Mengjun Shen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Mengjun Shen has a total of 60 co-authors. Here are the top co-authors by number of shared publications:
- Yin Wang
Department of Ultrasound, Shanghai Pulmonary Hospital, School of Medicine, Tongji University, Shanghai, China
5 shared publications Jon D. Blumenfeld
Department of Medicine, Weill Cornell Medicine
3 shared publications- Martin R. Prince
Department of Radiology, Weill Cornell Medicine, New York, NY
3 shared publications Chenglin Zhu
Department of Radiology, Weill Cornell Medicine, New York, NY
2 shared publications- Long Zhao
Department of Nuclear Medicine, Shanghai Pulmonary Hospital, School of Medicine, Tongji University, Shanghai, China
2 shared publications - Qing Xiong
Department of Radiology Weill Cornell Medicine New York New York USA
2 shared publications
Latest Publications
- 2 July 2026
- Published by Wolters Kluwer Health in Journal of Computer Assisted Tomography
- Vol. 50 (6)
Diagnostic Performance of Routine Abdominal MRI for Detecting Left Ventricular Hypertrophy in ADPKD
Objective: To evaluate the diagnostic performance of left ventricular (LV) volumetric parameters derived from routine abdominal MRI for detecting left ventricular hypertrophy (MRI-LVH) in patients with autosomal dominant polycystic kidney disease (ADPKD) who were previously diagnosed with LVH by echocardiography (Echo-LVH). Methods: This retrospective study reviewed 156 ADPKD patients (27 with LVH). The LV wall (LVW) and cavity were manually segmented on T2-weighted abdominal MRI images to calculate 3 height-adjusted parameters: LV wall volume (ht-LVWV), LV cavity volume (ht-LVCV), and total LV volume (ht-LVV). Diagnostic performance was assessed using receiver operating characteristic curve analysis, with Echo-LVH as the reference standard. Subgroup analyses were performed across sex, chronic kidney disease stage, Mayo Imaging Classification, and blood pressure status. Sensitivity analysis was adjusted for body surface area, age, sex, and systolic blood pressure. Results: Ht-LVWV demonstrated high diagnostic accuracy for Echo-LVH, with an area under the curve (AUC) of 0.82 (95% CI: 0.74-0.90), sensitivity=92%, and specificity=64% at a threshold of 60 mL/m. Ht-LVV also showed good performance (AUC=0.70), while ht-LVCV had inadequate diagnostic performance (AUC=0.61). Ht-LVWV maintained robust performance across all clinical subgroups. After adjusting for confounders, the AUC for ht-LVWV improved to 0.92 (95% CI: 0.85-0.98). Conclusion: Ht-LVWV measured on routine abdominal MRI is a promising biomarker for detecting LVH in ADPKD patients, with performance independent of key clinical variables. These findings support the potential for a diagnosis of MRI-LVH from standard ADPKD abdominal imaging protocols that are obtained routinely in the evaluation of ADPKD for renal and cardiovascular risk assessment in ADPKD.
- 2 July 2026
- Published by Wolters Kluwer Health in Journal of Computer Assisted Tomography
- Vol. 50 (6)
Diagnostic Performance of Routine Abdominal MRI for Detecting Left Ventricular Hypertrophy in ADPKD
Objective: To evaluate the diagnostic performance of left ventricular (LV) volumetric parameters derived from routine abdominal MRI for detecting left ventricular hypertrophy (MRI-LVH) in patients with autosomal dominant polycystic kidney disease (ADPKD) who were previously diagnosed with LVH by echocardiography (Echo-LVH). Methods: This retrospective study reviewed 156 ADPKD patients (27 with LVH). The LV wall (LVW) and cavity were manually segmented on T2-weighted abdominal MRI images to calculate 3 height-adjusted parameters: LV wall volume (ht-LVWV), LV cavity volume (ht-LVCV), and total LV volume (ht-LVV). Diagnostic performance was assessed using receiver operating characteristic curve analysis, with Echo-LVH as the reference standard. Subgroup analyses were performed across sex, chronic kidney disease stage, Mayo Imaging Classification, and blood pressure status. Sensitivity analysis was adjusted for body surface area, age, sex, and systolic blood pressure. Results: Ht-LVWV demonstrated high diagnostic accuracy for Echo-LVH, with an area under the curve (AUC) of 0.82 (95% CI: 0.74-0.90), sensitivity=92%, and specificity=64% at a threshold of 60 mL/m. Ht-LVV also showed good performance (AUC=0.70), while ht-LVCV had inadequate diagnostic performance (AUC=0.61). Ht-LVWV maintained robust performance across all clinical subgroups. After adjusting for confounders, the AUC for ht-LVWV improved to 0.92 (95% CI: 0.85-0.98). Conclusion: Ht-LVWV measured on routine abdominal MRI is a promising biomarker for detecting LVH in ADPKD patients, with performance independent of key clinical variables. These findings support the potential for a diagnosis of MRI-LVH from standard ADPKD abdominal imaging protocols that are obtained routinely in the evaluation of ADPKD for renal and cardiovascular risk assessment in ADPKD.
