Intra- and peritumoral MRI radiomics assisted in predicting radiochemotherapy response in metastatic cervical lymph nodes of nasopharyngeal cancer

Intra- and peritumoral MRI radiomics assisted in predicting radiochemotherapy response in metastatic cervical lymph nodes of nasopharyngeal cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Intra- and peritumoral MRI radiomics assisted in predicting radiochemotherapy response in metastatic cervical lymph nodes of nasopharyngeal cancer Hao Xu, Ai Wang, Chi Zhang, Jing Ren, Jieke Liu, Peng Zhou This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2519551/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 30 May, 2023 Read the published version in BMC Medical Imaging → Version 1 posted 9 You are reading this latest preprint version Abstract Background: In this investigation, intratumoral (Intra) and peritumoral (Peri) features obtained from MRI imaging were used to create and evaluate radiomic models for response prediction to radiochemotherapy of metastatic cervical lymph nodes in individuals with nasopharyngeal cancer (NPC). Methods: Retrospectively, we included 145 consecutive subjects with NPC, 102 in the training set and 43 in the validation set. A total of 5408 initial radiomic features were acquired from the metastatic cervical lymph node's Intra and Peri areas. Then, employing multivariate logistic regression analysis, the radiomic features were chosen and integrated with clinical characteristics to create predictive models. And at last, these developed prediction models were examined using sensitivity, specificity, accuracy, and the area under the curve (AUC) of receiver operating characteristics. Results: In the training and validation sets, there was no statistically significant variation in the AUC among the Intra radiomic signature, Peri radiomic signature, combined Intra and Peri radiomic signature, and combined Intra and Peri radiomic nomogram (all P > 0.05). With an AUC of 0.941 (0.877-0.978) in the training set and 0.783 (0.631-0.894) in the validation set, the combined Intra and Peri radiomic nomogram enabled good discrimination among the responders and non-responders groups. Conclusions: The early response of metastatic cervical lymph nodes to radiochemotherapy in individuals with NPC may be predicted by pretreatment radiomic models determined by the combined Intra and Peri features from MRI imaging, facilitating therapeutic interventions and clinical decision-making. Nasopharyngeal cancer Magnetic resonance imaging Radiomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Background Nasopharyngeal cancer (NPC) is the most frequent type of head and neck cancer in Southeast Asia and South China [ 1 – 3 ]. Approximately 86.4% of NPC patients develop lymph node metastasis at the cervical or/and retropharyngeal site before therapy according to radiologic criteria [ 4 ]. Intensity-modulated radiotherapy (IMRT) is the main and curative therapy for NPC because of its intrinsic anatomic restrictions and great radiosensitivity. Previous research revealed that chemotherapy could assist patients with advanced diseases [ 5 , 6 ], and concurrent chemoradiotherapy (CCRT) combined with IMRT was later designated as the major typical therapy for stage II-IVA NPC by National Comprehensive Cancer Network (NCCN) recommendations [ 7 ]. A significant predictive factor in determining survival outcomes in NPC is the cancer response to radiochemotherapy [ 8 ]. Not all patients, however, react as well to radiochemotherapy. Predicting response prior to treatment may result in more targeted and personalized treatment, avoiding unnecessary side effects and costs. However, there are no established biomarkers, and some have been proposed as research tools. Thus, there is an urgent need to identify an effective radiochemotherapy response predictor in patients with NPC. The ability of multimodality imaging biomarkers generated from computed tomography (CT), 18 F-fluorodeoxyglucose positron emission tomography, or MRI-DWI to distinguish between metastatic cervical lymph node responses to therapy has been shown in NPC [ 9 – 11 ]. Conventional MRI plays a vital function in the diagnosis and management of NPC and is extensively employed in the early recognition, diagnosis, staging, and assessment of therapy response. MRI images, with a high soft tissue resolution, not only contain anatomical information about the primary NPC lesion and its adjacent constructions but also reflect the intra-tumor characteristics [ 12 ]. Radiomics, which is extensively used in disease identification, differential diagnosis, prognosis prediction, and therapeutic response evaluation [ 12 – 16 ], is the high-throughput computer extraction of potentially unlimited quantities of quantitative imaging metrics. According to recent investigations, radiomic features derived from pretreatment MRI images or MRI-based radiomic nomograms might anticipate NPC patients' progression-free survival [ 17 , 18 ]. While lacking independent validation, the pretreatment MRI signature may predict early treatment response to initiation chemotherapy in individuals with NPC [ 19 ]. However, these previous investigations have primarily focused on the function of MRI-based radiomics in expecting the treatment response of primary tumors in NPC patients. Its importance in identifying the cervical lymph nodes' treatment response is yet unclear. The objective of the investigation was to establish and validate radiomic models for predicting response to radiochemotherapy of cervical metastatic lymph nodes of NPC employing Intra and Peri radiomic features obtained from MRI. Materials And Methods Patient cohort The retrospective investigation received approval from our ethics committee and institutional review board, and informed consent was not required. Between October 2016 and November 2019 at our institution, 185 consecutive patients with histologically confirmed NPC treated with induction chemotherapy (IC) and CCRT were initially recruited retrospectively. The study's inclusion criteria were as follows: (1) histopathologically proven NPC; (2) available pre-treatment T2-weighted imaging (T2WI) and contrast-enhanced T1-weighted imaging (CE-T1WI) images before biopsy; (3) available post-treatment T2WI and CE-T1WI images; (4) no treatment before baseline biopsy; and (5) available clinical variables such as age, sex, T-stage, N-stage, clinical stage, and lymph node gross tumor volume (GTV-ln). The following were the exclusion criteria: (1) the period between baseline MRI and initial treatment was further than 2 weeks (n = 14); (2) the existence of other malign tumors (n = 9); (3) missing clinicopathological information (n = 10); (4) poor image quality (n = 7). Finally, a total of 145 patients (mean age 47.2 ± 12.6, range 13–77, male = 117, female = 28, stage II = 2, stage III = 59, stage IV = 84) were enrolled and allocated to the training set (102 patients, October 2016 to December 2018) and validation set (43 patients, December 2018 to November 2019) [20]. A flowchart of the patient selection is shown in Fig. 1. Patients were staged based on the 8th version American Joint Committee on Cancer (AJCC) Tumor-Node-Metastasis (TNM) staging system [21]. From computerized medical records, demographic information and stages were extracted. MR images acquisition protocol All patients underwent nasopharyngeal and cervical region MRI examination with head-neck combined coils employing either a 1.5-T MR scanner (Avanto, Siemens, Germany) or a 3.0-T MR imaging scanner (Skyra, Siemens, Germany). Overall, 94/145 (64.8%) of the subjects had imaging done in a 1.5-T MRI scanner, and 51/145 (35.2%) had imaging done in a 3.0-T scanner. To keep away from magnetic exposure and motion artifacts, patients were instructed to eliminate all metal-containing items and lie supinely in the scanner before scanning. Before the MRI examination, all patients will be required to wear earplugs and headphones to decrease noise. The DICOM format images of axial fat-suppressed CE-T1WI (contrast agent = Gd-DTPA, Magnevist, Schering, Berlin, Germany; dose = 0.1 mmol/kg body weight) and T2WI scans of each case were retrieved from the picture archiving and communication system (Carestream, Ontario, Canada). Additional file 1 provided detailed image acquisition parameters (Additional file 1: S1). The inverse fourier transform with the linear filling was used to reconstruct all images from k-space. Evaluation of lymph nodes Multiple radiologic criteria were used to determine whether metastatic cervical lymph nodes were involved. These included: (1) regions of central necrosis/cystic necrosis (T2WI with a focal high signal intensity or CE-T1WI with low signal intensity/ with or without an adjacent border of enhancement), (2) extracapsular distribution in any size lymph node, including ambiguous nodal boundaries, irregular nodal capsular improvement, and infiltration into neighboring muscle or fat, (3) the shortest diameter of the cervical or medial retropharyngeal lymph node is 10 mm; the lateral retropharyngeal lymph node is 5 mm. In each patient, we chose the largest lymph node within the scanning range as the target lesion. The target lymph nodes were then assessed on T2WI for maximum axial diameter and minimum axial diameter long axis. The minimum axial diameter matched the node's largest diameter in the axial plane perpendicular to its maximum axial diameter. The Response Evaluation Criteria in Solid Tumors 1.1 (RECIST) was used to assess the treatment response [22]. The responders were defined as a reduction of at least 30% in the target lesions' maximum axial diameter. In contrast, non-responders had inadequate shrinking to qualify for the responders. Preprocessing of images All raw images were preprocessed utilizing the Artificial Intelligence Kit program (A.K., version 3.2.0, GE Healthcare, China) prior to segmentation because of the variability of CE-T1WI and T2WI acquisition parameters. Using a trilinear interpolation algorithm, all raw images were initially resliced to isotropic 1 mm 3 voxels. The center was employed to align the interpolation grid's location, and its measurements were rounded to the closest integer. Following the use of a Gaussian filter for anti-aliasing during down-sampling, the bias-field correction was also performed. Lastly, voxel intensity values were standardized utilizing a z-score normalization approach to remove the impact of various gray value ranges. Image segmentation The target lymph nodes on the axial T2WI and CE-T1WI images separately served as the basis for the region of interest (ROIs). A junior radiologist with 7 years of experience in head and neck imaging manually segmented cancer on the ITK-SNAP (version 3.8.0, http://www.itksnap.org), which was then examined by a senior radiologist with 20 years of experience in head and neck imaging. The ROIs were created for every lesion to completely include the target lymph node on every succeeding slice. Then, utilizing A.K. software, equidistant 3-dimensional dilation of the Intra sections by 2 mm was used to create the Peri area. To measure interobserver variability in manual segmentation, the senior radiologist re-delineated 30 target lymph nodes that were randomly chosen from the T2WI and CE-T1WI images, respectively. Two experienced radiologists were unaware of the target lymph nodes' reaction to the treatment. The repeatability of radiomic features was then evaluated using the intraclass correlation coefficient (ICC), and those with an ICC value greater than 0.75 were selected for the subsequent process. The typical MRI image and its related Intra and Peri ROIs are displayed in Fig. 2. Feature extraction Following image preprocessing and segmentation, features from the Intra and Peri ROIs were extracted using an open-source Python program (PyRadiomics, version 3.0, https://pyradiomics.readthedocs.io) [23] that followed the image biomarker standardisation initiative (IBSI) standard [24]. These features belonged to three categories: shape, first order, and texture features. Image intensity was discretized using a fixed bin width of 25 before texture features were estimated. For every patient, 2704 radiomic features (1352 from T2WI and CE-T1WI, respectively) were lastly extracted in the Intra and Peri ROI, respectively (see Additional file 1: Table S1 and Table S2 for a detailed list of the extracted features). Feature selection and radiomic signature building All radiomic features were first normalized employing z-scores and then selected in the training set for the construction of radiomic signature (Intra, Peri, and combined Intra and Peri). For feature selection, we performed a five-step procedure. We used the Mann-Whitney U test to make the initial selection from the training set. For the remaining significant features, the P-value threshold was set at 0.05. Then, to eliminate redundant features, Spearman correlation assessment and maximum relevance-minimum redundancy (mRMR) were performed successively. Features with Spearman correlation coefficient values greater than 0.9 were deemed redundant and were thus eliminated. Simultaneously, the higher-ranked 15 features associated with the reaction of metastatic lymph nodes to therapy were retained following mRMR. Thirdly, the least absolute shrinkage and selection operator (LASSO) technique with cross-validations was utilized to detect the most predictive features related to the reaction of metastatic lymph nodes to therapy [25]. Lastly, radiomic signature building depended on multivariate logistic regression with backward stepwise selection utilizing the likelihood ratio examination with Akaike's data criterion as the stopping rule [26]. The radiomic signatures were computed utilizing a linear combination of the chosen features weighted by their corresponding coefficients. To examine possible multicollinearity between radiomic signature features, Spearman's correlations were also employed. For feature selection and radiomic signature building, R software (version 4.0.0, http://www.r-project.org) was employed. Construction and validation of the radiomics-based model Univariate logistic regression analysis was utilized to recognize independent risk factors for differentiating responders from non-responders between radiomic signatures and clinical variables. Subsequently, to provide a more individualized predictive model, a radiomic nomogram combining the radiomic signature and significant independent risk factors was built in the training set, using the multivariable logistic regression analysis. For both the training and validation sets, a receiver operating characteristic curve (ROC) and area under the curve (AUC) calculations were made to assess the radiomic-based model's ability to discriminate. The threshold was established utilizing the maximum Youden index (sensitivity + specificity − 1) on the training set and this same threshold was used for the validation set. Calculations