He Zhang
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
Research Subjects
He Zhang has a total of 147 co-authors. Here are the top co-authors by number of shared publications:
- Guofu Zhang
Department of Radiology, the Obstetrics and Gynecology Hospital, Fudan University, P.R. China
11 shared publications Guofu Zhang
Department of Radiology, Obstetrics and Gynecology Hospital, Fudan University, Shanghai, People's Republic of China
8 shared publications- Guang Yang
Shanghai Key Laboratory of Magnetic Resonance, School of Physics and Electronic Science, East China Normal University, Shanghai, China
7 shared publications Yida Wang
Shanghai Key Laboratory of Magnetic Resonance, East China Normal University, Shanghai, People’s Republic of China
7 shared publicationsYan Ning
Department of Pathology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai, China
5 shared publicationsYuanyuan Lu
Department of Radiology, Shanghai First Maternity and Infant Hospital, Shanghai, China
5 shared publications
Latest Publications
- 13 October 2026
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- p. 1-12
Development and External Testing of an Interpretable MRI-Based Machine Learning Model for Differentiating Endometrial Stromal Sarcoma from Cellular Leiomyoma
This retrospective two-center diagnostic study was aimed at developing and externally testing an interpretable MRI-based machine learning model for differentiating endometrial stromal sarcoma (ESS) from cellular leiomyoma (CL). A total of 155 women with pathologically confirmed ESS or CL were assigned to a derivation cohort (n = 111) and an external testing cohort (n = 44). Clinical variables, qualitative MRI features, and apparent diffusion coefficient (ADC) histogram parameters were evaluated. Three MRI-derived predictors—irregular tumor margin, cystic areas on T2-weighted imaging, and mean ADC—were used to train six classifiers. The final support vector machine model with a radial basis function kernel was evaluated at a fixed threshold determined in the derivation cohort using receiver operating characteristic, calibration, decision curve, and Shapley additive explanations (SHAP) analyses. A reader study assessed the effect of model assistance. The model achieved areas under the curve of 0.843 (95% CI, 0.760–0.917) in the derivation cohort and 0.835 (95% CI, 0.669–0.960) in the external testing cohort. Calibration and decision curve analyses indicated acceptable performance and potential clinical utility. In an exploratory model-assisted reader analysis, junior radiologist specificity increased from 46.4 to 82.1% and accuracy from 52.3 to 77.3%. This interpretable three-feature model demonstrated promising performance across cohorts and may serve as an adjunctive tool for preoperative MRI assessment, although prospective multicenter validation is warranted.
- 13 October 2026
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- p. 1-12
Development and External Testing of an Interpretable MRI-Based Machine Learning Model for Differentiating Endometrial Stromal Sarcoma from Cellular Leiomyoma
This retrospective two-center diagnostic study was aimed at developing and externally testing an interpretable MRI-based machine learning model for differentiating endometrial stromal sarcoma (ESS) from cellular leiomyoma (CL). A total of 155 women with pathologically confirmed ESS or CL were assigned to a derivation cohort (n = 111) and an external testing cohort (n = 44). Clinical variables, qualitative MRI features, and apparent diffusion coefficient (ADC) histogram parameters were evaluated. Three MRI-derived predictors—irregular tumor margin, cystic areas on T2-weighted imaging, and mean ADC—were used to train six classifiers. The final support vector machine model with a radial basis function kernel was evaluated at a fixed threshold determined in the derivation cohort using receiver operating characteristic, calibration, decision curve, and Shapley additive explanations (SHAP) analyses. A reader study assessed the effect of model assistance. The model achieved areas under the curve of 0.843 (95% CI, 0.760–0.917) in the derivation cohort and 0.835 (95% CI, 0.669–0.960) in the external testing cohort. Calibration and decision curve analyses indicated acceptable performance and potential clinical utility. In an exploratory model-assisted reader analysis, junior radiologist specificity increased from 46.4 to 82.1% and accuracy from 52.3 to 77.3%. This interpretable three-feature model demonstrated promising performance across cohorts and may serve as an adjunctive tool for preoperative MRI assessment, although prospective multicenter validation is warranted.
