Shuangqing Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Shuangqing Chen has a total of 95 co-authors. Here are the top co-authors by number of shared publications:
Yu-Ying Shen
Department of Radiology, the Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou, China
8 shared publicationsQing Cai
Department of Radiology, the Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou, China
7 shared publicationsChen-Lu Liu
Stem cells research center, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou, 215001, China
5 shared publicationsGenji Bai
Department of Radiology, The Affiliated Huaian No. 1 People’s Hospital of Nanjing Medical University, Huaian, China
5 shared publicationsYiping Gao
5 shared publications- Jian Zheng
Division of Life Sciences and Medicine, School of Biomedical Engineering (Suzhou), University of Science and Technology of China, Hefei, China
4 shared publications
Latest Publications
- Open Access
- 6 November 2026
- Published by Springer Nature in Insights into Imaging
- Vol. 17 (1) , 284
MRI-based deep-learning radiomics nomogram for differentiating FIGO 2023 stage I from stage II–IV endometrial cancer: a multicenter study
Objectives: Preoperative identification of FIGO 2023 stage I endometrial cancer (EC) is important for treatment planning. We aimed to develop and validate a deep-learning radiomics nomogram (DLRN) based on multiparametric MRI to distinguish stage I EC from stage II–IV disease. Materials and methods: This study retrospectively enrolled patients from three centers. Patients from Centers 1 and 2 were randomly split into a training set ( n = 616), and an internal test set ( n = 158), and Center 3 served as an external test set ( n = 232). The DLRN integrated deep-learning features, intratumoral and peritumoral radiomics features, and clinical characteristics. The DLRN was compared with clinical, radiomics, and deep-learning models. Discrimination was evaluated using the area under the receiver operating characteristic curve (AUC), with AUCs compared using DeLong’s test. Results: A total of 1006 patients (mean age, 56.5 ± 9.7 years) were analyzed. In the internal test set, the DLRN achieved an AUC of 0.852, higher than the clinical (0.673; p = 0.001), original radiomics (0.765; p = 0.030), and combined radiomics (0.793; p = 0.050) models. In the external test set, the DLRN remained robust (AUC = 0.829) and outperformed the clinical, original radiomics, and combined radiomics models (AUC = 0.680, 0.681, and 0.743; all p < 0.005). With DLRN assistance, radiologists showed improved diagnostic performance. Conclusion: The MRI-based DLRN showed robust discrimination between FIGO 2023 stage I and stage II–IV EC across test sets, supporting its potential for preoperative risk stratification and decision support. Key Points: Question The introduction of the FIGO 2023 staging system has made the differentiation between stages more challenging, highlighting the need for better tools. Findings A new MRI-based deep-learning radiomics nomogram (DLRN) accurately identifies high-risk patients, significantly outperforming traditional and individual clinical prediction methods. Critical relevance statement This tool helps doctors make better preoperative treatment plans, potentially improving patient outcomes in endometrial cancer.
- 6 November 2026
- Published by Springer Nature in Insights into Imaging
- Vol. 17 (1) , 284
Open AccessMRI-based deep-learning radiomics nomogram for differentiating FIGO 2023 stage I from stage II–IV endometrial cancer: a multicenter study
Objectives: Preoperative identification of FIGO 2023 stage I endometrial cancer (EC) is important for treatment planning. We aimed to develop and validate a deep-learning radiomics nomogram (DLRN) based on multiparametric MRI to distinguish stage I EC from stage II–IV disease. Materials and methods: This study retrospectively enrolled patients from three centers. Patients from Centers 1 and 2 were randomly split into a training set ( n = 616), and an internal test set ( n = 158), and Center 3 served as an external test set ( n = 232). The DLRN integrated deep-learning features, intratumoral and peritumoral radiomics features, and clinical characteristics. The DLRN was compared with clinical, radiomics, and deep-learning models. Discrimination was evaluated using the area under the receiver operating characteristic curve (AUC), with AUCs compared using DeLong’s test. Results: A total of 1006 patients (mean age, 56.5 ± 9.7 years) were analyzed. In the internal test set, the DLRN achieved an AUC of 0.852, higher than the clinical (0.673; p = 0.001), original radiomics (0.765; p = 0.030), and combined radiomics (0.793; p = 0.050) models. In the external test set, the DLRN remained robust (AUC = 0.829) and outperformed the clinical, original radiomics, and combined radiomics models (AUC = 0.680, 0.681, and 0.743; all p < 0.005). With DLRN assistance, radiologists showed improved diagnostic performance. Conclusion: The MRI-based DLRN showed robust discrimination between FIGO 2023 stage I and stage II–IV EC across test sets, supporting its potential for preoperative risk stratification and decision support. Key Points: Question The introduction of the FIGO 2023 staging system has made the differentiation between stages more challenging, highlighting the need for better tools. Findings A new MRI-based deep-learning radiomics nomogram (DLRN) accurately identifies high-risk patients, significantly outperforming traditional and individual clinical prediction methods. Critical relevance statement This tool helps doctors make better preoperative treatment plans, potentially improving patient outcomes in endometrial cancer.