- 9 May 2026
- Published by Wiley in Journal of Magnetic Resonance Imaging
- Vol. 64 (3) , 864-866
Cardio‐Thoracic Ratio for Detecting Left Ventricular Hypertrophy on Abdominal MRI in ADPKD
- 9 May 2026
- Published by Wiley in Journal of Magnetic Resonance Imaging
- Vol. 64 (3) , 864-866
Cardio‐Thoracic Ratio for Detecting Left Ventricular Hypertrophy on Abdominal MRI in ADPKD
- 2 March 2026
- Published by Elsevier in Lung Cancer
- Vol. 215, 109353
18F-FDG PET/CT and receptor-positive circulating tumor cells-based machine learning model for predicting poorly differentiated lung adenocarcinoma
- 2 March 2026
- Published by Elsevier in Lung Cancer
- Vol. 215, 109353
18F-FDG PET/CT and receptor-positive circulating tumor cells-based machine learning model for predicting poorly differentiated lung adenocarcinoma
- Open Access
- 4 November 2025
- Published by Wolters Kluwer Health in Journal of the American Society of Nephrology
- Vol. 37 (5) , 995-1009
Automatically Measuring Kidney, Liver, and Cyst Volumes in Autosomal Dominant Polycystic Kidney Disease
Key Points: TraceOrg is a web-based tool that automatically labels kidney, liver, and cysts, reporting volumes and Mayo Imaging Classification. External validation showed high performance and good generalizability on Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease, Polycystic Kidney Disease-Research Resource Consortium, and other external datasets. Training on multiple pulse sequences enables TraceOrg to process images from a wide variety of protocols. Background: Kidney, liver, and cyst volumes are important for diagnosis, classification, and management of autosomal dominant polycystic kidney disease (ADPKD) but challenging to measure accurately and reproducibly. Here, we develop a web-based deep learning platform to automatically and robustly measure kidneys, liver, and cyst volumes in ADPKD. Methods: Magnetic resonance imaging (MRI) and computed tomography scans from patients with ADPKD ( n =611) and participants without ADPKD ( n =109) were used to train a 3D hybrid model combining U-Net and transformer elements for segmenting kidneys, liver, and cysts. The model is implemented as a web-based calculator at www.traceorg.com, providing segmentation labels, volumes, and Mayo Clinic Image Classification. Automatic browser anonymization of digital imaging and communications in medicine images ensures privacy. Internal validation was conducted on 70 MRIs for kidney and liver segmentations and 46 MRIs for cyst segmentations, and performance was compared with five open access segmentation models (TotalSegmentator, MRAnnotator, Kim, Woznicki, and Gregory–Kline). External validation was performed on one single-center dataset ( n =58), one multicenter dataset ( n =73), Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease 2 (CRISP, n =30), and Polycystic Kidney Disease-Research Resource Consortium (PKD-RRC, n =115) MRIs with T2-weighted and T1-weighted images. Results: After training on 720 participants (mean age=48±15, eGFR=74±32 ml/min per 1.73 m 2 and height-adjusted total kidney volume=826±772 ml/m), TraceOrg internal validation performance achieved high mean Dice scores of 0.97 (kidneys), 0.97 (liver), 0.93 (kidney cysts), and 0.82 (liver cysts) outperforming existing models for ADPKD. External validation showed strong performance with Dice scores of 0.92–0.94 (kidney), 0.87–0.96 (liver), 0.85 (kidney cysts), and 0.76–0.90 (liver cysts) for the single-center dataset and 0.95 (kidney) and 0.81 (kidney cysts) for the multicenter dataset. Compared with CRISP volumes measured by stereology, the mean absolute percent difference was 5.3% (kidneys, n =30), 11% (kidney cysts, n =30), and 5.5% (liver, n =22). Compared with PKD-RRC ( n =115), the mean absolute percent difference in total kidney volume was 4.9%. Conclusions: TraceOrg, a publicly available web-based tool, automatically measured kidney, liver, and cyst volumes from abdominal MRI in ADPKD with high accuracy compared with manual segmentations. Podcast: This article contains a podcast at https://dts.podtrac.com/redirect.mp3/www.asn-online.org/media/podcast/JASN/2026_02_03_ASN0000000904.mp3
- 4 November 2025
- Published by Wolters Kluwer Health in Journal of the American Society of Nephrology
- Vol. 37 (5) , 995-1009
Open AccessAutomatically Measuring Kidney, Liver, and Cyst Volumes in Autosomal Dominant Polycystic Kidney Disease