were also made for the relevant sensitivity, specificity, and accuracy. The above diagnostic measures' 95% confidence intervals (CIs) were computed utilizing the binomial exact approach. The Hosmer-Lemeshow goodness-of-fit analysis was employed to measure the calibration of the prediction model. Building calibration plots will also compute the calibration slope and calibration-in-the-large statistic (intercept) [27]. TRIPOD statement and statistical analysis We presented the results of this article based on the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) reporting checklist since the study was intended to serve as a prediction model building for diagnostic reasons [20]. SPSS (version 26.0, available at https://www.ibm.com) and MedCalc (version 18.2.1, available at https://www.medcalc.org) were utilized for all statistical analyses. The Shapiro-Wilk test was used to examine if quantitative data had a normal distribution. The clinical and demographic variables of the responders and non-responders were compared using an independent samples t-test, the Mann-Whitney U test, Fisher's exact analysis, or the chi-squared (χ2) test, as applicable. The AUCs of several prediction models in the training and validation sets were compared employing the Delong method [28]. Using McNemar's chi-square test, further comparisons of sensitivity and specificity in the training and validation sets were made [29]. Statistical significance was identified as a P-value< 0.05 (two-tailed). Results Baseline characteristics of the patient The clinical characteristics and sociodemographics of 145 patients enrolled in this investigation are listed in Table 1. The numbers of responders and non-responders in the training and validation sets were 56 and 46, and 24 and 19, respectively. There were no substantial variations in age, sex, T- stage, N-stage, and clinical stage between responders and non-responders in the training and validation sets (all P > 0.05). There was a substantial variation in the GTV-ln between responders and non-responders in the two sets ( P < 0.001 and = 0.047). Feature selection and radiomic signature development 5408 radiomic features were obtained (2704 in the Intra region and 2704 in the Peri region). After calculating the ICCs, the numbers of reproducible features (ICCs > 0.75) from the CE-T1WI and T2WI images in the Intra region and Peri region were 940 and 1076, and 971 and 975, respectively. Then, five feature selection methods like the Mann-Whitney U test, Spearman correlation assessment, mRMR procedure, LASSO, and multivariate logistic regression were utilized to obtain the radiomic features with the most strongly associated with responders. Finally, we selected 4, 4, and 5 features for Intra, Peri, and combined Intra and Peri radiomic signatures construction (Additional file 1: Table S3 and Table S4). A detailed description of the procedure for feature selection and LASSO results were provided in Additional file 1 (Additional file 1: Fig. S1 and Fig. S2). We also included the mean and standard deviation of these selected features in Additional file 1 (Additional file 1: Table S5) for z-score normalization prior to estimation. The radiomic signature of Intra was constructed using the following formula: ln (P/1-P) = -0.8765 – 0.8685 × CE-T1WI_firstorder_wavelet.HHH.Skewness + 1.2215 × CE-T1WI_GLDM_wavelet.HHL.LDLGLE + 2.8194 × T2WI_GLCM_LoG.sigma.1.5.mm.3D.ClusterShade – 2.1866 × T2WI_GLCM_wavelet.LLH.MaximumProbability, where P was the possibility of responders (threshold > 0.6406). The radiomic signature of Peri was constructed using the following formula: ln (P/1-P) = -0.0330 – 1.3038 × CE-T1WI_firstorder_LoG.sigma.0.5.mm.3D.Minimum + 1.8829 × CE-T1WI_GLRLM_wavelet.HHH.SRHGLE + 1.1131 × T2WI_GLDM_wavelet.LHL.SDHGLE – 1.3508 × T2WI_GLCM_wavelet.LLH.ClusterTendency, where P was the possibility of responders (threshold > 0.6899). The radiomic signature of combined Intra and Peri was constructed using the following formula: ln (P/1-P) = -0.0335 – 1.0301 × Intra_CE-T1WI_firstorder_LoG.sigma.0.5.mm.3D.90Percentile – 3.0175 × Intra_T2WI_GLCM_wavelet.LLH.MaximumProbability – 1.8770 × Peri_CE-T1WI_firstorder_LoG.sigma.0.5.mm.3D.Minimum + 1.3759 × Peri_CE-T1WI_GLRLM_wavelet.HHH. SRHGLE + 1.4714 × Peri_T2WI_GLRLM_wavelet.HLL.LRLGLE, where P was the possibility of responders (threshold > 0.3525). In Additional file 1 (Additional file 1: Fig. S3), the distribution of these radiomic signatures of responders and non-responders in the training and validation sets was displayed. A heatmap was utilized to show the Spearman correlations between specific features in these radiomic signatures (Additional file 1: Fig. S4). Construction, validation, and performance of radiomic models GTV-ln was shown to be an important independent factor for distinguishing responders from non-responders in a univariate logistic regression assessment ( P = 0.003), however, a subsequent multivariate logistic regression examination revealed that it was only substantial in the combined Intra and Peri dataset ( P = 0.036) (Table 2). So, the combined Intra and Peri radiomic nomogram was generated by incorporating the combined Intra and Peri radiomic signature and GTV-ln (Fig. 3a). The formula for the combined Intra and Peri radiomic nomogram calculation: ln (P/1-P) = -2.8790 + 0.0590 × GTV-ln + 1.0770 × combined Intra and Peri radiomic signature, where P is the possibility of responders (threshold > 0.4815). Lastly, four radiomic models (combined Intra and Peri radiomic nomogram, Intra radiomic signature, Peri radiomic signature, and combined Intra and Peri radiomic signature) for the prediction of responders were constructed. The combined Intra and Peri radiomic nomogram illustrated worthy performance in discriminating the responders and non-responders in patients with NPC, which yielded an AUC of 0.941 (95% confidence interval [CI]: 0.877-0.978), Sen of 92.9%, Spe of 84.8%, and ACC of 89.2% in the training set. The corresponding AUC, Sen, Spe, and ACC for the combined Intra and Peri radiomic nomogram in the validation set were 0.783 (95% CI: 0.631-0.894), 66.7%, 84.2%, and 74.4%, respectively (Table 3). Fig. 4. shows the radiomic models' ROC curves in both cohorts. Table 3 presents the AUC, Sen, Spe, and ACC of each radiomic model in both cohorts. Representative CE-T1WI and T2WI images of NPC patients as responders and non-responders were shown in Fig. 5. In training and validation sets, the combined Intra and Peri radiomic nomogram calibration curve exhibited excellent agreement between the predictive and observational probability of differentiating between responders and non-responders (Fig. 3b), and the Hosmer-Lemeshow assessment outcomes were non-significant ( P = 0.618 and 0.780). The Hosmer-Lemeshow test also yielded non-significant results in training and validation sets of the Intra radiomic signature ( P = 0.405 and 0.165), Peri radiomic signature ( P = 0.300 and 0.512), and combined Intra and Peri radiomic signature ( P = 0.610 and 0.158), which revealed no departure from the ideal fit (Additional file 1: Fig. S5). Radiomic models comparison and TRIPOD Depended on the DeLong analysis, there were no substantial variations in AUCs across the radiomic models in the training and validation sets (all P > 0.05) (Table 4). The combined Intra and Peri radiomic nomogram's sensitivity was greater than the Peri radiomic signature in training and validation sets, according to the McNemar analysis ( P = 0.009 and 0.045) (Table 4). However, no statistically significant differences in specificity between the radiomic models in the training and validation sets were found (all P > 0.05) (Table 4). According to the guidelines of the TRIPOD statement, the type of this investigation belongs to Type 2b. A full list of TRIPOD was provided as Additional file 1 (Additional file 1: Table S6). Discussion In this retrospective investigation, we developed and validated MRI-based radiomic models for predicting response to radiochemotherapy of cervical metastatic lymph nodes of NPC. These models included Intra, Peri, and a combination of Intra and Peri radiomic signatures all demonstrated satisfactory predictive value. Additionally, we demonstrated and verified a radiomic nomogram that incorporated the combined Intra and Peri radiomic signature with GTV-ln, which had an acceptable discrimination capability. Recent studies found that the response to radiochemotherapy was associated with clinical results in NPC. Peng et al [ 8 ]. discovered that the tumor response to radiochemotherapy was recognized as an independent prognostic variable of 4-year disease-free, overall, and locoregional relapse-free survival. In another investigation, Liu et al [ 30 ]. revealed that poor tumor response to radiochemotherapy could be a predictor of advanced-stage NPC. NPC is highly susceptible to regional lymph node metastasis [ 31 ]. Additionally, earlier findings indicate that the characteristics of lymph node metastasis were indicators of distant metastasis and had an effect on prognosis for overall, local recurrence, regional relapse, and disease-free survival [ 32 – 34 ]. Thus, it is essential to enhance the treatment impact of metastatic lymph nodes in order to improve the prognosis of NPC. Our findings may provide an effective treatment approach for metastatic lymph nodes prior to therapy. For patients with NPC, due to different levels of sensitivity to the treatment, patients at the same stage may have different treatment responses. Additionally, successful prevention of therapeutic side effects and disease control, however, requires careful consideration of radiotherapy dose and chemotherapy regimen. Using our predictive model, physicians could preliminarily predict the treatment impact on patients and take preventative measures in advance. Our findings suggest that elevating GTV-ln suitably during radiological treatment planning could be an effective alternative, particularly for these potentially poor responders. Moreover, at this time, radiation and chemotherapy are the main treatments for locoregionally progressed NPC [ 7 ]. Early detection of non-responders using our prediction model might assist to decrease related toxicities and costs by preventing excessively extended therapies and unsuccessful chemotherapy regimens. MRI radiomics has been frequently utilized to assess treatment response in NPC [ 17 , 35 , 36 ]. Wang et al [ 19 ]. discovered that a multisequence pretreatment MRI radiomics signature might predict early IC response. However, there was no additional validation cohort in this study. Liu et al [ 37 ]. confirmed that MRI radiomics has a great potential for predicting chemoradiotherapy response, but only 53 NPC patients were included in this study. In another study, Zhang et al [ 38 ]. demonstrate that radiomic features derived from T2WI and CE-T1WI data enable the prediction progression of NPC with a high accuracy. In the study, the constructed MRI-based radiomic models with peritumoral features may serve as a noninvasive and repeatable diagnostic tool for clinical application in predicting radiochemotherapy response in metastatic cervical lymph nodes of NPC. But further research is necessary before these findings can be applied clinically, because the sensitivity of the nomogram was 0.667 in the validation cohort. Cervical lymph nodes can be evaluated by MRI based on their size and morphologic aspect including central necrosis and extracapsular spread. These nodal characteristics, which were considered in our study, have been demonstrated to be important prognostic factors in NPC patients [ 34 , 39 , 40 ]. However, these morphologic findings cannot provide sufficient information for the accurate prediction of metastatic cervical lymph nodes' therapeutic reaction. Consequently, utilizing the Intra and Peri radiomic features may aid in predicting treatment response. Our investigation showed that the combined Intra and Peri radiomic nomogram achieved satisfactory predictive capability with AUCs of 0.941 and 0.783 in training and validation sets, respectively. Additionally, the McNemar test revealed that in training and validation sets, the combined Intra and Peri radiomic nomogram's sensitivity was greater than that of the Peri radiomic signature ( P = 0.009 and 0.045). In the combined Intra and Peri radiomic signature, the firstorder_90Percentile from the Intra area and firstorder_Minimum from the Peri area were negatively connected with responders, exhibiting that the responders had a low distribution of gray level intensities within the ROI of images than that of non-responders. The significant gray level co-occurrence matrix feature included MaximumProbability from the Intra area. The non-responders had a higher Maximum Probability, indicating greater incidences of the most main pair of adjacent gray values in non-responders than in responders. Besides, the remaining two texture features (short-run elevated gray level emphasis, long-run reduced gray level emphasis) of gray level run length matrix from the Peri area were also chosen to create the combined Intra and Peri radiomic signature. Our findings indicated that, compared to non-responders, the responders showed a larger joint distribution of shorter run lengths with greater gray-level values and long run lengths with lower gray-level values inside the Peri area. Our investigation has many additional restrictions. First, the heterogeneity in acquisition parameters of two different magnetic field strengths scanners may have affected the image texture. Thus, additional studies may be required to quantitatively investigate these effects. Secondly, other MRI method, especially DCE-MRI-based radiomic, has been employed in estimating treatment response for breast malignancy and their function in NPC requirements is to be further explored [ 41 ]. Thirdly, the single-center nature and small sample size of our study limit the generalizability of our models. Hence, a multi-center study with different imaging equipment and a large sample size in the future is required. In summary, the present investigation developed the MRI-based radiomic model that depends on combined Intra and Peri features to anticipate the radiochemotherapy response of metastatic cervical lymph nodes prior to therapy in NPC. This radiomic model may be useful in NPC patients for individualized risk stratification treatment and therapeutic decision-making. Abbreviations AUC: Area under the curve; AJCC: American joint committee on cancer; CI: Confdence internal; CCRT: Concurrent chemoradiotherapy; CT: Computed tomography; CE-T1WI: Contrast-enhanced T1-weighted imaging; GTV-ln: Lymph node gross tumor volume; Intratumoral: Intra; IC: Induction chemotherapy; IMRT: Intensity-modulated radiotherapy; ICC: Intraclass correlation coefficient; IBSI: Image biomarker standardisation initiative; LASSO: Least absolute shrinkage and selection operator; mRMR: Maximum relevance-minimum redundancy; NPC: Nasopharyngeal cancer; NCCN: National comprehensive cancer network; Peritumoral: Peri; RECIST: Response evaluation criteria in solid tumors; ROI: Region of interest; ROC: Receiver operating characteristic curve; T2WI: T2-weighted imaging; TNM: Tumor-node-metastasis; TRIPOD: Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis. Declarations Acknowledgements Not applicable. Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Hao Xu, Ai Wang, Chi Zhang, and Jing Ren. Hao Xu, Ai Wang, Jieke Liu, and Peng Zhou performed data interpretation and statistical analysis. The first draft of the manuscript was written by Hao Xu and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding This study was supported by the Sichuan Science and Technology Program (grant number 2021YFG0125). Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Ethics approval and consent to participate The research was approved and the requirement of informed consent from the patients was waived by Sichuan Cancer Hospital Ethic Committee because of the retrospective design of this study, and patients’ information was protected. The study was performed in accordance with the Declaration of Helsinki. Consent for publication Not applicable. Competing interests None of the authors declare conflict of interest. References Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A: Global cancer statistics, 2012. CA Cancer J Clin. 2015;65(2):87-108. Wei KR, Zheng RS, Zhang SW, Liang ZH, Li ZM, Chen WQ: Nasopharyngeal carcinoma incidence and mortality in China, 2013. Chin J Cancer. 2017;36(1):90. Chen YP, Chan ATC, Le QT, Blanchard P, Sun Y, Ma J: Nasopharyngeal carcinoma. Lancet. 2019;394(10192):64-80. Wang X, Hu C, Ying H, He X, Zhu G, Kong L, Ding J: Patterns of lymph node metastasis from nasopharyngeal carcinoma based on the 2013 updated consensus guidelines for neck node levels. Radiother Oncol. 2015;115(1):41-45. Chen QY, Wen YF, Guo L, Liu H, Huang PY, Mo HY, Li NW, Xiang YQ, Luo DH, Qiu F et al: Concurrent chemoradiotherapy vs radiotherapy alone in stage II nasopharyngeal carcinoma: phase III randomized trial. J Natl Cancer Inst. 2011;103(23):1761-70. Miao J, Wang L, Tan SH, Li JG, Yi J, Ong EHW, Tan LLY, Zhang Y, Gong X, Chen Q et al: Adjuvant Capecitabine Following Concurrent Chemoradiotherapy in Locoregionally Advanced Nasopharyngeal Carcinoma: A Randomized Clinical Trial. JAMA Oncol. 2022. Colevas AD, Yom SS, Pfister DG, Spencer S, Adelstein D, Adkins D, Brizel DM, Burtness B, Busse PM, Caudell JJ et al: NCCN Guidelines Insights: Head and Neck Cancers, Version 1.2018. J Natl Compr Canc Netw. 2018;16(5):479-90. Peng H, Chen L, Zhang Y, Li WF, Mao YP, Liu X, Zhang F, Guo R, Liu LZ, Tian L et al: The Tumour Response to Induction Chemotherapy has Prognostic Value for Long-Term Survival Outcomes after Intensity-Modulated Radiation Therapy in Nasopharyngeal Carcinoma. Sci Rep. 2016;6:24835. Zhang GY, Wang YJ, Liu JP, Zhou XH, Xu ZF, Chen XP, Xu T, Wei WH, Zhang Y, Huang Y: Pretreatment Diffusion-Weighted MRI Can Predict the Response to Neoadjuvant Chemotherapy in Patients with Nasopharyngeal Carcinoma. Biomed Res Int. 2015;2015:307943. Yen RF, Chen TH, Ting LL, Tzen KY, Pan MH, Hong RL: Early restaging whole-body (18)F-FDG PET during induction chemotherapy predicts clinical outcome in patients with locoregionally advanced nasopharyngeal carcinoma. Eur J Nucl Med Mol Imaging. 2005;32(10):1152-59. Yang Y, Wang M, Qiu K, Wang Y, Ma X: Computed tomography-based deep-learning prediction of induction chemotherapy treatment response in locally advanced nasopharyngeal carcinoma. Strahlenther Onkol. 2022;198(2):183-93. Lambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, Sanduleanu S, Larue R, Even AJG, Jochems A et al: Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749-62. Lambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, van Stiphout RG, Granton P, Zegers CM, Gillies R, Boellard R, Dekker A et al: Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer. 2012;48(4):441-46. Rogers W, Thulasi Seetha S, Refaee TAG, Lieverse RIY, Granzier RWY, Ibrahim A, Keek SA, Sanduleanu S, Primakov SP, Beuque MPL et al: Radiomics: from qualitative to quantitative imaging. Br J Radiol. 2020;93(1108):20190948. Gillies RJ, Kinahan PE, Hricak H: Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016;278(2):563-77. Verma V, Simone CB, 2nd, Krishnan S, Lin SH, Yang J, Hahn SM: The Rise of Radiomics and Implications for Oncologic Management. J Natl Cancer Inst. 2017;109(7). Zhao L, Gong J, Xi Y, Xu M, Li C, Kang X, Yin Y, Qin W, Yin H, Shi M: MRI-based radiomics nomogram may predict the response to induction chemotherapy and survival in locally advanced nasopharyngeal carcinoma. Eur Radiol. 2020;30(1):537-46. Mao J, Fang J, Duan X, Yang Z, Cao M, Zhang F, Lu L, Zhang X, Wu X, Ding Y et al: Predictive value of pretreatment MRI texture analysis in patients with primary nasopharyngeal carcinoma. Eur Radiol. 2019;29(8):4105-13. Wang G, He L, Yuan C, Huang Y, Liu Z, Liang C: Pretreatment MR imaging radiomics signatures for response prediction to induction chemotherapy in patients with nasopharyngeal carcinoma. Eur J Radiol. 2018;98:100-06. Moons KG, Altman DG, Reitsma JB, Ioannidis JP, Macaskill P, Steyerberg EW, Vickers AJ, Ransohoff DF, Collins GS: Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med. 2015; 162(1):W1-73. Tang LL, Chen YP, Mao YP, Wang ZX, Guo R, Chen L, Tian L, Lin AH, Li L, Sun Y et al: Validation of the 8th Edition of the UICC/AJCC Staging System for Nasopharyngeal Carcinoma From Endemic Areas in the Intensity-Modulated Radiotherapy Era. J Natl Compr Canc Netw. 2017;15(7):913-19. Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS, Rubinstein L, Verweij J, Van Glabbeke M, van Oosterom AT, Christian MC et al: New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst. 2000;92(3):205-16. van Griethuysen JJM, Fedorov A, Parmar C, Hosny A, Aucoin N, Narayan V, Beets-Tan RGH, Fillion-Robin JC, Pieper S, Aerts H: Computational Radiomics System to Decode the Radiographic Phenotype. Cancer Res. 2017;77(21):e104-e07. Zwanenburg A, Vallieres M, Abdalah MA, Aerts H, Andrearczyk V, Apte A, Ashrafinia S, Bakas S, Beukinga RJ, Boellaard R et al: The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping. Radiology. 2020;295(2):328-38. Sauerbrei W, Royston P, Binder H: Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Stat Med. 2007;26(30):5512-28. Pan W: Akaike's information criterion in generalized estimating equations. Biometrics. 2001;57(1):120-25. Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, Pencina MJ, Kattan MW: Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21(1):128-38. DeLong ER, DeLong DM, Clarke-Pearson DL: Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837-45. Hawass NE: Comparing the sensitivities and specificities of two diagnostic procedures performed on the same group of patients. Br J Radiol. 1997;70(832):360-66. Liu LT, Tang LQ, Chen QY, Zhang L, Guo SS, Guo L, Mo HY, Zhao C, Guo X, Cao KJ et al: The Prognostic Value of Plasma Epstein-Barr Viral DNA and Tumor Response to Neoadjuvant Chemotherapy in Advanced-Stage Nasopharyngeal Carcinoma. Int J Radiat Oncol Biol Phys. 2015;93(4):862-69. Sham JS, Choy D, Wei WI: Nasopharyngeal carcinoma: orderly neck node spread. Int J Radiat Oncol Biol Phys. 1990;19(4):929-33. Feng Y, Cao C, Hu Q, Chen X: Prognostic Value and Staging Classification of Lymph Nodal Necrosis in Nasopharyngeal Carcinoma after Intensity-Modulated Radiotherapy. Cancer Res Treat. 2019;51(3):1222-30. Xu Y, Chen X, Zhang M, Xiao Y, Zong J, Guo Q, Qiu S, Zheng W, Lin S, Pan J: Prognostic effect of parotid area lymph node metastases after preliminary diagnosis of nasopharyngeal carcinoma: a propensity score matching study. Cancer Med. 2017;6(10):2213-21. Zhang LL, Li JX, Zhou GQ, Tang LL, Ma J, Lin AH, Qi ZY, Sun Y: Influence of Cervical Node Necrosis of Different Grades on the Prognosis of Nasopharyngeal Carcinoma Patients Treated with Intensity-Modulated Radiotherapy. J Cancer. 2017;8(6):959-66. Bao D, Liu Z, Geng Y, Li L, Xu H, Zhang Y, Hu L, Zhao X, Zhao Y, Luo D: Baseline MRI-based radiomics model assisted predicting disease progression in nasopharyngeal carcinoma patients with complete response after treatment. Cancer Imaging. 2022;22(1):10. Duan W, Xiong B, Tian T, Zou X, He Z, Zhang L: Radiomics in Nasopharyngeal Carcinoma. Clin Med Insights Oncol. 2022;16:11795549221079186. Liu J, Mao Y, Li Z, Zhang D, Zhang Z, Hao S, Li B: Use of texture analysis based on contrast-enhanced MRI to predict treatment response to chemoradiotherapy in nasopharyngeal carcinoma. J Magn Reson Imaging. 2016;44(2):445-55. Zhang B, Ouyang F, Gu D, Dong Y, Zhang L, Mo X, Huang W, Zhang S: Advanced nasopharyngeal carcinoma: pre-treatment prediction of progression based on multi-parametric MRI radiomics. Oncotarget. 2017;8(42):72457-65. Lan M, Huang Y, Chen CY, Han F, Wu SX, Tian L, Zheng L, Lu TX: Prognostic Value of Cervical Nodal Necrosis in Nasopharyngeal Carcinoma: Analysis of 1800 Patients with Positive Cervical Nodal Metastasis at MR Imaging. Radiology. 2015;276(2):536-44. Maxwell JH, Ferris RL, Gooding W, Cunningham D, Mehta V, Kim S, Myers EN, Johnson J, Chiosea S: Extracapsular spread in head and neck carcinoma: impact of site and human papillomavirus status. Cancer. 2013;119(18):3302-8. Mani S, Chen Y, Li X, Arlinghaus L, Chakravarthy AB, Abramson V, Bhave SR, Levy MA, Xu H, Yankeelov TE: Machine learning for predicting the response of breast cancer to neoadjuvant chemotherapy. J Am Med Inform Assoc. 2013;20(4):688-95. Tables Table 1 Clinical characteristics of patients in the training and validation sets Characteristics Training set (n = 102) Validation set (n = 43) Responders (n = 56) Non-responders (n = 46) P Responders (n = 24) Non-responders (n = 19) P Age (mean ± SD) 45.4 ± 12.6 48.0 ± 13.4 0.310 47.5 ± 11.1 50.0 ± 12.4 0.492 Sex, n (%) 0.825 0.708 Female 10 (17.9) 9 (19.6) 6 (25.0) 3 (15.8) Male 46 (82.1) 37 (80.4) 18 (75.0) 16 (84.2) T-stage, n (%) 0.559 1.000 T1 6 (10.7) 2 (4.3) 1 (4.2) 0 (0.0) T2 16 (28.6) 16 (34.8) 5 (20.8) 4 (21.1) T3 19 (33.9) 13 (28.3) 13 (54.2) 11 (57.9) T4 15 (26.8) 15 (32.6) 5 (20.8) 4 (21.0) N-stage, n (%) 0.207 1.000 N1 4 (7.1) 0 (0.0) 1 (4.2) 1 (5.3) N2 30 (53.6) 28 (60.9) 13 (54.1) 11 (57.9) N3 22 (39.3) 18 (39.1) 10 (41.7) 7 (36.8) Clinical stage, n (%) 0.600 0.606 II 2 (3.6) 0 (0.0) 0 (0.0) 0 (0.0) III 20 (35.7) 16 (34.8) 12 (50.0) 11 (57.9) IV 34 (60.7) 30 (65.2) 12 (50.0) 8 (42.1) GTV-ln (Gy), median (IQR) 52.80 (46.20 - 60.00) 45.60 (36.07 - 52.35) < 0.001* 51.50 (42.07 - 54.22) 41.80 (33.00 - 52.30) 0.047* GTV-ln Lymph node gross tumor volume, IQR interquartile range. * indicates statistical significant difference Table 2 Univariate and multivariate logistic regression analyses for predictive factors of responders in the training set Variables Odds ratio 95% CI P Univariate logistic analysis GTV-ln 1.055 1.018-1.093 0.003 Intra radiomic signature 2.718 1.857-3.980 < 0.001 Peri radiomic signature 2.718 1.837-4.022 < 0.001 Combined Intra and Peri radiomic signature 2.718 1.835-4.026 < 0.001 Multivariate logistic analysis Intra radiomic nomogram GTV-ln 1.041 0.992-1.091 0.102 Radiomic signature 2.577 1.784-3.723 < 0.001 Peri radiomic nomogram GTV-ln 1.039 0.994-1.086 0.087 Radiomic signature 2.638 1.770-3.932 < 0.001 Combined Intra and Peri radiomic nomogram GTV-ln 1.061 1.004-1.121 0.036 Radiomic signature 2.935 1.867-4.613 < 0.001 GTV-ln Lymph node gross tumor volume, Intra intratumoral, Peri peritumoral Table 3 Diagnostic performance of radiomic models in the training and validation sets Models Cohorts AUC (95% CI) Sen Spe ACC Intra radiomic signature Training 0.910 (0.837, 0.958) 0.839 (0.716, 0.923) 0.891 (0.764, 0.963) 0.863 (0.780, 0.922) Validation 0.737 (0.580, 0.859) 0.625 (0.405, 0.812) 0.895 (0.668, 0.986) 0.744 (0.588, 0.864) Peri radiomic signature Training 0.887 (0.809, 0.941) 0.714 (0.577, 0.827) 0.934 (0.821, 0.986) 0.814 (0.724, 0.883) Validation 0.794 (0.643, 0.902) 0.375 (0.187, 0.594) 0.895 (0.668, 0.986) 0.605 (0.444, 0.750) Combined Intra and Peri radiomic signature Training 0.934 (0.867, 0.974) 0.946 (0.851, 0.988) 0.782 (0.636, 0.890) 0.873 (0.791, 0.930) Validation 0.774 (0.621, 0.887) 0.750 (0.532, 0.902) 0.737 (0.488, 0.908) 0.744 (0.588, 0.864) Combined Intra and Peri radiomic nomogram Training 0.941 (0.877, 0.978) 0.929 (0.827, 0.980) 0.848 (0.711, 0.936) 0.892 (0.815, 0.944) Validation 0.783 (0.631, 0.894) 0.667 (0.446, 0.843) 0.842 (0.604, 0.966) 0.744 (0.588, 0.864) AUC the area under the curve, Intra intratumoral, Peri peritumoral, CI confidence interval, Sen sensitivity, Spe specificity, ACC accuracy Table 4 Comparisons of AUC, sensitivity, and specificity between the combined Intra and Peri radiomic nomogram and the other radiomic models in the training and validation sets Comparisons AUC Sensitivity Specificity Z P c 2 P c 2 P Training set Intra radiomic signature 1.107 0.268 2.285 0.130 0.166 0.683 Peri radiomic signature 1.570 0.116 6.722 0.009 * 1.125 0.288 Combined Intra and Peri radiomic signature 0.762 0.446 0.000 0.999 1.333 0.248 Validation set Intra radiomic signature 0.639 0.522 0.000 0.999 0.000 0.999 Peri radiomic signature 0.167 0.867 4.000 0.045 * 0.000 0.999 Combined Intra and Peri radiomic signature 0.586 0.558 0.500 0.479 0.500 0.479 * Differences was significant at P < 0.05. AUC the area under the curve, Intra intratumoral, Peri peritumoral Additional Declarations No competing interests reported. Supplementary Files Additionalfile1.docx Cite Share Download PDF Status: Published Journal Publication published 30 May, 2023 Read the published version in BMC Medical Imaging → Version 1 posted Editorial decision: Major revision 17 Mar, 2023 Reviews received at journal 26 Feb, 2023 Reviewers agreed at journal 12 Feb, 2023 Reviewers agreed at journal 03 Feb, 2023 Reviewers invited by journal 03 Feb, 2023 Editor assigned by journal 02 Feb, 2023 Editor invited by journal 02 Feb, 2023 Submission checks completed at journal 02 Feb, 2023 First submitted to journal 27 Jan, 2023 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-2519551","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":172684720,"identity":"e0562932-c0be-4d3f-a9e3-bf6fc9599ca5","order_by":0,"name":"Hao Xu","email":"","orcid":"","institution":"Sichuan Cancer Hospital \u0026 Institute, University of Electronic Science and Technology of China","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Hao","middleName":"","lastName":"Xu","suffix":""},{"id":172684722,"identity":"6b8bee45-0d50-41a3-8876-0f6a67a2eeb3","order_by":1,"name":"Ai Wang","email":"","orcid":"","institution":"Sichuan Cancer Hospital \u0026 Institute, University of Electronic Science and Technology of China","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Ai","middleName":"","lastName":"Wang","suffix":""},{"id":172684724,"identity":"beea66d8-5c39-430e-bc9b-9b8f042fb9ba","order_by":2,"name":"Chi Zhang","email":"","orcid":"","institution":"Sichuan