- Open Access
- 1 June 2026
- Published by AME Publishing Company in Quantitative Imaging in Medicine and Surgery
- Vol. 16 (6) , 454
Computed tomography enterography-based deep learning radiomics models of intestinal lesions and perienteric fat for predicting Crohn’s disease activity: a multicenter cohort study
Background: Accurate assessment of Crohn’s disease (CD) activity is critical for improving patient prognosis, while current imaging methods lack quantitative objectivity. Therefore, this study aimed to develop an integrated model based on computed tomography enterography (CTE) using radiomics, deep learning, and clinical characteristics for noninvasively predicting CD activity. Methods: A total of 259 CD patients were retrospectively enrolled from three medical centers. With patients from one center serving as an independent external validation cohort (n=68), patients from the other two centers were randomly divided into a training (n=133) and a testing (n=58) cohort in a 7:3 ratio. Independent predictors were selected from clinical features to construct a clinical model. Based on the imaging features of intestinal lesions and perienteric adipose tissue, radiomics models were constructed using least absolute shrinkage and selection operator (LASSO) regression, and deep learning radiomics models were developed using features extracted by ResNet-18 and ResNet-34 architectures. Finally, an integrated model combining the radiomics, deep learning, and independent clinical predictors was constructed. The area under the receiver operating characteristic curve (AUC) was used to assess model performance. Results: The combined model demonstrated optimal performance, with an AUC of 0.984 [95% confidence interval (CI): 0.970–0.999], 0.902 (95% CI: 0.804–1.000) and 0.884 (95% CI: 0.809–0.960) in the training, testing and validation cohorts, respectively. Among other models, the deep learning radiomics model for intestinal lesions achieved the highest AUC of 0.938 (95% CI: 0.901–0.976), 0.865 (95% CI: 0.767–0.963) and 0.841 (95% CI: 0.745–0.938) in the training, testing and validation cohorts. No statistically significant difference was observed between the intestinal lesion-based radiomics model and the perienteric fat-based radiomics model. Conclusions: The combined model integrating the radiomics based on intestinal lesions and perienteric fat, deep learning, and clinical characteristics demonstrated the best performance in predicting CD activity and can provide novel insights for clinical decision-making.
- 1 June 2026
- Published by AME Publishing Company in Quantitative Imaging in Medicine and Surgery
- Vol. 16 (6) , 454
Open AccessComputed tomography enterography-based deep learning radiomics models of intestinal lesions and perienteric fat for predicting Crohn’s disease activity: a multicenter cohort study
Background: Accurate assessment of Crohn’s disease (CD) activity is critical for improving patient prognosis, while current imaging methods lack quantitative objectivity. Therefore, this study aimed to develop an integrated model based on computed tomography enterography (CTE) using radiomics, deep learning, and clinical characteristics for noninvasively predicting CD activity. Methods: A total of 259 CD patients were retrospectively enrolled from three medical centers. With patients from one center serving as an independent external validation cohort (n=68), patients from the other two centers were randomly divided into a training (n=133) and a testing (n=58) cohort in a 7:3 ratio. Independent predictors were selected from clinical features to construct a clinical model. Based on the imaging features of intestinal lesions and perienteric adipose tissue, radiomics models were constructed using least absolute shrinkage and selection operator (LASSO) regression, and deep learning radiomics models were developed using features extracted by ResNet-18 and ResNet-34 architectures. Finally, an integrated model combining the radiomics, deep learning, and independent clinical predictors was constructed. The area under the receiver operating characteristic curve (AUC) was used to assess model performance. Results: The combined model demonstrated optimal performance, with an AUC of 0.984 [95% confidence interval (CI): 0.970–0.999], 0.902 (95% CI: 0.804–1.000) and 0.884 (95% CI: 0.809–0.960) in the training, testing and validation cohorts, respectively. Among other models, the deep learning radiomics model for intestinal lesions achieved the highest AUC of 0.938 (95% CI: 0.901–0.976), 0.865 (95% CI: 0.767–0.963) and 0.841 (95% CI: 0.745–0.938) in the training, testing and validation cohorts. No statistically significant difference was observed between the intestinal lesion-based radiomics model and the perienteric fat-based radiomics model. Conclusions: The combined model integrating the radiomics based on intestinal lesions and perienteric fat, deep learning, and clinical characteristics demonstrated the best performance in predicting CD activity and can provide novel insights for clinical decision-making.
