objective fact and thus requires no further elaboration. Although both SF and NSE levels were elevated to varying degrees in NB patients, these two laboratory parameters exhibited no statistically significant differences in the diagnosis of lymph node metastasis in NB (\u003cem\u003eP\u0026thinsp;\u0026ge;\u003c/em\u003e\u0026thinsp;0.05). We hypothesize that this may be attributed to the incomplete correlation between the levels of these two biochemical indicators and the invasiveness of tumors to the surrounding and distant lymph nodes; in other words, these two indicators cannot fully reflect the biological behaviors of the tumor.\u003c/p\u003e \u003cp\u003eNumerous previous and recent studies have employed radiomics to predict the Ki-67 expression index in various tumors, including glioma, gastrointestinal stromal tumor, and breast cancer \u003csup\u003e[\u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e. In the present study, the Ki-67 expression index, as a biological indicator for evaluating tumor proliferative activity, was ultimately included in the construction of the clinical and nomogram predictive models after screening. This indicator has also been incorporated into the development of tumor predictive models in many previous studies. Qiu \u003csup\u003e[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]\u003c/sup\u003e, Zhang \u003csup\u003e[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]\u003c/sup\u003e et al. constructed a combined model based on radiomics combined with biochemical indicators such as Ki-67 and Her, which effectively predicted lymph node metastasis in breast cancer. Cozzi \u003csup\u003e[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/sup\u003e et al. confirmed that contrast-enhanced CT radiomics combined with indicators including Ki-67 could serve as an effective tool for predicting the subtypes and invasiveness of pulmonary neuroendocrine neoplasms. Chen \u003csup\u003e[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]\u003c/sup\u003e et al. established a model based on radiomics analysis of preoperative contrast-enhanced CT images combined with clinical indicators such as Ki-67 to predict vascular invasion (VI) in gastric cancer patients, thereby optimizing the clinical treatment of these patients.\u003c/p\u003e \u003cp\u003eThe aforementioned studies indicate that radiomics combined with Ki-67 has considerable research potential and clinical utility in evaluating tumor invasiveness, heterogeneity, and intratumoral and peritumoral microenvironmental characteristics, providing new insights and approaches for future clinical diagnosis and treatment.In summary, this is the rationale and original intention for our further investigation into tumor biological behaviors using radiomics information combined with clinical indicators, which further confirms the clinical practical value of the nomogram predictive model integrating radiomics and clinical indicators.\u003c/p\u003e \u003cp\u003eThis study has several limitations. First, it was a single-center, single-region retrospective analysis without external validation or multi-regional research, which represents a critical concern regarding the general applicability of radiomics models. In the future, we will further expand the sample size and conduct multi-center, multi-regional collaborations to improve the broad applicability of the predictive model. In addition, the methods used to determine Ki-67 expression in this study included preoperative core needle biopsy (CNB) and postoperative immunohistochemical analysis of surgical specimens, and these two approaches may yield inconsistent results \u003csup\u003e[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]\u003c/sup\u003e. A previous study on breast cancer demonstrated that the discrepancy between the Ki-67 index measured by preoperative biopsy and that determined by postoperative pathological examination could be as high as 10\u0026ndash;40% \u003csup\u003e[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/sup\u003e, which may be attributed to tumor heterogeneity and the limited tissue volume obtained during biopsy procedures. Third, to ensure the integrity of the data samples, children with NB who had received prior treatment were excluded from the study cohort. Therefore, the diagnostic accuracy of the predictive model constructed in this study for evaluating lymph node metastasis in pediatric patients with NB after treatment requires further investigation and validation.\u003c/p\u003e"},{"header":"8. Conclusion","content":"\u003cp\u003eIn conclusion, the nomogram integrating contrast-enhanced CT radiomics and clinical indicators has high practical value in the diagnostic assessment of lymph node metastasis in pediatric patients with peripheral NB. It can significantly improve the diagnostic accuracy of radiologists and provide clinically applicable evidence for the clinical evaluation and formulation of personalized treatment regimens for these patients.