- 14 July 2026
- Published by Bentham Science Publishers Ltd. in Current Medical Imaging
- Vol. 22 (1) , 1-12
Magnetic Resonance Spectrum Imaging of Endometrial Lesions: A Diagnostic Performance Comparative Study with Pathological Findings
Objective: To evaluate the diagnostic value and significance of magnetic resonance spectroscopy (MRS) in endometrial pathology Methods: In this retrospective study, the 1H-MRS records of patients with suspected abnormal endometrial conditions were analyzed. Single-voxel spectra acquired from the largest region of the lesions were evaluated. The peak heights and areas under the curves (AUCs) for different metabolic peaks (total choline (tCho), creatine (Cr), lactate, lipid, N-acetylaspartate (NAA)) were recorded. The patients were classified into two groups: benign (hyperplasia or polyp) and malignant. Statistical differences in each peak between the two groups were analyzed individually. Results: Tumor lesions showed higher tCho (p=0.006), Cr(p=0.007), and lipid peaks (p=0.042) than benign endometrial lesions, together with increased tCho/Cr (p=0.021) and tCho/NAA ratios (p=0.016). In endometrial cancer, higher tCho(AUC=0.64), lipid (AUC=0.618), and tCho/Cr (AUC=0.649) were associated with higher tumor grade. Lipid and tCho/NAA also demonstrated value for differentiating endometrioid from nonendometrioid subtypes, with AUCs of 0.761 and 0.756, respectively. Lactate and an NAA-like peak were detectable in a subset of endometrial lesions. Multivariable logistic regression revealed that the incorporation of the proton MRS tCho metric into a baseline clinico-radiological model (age, body mass index (BMI), cancer antigen 125 (CA125)) significantly enhanced diagnostic performance (AUC = 0.797). Discussion: tCho- and lipid-related MRS metrics may aid non-invasive characterization of malignancy and tumor aggressiveness, supporting preoperative subtype assessment and risk stratification. Conclusion: MRS reflects the basic metabolic characteristics of tissues and aids in determining the nature and invasiveness of endometrial lesions.
- 14 July 2026
- Published by Bentham Science Publishers Ltd. in Current Medical Imaging
- Vol. 22 (1) , 1-12
Magnetic Resonance Spectrum Imaging of Endometrial Lesions: A Diagnostic Performance Comparative Study with Pathological Findings
Objective: To evaluate the diagnostic value and significance of magnetic resonance spectroscopy (MRS) in endometrial pathology Methods: In this retrospective study, the 1H-MRS records of patients with suspected abnormal endometrial conditions were analyzed. Single-voxel spectra acquired from the largest region of the lesions were evaluated. The peak heights and areas under the curves (AUCs) for different metabolic peaks (total choline (tCho), creatine (Cr), lactate, lipid, N-acetylaspartate (NAA)) were recorded. The patients were classified into two groups: benign (hyperplasia or polyp) and malignant. Statistical differences in each peak between the two groups were analyzed individually. Results: Tumor lesions showed higher tCho (p=0.006), Cr(p=0.007), and lipid peaks (p=0.042) than benign endometrial lesions, together with increased tCho/Cr (p=0.021) and tCho/NAA ratios (p=0.016). In endometrial cancer, higher tCho(AUC=0.64), lipid (AUC=0.618), and tCho/Cr (AUC=0.649) were associated with higher tumor grade. Lipid and tCho/NAA also demonstrated value for differentiating endometrioid from nonendometrioid subtypes, with AUCs of 0.761 and 0.756, respectively. Lactate and an NAA-like peak were detectable in a subset of endometrial lesions. Multivariable logistic regression revealed that the incorporation of the proton MRS tCho metric into a baseline clinico-radiological model (age, body mass index (BMI), cancer antigen 125 (CA125)) significantly enhanced diagnostic performance (AUC = 0.797). Discussion: tCho- and lipid-related MRS metrics may aid non-invasive characterization of malignancy and tumor aggressiveness, supporting preoperative subtype assessment and risk stratification. Conclusion: MRS reflects the basic metabolic characteristics of tissues and aids in determining the nature and invasiveness of endometrial lesions.
