Xiuming Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xiuming Wu has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Peizhong Liu
College of Engineering, Huaqiao University, Quanzhou, China
3 shared publications - Guorong Lyu
Department of Ultrasound, Second Affiliated Hospital of Fujian Medical University, Quanzhou, China
2 shared publications - Jiansong Zhang
School of Computer and Software, Shenzhen University, Shenzhen, China
2 shared publications - Shunlan Liu
Department of Ultrasound, the Second Affiliated Hospital of Fujian Medical University, Quanzhou 362000, China
2 shared publications - Yongjian Chen
Department of Ultrasound, The Second Affiliated Hospital of Fujian Medical University, Quanzhou, China
2 shared publications - Bingtian Dong
Department of Ultrasound, The First Affiliated Hospital of Anhui Medical University, Hefei, China
1 shared publication
Latest Publications
- Open Access
- 31 March 2026
- Published by MDPI AG in Bioengineering
- Vol. 13 (4) , 403
SEG-FAUSP: Anatomical Structure Segmentation of the Standard Sections of Fetal Abdominal Ultrasounds
This study addresses the challenge of the difficult identification of organ structures in the standard sections of fetal abdominal ultrasounds. A deep learning-based multi-task model named SEG-FAUSP was developed to segment the core anatomical structures of seven key fetal abdominal ultrasound sections. We collected fetal abdominal ultrasound images from pregnant women in the mid-pregnancy period (18–24 weeks) using various mainstream ultrasound devices, and professional physicians annotated key anatomical structures (e.g., umbilical veins, gastric bubbles, spine) in the images. Based on an improved deep learning framework, the model accurately segments and locates the target organ structures through a parallel dual-branch semantic segmentation network, which avoids the over-reliance on large-scale pre-trained data in traditional methods. Experimental results show that the model achieves excellent performance in anatomical structure segmentation, with the intersection over union of the bladder and gastric bubble both reaching above 0.84; its segmentation accuracy for complex structures such as the inferior vena cava is also significantly superior to the baseline model. As an end-to-end model, it simplifies the clinical interpretation process of fetal abdominal ultrasound, reduces the risk of missed diagnoses caused by unclear organ identification, provides an efficient auxiliary tool for prenatal screening in grassroots medical institutions, and is of great significance for improving the quality of newborns.
- 31 March 2026
- Published by MDPI AG in Bioengineering
- Vol. 13 (4) , 403
Open AccessSEG-FAUSP: Anatomical Structure Segmentation of the Standard Sections of Fetal Abdominal Ultrasounds
This study addresses the challenge of the difficult identification of organ structures in the standard sections of fetal abdominal ultrasounds. A deep learning-based multi-task model named SEG-FAUSP was developed to segment the core anatomical structures of seven key fetal abdominal ultrasound sections. We collected fetal abdominal ultrasound images from pregnant women in the mid-pregnancy period (18–24 weeks) using various mainstream ultrasound devices, and professional physicians annotated key anatomical structures (e.g., umbilical veins, gastric bubbles, spine) in the images. Based on an improved deep learning framework, the model accurately segments and locates the target organ structures through a parallel dual-branch semantic segmentation network, which avoids the over-reliance on large-scale pre-trained data in traditional methods. Experimental results show that the model achieves excellent performance in anatomical structure segmentation, with the intersection over union of the bladder and gastric bubble both reaching above 0.84; its segmentation accuracy for complex structures such as the inferior vena cava is also significantly superior to the baseline model. As an end-to-end model, it simplifies the clinical interpretation process of fetal abdominal ultrasound, reduces the risk of missed diagnoses caused by unclear organ identification, provides an efficient auxiliary tool for prenatal screening in grassroots medical institutions, and is of great significance for improving the quality of newborns.
- Open Access
OTC-NET: A Multimodal Method for Accurate Diagnosis of Ovarian Cancer in O-RADS Category 4 Masses
Background: Ovarian cancer is the deadliest female reproductive malignancy. Accurate preoperative differentiation of benign and malignant ovarian masses is critical for appropriate treatment. O-RADS category 4 lesions present a wide range of malignant risk, challenging radiologists. Ultrasonic images are the primary focus of current deep learning models, with no consideration for clinical data. Methods: We proposed OTC-NET, a model that uses multimodal data for classification, which combines ultrasound images and clinical information to improve the classification ability of O-RADS 4 ovarian masses. Results: OTC-NET outperforms seven deep learning models and three radiologists of varying experience, with AUC significantly higher than junior (p < 0.001), intermediate (p < 0.01), and senior (p < 0.05) radiologists. Additionally, OTC-NET–assisted diagnosis notably improves AUC and accuracy of junior and senior radiologists (p < 0.05). Conclusions: These results indicate that OTC-NET provides superior diagnostic accuracy and has strong potential for clinical application.