- Open Access
- 11 December 2025
- Published by Springer Nature in Journal of Translational Medicine
- Vol. 24 (1)
Radiogenomic MRI biomarkers for noninvasive prediction of GPC3 expression and tumor microenvironment in hepatocellular carcinoma
Glypican-3 (GPC3) is frequently overexpressed in hepatocellular carcinoma (HCC) and plays a key role in immune and metabolic remodeling of the tumor microe
- 11 December 2025
- Published by Springer Nature in Journal of Translational Medicine
- Vol. 24 (1)
Open AccessRadiogenomic MRI biomarkers for noninvasive prediction of GPC3 expression and tumor microenvironment in hepatocellular carcinoma
Glypican-3 (GPC3) is frequently overexpressed in hepatocellular carcinoma (HCC) and plays a key role in immune and metabolic remodeling of the tumor microe
- Open Access
- 11 November 2025
- Published by Springer Nature in BMC Medical Imaging
- Vol. 25 (1) , 1-12
A nomogram based on dual-layer spectral detector CT-derived 40KeV virtual monoenergetic images for preoperative prediction of simultaneous distant metastasis in colorectal cancer
To establish and validate a nomogram combining dual-layer spectral detector CT(DLSCT)-derived 40KeV virtual monoenergetic images (VMI) of radiomics features, spectral parameters and clinical features for preoperative prediction of simultaneous distant metastases (SDM) in colorectal cancer (CRC). We retrospectively included 137 patients [SDM+ (n = 51); SDM−(n = 86)] with pathologically confirmed CRC who attended two hospitals between June 2022 and April 2023. Patients were divided into a training group (n = 90, from hospital A) and an external validation group (n = 47, from hospital B). Clinical characteristics and spectral parameters of arterial phase (AP), venous phase (VP) and delayed phase (DP) were collected to establish a clinical model. Radiomics modeling by extracting radiomics features in the three-dimensional region of interest of the 40 KeV-VMI. Combining radiomics scores, clinical features, and spectral parameters to create a nomogram. The performance of each model was assessed by the area under the curve (AUC), plotting calibration curves and decision curve analysis (DCA). In the training group, the nomogram (AUC = 0.938) was remarkably better than that of the radiomics models. In the external validation group, the nomogram (AUC = 0.930) was remarkably superior to that of the DP model and the clinical model. In the vast majority of threshold probabilities, the nomogram had a better critical net benefit than the other four models in predicting SDM of CRC. The nomogram incorporating radiomics features of DLSCT-derived 40KeV-VMI, spectral parameters and clinical features showed excellent predictive performance in preoperatively predicting SDM in CRC, which can help clinicians make accurate individualized treatment plans.