- Open Access
- 24 October 2025
- Published by Springer Nature in Archives of Gynecology and Obstetrics
- Vol. 312 (6) , 2279-2288
Comparative efficacy of MRI versus ultrasound in prenatal diagnosis of placenta accreta spectrum disorders: a single-tertiary center study
Aim This study aimed to compare the diagnostic accuracy of ultrasound (US) and magnetic resonance imaging (MRI) in the detection of placenta accreta spectrum (PAS) disorders in a single tertiary care setting. Materials and methods All eligible participants prospectively underwent both US and MRI examinations prior to delivery at our institution. Two experienced radiologists independently interpreted both the US and MRI findings. The interobserver agreement was assessed using the intraclass correlation coefficient (ICC) in a subset of twenty randomly selected participants. The diagnostic performance of each imaging modality was evaluated against the gold standard of surgical or pathological findings. Results In a cohort of 131 pregnancies, the diagnostic performance of MRI and US was found to be comparable, with an accuracy (ACC) of 0.718 for MRI and 0.702 for US, respectively. In the subgroup of placental accrete cases, MRI demonstrated modestly higher performance than US (p = 0.023), while in the normal, increta, or percreta groups, the performance was equivalent. The ICC for both MRI and US was moderate, with values of 0.663 and 0.600, respectively. The most common causes of misdiagnosis for US were challenging patient positioning for US acquisition, thickened placenta, and prior cesarean section surgery. For MRI, the primary reason was the morphological changes occurring between the time of scanning and delivery. Conclusion Our findings indicate that MRI provides incremental value in the assessment of suspected placental accrete cases. Overall, MRI and US showed equivalent diagnostic performance for the confirmation of PAS disorders.
- 24 October 2025
- Published by Springer Nature in Archives of Gynecology and Obstetrics
- Vol. 312 (6) , 2279-2288
Open AccessComparative efficacy of MRI versus ultrasound in prenatal diagnosis of placenta accreta spectrum disorders: a single-tertiary center study
Aim This study aimed to compare the diagnostic accuracy of ultrasound (US) and magnetic resonance imaging (MRI) in the detection of placenta accreta spectrum (PAS) disorders in a single tertiary care setting. Materials and methods All eligible participants prospectively underwent both US and MRI examinations prior to delivery at our institution. Two experienced radiologists independently interpreted both the US and MRI findings. The interobserver agreement was assessed using the intraclass correlation coefficient (ICC) in a subset of twenty randomly selected participants. The diagnostic performance of each imaging modality was evaluated against the gold standard of surgical or pathological findings. Results In a cohort of 131 pregnancies, the diagnostic performance of MRI and US was found to be comparable, with an accuracy (ACC) of 0.718 for MRI and 0.702 for US, respectively. In the subgroup of placental accrete cases, MRI demonstrated modestly higher performance than US (p = 0.023), while in the normal, increta, or percreta groups, the performance was equivalent. The ICC for both MRI and US was moderate, with values of 0.663 and 0.600, respectively. The most common causes of misdiagnosis for US were challenging patient positioning for US acquisition, thickened placenta, and prior cesarean section surgery. For MRI, the primary reason was the morphological changes occurring between the time of scanning and delivery. Conclusion Our findings indicate that MRI provides incremental value in the assessment of suspected placental accrete cases. Overall, MRI and US showed equivalent diagnostic performance for the confirmation of PAS disorders.
- 4 August 2025
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (2) , 1547-1557
Development and Validation of an Explainable MRI-Based Habitat Radiomics Model for Predicting p53-Abnormal Endometrial Cancer: A Multicentre Feasibility Study
We developed an MRI-based habitat radiomics model (HRM) to predict p53-abnormal (p53abn) molecular subtypes of endometrial cancer (EC). Patients with pathologically confirmed EC were retrospectively enrolled from three hospitals and categorized into a training cohort (n = 270), test cohort 1 (n = 70), and test cohort 2 (n = 154). The tumour was divided into habitat sub-regions using diffusion-weighted imaging (DWI) and contrast-enhanced (CE) images with the K-means algorithm. Radiomics features were extracted from T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), DWI, and CE images. Three machine learning classifiers-logistic regression, support vector machines, and random forests-were applied to develop predictive models for p53abn EC. Model performance was validated using receiver operating characteristic (ROC) curves, and the model with the best predictive performance was selected as the HRM. A whole-region radiomics model (WRM) was also constructed, and a clinical model (CM) with five clinical features was developed. The SHApley Additive ExPlanations (SHAP) method was used to explain the outputs of the models. DeLong's test evaluated and compared the performance across the cohorts. A total of 1920 habitat radiomics features were considered. Eight features were selected for the HRM, ten for the WRM, and three clinical features for the CM. The HRM achieved the highest AUC: 0.855 (training), 0.769 (test1), and 0.766 (test2). The AUCs of the WRM were 0.707 (training), 0.703 (test1), and 0.738 (test2). The AUCs of the CM were 0.709 (training), 0.641 (test1), and 0.665 (test2). The MRI-based HRM successfully predicted p53abn EC. The results indicate that habitat combined with machine learning, radiomics, and SHAP can effectively predict p53abn EC, providing clinicians with intuitive insights and interpretability regarding the impact of risk factors in the model.