Key Points: TraceOrg is a web-based tool that automatically labels kidney, liver, and cysts, reporting volumes and Mayo Imaging Classification. External validation showed high performance and good generalizability on Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease, Polycystic Kidney Disease-Research Resource Consortium, and other external datasets. Training on multiple pulse sequences enables TraceOrg to process images from a wide variety of protocols. Background: Kidney, liver, and cyst volumes are important for diagnosis, classification, and management of autosomal dominant polycystic kidney disease (ADPKD) but challenging to measure accurately and reproducibly. Here, we develop a web-based deep learning platform to automatically and robustly measure kidneys, liver, and cyst volumes in ADPKD. Methods: Magnetic resonance imaging (MRI) and computed tomography scans from patients with ADPKD ( n =611) and participants without ADPKD ( n =109) were used to train a 3D hybrid model combining U-Net and transformer elements for segmenting kidneys, liver, and cysts. The model is implemented as a web-based calculator at www.traceorg.com, providing segmentation labels, volumes, and Mayo Clinic Image Classification. Automatic browser anonymization of digital imaging and communications in medicine images ensures privacy. Internal validation was conducted on 70 MRIs for kidney and liver segmentations and 46 MRIs for cyst segmentations, and performance was compared with five open access segmentation models (TotalSegmentator, MRAnnotator, Kim, Woznicki, and Gregory–Kline). External validation was performed on one single-center dataset ( n =58), one multicenter dataset ( n =73), Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease 2 (CRISP, n =30), and Polycystic Kidney Disease-Research Resource Consortium (PKD-RRC, n =115) MRIs with T2-weighted and T1-weighted images. Results: After training on 720 participants (mean age=48±15, eGFR=74±32 ml/min per 1.73 m 2 and height-adjusted total kidney volume=826±772 ml/m), TraceOrg internal validation performance achieved high mean Dice scores of 0.97 (kidneys), 0.97 (liver), 0.93 (kidney cysts), and 0.82 (liver cysts) outperforming existing models for ADPKD. External validation showed strong performance with Dice scores of 0.92–0.94 (kidney), 0.87–0.96 (liver), 0.85 (kidney cysts), and 0.76–0.90 (liver cysts) for the single-center dataset and 0.95 (kidney) and 0.81 (kidney cysts) for the multicenter dataset. Compared with CRISP volumes measured by stereology, the mean absolute percent difference was 5.3% (kidneys, n =30), 11% (kidney cysts, n =30), and 5.5% (liver, n =22). Compared with PKD-RRC ( n =115), the mean absolute percent difference in total kidney volume was 4.9%. Conclusions: TraceOrg, a publicly available web-based tool, automatically measured kidney, liver, and cyst volumes from abdominal MRI in ADPKD with high accuracy compared with manual segmentations. Podcast: This article contains a podcast at https://dts.podtrac.com/redirect.mp3/www.asn-online.org/media/podcast/JASN/2026_02_03_ASN0000000904.mp3
- 8 March 2025
- Published by Elsevier in Clinical Radiology
- Vol. 85, 106867
Radiomics analysis of 18F-FDG PET/CT for visceral pleural invasion in non-small cell lung cancer with pleural attachment
- 8 March 2025
- Published by Elsevier in Clinical Radiology
- Vol. 85, 106867
Radiomics analysis of 18F-FDG PET/CT for visceral pleural invasion in non-small cell lung cancer with pleural attachment
- 26 July 2024
- Published by Elsevier in Ultrasound in Medicine & Biology
- Vol. 50 (10) , 1585-1592
B-mode Ultrasound Related Combined Predictive Indicators for Benign Versus Malignant Subpleural Pulmonary Lesions
- 26 July 2024
- Published by Elsevier in Ultrasound in Medicine & Biology
- Vol. 50 (10) , 1585-1592
B-mode Ultrasound Related Combined Predictive Indicators for Benign Versus Malignant Subpleural Pulmonary Lesions
- 13 August 2021
- Published by Wiley in Journal of Ultrasound in Medicine
- Vol. 41 (5) , 1147-1157
Application of Contrast‐Enhanced Ultrasound in the Differential Diagnosis of Benign and Malignant Subpleural Pulmonary Lesions
Objective: To explore the clinical value of contrast‐enhanced ultrasound (CEUS) in the differential diagnosis of benign and malignant subpleural pulmonary lesions (SPLs).Methods: Among 959 patients with SPLs who were scheduled to undergo ultrasound‐guided puncture in our department between January 2019 and June 2019, 506 patients were included and their B‐mode ultrasound and CEUS features, including the lesion's location, size, margin, echo, perfusion pattern of ultrasound contrast agent, degree of enhancement, homogeneity, vascular signs, and necrosis, were retrospectively investigated. All malignant cases were diagnosed by pathology, while benign cases were diagnosed by two respiratory physicians after comprehensive analysis of pathology, etiology, imaging, and clinical symptoms. Statistical differences in these features between the benign and malignant groups were then analyzed.Results: There were 506 cases in this study, including 219 benign cases and 287 malignant cases. Among them, 351 were males and 155 were females, with an average age of 59 ± 16 years. There were statistically significant differences between benign and malignant groups in the perfusion pattern, the degree of enhancement, and vascular signs. The features of the malignant group included local‐to‐whole perfusion pattern, hypo‐enhancement, and curly hair sign, while those of the benign group included a centrifugal perfusion pattern, iso‐enhancement and hyper‐enhancement, and dendritic sign. There was no statistically significant difference between the two groups in homogeneity and necrosis.Conclusions: CEUS enhancement mode is different between benign and malignant SPLs, which can provide supplementary information for the differential diagnosis of SPLs in the existing imaging diagnosis.