- 11 November 2025
- Published by Springer Nature in BMC Medical Imaging
- Vol. 25 (1) , 1-12
Open AccessA nomogram based on dual-layer spectral detector CT-derived 40KeV virtual monoenergetic images for preoperative prediction of simultaneous distant metastasis in colorectal cancer
To establish and validate a nomogram combining dual-layer spectral detector CT(DLSCT)-derived 40KeV virtual monoenergetic images (VMI) of radiomics features, spectral parameters and clinical features for preoperative prediction of simultaneous distant metastases (SDM) in colorectal cancer (CRC). We retrospectively included 137 patients [SDM+ (n = 51); SDM−(n = 86)] with pathologically confirmed CRC who attended two hospitals between June 2022 and April 2023. Patients were divided into a training group (n = 90, from hospital A) and an external validation group (n = 47, from hospital B). Clinical characteristics and spectral parameters of arterial phase (AP), venous phase (VP) and delayed phase (DP) were collected to establish a clinical model. Radiomics modeling by extracting radiomics features in the three-dimensional region of interest of the 40 KeV-VMI. Combining radiomics scores, clinical features, and spectral parameters to create a nomogram. The performance of each model was assessed by the area under the curve (AUC), plotting calibration curves and decision curve analysis (DCA). In the training group, the nomogram (AUC = 0.938) was remarkably better than that of the radiomics models. In the external validation group, the nomogram (AUC = 0.930) was remarkably superior to that of the DP model and the clinical model. In the vast majority of threshold probabilities, the nomogram had a better critical net benefit than the other four models in predicting SDM of CRC. The nomogram incorporating radiomics features of DLSCT-derived 40KeV-VMI, spectral parameters and clinical features showed excellent predictive performance in preoperatively predicting SDM in CRC, which can help clinicians make accurate individualized treatment plans.
- 9 October 2025
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (4) , 1-13
Noninvasive Multi-Omics Radiomic Model Integrating scRNA-seq and Bulk RNA-seq for Hepatocellular Carcinoma Prognosis
Hepatocellular carcinoma (HCC) is characterized by high heterogeneity and a complex tumor microenvironment, which challenges conventional prognostic approaches. We developed a machine learning (ML)-based radiomic prognostic model that integrates single-cell and bulk RNA sequencing to improve risk stratification in HCC patients. scRNA-seq analysis was performed, excluding cells with < 200 or > 7500 detected genes or > 20% mitochondrial genes. Dimensionality reduction And clustering identified 2317 co-heterogeneous genes across six cell types. A nine-gene TME signature, based on the intersection with TCGA data, was used to stratify survival risk. We screened radiomic features strongly correlated with TME scores and developed a support vector machine model. Feature importance was assessed by SHAP analysis, and model performance was validated using Cox regression and nomogram analysis. Patients with higher TME risk scores had significantly reduced survival (HR: 2.13, 95% CI: 1.42–3.21, p < 0.001). The SVM model, based on four selected radiomic features, achieved high prognostic accuracy (area under the curve (AUC) = 0.85; C-index = 0.78), and its predictions aligned with nomogram survival estimates. By integrating molecular and imaging data, this radiomic model shows promising prognostic performance and may provide a non-invasive framework for HCC patient stratification.
- 9 October 2025
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (4) , 1-13
Noninvasive Multi-Omics Radiomic Model Integrating scRNA-seq and Bulk RNA-seq for Hepatocellular Carcinoma Prognosis
Hepatocellular carcinoma (HCC) is characterized by high heterogeneity and a complex tumor microenvironment, which challenges conventional prognostic approaches. We developed a machine learning (ML)-based radiomic prognostic model that integrates single-cell and bulk RNA sequencing to improve risk stratification in HCC patients. scRNA-seq analysis was performed, excluding cells with < 200 or > 7500 detected genes or > 20% mitochondrial genes. Dimensionality reduction And clustering identified 2317 co-heterogeneous genes across six cell types. A nine-gene TME signature, based on the intersection with TCGA data, was used to stratify survival risk. We screened radiomic features strongly correlated with TME scores and developed a support vector machine model. Feature importance was assessed by SHAP analysis, and model performance was validated using Cox regression and nomogram analysis. Patients with higher TME risk scores had significantly reduced survival (HR: 2.13, 95% CI: 1.42–3.21, p < 0.001). The SVM model, based on four selected radiomic features, achieved high prognostic accuracy (area under the curve (AUC) = 0.85; C-index = 0.78), and its predictions aligned with nomogram survival estimates. By integrating molecular and imaging data, this radiomic model shows promising prognostic performance and may provide a non-invasive framework for HCC patient stratification.