- 4 August 2025
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (2) , 1547-1557
Development and Validation of an Explainable MRI-Based Habitat Radiomics Model for Predicting p53-Abnormal Endometrial Cancer: A Multicentre Feasibility Study
We developed an MRI-based habitat radiomics model (HRM) to predict p53-abnormal (p53abn) molecular subtypes of endometrial cancer (EC). Patients with pathologically confirmed EC were retrospectively enrolled from three hospitals and categorized into a training cohort (n = 270), test cohort 1 (n = 70), and test cohort 2 (n = 154). The tumour was divided into habitat sub-regions using diffusion-weighted imaging (DWI) and contrast-enhanced (CE) images with the K-means algorithm. Radiomics features were extracted from T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), DWI, and CE images. Three machine learning classifiers-logistic regression, support vector machines, and random forests-were applied to develop predictive models for p53abn EC. Model performance was validated using receiver operating characteristic (ROC) curves, and the model with the best predictive performance was selected as the HRM. A whole-region radiomics model (WRM) was also constructed, and a clinical model (CM) with five clinical features was developed. The SHApley Additive ExPlanations (SHAP) method was used to explain the outputs of the models. DeLong's test evaluated and compared the performance across the cohorts. A total of 1920 habitat radiomics features were considered. Eight features were selected for the HRM, ten for the WRM, and three clinical features for the CM. The HRM achieved the highest AUC: 0.855 (training), 0.769 (test1), and 0.766 (test2). The AUCs of the WRM were 0.707 (training), 0.703 (test1), and 0.738 (test2). The AUCs of the CM were 0.709 (training), 0.641 (test1), and 0.665 (test2). The MRI-based HRM successfully predicted p53abn EC. The results indicate that habitat combined with machine learning, radiomics, and SHAP can effectively predict p53abn EC, providing clinicians with intuitive insights and interpretability regarding the impact of risk factors in the model.
- 24 May 2025
- Published by Springer Nature in Abdominal Radiology
- Vol. 50 (12) , 6017-6030
Preoperative risk assessment of invasive endometrial cancer using MRI-based radiomics: a systematic review and meta-analysis
Objective Image-derived machine learning (ML) is a robust and growing field in diagnostic imaging systems for both clinicians and radiologists. Accurate preoperative radiological evaluation of the invasive ability of endometrial cancer (EC) can increase the degree of clinical benefit. The present study aimed to investigate the diagnostic performance of magnetic resonance imaging (MRI)-derived artificial intelligence for accurate preoperative assessment of the invasive risk. Methods The PubMed, Embase, Cochrane Library and Web of Science databases were searched, and pertinent English-language papers were collected. The pooled sensitivity, specificity, diagnostic odds ratio (DOR), and positive and negative likelihood ratios (PLR and NLR, respectively) of all the papers were calculated using Stata software. The results were plotted on a summary receiver operating characteristic (SROC) curve, publication bias and threshold effects were evaluated, and meta-regression and subgroup analyses were conducted to explore the possible causes of intratumoral heterogeneity. Results MRI-based radiomics revealed pooled sensitivity (SEN) and specificity (SPE) values of 0.85 and 0.82 for the prediction of high-grade EC; 0.80 and 0.85 for deep myometrial invasion (DMI); 0.85 and 0.73 for lymphovascular space invasion (LVSI); 0.79 and 0.85 for microsatellite instability (MSI); and 0.90 and 0.72 for lymph node metastasis (LNM), respectively. For LVSI prediction and high-grade histological analysis, meta-regression revealed that the image segmentation and MRI-based radiomics modeling contributed to heterogeneity (p = 0.003 and 0.04). Conclusion Through a systematic review and meta-analysis of the reported literature, preoperative MRI-derived ML could help clinicians accurately evaluate EC risk factors, potentially guiding individual treatment thereafter.