- 13 August 2021
- Published by Wiley in Journal of Ultrasound in Medicine
- Vol. 41 (5) , 1147-1157
Application of Contrast‐Enhanced Ultrasound in the Differential Diagnosis of Benign and Malignant Subpleural Pulmonary Lesions
Objective: To explore the clinical value of contrast‐enhanced ultrasound (CEUS) in the differential diagnosis of benign and malignant subpleural pulmonary lesions (SPLs).Methods: Among 959 patients with SPLs who were scheduled to undergo ultrasound‐guided puncture in our department between January 2019 and June 2019, 506 patients were included and their B‐mode ultrasound and CEUS features, including the lesion's location, size, margin, echo, perfusion pattern of ultrasound contrast agent, degree of enhancement, homogeneity, vascular signs, and necrosis, were retrospectively investigated. All malignant cases were diagnosed by pathology, while benign cases were diagnosed by two respiratory physicians after comprehensive analysis of pathology, etiology, imaging, and clinical symptoms. Statistical differences in these features between the benign and malignant groups were then analyzed.Results: There were 506 cases in this study, including 219 benign cases and 287 malignant cases. Among them, 351 were males and 155 were females, with an average age of 59 ± 16 years. There were statistically significant differences between benign and malignant groups in the perfusion pattern, the degree of enhancement, and vascular signs. The features of the malignant group included local‐to‐whole perfusion pattern, hypo‐enhancement, and curly hair sign, while those of the benign group included a centrifugal perfusion pattern, iso‐enhancement and hyper‐enhancement, and dendritic sign. There was no statistically significant difference between the two groups in homogeneity and necrosis.Conclusions: CEUS enhancement mode is different between benign and malignant SPLs, which can provide supplementary information for the differential diagnosis of SPLs in the existing imaging diagnosis.
- 1 October 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 40, 1580-1585
Stacked sparse autoencoder based fault detection and location method for modular five-level converters
This paper presents a novel method for fault detection and location in modular five-level converters (MFLC) based on stacked sparse autoencoder (SSAE). SSAE is composed of multiple SAE and a softmax classifier. The capacitor voltage signals of all sub-modules (SMs) in the MFLC circuit are combined into a multi-channel signal. By moving window along the multi-channel signal, a set of signal segments are acquired. Then these segments are flattened into vectors and used as SSAE model's input. The SAE can unsupervised learn features from the input data and the fault detection and location problem is converted into a classification problem. In order to improve the anti-noise performance of the proposed method, the stacked denoising SAE (SDSAE) is implemented. Results show that the proposed method can quickly and accurately detect and locate the faulty SM. Comparison with existing methods shows that the proposed method has higher robustness and generalization ability, and is practical for online MFLC protection.
- 1 October 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 40, 1580-1585
Stacked sparse autoencoder based fault detection and location method for modular five-level converters
This paper presents a novel method for fault detection and location in modular five-level converters (MFLC) based on stacked sparse autoencoder (SSAE). SSAE is composed of multiple SAE and a softmax classifier. The capacitor voltage signals of all sub-modules (SMs) in the MFLC circuit are combined into a multi-channel signal. By moving window along the multi-channel signal, a set of signal segments are acquired. Then these segments are flattened into vectors and used as SSAE model's input. The SAE can unsupervised learn features from the input data and the fault detection and location problem is converted into a classification problem. In order to improve the anti-noise performance of the proposed method, the stacked denoising SAE (SDSAE) is implemented. Results show that the proposed method can quickly and accurately detect and locate the faulty SM. Comparison with existing methods shows that the proposed method has higher robustness and generalization ability, and is practical for online MFLC protection.