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective study complied with the Declaration of Helsinki and was approved by the Ethics Committee of the Children\u0026apos;s Hospital Affiliated to Shandong University (Children\u0026apos;s Hospital of Jinan) (No.SDFE-IRB/T-2026014). Written informed consent was waived by the ethics committee of the Children\u0026apos;s Hospital Affiliated to Shandong University.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003ch2\u003eCompeting interests\u003c/h2\u003e\n\u003cp\u003eThe authors declare that they have no conflicts of interest.\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eNo funding.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eWB (1): Conceptualization, Methodology, Investigation, Writing - Original Draft.GY (2): Data Curation.ZD (3): Investigation, Data Collection, Validation.JY (4): Data Curation.BQ (5): Visualization.JZ (1)*: Conceptualization, Supervision. Corresponding author.GF (2)*: Writing - Review \u0026amp; Editing. Corresponding author.All authors have read and approved the final manuscript.\u003c/p\u003e\n\u003ch2\u003eAcknowledgements\u003c/h2\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eThe data are not publicly available because they contain information that could compromise research participant privacy or consent but are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMaris JM. Recent advances in neuroblastoma. N Engl J Med. 2010;362:2202\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShimada H, DeLellis RA, Marx A. Neuroblastic tumours of the adrenal gland. In: Lloyd RV, Osamura RY, Kloppel \u0026uml; G, Rosai J, editors. WHO Classification of Tumours of Endocrine Organs. 4th ed. Lyon: IARC; 2017. pp. 196\u0026ndash;203.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan KJ, Chung DH. Neuroblastoma: tumor biology and its implications for staging and treatment[J]. Child (Basel). 2019;6(1):12\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoodman MTGJ, Smith MA, Olshan AF. Sympathetic nervous system tumors. Cancer incidence and survival among children and adolescents: United States SEER Program, 1975\u0026ndash;1995. Bethesda, MD; 1999.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKholodenko IV, Kalinovsky DV, Doronin SM, Deyev II, Kholodenko RV. Neuroblastoma Origin and Therapeutic Targets for Immunotherapy. J Immunol Res 2018 (2018) 7394268.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLouis CU, Shohet JM. Neuroblastoma: molecular pathogenesis and therapy. Annu Rev Med. 2015;66:49\u0026ndash;63.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSantiago T, Polanco AC, Fuentes-Alabi S, et al. Multinational retrospective central pathology review of neuroblastoma: lessons learned to establish a regional pathology referral center in resource-limited settings. Arch Pathol Lab Med. 2021;145:214\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLa Quaglia MP. Surgical management of neuroblastoma. Semin Pediatr Surg. 2001;10(3):132\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark JR, Bagatell R, Cohn SL, et al. Revisions to the International Neuroblastoma Response Criteria: A Consensus Statement From the National Cancer Institute Clinical Trials Planning Meeting. J Clin Oncol. 2017;35(22):2580\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBagatell R, Park JR, Acharya S, Neuroblastoma, et al. Version 2.2024, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2024;22(6):413\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVoss SD. Staging and following common pediatric malignancies: MRI versus CT versus functional imaging. Pediatr Radiol. 2018;48(9):1324\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbbaspour E, Karimzadhagh S, Monsef A, Joukar F, Mansour-Ghanaei F, Hassanipour S. Application of radiomics for preoperative prediction of lymph node metastasis in colorectal cancer: a systematic review and meta-analysis. Int J Surg. 2024;110(6):3795\u0026ndash;813.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Q, Lin Y, Ding C, et al. Multi-modality radiomics model predicts axillary lymph node metastasis of breast cancer using MRI and mammography. Eur Radiol. 2024;34(9):6121\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi S, Wang Y, Sun Y, et al. Both intra- and peri-tumoral radiomics signatures can be used to predict lymphatic vascular space invasion and lymphatic metastasis positive status from endometrial cancer MR imaging. Abdom Radiol (NY). 2024;49(11):4140\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSong J, Yin Y, Wang H, Chang Z, Liu Z, Cui L. A review of original articles published in the emerging field of radiomics. Eur J Radiol. 2020;127:108991.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZwanenburg A, Valli\u0026egrave;res M, Abdalah MA, et al. The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping. Radiology. 2020;295(2):328\u0026ndash;38.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShin J, Seo N, Baek S-E et al. MRI Radiomics Model Predicts Pathologic Complete Response of Rectal Cancer Following Chemoradiotherapy. Radiology.2022;303(2):351\u0026ndash;358.