- 24 May 2025
- Published by Springer Nature in Abdominal Radiology
- Vol. 50 (12) , 6017-6030
Preoperative risk assessment of invasive endometrial cancer using MRI-based radiomics: a systematic review and meta-analysis
Objective Image-derived machine learning (ML) is a robust and growing field in diagnostic imaging systems for both clinicians and radiologists. Accurate preoperative radiological evaluation of the invasive ability of endometrial cancer (EC) can increase the degree of clinical benefit. The present study aimed to investigate the diagnostic performance of magnetic resonance imaging (MRI)-derived artificial intelligence for accurate preoperative assessment of the invasive risk. Methods The PubMed, Embase, Cochrane Library and Web of Science databases were searched, and pertinent English-language papers were collected. The pooled sensitivity, specificity, diagnostic odds ratio (DOR), and positive and negative likelihood ratios (PLR and NLR, respectively) of all the papers were calculated using Stata software. The results were plotted on a summary receiver operating characteristic (SROC) curve, publication bias and threshold effects were evaluated, and meta-regression and subgroup analyses were conducted to explore the possible causes of intratumoral heterogeneity. Results MRI-based radiomics revealed pooled sensitivity (SEN) and specificity (SPE) values of 0.85 and 0.82 for the prediction of high-grade EC; 0.80 and 0.85 for deep myometrial invasion (DMI); 0.85 and 0.73 for lymphovascular space invasion (LVSI); 0.79 and 0.85 for microsatellite instability (MSI); and 0.90 and 0.72 for lymph node metastasis (LNM), respectively. For LVSI prediction and high-grade histological analysis, meta-regression revealed that the image segmentation and MRI-based radiomics modeling contributed to heterogeneity (p = 0.003 and 0.04). Conclusion Through a systematic review and meta-analysis of the reported literature, preoperative MRI-derived ML could help clinicians accurately evaluate EC risk factors, potentially guiding individual treatment thereafter.
- Open Access
- 14 March 2025
- Published by Frontiers Media SA in Frontiers in Oncology
- Vol. 15, 1472250
The qualitative and quantitative characteristics of serous endometrial carcinoma on MRI: applying a novel nomogram for predicting an aggressive histological type
Objectives: To comprehensively describe MRI characteristics of serous endometrial carcinoma (SEC) and distinguish SEC from endometrioid endometrial carcinoma (EEC).Methods: We retrospectively recruited 62 patients from a tertiary center with pathologically proven endometrioid cancers (37 SEC and 25 EEC) as the training set. MRI image interpretation was blindly interpreted by two experienced radiologists with consensus reading. Both qualitative and quantitative characteristics on MRI were recorded case by case. Histological findings were retrieved from the hospital information system. Fifty-four samples (27 SEC and 27 EEC) from the external hospital were treated as the testing set.Results: The qualitative MRI characteristics had no statistical difference between the SEC and EEC groups in the training set. SEC more often invaded the deep myometrium than EEC (p = 0.03). The signal intensity (SI)T2Ratio, SIcontrastRatio, LesionareaRatio, and VolumeareaRatio in the SEC group were 1.35 ± 0.36, 0.77 ± 0.18, 0.25 ± 0.24, and 0.22 ± 0.26, respectively. The SIT2Ratio, SIcontrastRatio, and VolumeareaRatio showed statistically significant differences between SEC and EEC (p < 0.05). The highest discriminative index for distinguishing SEC from EEC was SIcontrastRatio with an area under the curve (AUC) of 0.7533 (95% CI: 0.627–0.878). A predictive nomogram achieved an AUC of 0.814 (95% CI: 0.614–0.968), a sensitivity of 1.0, and a specificity of 0.60 in the testing set.Conclusions: This study developed and validated a nomogram model to predict SEC patients based on clinical and quantitative MRI features, which can be used in distinguishing SEC from EEC.