Cancer Hospital \u0026 Institute, University of Electronic Science and Technology of China","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Chi","middleName":"","lastName":"Zhang","suffix":""},{"id":172684725,"identity":"73af1ea7-0966-4e81-aa3a-e1e3084c0bd1","order_by":3,"name":"Jing Ren","email":"","orcid":"","institution":"Sichuan Cancer Hospital \u0026 Institute, University of Electronic Science and Technology of China","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jing","middleName":"","lastName":"Ren","suffix":""},{"id":172684726,"identity":"54e686ac-907d-4258-a068-c4095e588d47","order_by":4,"name":"Jieke Liu","email":"","orcid":"","institution":"Sichuan Cancer Hospital \u0026 Institute, University of Electronic Science and Technology of China","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jieke","middleName":"","lastName":"Liu","suffix":""},{"id":172684729,"identity":"902dff02-9835-42d7-84c8-5ca3ac75f0af","order_by":5,"name":"Peng Zhou","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1UlEQVRIiWNgGAWjYJACZiBOAFIHDnyoIE0LW+LBGWdI08JjfJi3hQjl5uy9h18XttXl8UvkfDjA28Agzy92AL8Wy55zadYz29iKJWfkbjgguYPBcObsBPxaDG7kmBnztvEkbrgB1GJ4hiHB4DYhLfffgLRIALXkPDiQ2EaMlhs8xo952wxAWhgOHCRGi2VPjhkzz7mExJk9zwwONpyRIOwXc/Yzxp95yuoS+9mTH3/+U2Ejzy9NyGHAKJRgZIPzJfArh2ph/sDwh7DCUTAKRsEoGMEAAKd6SZcozodgAAAAAElFTkSuQmCC","orcid":"","institution":"Sichuan Cancer Hospital \u0026 Institute, University of Electronic Science and Technology of China","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Peng","middleName":"","lastName":"Zhou","suffix":""}],"badges":[],"createdAt":"2023-01-27 08:44:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2519551/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2519551/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12880-023-01026-1","type":"published","date":"2023-05-30T21:00:40+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":32443400,"identity":"9c442e93-ed71-48fa-8f76-9b3311692d31","added_by":"auto","created_at":"2023-02-03 15:55:39","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1109839,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart shows patient selection. NPC, nasopharyngeal carcinoma; CE-T1WI, contrast-enhanced T1-weighted imaging; T2WI, T2-weighted imaging\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-2519551/v1/db3ce857c6218221e8053ec9.png"},{"id":32443401,"identity":"2bffccd4-907a-4966-b674-c8b3f2b23e46","added_by":"auto","created_at":"2023-02-03 15:55:39","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":8686877,"visible":true,"origin":"","legend":"\u003cp\u003eRepresentative slice of the MRI images and the corresponding Intra and Peri regions of interest (ROIs). The Intra ROI (red regions) drawn by a junior radiologist with the Peri ROI (yellow regions) generated by equidistant 3-dimensional dilation of the Intra regions with 2 mm. (\u003cstrong\u003ea, b\u003c/strong\u003e) axial pretreatment CE-T1WI images of a 35-year-old female. (\u003cstrong\u003ec, d\u003c/strong\u003e) axial pretreatment T2WI images of a 53-year-old male. Intra, intratumoral; Peri, peritumoral\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-2519551/v1/28840d47f84d0814960b7e5e.png"},{"id":32443398,"identity":"3a5eaff2-3398-494c-9fa6-07f062113b98","added_by":"auto","created_at":"2023-02-03 15:55:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1329735,"visible":true,"origin":"","legend":"\u003cp\u003e(\u003cstrong\u003ea\u003c/strong\u003e) The radiomic nomogram was developed by integrating the combined Intra and Peri radiomic signature and GTV-ln in the training set. The different values of each variable corresponds to a point at the top of the graph, while the sum of points of all variables corresponds to a total point. Drawing a line from the total points to the bottom line is the probability of responders. (\u003cstrong\u003eb\u003c/strong\u003e) Calibration curves of the radiomic nomogram for predicting treatment response in the training and in the validation sets, respectively. Intra, intratumoral; Peri, peritumoral; GTV-ln, Lymph node gross tumor volume\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-2519551/v1/bdc51bfff40dd50f0ff8be03.png"},{"id":32444652,"identity":"70f81f47-3005-4a87-94cd-4f84ab2b30cf","added_by":"auto","created_at":"2023-02-03 16:03:39","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1044867,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic curves of radiomic models for predicting treatment response in the training set (\u003cstrong\u003ea\u003c/strong\u003e) and validation set (\u003cstrong\u003eb\u003c/strong\u003e), respectively. Intra, intratumoral; Peri, peritumoral\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-2519551/v1/e7479c72a6ba6f127ccc55a1.png"},{"id":32443403,"identity":"7ab82194-26a2-4f18-84ee-a55a579b4e68","added_by":"auto","created_at":"2023-02-03 15:55:39","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":12855561,"visible":true,"origin":"","legend":"\u003cp\u003eRepresentative slice of T2WI and CE-T1WI of metastatic cervical lymph nodes (red arrows) before and after radiochemotherapy. (\u003cstrong\u003ea\u003c/strong\u003e) A 32-year-old male in the responder group, radiomic nomogram of combined Intra and Peri is 5.739, and the probability to be a responder is 0.996; (\u003cstrong\u003eb\u003c/strong\u003e) A 45-year-old male in the non-responder group, radiomic nomogram of combined Intra and Peri is -4.037, and the probability to be a responder is 0.017. The threshold of probability is 0.481. CE-T1WI, contrast-enhanced T1-weighted imaging; T2WI, T2-weighted imaging; Intra intratumoral; Peri, peritumoral\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-2519551/v1/9d551c1b99c31eafa528d2ff.png"},{"id":44732430,"identity":"96bcdc7b-445c-4084-812d-9a7dd8d1be3f","added_by":"auto","created_at":"2023-10-16 21:55:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2611216,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2519551/v1/a03f2993-7e46-4d5c-9e20-ff9edef94dd2.pdf"},{"id":32443402,"identity":"6448f6c3-1253-4052-ac89-c2125f8eff3a","added_by":"auto","created_at":"2023-02-03 15:55:39","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1526655,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile1.docx","url":"https://assets-eu.researchsquare.com/files/rs-2519551/v1/42320d11c60c1c0c88c898b7.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Intra- and peritumoral MRI radiomics assisted in predicting radiochemotherapy response in metastatic cervical lymph nodes of nasopharyngeal cancer","fulltext":[{"header":"Background","content":"\u003cp\u003eNasopharyngeal cancer (NPC) is the most frequent type of head and neck cancer in Southeast Asia and South China [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Approximately 86.4% of NPC patients develop lymph node metastasis at the cervical or/and retropharyngeal site before therapy according to radiologic criteria [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Intensity-modulated radiotherapy (IMRT) is the main and curative therapy for NPC because of its intrinsic anatomic restrictions and great radiosensitivity. Previous research revealed that chemotherapy could assist patients with advanced diseases [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], and concurrent chemoradiotherapy (CCRT) combined with IMRT was later designated as the major typical therapy for stage II-IVA NPC by National Comprehensive Cancer Network (NCCN) recommendations [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. A significant predictive factor in determining survival outcomes in NPC is the cancer response to radiochemotherapy [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Not all patients, however, react as well to radiochemotherapy. Predicting response prior to treatment may result in more targeted and personalized treatment, avoiding unnecessary side effects and costs. However, there are no established biomarkers, and some have been proposed as research tools. Thus, there is an urgent need to identify an effective radiochemotherapy response predictor in patients with NPC.\u003c/p\u003e \u003cp\u003eThe ability of multimodality imaging biomarkers generated from computed tomography (CT), \u003csup\u003e18\u003c/sup\u003eF-fluorodeoxyglucose positron emission tomography, or MRI-DWI to distinguish between metastatic cervical lymph node responses to therapy has been shown in NPC [\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Conventional MRI plays a vital function in the diagnosis and management of NPC and is extensively employed in the early recognition, diagnosis, staging, and assessment of therapy response. MRI images, with a high soft tissue resolution, not only contain anatomical information about the primary NPC lesion and its adjacent constructions but also reflect the intra-tumor characteristics [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Radiomics, which is extensively used in disease identification, differential diagnosis, prognosis prediction, and therapeutic response evaluation [\u003cspan additionalcitationids=\"CR13 CR14 CR15\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], is the high-throughput computer extraction of potentially unlimited quantities of quantitative imaging metrics. According to recent investigations, radiomic features derived from pretreatment MRI images or MRI-based radiomic nomograms might anticipate NPC patients' progression-free survival [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. While lacking independent validation, the pretreatment MRI signature may predict early treatment response to initiation chemotherapy in individuals with NPC [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. However, these previous investigations have primarily focused on the function of MRI-based radiomics in expecting the treatment response of primary tumors in NPC patients. Its importance in identifying the cervical lymph nodes' treatment response is yet unclear.\u003c/p\u003e \u003cp\u003eThe objective of the investigation was to establish and validate radiomic models for predicting response to radiochemotherapy of cervical metastatic lymph nodes of NPC employing Intra and Peri radiomic features obtained from MRI.\u003c/p\u003e"},{"header":"Materials And Methods","content":"\u003cp\u003e\u003cstrong\u003ePatient cohort\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe retrospective investigation received approval from our ethics committee and institutional review board, and informed consent was not required. Between October 2016 and November 2019 at our institution, 185 consecutive patients with histologically confirmed NPC treated with induction chemotherapy (IC) and CCRT were initially recruited retrospectively. The study\u0026apos;s inclusion criteria were as follows: (1) histopathologically proven NPC; (2) available pre-treatment T2-weighted imaging (T2WI) and contrast-enhanced T1-weighted imaging (CE-T1WI) images before biopsy; (3) available post-treatment T2WI and CE-T1WI images; (4) no treatment before baseline biopsy; and (5) available clinical variables such as age, sex, T-stage, N-stage, clinical stage, and lymph node gross tumor volume (GTV-ln). The following were the exclusion criteria: (1) the period between baseline MRI and initial treatment was further than 2 weeks (n = 14); (2) the existence of other malign tumors (n = 9); (3) missing clinicopathological information (n = 10); (4) poor image quality (n = 7). Finally, a total of 145 patients (mean age 47.2 \u0026plusmn; 12.6, range 13\u0026ndash;77, male = 117, female = 28, stage II = 2, stage III = 59, stage IV = 84) were enrolled and allocated to the training set (102 patients, October 2016 to December 2018) and validation set (43 patients, December 2018 to November 2019) [20]. A flowchart of the patient selection is shown in Fig. 1. Patients were staged based on the 8th version American Joint Committee on Cancer (AJCC) Tumor-Node-Metastasis (TNM) staging system [21]. From computerized medical records, demographic information and stages were extracted.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMR images acquisition protocol\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll patients underwent nasopharyngeal and cervical region MRI examination with head-neck combined coils employing either a 1.5-T MR scanner (Avanto, Siemens, Germany) or a 3.0-T MR imaging scanner (Skyra, Siemens, Germany). Overall, 94/145 (64.8%) of the subjects had imaging done in a 1.5-T MRI scanner, and 51/145 (35.2%) had imaging done in a 3.0-T scanner. To keep away from magnetic exposure and motion artifacts, patients were instructed to eliminate all metal-containing items and lie supinely in the scanner before scanning. Before the MRI examination, all patients will be required to wear earplugs and headphones to decrease noise. The DICOM format images of axial fat-suppressed CE-T1WI (contrast agent = Gd-DTPA, Magnevist, Schering, Berlin, Germany; dose = 0.1 mmol/kg body weight) and T2WI scans of each case were retrieved from the picture archiving and communication system (Carestream, Ontario, Canada). Additional file 1 provided detailed image acquisition parameters (Additional file 1: S1). The inverse fourier transform with the linear filling was used to reconstruct all images from k-space.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEvaluation of lymph nodes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMultiple radiologic criteria were used to determine whether metastatic cervical lymph nodes were involved. These included: (1) regions of central necrosis/cystic necrosis (T2WI with a focal high signal intensity or CE-T1WI with low signal intensity/ with or without an adjacent border of enhancement), (2) extracapsular distribution in any size lymph node, including ambiguous nodal boundaries, irregular nodal capsular improvement, and infiltration into neighboring muscle or fat, (3) the shortest diameter of the cervical or medial retropharyngeal lymph node is 10 mm; the lateral retropharyngeal lymph node is 5 mm.\u003c/p\u003e\n\u003cp\u003eIn each patient, we chose the largest lymph node within the scanning range as the target lesion. The target lymph nodes were then assessed on T2WI for maximum axial diameter and minimum axial diameter long axis. The minimum axial diameter matched the node\u0026apos;s largest diameter in the axial plane perpendicular to its maximum axial diameter. The Response Evaluation Criteria in Solid Tumors 1.1 (RECIST) was used to assess the treatment response [22]. The responders were defined as a reduction of at least 30% in the target lesions\u0026apos; maximum axial diameter. In contrast, non-responders had inadequate shrinking to qualify for the responders.