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFornacon-Wood I, Faivre-Finn C, O'Connor JPB, Price GJ. Radiomics as a personalized medicine tool in lung cancer: Separating the hope from the hype. Lung Cancer. 2020;146:197\u0026ndash;208.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoska IO, Ozcan HN, Tan AA, et al. Radiomics in differential diagnosis of Wilms tumor and neuroblastoma with adrenal location in children. Eur Radiol. 2024;34(8):5016\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang H, Xie M, Chen X, et al. Radiomics analysis of contrast-enhanced computed tomography in predicting the International Neuroblastoma Pathology Classification in neuroblastoma. Insights Imaging. 2023;14(1):106. Published 2023 Jun 14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Jia Y, Hou C, et al. Pathological prognosis classification of patients with neuroblastoma using computational pathology analysis. Comput Biol Med. 2022;149:105980.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHao J, Sang J, Xu X, Bao A. Diagnostic value of CT and MRI combined with serum LDH, NSE, CEA, and MYCN in pediatric neuroblastoma. World J Surg Oncol. 2023;21(1):251.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang W, Knovich MA, Coffman LG, Torti FM, Torti SV. Serum ferritin:past, present and future. Biochim Biophys Acta. 2010;1800(8):760\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEvans AE, D\u0026rsquo;angio GJ, Propert K, Anderson J, Hann HL. Prognostic factors in neuroblastoma. Cancer. 1987;59:1853\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoroz V, Machin D, Hero B, et al. The prognostic strength of serum LDH and serum ferritin in children with neuroblastoma: A report from the International Neuroblastoma Risk Group (INRG) project. Pediatr Blood Cancer. 2020;67(8):e28359.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOng DS, Wang L, Zhu Y, Ho B, Ding JL. The response of ferritin to LPS and acute phase of Pseudomonas infection. J Endotoxin Res. 2005;11(5):267\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBi S, Li J, Wang T, Man F, Zhang P, Hou F, Wang H, Hao D. Multi-parametric MRI-based radiomics signature for preoperative prediction of Ki-67 proliferation status in sinonasal malignancies: a two-centre study. Eur Radiol. 2022;32(10):6933\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaeshima AM, Taniguchi H, Hori Y, Ida H, Hosoba R, Makita S, Fukuhara S, Munakata W, Suzuki T, Maruyama D, et al. Diagnostic utility and prognostic significance of the Ki-67 labeling index in diffuse large B-cell lymphoma transformed from follicular lymphoma: a study of 76 patients. Pathol Int. 2021;71(10):674\u0026ndash;81.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNi J, Zhang H, Yang Q, et al. Machine-Learning and Radiomics-Based Preoperative Prediction of Ki-67 Expression in Glioma Using MRI Data. Acad Radiol. 2024;31(8):3397\u0026ndash;405.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCai W, Guo K, Chen Y, Shi Y, Chen J. Sub-regional CT Radiomics for the Prediction of Ki-67 Proliferation Index in Gastrointestinal Stromal Tumors: A Multi-center Study. Acad Radiol. 2024;31(12):4974\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu B, Xu Y, Gong H, Tang L, Li H. Radiomics Analysis of Intratumoral and Various Peritumoral Regions From Automated Breast Volume Scanning for Accurate Ki-67 Prediction in Breast Cancer Using Machine Learning. Acad Radiol. 2025;32(2):651\u0026ndash;63.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQiu X, Fu Y, Ye Y, Wang Z, Cao C. A Nomogram Based on Molecular Biomarkers and Radiomics to Predict Lymph Node Metastasis in Breast Cancer. Front Oncol. 2022;12:790076. Published 2022 Mar 15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang D, Shen M, Zhang L, He X, Huang X. Establishment of an interpretable MRI radiomics-based machine learning model capable of predicting axillary lymph node metastasis in invasive breast cancer. Sci Rep. 2025;15(1):26030.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCozzi D, Bicci E, Cavigli E, et al. Radiomics in pulmonary neuroendocrine tumours (NETs). Radiol Med. 2022;127(6):609\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Z, Zhang G, Liu Y, Zhu K. Radiomics analysis in predicting vascular invasion in gastric cancer based on enhanced CT: a preliminary study. BMC Cancer. 2024;24(1):1020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDowsett M, Nielsen TO, A\u0026rsquo;Hern R, et al. Assessment of Ki67 in breast cancer: recommendations from the international Ki67 in breast cancer working group. J Natl Cancer Inst. 2011;103(22):1656\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRossi C, Fraticelli S, Fanizza M, et al. Concordance of immunohistochemistry for predictive and prognostic factors in breast cancer between biopsy and surgical excision: a single-centre experience and review of the literature. Breast Cancer Res Treat. 