- 14 March 2025
- Published by Frontiers Media SA in Frontiers in Oncology
- Vol. 15, 1472250
Open AccessThe qualitative and quantitative characteristics of serous endometrial carcinoma on MRI: applying a novel nomogram for predicting an aggressive histological type
Objectives: To comprehensively describe MRI characteristics of serous endometrial carcinoma (SEC) and distinguish SEC from endometrioid endometrial carcinoma (EEC).Methods: We retrospectively recruited 62 patients from a tertiary center with pathologically proven endometrioid cancers (37 SEC and 25 EEC) as the training set. MRI image interpretation was blindly interpreted by two experienced radiologists with consensus reading. Both qualitative and quantitative characteristics on MRI were recorded case by case. Histological findings were retrieved from the hospital information system. Fifty-four samples (27 SEC and 27 EEC) from the external hospital were treated as the testing set.Results: The qualitative MRI characteristics had no statistical difference between the SEC and EEC groups in the training set. SEC more often invaded the deep myometrium than EEC (p = 0.03). The signal intensity (SI)T2Ratio, SIcontrastRatio, LesionareaRatio, and VolumeareaRatio in the SEC group were 1.35 ± 0.36, 0.77 ± 0.18, 0.25 ± 0.24, and 0.22 ± 0.26, respectively. The SIT2Ratio, SIcontrastRatio, and VolumeareaRatio showed statistically significant differences between SEC and EEC (p < 0.05). The highest discriminative index for distinguishing SEC from EEC was SIcontrastRatio with an area under the curve (AUC) of 0.7533 (95% CI: 0.627–0.878). A predictive nomogram achieved an AUC of 0.814 (95% CI: 0.614–0.968), a sensitivity of 1.0, and a specificity of 0.60 in the testing set.Conclusions: This study developed and validated a nomogram model to predict SEC patients based on clinical and quantitative MRI features, which can be used in distinguishing SEC from EEC.
- Open Access
- 10 March 2025
- Published by Bentham Science Publishers Ltd. in Current Medical Imaging
- Vol. 21, 1-6
Intracranial Structural Malformations in Children in Tibet: CT and MRI Findings in a Single Tertiary Center
Objectives: The objective of this study was to summarize the findings of children’s intracranial congenital or developmental malformations found during imaging procedures in the Tibetan plateau. Methods: We retrospectively reviewed the imaging data of the suspected patients who presented with the central nervous system (CNS) malformations and were enrolled either through the clinic or after ultrasound examinations between June 2019 and June 2023 in our institution. All imaging data were interpreted by two experienced radiologists through consensus reading. Results: In this study, we recruited 36 patients, including two neonates, 17 infants and 17 children. Seven cases underwent an MRI examination, while the others had a CT scan. Polygyria and pachygyria malformation were the most common type of congenital neurological malformations (7 cases, 31.8%), followed by cystic changes of the cerebral parenchyma (3 cases, 13.6%). Cerebral atrophy was the most common type of secondary CNS abnormality(8 cases, 57.1%), followed by communicative hydrocephalus (3 cases, 21.4%). Five patients in the congenital group and 4 patients in the secondary group had complex malformations. In the current study group, there were 8 deaths, 12 cases with neurological sequelae, 1 case with normal development, and 15 cases lost to follow-up. There were no significant differences between the primary and secondary CNS groups in terms of the outcome for both the infants and children groups. Conclusions: CNS malformations in the Tibetan Plateau are associated with high mortality and morbidity rates. Better utilization of imaging modalities could help design tailored treatments as early as possible.
- 10 March 2025
- Published by Bentham Science Publishers Ltd. in Current Medical Imaging
- Vol. 21, 1-6
Open AccessIntracranial Structural Malformations in Children in Tibet: CT and MRI Findings in a Single Tertiary Center
Objectives: The objective of this study was to summarize the findings of children’s intracranial congenital or developmental malformations found during imaging procedures in the Tibetan plateau. Methods: We retrospectively reviewed the imaging data of the suspected patients who presented with the central nervous system (CNS) malformations and were enrolled either through the clinic or after ultrasound examinations between June 2019 and June 2023 in our institution. All imaging data were interpreted by two experienced radiologists through consensus reading. Results: In this study, we recruited 36 patients, including two neonates, 17 infants and 17 children. Seven cases underwent an MRI examination, while the others had a CT scan. Polygyria and pachygyria malformation were the most common type of congenital neurological malformations (7 cases, 31.8%), followed by cystic changes of the cerebral parenchyma (3 cases, 13.6%). Cerebral atrophy was the most common type of secondary CNS abnormality(8 cases, 57.1%), followed by communicative hydrocephalus (3 cases, 21.4%). Five patients in the congenital group and 4 patients in the secondary group had complex malformations. In the current study group, there were 8 deaths, 12 cases with neurological sequelae, 1 case with normal development, and 15 cases lost to follow-up. There were no significant differences between the primary and secondary CNS groups in terms of the outcome for both the infants and children groups. Conclusions: CNS malformations in the Tibetan Plateau are associated with high mortality and morbidity rates. Better utilization of imaging modalities could help design tailored treatments as early as possible.