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePreprocessing of images\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll raw\u0026nbsp;images\u0026nbsp;were preprocessed utilizing the Artificial Intelligence Kit program (A.K., version 3.2.0, GE Healthcare, China) prior to segmentation because of the variability of CE-T1WI and T2WI acquisition parameters. Using a trilinear interpolation algorithm, all raw images were initially resliced to isotropic 1 mm\u003csup\u003e3\u003c/sup\u003e voxels. The center was employed to align the interpolation grid\u0026apos;s location, and its measurements were rounded to the closest integer. Following the use of a Gaussian filter for anti-aliasing during down-sampling, the bias-field correction was also performed. Lastly, voxel intensity values were standardized utilizing a z-score normalization approach to remove the impact of various gray value ranges.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImage segmentation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe target lymph nodes on the axial T2WI and CE-T1WI images separately served as the basis for the region of interest (ROIs). A junior radiologist with 7 years of experience in head and neck imaging manually segmented cancer on the ITK-SNAP (version 3.8.0, http://www.itksnap.org), which was then examined by a senior radiologist with 20 years of experience in head and neck imaging. The ROIs were created for every lesion to completely include the target lymph node on every succeeding slice. Then, utilizing A.K. software, equidistant 3-dimensional dilation of the Intra sections by 2 mm was used to create the Peri area. To measure interobserver variability in manual segmentation, the senior radiologist re-delineated 30 target lymph nodes that were randomly chosen from the T2WI and CE-T1WI images, respectively. Two experienced radiologists were unaware of the target lymph nodes\u0026apos; reaction to the treatment. The repeatability of radiomic features was then evaluated using the intraclass correlation coefficient (ICC), and those with an ICC value greater than 0.75 were selected for the subsequent process. The typical MRI image and its related Intra and Peri ROIs are displayed in Fig. 2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFeature extraction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing image preprocessing and segmentation, features from the Intra and Peri ROIs were extracted using an open-source Python program (PyRadiomics, version 3.0, https://pyradiomics.readthedocs.io) [23] that followed the image biomarker standardisation initiative (IBSI) standard [24]. These features belonged to three categories: shape, first order, and texture features. Image intensity was discretized using a fixed bin width of 25 before texture features were estimated. For every patient, 2704 radiomic features (1352 from T2WI and CE-T1WI, respectively) were lastly extracted in the Intra and Peri ROI, respectively (see Additional file 1: Table S1 and Table S2 for a detailed list of the extracted features).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFeature selection and radiomic signature building\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll radiomic features were first normalized employing z-scores and then selected in the training set for the construction of radiomic signature (Intra, Peri, and combined Intra and Peri). For feature selection, we performed a five-step procedure. We used the Mann-Whitney U test to make the initial selection from the training set. For the remaining significant features, the P-value threshold was set at 0.05. Then, to eliminate redundant features, Spearman correlation assessment and maximum relevance-minimum redundancy (mRMR) were performed successively. Features with Spearman correlation coefficient values greater than 0.9 were deemed redundant and were thus eliminated. Simultaneously, the higher-ranked 15 features associated with the reaction of metastatic lymph nodes to therapy were retained following mRMR. Thirdly, the least absolute shrinkage and selection operator (LASSO) technique with cross-validations was utilized to detect the most predictive features related to the reaction of metastatic lymph nodes to therapy [25]. Lastly, radiomic signature building depended on multivariate logistic regression with backward stepwise selection utilizing the likelihood ratio examination with Akaike\u0026apos;s data criterion as the stopping rule [26]. The radiomic signatures were computed utilizing a linear combination of the chosen features weighted by their corresponding coefficients. To examine possible multicollinearity between radiomic signature features, Spearman\u0026apos;s correlations were also employed. For feature selection and radiomic signature building, R software (version 4.0.0, http://www.r-project.org) was employed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConstruction and validation of the radiomics-based model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUnivariate logistic regression analysis was utilized to recognize independent risk factors for differentiating responders from non-responders between radiomic signatures and\u0026nbsp;clinical\u0026nbsp;variables. Subsequently, to provide a more individualized predictive model, a radiomic nomogram combining the radiomic signature and significant independent risk factors was built in the training set, using the multivariable logistic regression analysis.\u003c/p\u003e\n\u003cp\u003eFor both the training and validation sets, a receiver operating characteristic curve (ROC) and area under the curve (AUC) calculations were made to assess the radiomic-based model\u0026apos;s ability to discriminate. The threshold was established utilizing the maximum Youden index (sensitivity + specificity \u0026minus; 1) on the training set and this same threshold was used for the validation set. Calculations were also made for the relevant sensitivity, specificity, and accuracy. The above diagnostic measures\u0026apos; 95% confidence intervals (CIs) were computed utilizing the binomial exact approach. The Hosmer-Lemeshow goodness-of-fit analysis was employed to measure the calibration of the prediction model. Building calibration plots will also compute the calibration slope and calibration-in-the-large statistic (intercept)\u0026nbsp;[27].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTRIPOD statement and statistical analysis\u0026nbsp;\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe presented the results of this article based on the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) reporting checklist since the study was intended to serve as a prediction model building for diagnostic reasons [20]. SPSS (version 26.0, available at https://www.ibm.com) and MedCalc (version 18.2.1, available at https://www.medcalc.org) were utilized for all statistical analyses. The Shapiro-Wilk test was used to examine if quantitative data had a normal distribution. The clinical and demographic variables of the responders and non-responders were compared using an independent samples t-test, the Mann-Whitney U test, Fisher\u0026apos;s exact analysis, or the chi-squared (\u0026chi;2) test, as applicable. The AUCs of several prediction models in the training and validation sets were compared employing the Delong method [28]. Using McNemar\u0026apos;s chi-square test, further comparisons of sensitivity and specificity in the training and validation sets were made [29]. Statistical significance was identified as a P-value\u0026lt;\u0026thinsp;0.05 (two-tailed).\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eBaseline characteristics of the patient\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe clinical characteristics and sociodemographics of 145 patients enrolled in this investigation are listed in Table 1. The numbers of responders and non-responders in the training and validation sets were 56 and 46, and 24 and 19, respectively. There were no substantial variations in age, sex, T- stage, N-stage, and\u0026nbsp;clinical\u0026nbsp;stage between responders and non-responders in the training and validation sets (all \u003cem\u003eP\u003c/em\u003e \u0026gt; 0.05). There was a substantial variation in the GTV-ln between responders and non-responders in the two sets (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001 and = 0.047).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFeature selection and radiomic signature development\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e5408 radiomic features were obtained (2704 in the Intra region and 2704 in the Peri region). After calculating the ICCs, the numbers of reproducible features (ICCs \u0026gt; 0.75) from the CE-T1WI and T2WI images in the Intra region and Peri region were 940 and 1076, and 971 and 975, respectively. Then, five feature selection methods like the Mann-Whitney U test, Spearman correlation assessment, mRMR procedure, LASSO, and multivariate logistic regression were utilized to obtain the radiomic features with the most strongly associated with responders. Finally, we selected 4, 4, and 5 features for Intra, Peri, and combined Intra and Peri radiomic signatures construction (Additional file\u0026nbsp;1: Table S3 and Table S4). A detailed description of the procedure for feature selection and LASSO results were provided in Additional file 1 (Additional file\u0026nbsp;1: Fig. S1 and Fig. S2). We also included the mean and standard deviation of these selected features in Additional file 1 (Additional file\u0026nbsp;1: Table S5) for z-score normalization prior to estimation.\u003c/p\u003e\n\u003cp\u003eThe radiomic signature of Intra was constructed using the following formula: ln (P/1-P) = -0.8765 \u0026ndash; 0.8685\u0026nbsp;\u0026times;\u0026nbsp;CE-T1WI_firstorder_wavelet.HHH.Skewness + 1.2215\u0026nbsp;\u0026times;\u0026nbsp;CE-T1WI_GLDM_wavelet.HHL.LDLGLE + 2.8194\u0026nbsp;\u0026times;\u0026nbsp;T2WI_GLCM_LoG.sigma.1.5.mm.3D.ClusterShade\u0026nbsp;\u0026ndash; 2.1866\u0026nbsp;\u0026times;\u0026nbsp;T2WI_GLCM_wavelet.LLH.MaximumProbability,\u0026nbsp;where P was the possibility of responders (threshold \u0026gt; 0.6406).\u003c/p\u003e\n\u003cp\u003eThe radiomic signature of Peri was constructed using the following formula: ln (P/1-P) = -0.0330 \u0026ndash; 1.3038\u0026nbsp;\u0026times;\u0026nbsp;CE-T1WI_firstorder_LoG.sigma.0.5.mm.3D.Minimum + 1.8829\u0026nbsp;\u0026times;\u0026nbsp;CE-T1WI_GLRLM_wavelet.HHH.SRHGLE + 1.1131\u0026nbsp;\u0026times;\u0026nbsp;T2WI_GLDM_wavelet.LHL.SDHGLE\u0026nbsp;\u0026ndash; 1.3508\u0026nbsp;\u0026times;\u0026nbsp;T2WI_GLCM_wavelet.LLH.ClusterTendency,\u0026nbsp;where P was the possibility of responders (threshold \u0026gt; 0.6899).\u003c/p\u003e\n\u003cp\u003eThe radiomic signature of combined Intra and Peri was constructed using the following formula: ln (P/1-P) = -0.0335 \u0026ndash; 1.0301\u0026nbsp;\u0026times;\u0026nbsp;Intra_CE-T1WI_firstorder_LoG.sigma.0.5.mm.3D.90Percentile\u0026nbsp;\u0026ndash;\u0026nbsp;3.0175\u0026nbsp;\u0026times;\u0026nbsp;Intra_T2WI_GLCM_wavelet.LLH.MaximumProbability\u0026nbsp;\u0026ndash;\u0026nbsp;1.8770\u0026nbsp;\u0026times;\u0026nbsp;Peri_CE-T1WI_firstorder_LoG.sigma.0.5.mm.3D.Minimum +\u0026nbsp;1.3759\u0026nbsp;\u0026times;\u0026nbsp;Peri_CE-T1WI_GLRLM_wavelet.HHH. SRHGLE +\u0026nbsp;1.4714\u0026nbsp;\u0026times;\u0026nbsp;Peri_T2WI_GLRLM_wavelet.HLL.LRLGLE,\u0026nbsp;where P was the possibility of responders (threshold \u0026gt; 0.3525).\u003c/p\u003e\n\u003cp\u003eIn Additional file 1 (Additional file 1: Fig. S3), the distribution of these radiomic signatures of responders and non-responders in the training and validation sets was displayed. A heatmap was utilized to show the Spearman correlations between specific features in these radiomic signatures (Additional file 1: Fig. S4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConstruction, validation, and performance of radiomic models\u003c/strong\u003e\u003c/p\u003e\n\u003cp skip=\"true\"\u003eGTV-ln was shown to be an important independent factor for distinguishing responders from non-responders in a univariate logistic regression assessment (\u003cem\u003eP\u003c/em\u003e = 0.003), however, a subsequent multivariate logistic regression examination revealed that it was only substantial in the combined Intra and Peri dataset (\u003cem\u003eP\u003c/em\u003e = 0.036) (Table 2). So, the combined Intra and Peri radiomic nomogram was generated by incorporating the combined Intra and Peri radiomic signature and GTV-ln (Fig. 3a). The formula for the combined Intra and Peri radiomic nomogram calculation: ln (P/1-P) = -2.8790 + 0.0590 \u0026times; GTV-ln + 1.0770 \u0026times; combined Intra and Peri radiomic signature, where P is the possibility of responders (threshold \u0026gt; 0.4815).\u003c/p\u003e\n\u003cp skip=\"true\"\u003eLastly, four radiomic models (combined Intra and Peri radiomic nomogram, Intra radiomic signature, Peri radiomic signature, and combined Intra and Peri radiomic signature) for the prediction of responders were constructed. The combined Intra and Peri radiomic nomogram illustrated worthy performance in discriminating the responders and non-responders in patients with NPC, which yielded an AUC of 0.941 (95% confidence interval [CI]: 0.877-0.978), Sen of 92.9%, Spe of 84.8%, and ACC of 89.2% in the training set. The corresponding AUC, Sen, Spe, and ACC for the combined Intra and Peri radiomic nomogram in the validation set were 0.783 (95% CI: 0.631-0.894), 66.7%, 84.2%, and 74.4%, respectively (Table 3). Fig. 4. shows the radiomic models\u0026apos; ROC curves in both cohorts. Table 3 presents the AUC, Sen, Spe, and ACC of each radiomic model in both cohorts. Representative CE-T1WI and T2WI images of NPC patients as responders and non-responders were shown in Fig. 5.\u003c/p\u003e\n\u003cp skip=\"true\"\u003eIn training and validation sets, the combined Intra and Peri radiomic nomogram calibration curve exhibited excellent agreement between the predictive and observational probability of differentiating between responders and non-responders (Fig. 3b), and the Hosmer-Lemeshow assessment outcomes were non-significant (\u003cem\u003eP\u003c/em\u003e = 0.618 and 0.780). The Hosmer-Lemeshow test also yielded non-significant results in training and validation sets of the Intra radiomic signature (\u003cem\u003eP \u003c/em\u003e= 0.405 and 0.165), Peri radiomic signature (\u003cem\u003eP \u003c/em\u003e= 0.300 and 0.512), and combined Intra and Peri radiomic signature (\u003cem\u003eP \u003c/em\u003e= 0.610 and 0.158), which revealed no departure from the ideal fit (Additional file 1: Fig. S5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRadiomic models comparison and TRIPOD\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDepended on the DeLong analysis, there were no substantial variations in AUCs across the radiomic models in the training and validation sets (all \u003cem\u003eP\u003c/em\u003e \u0026gt; 0.05) (Table 4). The combined Intra and Peri radiomic nomogram\u0026apos;s sensitivity was greater than the Peri radiomic signature in training and validation sets, according to the McNemar analysis (\u003cem\u003eP\u003c/em\u003e = 0.009 and 0.045) (Table 4). However, no statistically significant differences in specificity between the radiomic models in the training and validation sets were found (all \u003cem\u003eP\u003c/em\u003e \u0026gt; 0.05) (Table 4).