- Open Access
- 17 February 2025
- Published by MDPI AG in Bioengineering
- Vol. 12 (2) , 196
Cross-Attention Adaptive Feature Pyramid Network with Uncertainty Boundary Modeling for Mass Detection in Digital Breast Tomosynthesis
Computer-aided detection (CADe) of masses in digital breast tomosynthesis (DBT) is crucial for early breast cancer diagnosis. However, the variability in the size and morphology of breast masses and their resemblance to surrounding tissues present significant challenges. Current CNN-based CADe methods, particularly those that use Feature Pyramid Networks (FPN), often fail to integrate multi-scale information effectively and struggle to handle dense glandular tissue with high-density or iso-density mass lesions due to the unidirectional integration and progressive attenuation of features, leading to high false positive rates. Additionally, the commonly indistinct boundaries of breast masses introduce uncertainty in boundary localization, which makes traditional Dirac boundary modeling insufficient for precise boundary regression. To address these issues, we propose the CU-Net network, which efficiently fuses multi-scale features and accurately models blurred boundaries. Specifically, the CU-Net introduces the Cross-Attention Adaptive Feature Pyramid Network (CA-FPN), which enhances the effectiveness and accuracy of feature interactions through a cross-attention mechanism to capture global correlations across multi-scale feature maps. Simultaneously, the Breast Density Perceptual Module (BDPM) incorporates breast density information to weight intermediate features, thereby improving the network’s focus on dense breast regions susceptible to false positives. For blurred mass boundaries, we introduce Uncertainty Boundary Modeling (UBM) to model the positional distribution function of predicted bounding boxes for masses with uncertain boundaries. In comparative experiments on an in-house clinical DBT dataset and the BCS-DBT dataset, the proposed method achieved sensitivities of 89.68% and 72.73% at 2 false positives per DBT volume (FPs/DBT), respectively, significantly outperforming existing state-of-the-art detection methods. This method offers clinicians rapid, accurate, and objective diagnostic assistance, demonstrating substantial potential for clinical application.
- 17 February 2025
- Published by MDPI AG in Bioengineering
- Vol. 12 (2) , 196
Open AccessCross-Attention Adaptive Feature Pyramid Network with Uncertainty Boundary Modeling for Mass Detection in Digital Breast Tomosynthesis
Computer-aided detection (CADe) of masses in digital breast tomosynthesis (DBT) is crucial for early breast cancer diagnosis. However, the variability in the size and morphology of breast masses and their resemblance to surrounding tissues present significant challenges. Current CNN-based CADe methods, particularly those that use Feature Pyramid Networks (FPN), often fail to integrate multi-scale information effectively and struggle to handle dense glandular tissue with high-density or iso-density mass lesions due to the unidirectional integration and progressive attenuation of features, leading to high false positive rates. Additionally, the commonly indistinct boundaries of breast masses introduce uncertainty in boundary localization, which makes traditional Dirac boundary modeling insufficient for precise boundary regression. To address these issues, we propose the CU-Net network, which efficiently fuses multi-scale features and accurately models blurred boundaries. Specifically, the CU-Net introduces the Cross-Attention Adaptive Feature Pyramid Network (CA-FPN), which enhances the effectiveness and accuracy of feature interactions through a cross-attention mechanism to capture global correlations across multi-scale feature maps. Simultaneously, the Breast Density Perceptual Module (BDPM) incorporates breast density information to weight intermediate features, thereby improving the network’s focus on dense breast regions susceptible to false positives. For blurred mass boundaries, we introduce Uncertainty Boundary Modeling (UBM) to model the positional distribution function of predicted bounding boxes for masses with uncertain boundaries. In comparative experiments on an in-house clinical DBT dataset and the BCS-DBT dataset, the proposed method achieved sensitivities of 89.68% and 72.73% at 2 false positives per DBT volume (FPs/DBT), respectively, significantly outperforming existing state-of-the-art detection methods. This method offers clinicians rapid, accurate, and objective diagnostic assistance, demonstrating substantial potential for clinical application.