- Open Access
- 22 November 2024
- Published by JMIR Publications Inc. in JMIR Cancer
- Vol. 11 (1) , e69057
Machine Learning for Preoperative Assessment and Postoperative Prediction in Cervical Cancer: A Multicenter Retrospective Model Integrating MRI and Clinicopathological Data (Preprint)
Background: Machine learning (ML) has been increasingly applied to cervical cancer (CC) research. However, few studies have combined both clinical parameters and imaging data. Additionally, there remains an urgent need for more robust and accurate preoperative assessment of parametrial invasion and lymph node metastasis, as well as postoperative prognosis prediction. Objective: To develop an integrated ML model combining clinicopathological variables and MR images features for: (1) preoperative detection of parametrial invasion and lymph node metastasis; and (2) postoperative recurrence and survival prediction. Methods: Retrospective data from 250 patients with CC (2014-2022; at two tertiary hospitals) were analyzed. Variables were assessed for their predictive value regarding parametrial invasion, lymph node metastasis, survival, and recurrence using seven ML models: K-nearest neighbors (KNN), Support vector machine (SVM), Decision Tree (DT), Random forest (RF), Balanced RF, Weighted DT, and Weighted KNN. Performance was assessed via 5-fold cross-validation using accuracy, sensitivity, specificity, precision, F1-score, and area under the receiver operating characteristic curve (AUC). The optimal models were deployed in an AI-assisted contouring and prognosis prediction system. Results: Among 250 women, there were 11 deaths and 24 recurrences. (1) For preoperative evaluation, the integrated model using balanced RF achieved optimal performance (sensitivity 0.81, specificity 0.85) for parametrial invasion, while weighted KNN achieved best performance for lymph node metastasis (sensitivity 0.98, AUC 0.72). (2) For postoperative prognosis, weighted KNN also demonstrated high accuracy for recurrence (accuracy 0.94, AUC 0.86) and mortality (accuracy 0.97, AUC 0.77), with relatively balanced sensitivity of 0.80 and 0.33, respectively. (3) An AI-assisted contouring and prognosis prediction system was developed to support preoperative evaluation, and postoperative prognosis prediction. Conclusions: The integration of clinical data and MR images provides enhanced diagnostic capability for preoperatively detection of parametrial invasion and lymph node metastasis detection, as well as prognostic capability for predicting recurrence and mortality for CC, facilitating personalized, precise treatment strategies.
- 22 November 2024
- Published by JMIR Publications Inc. in JMIR Cancer
- Vol. 11 (1) , e69057
Open AccessMachine Learning for Preoperative Assessment and Postoperative Prediction in Cervical Cancer: A Multicenter Retrospective Model Integrating MRI and Clinicopathological Data (Preprint)
Background: Machine learning (ML) has been increasingly applied to cervical cancer (CC) research. However, few studies have combined both clinical parameters and imaging data. Additionally, there remains an urgent need for more robust and accurate preoperative assessment of parametrial invasion and lymph node metastasis, as well as postoperative prognosis prediction. Objective: To develop an integrated ML model combining clinicopathological variables and MR images features for: (1) preoperative detection of parametrial invasion and lymph node metastasis; and (2) postoperative recurrence and survival prediction. Methods: Retrospective data from 250 patients with CC (2014-2022; at two tertiary hospitals) were analyzed. Variables were assessed for their predictive value regarding parametrial invasion, lymph node metastasis, survival, and recurrence using seven ML models: K-nearest neighbors (KNN), Support vector machine (SVM), Decision Tree (DT), Random forest (RF), Balanced RF, Weighted DT, and Weighted KNN. Performance was assessed via 5-fold cross-validation using accuracy, sensitivity, specificity, precision, F1-score, and area under the receiver operating characteristic curve (AUC). The optimal models were deployed in an AI-assisted contouring and prognosis prediction system. Results: Among 250 women, there were 11 deaths and 24 recurrences. (1) For preoperative evaluation, the integrated model using balanced RF achieved optimal performance (sensitivity 0.81, specificity 0.85) for parametrial invasion, while weighted KNN achieved best performance for lymph node metastasis (sensitivity 0.98, AUC 0.72). (2) For postoperative prognosis, weighted KNN also demonstrated high accuracy for recurrence (accuracy 0.94, AUC 0.86) and mortality (accuracy 0.97, AUC 0.77), with relatively balanced sensitivity of 0.80 and 0.33, respectively. (3) An AI-assisted contouring and prognosis prediction system was developed to support preoperative evaluation, and postoperative prognosis prediction. Conclusions: The integration of clinical data and MR images provides enhanced diagnostic capability for preoperatively detection of parametrial invasion and lymph node metastasis detection, as well as prognostic capability for predicting recurrence and mortality for CC, facilitating personalized, precise treatment strategies.