2023;198(3):573\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"cancer-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"caig","sideBox":"Learn more about [Cancer Imaging](https://cancerimagingjournal.biomedcentral.com/)","snPcode":"40644","submissionUrl":"https://submission.nature.com/new-submission/40644/3","title":"Cancer Imaging","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Children, Neuroblastoma, Lymph node metastasis, Radiomics, Clinical indicators, Ki-67","lastPublishedDoi":"10.21203/rs.3.rs-8820742/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8820742/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjective\u003c/h2\u003e \u003cp\u003eTo construct a machine learning model using contrast-enhanced CT radiomics and clinical indicators, aiming to improve the diagnostic accuracy for lymph node metastasis in children with peripheral neuroblastoma and provide practical evidence for clinical diagnosis and treatment.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA total of children with pathologically confirmed neuroblastoma were retrospectively enrolled between February 2014 and December 2024, and then randomly divided into a training set and a test set via random sampling. Radiomics features were extracted separately from CT images of the arterial phase and venous phase. In the training set, four radiomics models, one clinical model, and one combined model incorporating radiomics and clinical features were constructed respectively using the filtered radiomics features and clinical features. All models were validated against the pathological reference standard in the test set, and the area under the receiver operating characteristic curve of each model was calculated. The clinical utility of each model was evaluated using the decision curve analysis curve. The optimal model was visualized with a nomogram, and the diagnostic gain of the nomogram for evaluating lymph node metastasis in neuroblastoma children was quantified via human-machine comparison.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eA total of 225 children with neuroblastoma were enrolled in this study, (with a mean age of 2.23\u0026thinsp;\u0026plusmn;\u0026thinsp;2.34 years and an age range of 0\u0026ndash;13 years). All subjects were randomly divided into a training set (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;157) and a test set (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;68) at a ratio of 7:3. Compared with four radiomics models (Arterial phase, Venous phase, Delta-Absolute, Delta-Relative) and one clinical model (Ki-67), the nomogram integrating radiomics and clinical features (Arterial phase\u0026thinsp;+\u0026thinsp;Delta-Relative\u0026thinsp;+\u0026thinsp;Ki-67) exhibited superior diagnostic performance in evaluating lymph node metastasis in pediatric peripheral neuroblastoma. The AUC values of the nomogram reached 0.937 and 0.829 in the training set and validation set, respectively. In the human-machine comparison experiment, the diagnostic accuracy of radiologists for lymph node metastasis in neuroblastoma children was improved by 21% when assisted by the nomogram.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThe nomogram combining contrast-enhanced CT radiomics and clinical indicators has significant diagnostic value in evaluating lymph node metastasis in pediatric patients with peripheral neuroblastoma. Moreover, it can substantially improve the diagnostic accuracy of radiologists with different levels of clinical experience.\u003c/p\u003e","manuscriptTitle":"Application Value of a Nomogram Integrating Contrast-enhanced CT Radiomics and Clinical Indicators in Evaluating Lymph Node Metastasis in Pediatric Peripheral Neuroblastoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-08 17:33:04","doi":"10.21203/rs.3.rs-8820742/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-04T22:06:31+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-10T12:21:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"44134132263771030492158946740382555475","date":"2026-03-10T14:03:53+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-02T17:58:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-11T16:58:55+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-11T11:21:25+00:00","index":"","fulltext":""},{"type":"submitted","content":"Cancer Imaging","date":"2026-02-08T10:00:23+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"cancer-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"caig","sideBox":"Learn more about [Cancer Imaging](https://cancerimagingjournal.biomedcentral.com/)","snPcode":"40644","submissionUrl":"https://submission.nature.com/new-submission/40644/3","title":"Cancer Imaging","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"97353309-2ae0-4dbc-b8b5-2741272399ac","owner":[],"postedDate":"March 8th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-04T22:06:31+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-18T12:38:34+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-08 17:33:04","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8820742","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8820742","identity":"rs-8820742","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}