- 1 January 2025
- Published by e-Century Publishing Corporation in American Journal of Translational Research
- Vol. 17 (1) , 162-177
Delta/pre-radiomics based on enhanced CT predicts complete response in locally advanced esophageal squamous cell carcinoma
Objectives: This study aimed to evaluate the effectiveness of neoadjuvant immunochemotherapy (NIC) in patients diagnosed with locally advanced esophageal squamous cell carcinoma (LAESCC), by assessing the performance of models that utilize enhanced computed tomography (CT) images at the pre, post, and delta/pre group stages. Methods: A total of 225 patients were included in our study and randomly divided into a training set (n = 157) and test set I (n = 68). In addition, we conducted a test set II involving 60 patients from another center. We obtained omics features from CT images before and after NIC. Then, the delta radiomics features were obtained by calculating the differences between the post and pre group features, which was then divided by the pre group features to obtain the delta/pre group. Imaging and clinicopathological data were collected in the two centers according to the same inclusion and exclusion criteria. The tumor regression grading (TRG) system was used according to the Japanese Esophageal Cancer (11th edition). Three sets of models were built and their performance was assessed using receiver operating characteristic (ROC) curve, confusion matrix, and calibration curve. The clinical utility of the model was evaluated through decision curve analysis and nomogram. Results: The area under the curve value of the delta/pre-radiomics (Rad) score model was 0.876 in the training set and 0.827 and 0.749 in the two test sets, which was significantly higher than that in the pre and post Rad score models. The radiomics nomogram was constructed using Rad scores derived from the post model, delta/pre model, Ki67, P53, and the pathological stage of lymph node after neoadjuvant therapy (ypN), demonstrating robust performance. The internal correction curve (apparent) and the external correction curve (bias-corrected) exhibited negligible deviations from the ideal curve, thereby demonstrating a high level of similarity. Conclusion: Nomogram, based on delta/pre-enhanced CT features and clinical risk indicators, is a non-invasive tool to predict therapeutic effects in patients with LAESCC after NIC.
- 1 January 2025
- Published by e-Century Publishing Corporation in American Journal of Translational Research
- Vol. 17 (1) , 162-177
Delta/pre-radiomics based on enhanced CT predicts complete response in locally advanced esophageal squamous cell carcinoma
Objectives: This study aimed to evaluate the effectiveness of neoadjuvant immunochemotherapy (NIC) in patients diagnosed with locally advanced esophageal squamous cell carcinoma (LAESCC), by assessing the performance of models that utilize enhanced computed tomography (CT) images at the pre, post, and delta/pre group stages. Methods: A total of 225 patients were included in our study and randomly divided into a training set (n = 157) and test set I (n = 68). In addition, we conducted a test set II involving 60 patients from another center. We obtained omics features from CT images before and after NIC. Then, the delta radiomics features were obtained by calculating the differences between the post and pre group features, which was then divided by the pre group features to obtain the delta/pre group. Imaging and clinicopathological data were collected in the two centers according to the same inclusion and exclusion criteria. The tumor regression grading (TRG) system was used according to the Japanese Esophageal Cancer (11th edition). Three sets of models were built and their performance was assessed using receiver operating characteristic (ROC) curve, confusion matrix, and calibration curve. The clinical utility of the model was evaluated through decision curve analysis and nomogram. Results: The area under the curve value of the delta/pre-radiomics (Rad) score model was 0.876 in the training set and 0.827 and 0.749 in the two test sets, which was significantly higher than that in the pre and post Rad score models. The radiomics nomogram was constructed using Rad scores derived from the post model, delta/pre model, Ki67, P53, and the pathological stage of lymph node after neoadjuvant therapy (ypN), demonstrating robust performance. The internal correction curve (apparent) and the external correction curve (bias-corrected) exhibited negligible deviations from the ideal curve, thereby demonstrating a high level of similarity. Conclusion: Nomogram, based on delta/pre-enhanced CT features and clinical risk indicators, is a non-invasive tool to predict therapeutic effects in patients with LAESCC after NIC.