\u003c/p\u003e\n\u003cp\u003eAccording to the guidelines of the TRIPOD statement, the type of this investigation belongs to Type 2b. A full list of TRIPOD was provided as Additional file 1 (Additional file 1: Table S6).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this retrospective investigation, we developed and validated MRI-based radiomic models for predicting response to radiochemotherapy of cervical metastatic lymph nodes of NPC. These models included Intra, Peri, and a combination of Intra and Peri radiomic signatures all demonstrated satisfactory predictive value. Additionally, we demonstrated and verified a radiomic nomogram that incorporated the combined Intra and Peri radiomic signature with GTV-ln, which had an acceptable discrimination capability.\u003c/p\u003e \u003cp\u003eRecent studies found that the response to radiochemotherapy was associated with clinical results in NPC. Peng et al [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. discovered that the tumor response to radiochemotherapy was recognized as an independent prognostic variable of 4-year disease-free, overall, and locoregional relapse-free survival. In another investigation, Liu et al [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. revealed that poor tumor response to radiochemotherapy could be a predictor of advanced-stage NPC. NPC is highly susceptible to regional lymph node metastasis [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Additionally, earlier findings indicate that the characteristics of lymph node metastasis were indicators of distant metastasis and had an effect on prognosis for overall, local recurrence, regional relapse, and disease-free survival [\u003cspan additionalcitationids=\"CR33\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Thus, it is essential to enhance the treatment impact of metastatic lymph nodes in order to improve the prognosis of NPC. Our findings may provide an effective treatment approach for metastatic lymph nodes prior to therapy. For patients with NPC, due to different levels of sensitivity to the treatment, patients at the same stage may have different treatment responses. Additionally, successful prevention of therapeutic side effects and disease control, however, requires careful consideration of radiotherapy dose and chemotherapy regimen. Using our predictive model, physicians could preliminarily predict the treatment impact on patients and take preventative measures in advance. Our findings suggest that elevating GTV-ln suitably during radiological treatment planning could be an effective alternative, particularly for these potentially poor responders. Moreover, at this time, radiation and chemotherapy are the main treatments for locoregionally progressed NPC [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Early detection of non-responders using our prediction model might assist to decrease related toxicities and costs by preventing excessively extended therapies and unsuccessful chemotherapy regimens.\u003c/p\u003e \u003cp\u003eMRI radiomics has been frequently utilized to assess treatment response in NPC [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Wang et al [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. discovered that a multisequence pretreatment MRI radiomics signature might predict early IC response. However, there was no additional validation cohort in this study. Liu et al [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. confirmed that MRI radiomics has a great potential for predicting chemoradiotherapy response, but only 53 NPC patients were included in this study. In another study, Zhang et al [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. demonstrate that radiomic features derived from T2WI and CE-T1WI data enable the prediction progression of NPC with a high accuracy. In the study, the constructed MRI-based radiomic models with peritumoral features may serve as a noninvasive and repeatable diagnostic tool for clinical application in predicting radiochemotherapy response in metastatic cervical lymph nodes of NPC. But further research is necessary before these findings can be applied clinically, because the sensitivity of the nomogram was 0.667 in the validation cohort.\u003c/p\u003e \u003cp\u003eCervical lymph nodes can be evaluated by MRI based on their size and morphologic aspect including central necrosis and extracapsular spread. These nodal characteristics, which were considered in our study, have been demonstrated to be important prognostic factors in NPC patients [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. However, these morphologic findings cannot provide sufficient information for the accurate prediction of metastatic cervical lymph nodes' therapeutic reaction. Consequently, utilizing the Intra and Peri radiomic features may aid in predicting treatment response. Our investigation showed that the combined Intra and Peri radiomic nomogram achieved satisfactory predictive capability with AUCs of 0.941 and 0.783 in training and validation sets, respectively. Additionally, the McNemar test revealed that in training and validation sets, the combined Intra and Peri radiomic nomogram's sensitivity was greater than that of the Peri radiomic signature (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.009 and 0.045).\u003c/p\u003e \u003cp\u003eIn the combined Intra and Peri radiomic signature, the firstorder_90Percentile from the Intra area and firstorder_Minimum from the Peri area were negatively connected with responders, exhibiting that the responders had a low distribution of gray level intensities within the ROI of images than that of non-responders. The significant gray level co-occurrence matrix feature included MaximumProbability from the Intra area. The non-responders had a higher Maximum Probability, indicating greater incidences of the most main pair of adjacent gray values in non-responders than in responders. Besides, the remaining two texture features (short-run elevated gray level emphasis, long-run reduced gray level emphasis) of gray level run length matrix from the Peri area were also chosen to create the combined Intra and Peri radiomic signature. Our findings indicated that, compared to non-responders, the responders showed a larger joint distribution of shorter run lengths with greater gray-level values and long run lengths with lower gray-level values inside the Peri area.\u003c/p\u003e \u003cp\u003eOur investigation has many additional restrictions. First, the heterogeneity in acquisition parameters of two different magnetic field strengths scanners may have affected the image texture. Thus, additional studies may be required to quantitatively investigate these effects. Secondly, other MRI method, especially DCE-MRI-based radiomic, has been employed in estimating treatment response for breast malignancy and their function in NPC requirements is to be further explored [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. Thirdly, the single-center nature and small sample size of our study limit the generalizability of our models. Hence, a multi-center study with different imaging equipment and a large sample size in the future is required.\u003c/p\u003e \u003cp\u003eIn summary, the present investigation developed the MRI-based radiomic model that depends on combined Intra and Peri features to anticipate the radiochemotherapy response of metastatic cervical lymph nodes prior to therapy in NPC. This radiomic model may be useful in NPC patients for individualized risk stratification treatment and therapeutic decision-making.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAUC: Area under the curve; AJCC: American joint committee on cancer; CI: Confdence internal; CCRT: Concurrent chemoradiotherapy; CT: Computed tomography; CE-T1WI: Contrast-enhanced T1-weighted imaging; GTV-ln: Lymph node gross tumor volume; Intratumoral: Intra; IC: Induction chemotherapy; IMRT: Intensity-modulated radiotherapy; ICC: Intraclass correlation coefficient; IBSI: Image biomarker standardisation initiative; LASSO: Least absolute shrinkage and selection operator; mRMR: Maximum relevance-minimum redundancy; NPC: Nasopharyngeal cancer; NCCN: National comprehensive cancer network; Peritumoral: Peri; RECIST: Response evaluation criteria in solid tumors; ROI: Region of interest; ROC: Receiver operating characteristic curve; T2WI: T2-weighted imaging; TNM: Tumor-node-metastasis; TRIPOD: Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Hao Xu, Ai Wang, Chi Zhang, and Jing Ren. Hao Xu, Ai Wang, Jieke Liu, and Peng Zhou performed data interpretation and statistical analysis. The first draft of the manuscript was written by Hao Xu and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Sichuan Science and Technology Program (grant number 2021YFG0125).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe research was approved and the requirement of informed consent from the patients was waived by Sichuan Cancer Hospital Ethic Committee because of the retrospective design of this study, and patients\u0026rsquo; information was protected. The study was performed in accordance with the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone of the authors declare conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eTorre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A: Global cancer statistics, 2012. CA Cancer J Clin. 2015;65(2):87-108.\u003c/li\u003e\n\u003cli\u003eWei KR, Zheng RS, Zhang SW, Liang ZH, Li ZM, Chen WQ: Nasopharyngeal carcinoma incidence and mortality in China, 2013. Chin J Cancer. 2017;36(1):90.\u003c/li\u003e\n\u003cli\u003eChen YP, Chan ATC, Le QT, Blanchard P, Sun Y, Ma J: Nasopharyngeal carcinoma. Lancet. 2019;394(10192):64-80.\u003c/li\u003e\n\u003cli\u003eWang X, Hu C, Ying H, He X, Zhu G, Kong L, Ding J: Patterns of lymph node metastasis from nasopharyngeal carcinoma based on the 2013 updated consensus guidelines for neck node levels. Radiother Oncol. 2015;115(1):41-45.\u003c/li\u003e\n\u003cli\u003eChen QY, Wen YF, Guo L, Liu H, Huang PY, Mo HY, Li NW, Xiang YQ, Luo DH, Qiu F et al: Concurrent chemoradiotherapy vs radiotherapy alone in stage II nasopharyngeal carcinoma: phase III randomized trial. J Natl Cancer Inst. 2011;103(23):1761-70.\u003c/li\u003e\n\u003cli\u003eMiao J, Wang L, Tan SH, Li JG, Yi J, Ong EHW, Tan LLY, Zhang Y, Gong X, Chen Q et al: Adjuvant Capecitabine Following Concurrent Chemoradiotherapy in Locoregionally Advanced Nasopharyngeal Carcinoma: A Randomized Clinical Trial. JAMA Oncol. 2022.\u003c/li\u003e\n\u003cli\u003eColevas AD, Yom SS, Pfister DG, Spencer S, Adelstein D, Adkins D, Brizel DM, Burtness B, Busse PM, Caudell JJ et al: NCCN Guidelines Insights: Head and Neck Cancers, Version 1.2018. J Natl Compr Canc Netw. 2018;16(5):479-90.\u003c/li\u003e\n\u003cli\u003ePeng H, Chen L, Zhang Y, Li WF, Mao YP, Liu X, Zhang F, Guo R, Liu LZ, Tian L et al: The Tumour Response to Induction Chemotherapy has Prognostic Value for Long-Term Survival Outcomes after Intensity-Modulated Radiation Therapy in Nasopharyngeal Carcinoma. Sci Rep. 2016;6:24835.\u003c/li\u003e\n\u003cli\u003eZhang GY, Wang YJ, Liu JP, Zhou XH, Xu ZF, Chen XP, Xu T, Wei WH, Zhang Y, Huang Y: Pretreatment Diffusion-Weighted MRI Can Predict the Response to Neoadjuvant Chemotherapy in Patients with Nasopharyngeal Carcinoma. Biomed Res Int. 2015;2015:307943.\u003c/li\u003e\n\u003cli\u003eYen RF, Chen TH, Ting LL, Tzen KY, Pan MH, Hong RL: Early restaging whole-body (18)F-FDG PET during induction chemotherapy predicts clinical outcome in patients with locoregionally advanced nasopharyngeal carcinoma. Eur J Nucl Med Mol Imaging. 2005;32(10):1152-59.\u003c/li\u003e\n\u003cli\u003eYang Y, Wang M, Qiu K, Wang Y, Ma X: Computed tomography-based deep-learning prediction of induction chemotherapy treatment response in locally advanced nasopharyngeal carcinoma. Strahlenther Onkol. 2022;198(2):183-93.\u003c/li\u003e\n\u003cli\u003eLambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, Sanduleanu S, Larue R, Even AJG, Jochems A et al: Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749-62.\u003c/li\u003e\n\u003cli\u003eLambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, van Stiphout RG, Granton P, Zegers CM, Gillies R, Boellard R, Dekker A et al: Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer. 2012;48(4):441-46.\u003c/li\u003e\n\u003cli\u003eRogers W, Thulasi Seetha S, Refaee TAG, Lieverse RIY, Granzier RWY, Ibrahim A, Keek SA, Sanduleanu S, Primakov SP, Beuque MPL et al: Radiomics: from qualitative to quantitative imaging. Br J Radiol. 2020;93(1108):20190948.\u003c/li\u003e\n\u003cli\u003eGillies RJ, Kinahan PE, Hricak H: Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016;278(2):563-77.\u003c/li\u003e\n\u003cli\u003eVerma V, Simone CB, 2nd, Krishnan S, Lin SH, Yang J, Hahn SM: The Rise of Radiomics and Implications for Oncologic Management. J Natl Cancer Inst. 2017;109(7).\u003c/li\u003e\n\u003cli\u003eZhao L, Gong J, Xi Y, Xu M, Li C, Kang X, Yin Y, Qin W, Yin H, Shi M: MRI-based radiomics nomogram may predict the response to induction chemotherapy and survival in locally advanced nasopharyngeal carcinoma. Eur Radiol. 2020;30(1):537-46.\u003c/li\u003e\n\u003cli\u003eMao J, Fang J, Duan X, Yang Z, Cao M, Zhang F, Lu L, Zhang X, Wu X, Ding Y et al: Predictive value of pretreatment MRI texture analysis in patients with primary nasopharyngeal carcinoma. Eur Radiol. 2019;29(8):4105-13.\u003c/li\u003e\n\u003cli\u003eWang G, He L, Yuan C, Huang Y, Liu Z, Liang C: Pretreatment MR imaging radiomics signatures for response prediction to induction chemotherapy in patients with nasopharyngeal carcinoma. Eur J Radiol. 2018;98:100-06.\u003c/li\u003e\n\u003cli\u003eMoons KG, Altman DG, Reitsma JB, Ioannidis JP, Macaskill P, Steyerberg EW, Vickers AJ, Ransohoff DF, Collins GS: Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med. 2015; 162(1):W1-73.\u003c/li\u003e\n\u003cli\u003eTang LL, Chen YP, Mao YP, Wang ZX, Guo R, Chen L, Tian L, Lin AH, Li L, Sun Y et al: Validation of the 8th Edition of the UICC/AJCC Staging System for Nasopharyngeal Carcinoma From Endemic Areas in the Intensity-Modulated Radiotherapy Era. J Natl Compr Canc Netw. 2017;15(7):913-19.\u003c/li\u003e\n\u003cli\u003eTherasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS, Rubinstein L, Verweij J, Van Glabbeke M, van Oosterom AT, Christian MC et al: New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst. 2000;92(3):205-16.\u003c/li\u003e\n\u003cli\u003evan Griethuysen JJM, Fedorov A, Parmar C, Hosny A, Aucoin N, Narayan V, Beets-Tan RGH, Fillion-Robin JC, Pieper S, Aerts H: Computational Radiomics System to Decode the Radiographic Phenotype. Cancer Res. 2017;77(21):e104-e07.