- 25 June 2024
- Published by Springer Nature in Abdominal Radiology
- Vol. 49 (11) , 4140-4150
Both intra- and peri-tumoral radiomics signatures can be used to predict lymphatic vascular space invasion and lymphatic metastasis positive status from endometrial cancer MR imaging
Objectives To identify lymphatic vascular space invasion (LVSI) and lymphatic node metastasis (LNM) status of endometrial cancer (EC) patients, using radiomics based on MRI images. Methods Five hundred and ninety-eight EC patients between January 2015 and September 2020 from two institutions were retrospectively included. Tumoral regions on DWI, T1CE, and T2W images were manually outlined. Radiomics features were extracted from tumor region and peri-tumor region of different thicknesses. We established sub-models to select features from each smaller category. Using this method, we separately constructed radiomic signatures for intra-tumoral and peri-tumoral images using different sequences. We constructed intra-tumoral and peri-tumoral models by combining their features, and a multi-sequence model by combining logits. Models were trained with 397 patients and validated with 170 internal and 31 external patients. Results For LVSI positive/LNM positive status identification, the multi-parameter MRI radiomics model achieved the area under curve (AUC) values of 0.771 (95%CI: [0.692–0.849])/0.801 (95%CI: [0.704, 0.898]) and 0.864 (95%CI: [0.728–1.000])/0.976 (95%CI: [0.919, 1.000]) in internal and external test cohorts, respectively. Conclusions Intra-tumoral and peri-tumoral radiomics signatures based on mpMRI can both be used to identify LVSI or LNM status in EC patients non-invasively. Further studies on LVSI and LNM should pay attention to both of them. Graphical Abstract
- 25 June 2024
- Published by Springer Nature in Abdominal Radiology
- Vol. 49 (11) , 4140-4150
Both intra- and peri-tumoral radiomics signatures can be used to predict lymphatic vascular space invasion and lymphatic metastasis positive status from endometrial cancer MR imaging
Objectives To identify lymphatic vascular space invasion (LVSI) and lymphatic node metastasis (LNM) status of endometrial cancer (EC) patients, using radiomics based on MRI images. Methods Five hundred and ninety-eight EC patients between January 2015 and September 2020 from two institutions were retrospectively included. Tumoral regions on DWI, T1CE, and T2W images were manually outlined. Radiomics features were extracted from tumor region and peri-tumor region of different thicknesses. We established sub-models to select features from each smaller category. Using this method, we separately constructed radiomic signatures for intra-tumoral and peri-tumoral images using different sequences. We constructed intra-tumoral and peri-tumoral models by combining their features, and a multi-sequence model by combining logits. Models were trained with 397 patients and validated with 170 internal and 31 external patients. Results For LVSI positive/LNM positive status identification, the multi-parameter MRI radiomics model achieved the area under curve (AUC) values of 0.771 (95%CI: [0.692–0.849])/0.801 (95%CI: [0.704, 0.898]) and 0.864 (95%CI: [0.728–1.000])/0.976 (95%CI: [0.919, 1.000]) in internal and external test cohorts, respectively. Conclusions Intra-tumoral and peri-tumoral radiomics signatures based on mpMRI can both be used to identify LVSI or LNM status in EC patients non-invasively. Further studies on LVSI and LNM should pay attention to both of them. Graphical Abstract