- Open Access
- 22 November 2024
- Published by Springer Nature in Insights into Imaging
- Vol. 15 (1) , 1-11
Intratumoral and peritumoral MRI-based radiomics for predicting extrapelvic peritoneal metastasis in epithelial ovarian cancer
Objectives: To investigate the potential of intratumoral and peritumoral radiomics derived from T2-weighted MRI to preoperatively predict extrapelvic peritoneal metastasis (EPM) in patients with epithelial ovarian cancer (EOC). Methods: In this retrospective study, 488 patients from four centers were enrolled and divided into training (n = 245), internal test (n = 105), and external test (n = 138) sets. Intratumoral and peritumoral models were constructed based on radiomics features extracted from the corresponding regions. A combined intratumoral and peritumoral model was developed via a feature-level fusion. An ensemble model was created by integrating this combined model with specific independent clinical predictors. The robustness and generalizability of these models were assessed using tenfold cross-validation and both internal and external testing. Model performance was evaluated by the area under the receiver operating characteristic curve (AUC). The Shapley Additive Explanation method was employed for model interpretation. Results: The ensemble model showed superior performance across the tenfold cross-validation, with the highest mean AUC of 0.844 ± 0.063. On the internal test set, the peritumoral and ensemble models significantly outperformed the intratumoral model (AUC = 0.786 and 0.832 vs. 0.652, p = 0.007 and p < 0.001, respectively). On the external test set, the AUC of the ensemble model significantly exceeded those of the intratumoral and peritumoral models (0.843 vs. 0.750 and 0.789, p = 0.008 and 0.047, respectively). Conclusion: Peritumoral radiomics provide more informative insights about EPM than intratumoral radiomics. The ensemble model based on MRI has the potential to preoperatively predict EPM in EOC patients. Critical relevance statement: Integrating both intratumoral and peritumoral radiomics information based on MRI with clinical characteristics is a promising noninvasive method to predict EPM to guide preoperative clinical decision-making for EOC patients. Key Points: Peritumoral radiomics can provide valuable information about extrapelvic peritoneal metastasis in epithelial ovarian cancer. The ensemble model demonstrated satisfactory performance in predicting extrapelvic peritoneal metastasis. Combining intratumoral and peritumoral MRI radiomics contributes to clinical decision-making in epithelial ovarian cancer. Graphical
- 22 November 2024
- Published by Springer Nature in Insights into Imaging
- Vol. 15 (1) , 1-11
Open AccessIntratumoral and peritumoral MRI-based radiomics for predicting extrapelvic peritoneal metastasis in epithelial ovarian cancer
Objectives: To investigate the potential of intratumoral and peritumoral radiomics derived from T2-weighted MRI to preoperatively predict extrapelvic peritoneal metastasis (EPM) in patients with epithelial ovarian cancer (EOC). Methods: In this retrospective study, 488 patients from four centers were enrolled and divided into training (n = 245), internal test (n = 105), and external test (n = 138) sets. Intratumoral and peritumoral models were constructed based on radiomics features extracted from the corresponding regions. A combined intratumoral and peritumoral model was developed via a feature-level fusion. An ensemble model was created by integrating this combined model with specific independent clinical predictors. The robustness and generalizability of these models were assessed using tenfold cross-validation and both internal and external testing. Model performance was evaluated by the area under the receiver operating characteristic curve (AUC). The Shapley Additive Explanation method was employed for model interpretation. Results: The ensemble model showed superior performance across the tenfold cross-validation, with the highest mean AUC of 0.844 ± 0.063. On the internal test set, the peritumoral and ensemble models significantly outperformed the intratumoral model (AUC = 0.786 and 0.832 vs. 0.652, p = 0.007 and p < 0.001, respectively). On the external test set, the AUC of the ensemble model significantly exceeded those of the intratumoral and peritumoral models (0.843 vs. 0.750 and 0.789, p = 0.008 and 0.047, respectively). Conclusion: Peritumoral radiomics provide more informative insights about EPM than intratumoral radiomics. The ensemble model based on MRI has the potential to preoperatively predict EPM in EOC patients. Critical relevance statement: Integrating both intratumoral and peritumoral radiomics information based on MRI with clinical characteristics is a promising noninvasive method to predict EPM to guide preoperative clinical decision-making for EOC patients. Key Points: Peritumoral radiomics can provide valuable information about extrapelvic peritoneal metastasis in epithelial ovarian cancer. The ensemble model demonstrated satisfactory performance in predicting extrapelvic peritoneal metastasis. Combining intratumoral and peritumoral MRI radiomics contributes to clinical decision-making in epithelial ovarian cancer. Graphical