\u003c/li\u003e\n\u003cli\u003eZwanenburg A, Vallieres M, Abdalah MA, Aerts H, Andrearczyk V, Apte A, Ashrafinia S, Bakas S, Beukinga RJ, Boellaard R et al: The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping. Radiology. 2020;295(2):328-38.\u003c/li\u003e\n\u003cli\u003eSauerbrei W, Royston P, Binder H: Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Stat Med. 2007;26(30):5512-28.\u003c/li\u003e\n\u003cli\u003ePan W: Akaike\u0026apos;s information criterion in generalized estimating equations. Biometrics. 2001;57(1):120-25.\u003c/li\u003e\n\u003cli\u003eSteyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, Pencina MJ, Kattan MW: Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21(1):128-38.\u003c/li\u003e\n\u003cli\u003eDeLong ER, DeLong DM, Clarke-Pearson DL: Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837-45.\u003c/li\u003e\n\u003cli\u003eHawass NE: Comparing the sensitivities and specificities of two diagnostic procedures performed on the same group of patients. Br J Radiol. 1997;70(832):360-66.\u003c/li\u003e\n\u003cli\u003eLiu LT, Tang LQ, Chen QY, Zhang L, Guo SS, Guo L, Mo HY, Zhao C, Guo X, Cao KJ et al: The Prognostic Value of Plasma Epstein-Barr Viral DNA and Tumor Response to Neoadjuvant Chemotherapy in Advanced-Stage Nasopharyngeal Carcinoma. Int J Radiat Oncol Biol Phys. 2015;93(4):862-69.\u003c/li\u003e\n\u003cli\u003eSham JS, Choy D, Wei WI: Nasopharyngeal carcinoma: orderly neck node spread. Int J Radiat Oncol Biol Phys. 1990;19(4):929-33.\u003c/li\u003e\n\u003cli\u003eFeng Y, Cao C, Hu Q, Chen X: Prognostic Value and Staging Classification of Lymph Nodal Necrosis in Nasopharyngeal Carcinoma after Intensity-Modulated Radiotherapy. Cancer Res Treat. 2019;51(3):1222-30.\u003c/li\u003e\n\u003cli\u003eXu Y, Chen X, Zhang M, Xiao Y, Zong J, Guo Q, Qiu S, Zheng W, Lin S, Pan J: Prognostic effect of parotid area lymph node metastases after preliminary diagnosis of nasopharyngeal carcinoma: a propensity score matching study. Cancer Med. 2017;6(10):2213-21.\u003c/li\u003e\n\u003cli\u003eZhang LL, Li JX, Zhou GQ, Tang LL, Ma J, Lin AH, Qi ZY, Sun Y: Influence of Cervical Node Necrosis of Different Grades on the Prognosis of Nasopharyngeal Carcinoma Patients Treated with Intensity-Modulated Radiotherapy. J Cancer. 2017;8(6):959-66.\u003c/li\u003e\n\u003cli\u003eBao D, Liu Z, Geng Y, Li L, Xu H, Zhang Y, Hu L, Zhao X, Zhao Y, Luo D: Baseline MRI-based radiomics model assisted predicting disease progression in nasopharyngeal carcinoma patients with complete response after treatment. Cancer Imaging. 2022;22(1):10.\u003c/li\u003e\n\u003cli\u003eDuan W, Xiong B, Tian T, Zou X, He Z, Zhang L: Radiomics in Nasopharyngeal Carcinoma. Clin Med Insights Oncol. 2022;16:11795549221079186.\u003c/li\u003e\n\u003cli\u003eLiu J, Mao Y, Li Z, Zhang D, Zhang Z, Hao S, Li B: Use of texture analysis based on contrast-enhanced MRI to predict treatment response to chemoradiotherapy in nasopharyngeal carcinoma. J Magn Reson Imaging. 2016;44(2):445-55.\u003c/li\u003e\n\u003cli\u003eZhang B, Ouyang F, Gu D, Dong Y, Zhang L, Mo X, Huang W, Zhang S: Advanced nasopharyngeal carcinoma: pre-treatment prediction of progression based on multi-parametric MRI radiomics. Oncotarget. 2017;8(42):72457-65.\u003c/li\u003e\n\u003cli\u003eLan M, Huang Y, Chen CY, Han F, Wu SX, Tian L, Zheng L, Lu TX: Prognostic Value of Cervical Nodal Necrosis in Nasopharyngeal Carcinoma: Analysis of 1800 Patients with Positive Cervical Nodal Metastasis at MR Imaging. Radiology. 2015;276(2):536-44.\u003c/li\u003e\n\u003cli\u003eMaxwell JH, Ferris RL, Gooding W, Cunningham D, Mehta V, Kim S, Myers EN, Johnson J, Chiosea S: Extracapsular spread in head and neck carcinoma: impact of site and human papillomavirus status. Cancer. 2013;119(18):3302-8.\u003c/li\u003e\n\u003cli\u003eMani S, Chen Y, Li X, Arlinghaus L, Chakravarthy AB, Abramson V, Bhave SR, Levy MA, Xu H, Yankeelov TE: Machine learning for predicting the response of breast cancer to neoadjuvant chemotherapy. J Am Med Inform Assoc. 2013;20(4):688-95.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1\u0026nbsp;\u003c/strong\u003eClinical characteristics of patients in the training and validation sets\u003c/p\u003e\n\u003ctable align=\"\" border=\"1\" cellpadding=\"0\" cellspacing=\"0\" width=\"699\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"16.165951359084406%\"\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" width=\"41.9170243204578%\"\u003e\n \u003cp\u003e\u003cstrong\u003eTraining set\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n = 102)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" width=\"41.9170243204578%\"\u003e\n \u003cp\u003e\u003cstrong\u003eValidation set\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n = 43)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.747440273037544%\"\u003e\n \u003cp\u003e\u003cstrong\u003eResponders\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n = 56)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.98976109215017%\"\u003e\n \u003cp\u003e\u003cstrong\u003eNon-responders\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n = 46)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.262798634812286%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.747440273037544%\"\u003e\n \u003cp\u003e\u003cstrong\u003eResponders\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n = 24)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.525597269624573%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eNon-responders\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n = 19)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.726962457337883%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eAge (mean \u0026plusmn; SD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e45.4\u0026nbsp;\u0026plusmn; 12.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e48.0\u0026nbsp;\u0026plusmn; 13.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e0.310\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e47.5\u0026nbsp;\u0026plusmn; 11.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e50.0\u0026nbsp;\u0026plusmn; 12.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e0.492\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eSex, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e0.825\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e0.708\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eFemale \u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e10 (17.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e9 (19.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e6 (25.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e3 (15.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e46 (82.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e37 (80.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e18 (75.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e16 (84.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eT-stage, n (%)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e0.559\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eT1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e6 (10.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e2 (4.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e1 (4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eT2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e16 (28.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e16 (34.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e5 (20.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e4 (21.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eT3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e19 (33.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e13 (28.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e13 (54.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e11 (57.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eT4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e15 (26.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e15 (32.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e5 (20.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e4 (21.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eN-stage, n (%)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e0.207\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eN1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e4 (7.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e1 (4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e1 (5.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eN2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e30 (53.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e28 (60.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e13 (54.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e11 (57.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eN3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e22 (39.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e18 (39.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e10 (41.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e7 (36.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eClinical stage, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e0.600\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e0.606\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eII\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e2 (3.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eIII\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e20 (35.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e16 (34.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e12 (50.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e11 (57.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eIV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e34 (60.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e30 (65.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e12 (50.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e8 (42.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.165951359084406%\"\u003e\n \u003cp\u003eGTV-ln\u0026nbsp;(Gy),\u003c/p\u003e\n \u003cp\u003emedian (IQR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e52.80\u003c/p\u003e\n \u003cp\u003e(46.20 - 60.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.59656652360515%\"\u003e\n \u003cp\u003e45.60\u003c/p\u003e\n \u003cp\u003e(36.07 - 52.35)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.44206008583691%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt; 0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.878397711015737%\"\u003e\n \u003cp\u003e51.50\u003c/p\u003e\n \u003cp\u003e(42.07 - 54.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.88412017167382%\"\u003e\n \u003cp\u003e41.80\u003c/p\u003e\n \u003cp\u003e(33.00 - 52.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"8.15450643776824%\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.047*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eGTV-ln Lymph node gross tumor volume, IQR interquartile range. * indicates statistical significant difference\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2\u0026nbsp;\u003c/strong\u003eUnivariate and multivariate logistic regression analyses for predictive factors of responders in the training set\u003c/p\u003e\n\u003cdiv align=\"\"\u003e\n \u003ctable border=\"1\" cellpadding=\"0\" cellspacing=\"0\" width=\"548\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.021897810218977%\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariables\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e\u003cstrong\u003eOdds ratio\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.021897810218977%\"\u003e\n \u003cp\u003eUnivariate logistic analysis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.021897810218977%\"\u003e\n \u003cp\u003e\u0026nbsp; GTV-ln\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e1.055\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e1.018-1.093\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.021897810218977%\"\u003e\n \u003cp\u003eIntra radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e2.718\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e1.857-3.980\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.021897810218977%\"\u003e\n \u003cp\u003ePeri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e2.718\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e1.837-4.022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.021897810218977%\"\u003e\n \u003cp\u003eCombined Intra and Peri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e2.718\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e1.835-4.026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.021897810218977%\"\u003e\n \u003cp\u003eMultivariate logistic analysis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"31.021897810218977%\"\u003e\n \u003cp\u003eIntra radiomic nomogram\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003eGTV-ln\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e1.041\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e0.992-1.091\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e0.102\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.94708994708995%\"\u003e\n \u003cp\u003eRadiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.105820105820104%\"\u003e\n \u003cp\u003e2.577\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.486772486772487%\"\u003e\n \u003cp\u003e1.784-3.723\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.46031746031746%\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"31.021897810218977%\"\u003e\n \u003cp\u003ePeri radiomic nomogram\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003eGTV-ln\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e1.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e0.994-1.086\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e0.087\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.94708994708995%\"\u003e\n \u003cp\u003eRadiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.105820105820104%\"\u003e\n \u003cp\u003e2.638\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.486772486772487%\"\u003e\n \u003cp\u003e1.770-3.932\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.46031746031746%\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"31.021897810218977%\"\u003e\n \u003cp\u003eCombined Intra and Peri radiomic nomogram\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"27.554744525547445%\"\u003e\n \u003cp\u003eGTV-ln\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.86861313868613%\"\u003e\n \u003cp\u003e1.061\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.510948905109489%\"\u003e\n \u003cp\u003e1.004-1.121\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.043795620437956%\"\u003e\n \u003cp\u003e0.036\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.94708994708995%\"\u003e\n \u003cp\u003eRadiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.105820105820104%\"\u003e\n \u003cp\u003e2.935\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.486772486772487%\"\u003e\n \u003cp\u003e1.867-4.613\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.46031746031746%\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eGTV-ln Lymph node gross tumor volume, Intra intratumoral, Peri peritumoral\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3\u003c/strong\u003e Diagnostic performance of radiomic models in the training and validation sets\u003c/p\u003e\n\u003cdiv align=\"\"\u003e\n \u003ctable border=\"1\" cellpadding=\"0\" cellspacing=\"0\" width=\"661\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"28.54984894259819%\"\u003e\n \u003cp\u003e\u003cstrong\u003eModels\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.48036253776435%\"\u003e\n \u003cp\u003e\u003cstrong\u003eCohorts\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.163141993957705%\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC (95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSen\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpe\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"14.19939577039275%\"\u003e\n \u003cp\u003e\u003cstrong\u003eACC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"28.54984894259819%\"\u003e\n \u003cp\u003eIntra radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.48036253776435%\"\u003e\n \u003cp\u003eTraining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.163141993957705%\"\u003e\n \u003cp\u003e0.910\u003c/p\u003e\n \u003cp\u003e(0.837, 0.958)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.839\u003c/p\u003e\n \u003cp\u003e(0.716, 0.923)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.891\u003c/p\u003e\n \u003cp\u003e(0.764, 0.963)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"14.19939577039275%\"\u003e\n \u003cp\u003e0.863\u003c/p\u003e\n \u003cp\u003e(0.780, 0.922)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.067653276955603%\"\u003e\n \u003cp\u003eValidation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.621564482029598%\"\u003e\n \u003cp\u003e0.737\u003c/p\u003e\n \u003cp\u003e(0.580, 0.859)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.625\u003c/p\u003e\n \u003cp\u003e(0.405, 0.812)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.895\u003c/p\u003e\n \u003cp\u003e(0.668, 0.986)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"19.873150105708245%\"\u003e\n \u003cp\u003e0.744\u003c/p\u003e\n \u003cp\u003e(0.588, 0.864)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"28.54984894259819%\"\u003e\n \u003cp\u003ePeri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.48036253776435%\"\u003e\n \u003cp\u003eTraining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.163141993957705%\"\u003e\n \u003cp\u003e0.887\u003c/p\u003e\n \u003cp\u003e(0.809, 0.941)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.714\u003c/p\u003e\n \u003cp\u003e(0.577, 0.827)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.934\u003c/p\u003e\n \u003cp\u003e(0.821, 0.986)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"14.19939577039275%\"\u003e\n \u003cp\u003e0.814\u003c/p\u003e\n \u003cp\u003e(0.724, 0.883)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.067653276955603%\"\u003e\n \u003cp\u003eValidation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.621564482029598%\"\u003e\n \u003cp\u003e0.794\u003c/p\u003e\n \u003cp\u003e(0.643, 0.902)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.375\u003c/p\u003e\n \u003cp\u003e(0.187, 0.594)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.895\u003c/p\u003e\n \u003cp\u003e(0.668, 0.986)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"19.873150105708245%\"\u003e\n \u003cp\u003e0.605\u003c/p\u003e\n \u003cp\u003e(0.444, 0.750)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"28.54984894259819%\"\u003e\n \u003cp\u003eCombined Intra and Peri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.48036253776435%\"\u003e\n \u003cp\u003eTraining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.163141993957705%\"\u003e\n \u003cp\u003e0.934\u003c/p\u003e\n \u003cp\u003e(0.867, 0.974)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.946\u003c/p\u003e\n \u003cp\u003e(0.851, 0.988)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.782\u003c/p\u003e\n \u003cp\u003e(0.636, 0.890)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"14.19939577039275%\"\u003e\n \u003cp\u003e0.873\u003c/p\u003e\n \u003cp\u003e(0.791, 0.930)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.067653276955603%\"\u003e\n \u003cp\u003eValidation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.621564482029598%\"\u003e\n \u003cp\u003e0.774\u003c/p\u003e\n \u003cp\u003e(0.621, 0.887)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.750\u003c/p\u003e\n \u003cp\u003e(0.532, 0.902)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.737\u003c/p\u003e\n \u003cp\u003e(0.488, 0.908)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"19.873150105708245%\"\u003e\n \u003cp\u003e0.744\u003c/p\u003e\n \u003cp\u003e(0.588, 0.864)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"28.54984894259819%\"\u003e\n \u003cp\u003eCombined Intra and Peri radiomic nomogram\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.48036253776435%\"\u003e\n \u003cp\u003eTraining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.163141993957705%\"\u003e\n \u003cp\u003e0.941\u003c/p\u003e\n \u003cp\u003e(0.877, 0.978)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.929\u003c/p\u003e\n \u003cp\u003e(0.827, 0.980)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.803625377643504%\"\u003e\n \u003cp\u003e0.848\u003c/p\u003e\n \u003cp\u003e(0.711, 0.936)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"14.19939577039275%\"\u003e\n \u003cp\u003e0.892\u003c/p\u003e\n \u003cp\u003e(0.815, 0.944)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.067653276955603%\"\u003e\n \u003cp\u003eValidation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"22.621564482029598%\"\u003e\n \u003cp\u003e0.783\u003c/p\u003e\n \u003cp\u003e(0.631, 0.894)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.667\u003c/p\u003e\n \u003cp\u003e(0.446, 0.843)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.718816067653275%\"\u003e\n \u003cp\u003e0.842\u003c/p\u003e\n \u003cp\u003e(0.604, 0.966)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"19.873150105708245%\"\u003e\n \u003cp\u003e0.744\u003c/p\u003e\n \u003cp\u003e(0.588, 0.864)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eAUC the area under the curve, Intra intratumoral, Peri peritumoral, CI confidence interval, Sen sensitivity, Spe specificity, ACC accuracy\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4\u003c/strong\u003e Comparisons of AUC, sensitivity, and specificity between the combined Intra and Peri radiomic nomogram and the other radiomic models in the training and validation sets\u003c/p\u003e\n\u003cdiv align=\"\"\u003e\n \u003ctable border=\"1\" cellpadding=\"0\" cellspacing=\"0\" width=\"558\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"35.483870967741936%\"\u003e\n \u003cp\u003e\u003cstrong\u003eComparisons\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" width=\"21.50537634408602%\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" width=\"21.50537634408602%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" width=\"21.50537634408602%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"16.666666666666668%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eZ\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.666666666666668%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.666666666666668%\"\u003e\n \u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e2\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.666666666666668%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"16.666666666666668%\"\u003e\n \u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e2\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"16.666666666666668%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003eTraining set\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003eIntra radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e1.107\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.268\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e2.285\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.166\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.683\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003ePeri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e1.570\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.116\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e6.722\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.009\u003csup\u003e*\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e1.125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.288\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003eCombined Intra and Peri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.762\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.446\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e1.333\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.248\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003eValidation set\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003eIntra radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.639\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.522\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003ePeri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.167\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.867\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e4.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.045\u003csup\u003e*\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.483870967741936%\"\u003e\n \u003cp\u003eCombined Intra and Peri radiomic signature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.586\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.558\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.500\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.479\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.500\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.75268817204301%\"\u003e\n \u003cp\u003e0.479\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003csup\u003e*\u003c/sup\u003eDifferences was significant at \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05. AUC the area under the curve, Intra intratumoral, Peri peritumoral\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmim","sideBox":"Learn more about [BMC Medical Imaging](http://bmcmedimaging.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmim/default.aspx","title":"BMC Medical Imaging","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Nasopharyngeal cancer, Magnetic resonance imaging, Radiomics","lastPublishedDoi":"10.21203/rs.3.rs-2519551/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2519551/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eIn this investigation, intratumoral (Intra) and peritumoral (Peri) features obtained from MRI imaging were used to create and evaluate radiomic models for response prediction to radiochemotherapy of metastatic cervical lymph nodes in individuals with nasopharyngeal cancer (NPC).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eRetrospectively, we included 145 consecutive subjects with NPC, 102 in the training set and 43 in the validation set. A total of 5408 initial radiomic features were acquired from the metastatic cervical lymph node's Intra and Peri areas. Then, employing multivariate logistic regression analysis, the radiomic features were chosen and integrated with clinical characteristics to create predictive models. And at last, these developed prediction models were examined using sensitivity, specificity, accuracy, and the area under the curve (AUC) of receiver operating characteristics.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eIn the training and validation sets, there was no statistically significant variation in the AUC among the Intra radiomic signature, Peri radiomic signature, combined Intra and Peri radiomic signature, and combined Intra and Peri radiomic nomogram (all \u003cem\u003eP\u003c/em\u003e \u0026gt; 0.05). With an AUC of 0.941 (0.877-0.978) in the training set and 0.783 (0.631-0.894) in the validation set, the combined Intra and Peri radiomic nomogram enabled good discrimination among the responders and non-responders groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eThe early response of metastatic cervical lymph nodes to radiochemotherapy in individuals with NPC may be predicted by pretreatment radiomic models determined by the combined Intra and Peri features from MRI imaging, facilitating therapeutic interventions and clinical decision-making.\u003c/p\u003e","manuscriptTitle":"Intra- and peritumoral MRI radiomics assisted in predicting radiochemotherapy response in metastatic cervical lymph nodes of nasopharyngeal cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2023-02-03 15:55:34","doi":"10.21203/rs.3.rs-2519551/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major revision","date":"2023-03-17T12:11:07+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2023-02-26T07:50:29+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"190ed549-1ad1-4428-a6ce-fbda5cd442b0","date":"2023-02-12T07:45:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"d6142ef9-ac9f-4f05-b0d3-5cd258deb426","date":"2023-02-03T07:48:19+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2023-02-03T07:07:54+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2023-02-03T02:14:30+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2023-02-02T07:14:19+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2023-02-02T07:11:16+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Imaging","date":"2023-01-27T08:29:39+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmim","sideBox":"Learn more about [BMC Medical Imaging](http://bmcmedimaging.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmim/default.aspx","title":"BMC Medical Imaging","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"94cab30b-fd8d-4367-9562-0e840406177b","owner":[],"postedDate":"February 3rd, 2023","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2023-10-16T21:39:41+00:00","versionOfRecord":{"articleIdentity":"rs-2519551","link":"https://doi.org/10.1186/s12880-023-01026-1","journal":{"identity":"bmc-medical-imaging","isVorOnly":false,"title":"BMC Medical Imaging"},"publishedOn":"2023-05-30 21:00:40","publishedOnDateReadable":"May 30th, 2023"},"versionCreatedAt":"2023-02-03 15:55:34","video":"","vorDoi":"10.1186/s12880-023-01026-1","vorDoiUrl":"https://doi.org/10.1186/s12880-023-01026-1","workflowStages":[]},"version":"v1","identity":"rs-2519551","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2519551","identity":"rs-2519551","version":["v1"]},"buildId":"WrCJVZZCHTDjtuVLN7oU0","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}