Open AccessOTC-NET: A Multimodal Method for Accurate Diagnosis of Ovarian Cancer in O-RADS Category 4 Masses
Background: Ovarian cancer is the deadliest female reproductive malignancy. Accurate preoperative differentiation of benign and malignant ovarian masses is critical for appropriate treatment. O-RADS category 4 lesions present a wide range of malignant risk, challenging radiologists. Ultrasonic images are the primary focus of current deep learning models, with no consideration for clinical data. Methods: We proposed OTC-NET, a model that uses multimodal data for classification, which combines ultrasound images and clinical information to improve the classification ability of O-RADS 4 ovarian masses. Results: OTC-NET outperforms seven deep learning models and three radiologists of varying experience, with AUC significantly higher than junior (p < 0.001), intermediate (p < 0.01), and senior (p < 0.05) radiologists. Additionally, OTC-NET–assisted diagnosis notably improves AUC and accuracy of junior and senior radiologists (p < 0.05). Conclusions: These results indicate that OTC-NET provides superior diagnostic accuracy and has strong potential for clinical application.
- 13 May 2025
- Published by Springer Nature in Pediatric Radiology
- Vol. 55 (11) , 2428-2441
Prenatal ultrasound prediction of coarctation of the aorta: a nomogram for risk stratification
Background Coarctation of the aorta is one of the most common congenital heart malformations, accounting for approximately 7% of all live births with congenital heart disease. It is crucial to make a definitive prenatal diagnosis as it can inform clinical treatment decisions. Objective The diagnostic criteria for coarctation of the aorta are still controversial, and there is currently no risk nomogram available to assess the probability of coarctation of the aorta using routine ultrasound parameters. We explored the prenatal diagnostic efficacy of ultrasound parameters and established a nomogram for coarctation of the aorta. Materials and methods A total of 101 fetuses with suspected coarctation of the aorta diagnosed by prenatal ultrasound from July 2015 to June 2023 were collected retrospectively. The patients were divided into two groups according to the diagnostic results: a normal group (n=42; gestational weeks, 28.5±6.0) and a coarctation of the aorta group (n=59; gestational weeks, 26.7±5.1). Univariate and multivariate logistic regression analyses were used to identify echocardiographic predictors of coarctation of the aorta. Moreover, the patients were divided into a training set and a validation set in a ratio of 8:2, and a nomogram for the prenatal diagnosis of coarctation of the aorta was established using R. Results (1) Aortic isthmus, aortic isthmus z-score, ascending aorta, ascending aorta z-score, pulmonary artery, pulmonary artery z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia were the predictive markers of coarctation of the aorta in the univariate logistic regression analysis (Pz-score, ascending aorta z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia were identified as the final predictors after multivariate logistic regression analysis (Pz-score, ascending aorta z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia, demonstrated a larger area under the receiver operating characteristic curve (AUC) (AUC=0.96, sensitivity=93.22%, specificity=88.10%) than aortic isthmus z-score alone (AUC=0.77, sensitivity=77.97%, specificity=71.43%), ascending aorta z-score alone (AUC=0.78, sensitivity=54.24%, specificity=90.48%), pulmonary artery/ascending aorta ratio alone (AUC=0.68, sensitivity=72.88%, specificity=54.76%), aortic arch dysplasia alone (AUC=0.70, sensitivity=66.10%, specificity=73.81%), and persistent left superior vena cava alone (AUC=0.72, sensitivity=79.66%, specificity=64.29%). The nomogram, which was constructed with these parameters, also exhibited excellent calibration curves and a good decision curve analysis curve. Conclusions The nomogram established by aortic isthmus z-score, ascending aorta z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia demonstrated excellent efficacy in the prenatal diagnosis of coarctation of the aorta. Graphical Abstract
- 13 May 2025
- Published by Springer Nature in Pediatric Radiology
- Vol. 55 (11) , 2428-2441
Prenatal ultrasound prediction of coarctation of the aorta: a nomogram for risk stratification
Background Coarctation of the aorta is one of the most common congenital heart malformations, accounting for approximately 7% of all live births with congenital heart disease. It is crucial to make a definitive prenatal diagnosis as it can inform clinical treatment decisions. Objective The diagnostic criteria for coarctation of the aorta are still controversial, and there is currently no risk nomogram available to assess the probability of coarctation of the aorta using routine ultrasound parameters. We explored the prenatal diagnostic efficacy of ultrasound parameters and established a nomogram for coarctation of the aorta. Materials and methods A total of 101 fetuses with suspected coarctation of the aorta diagnosed by prenatal ultrasound from July 2015 to June 2023 were collected retrospectively. The patients were divided into two groups according to the diagnostic results: a normal group (n=42; gestational weeks, 28.5±6.0) and a coarctation of the aorta group (n=59; gestational weeks, 26.7±5.1). Univariate and multivariate logistic regression analyses were used to identify echocardiographic predictors of coarctation of the aorta. Moreover, the patients were divided into a training set and a validation set in a ratio of 8:2, and a nomogram for the prenatal diagnosis of coarctation of the aorta was established using R. Results (1) Aortic isthmus, aortic isthmus z-score, ascending aorta, ascending aorta z-score, pulmonary artery, pulmonary artery z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia were the predictive markers of coarctation of the aorta in the univariate logistic regression analysis (Pz-score, ascending aorta z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia were identified as the final predictors after multivariate logistic regression analysis (Pz-score, ascending aorta z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia, demonstrated a larger area under the receiver operating characteristic curve (AUC) (AUC=0.96, sensitivity=93.22%, specificity=88.10%) than aortic isthmus z-score alone (AUC=0.77, sensitivity=77.97%, specificity=71.43%), ascending aorta z-score alone (AUC=0.78, sensitivity=54.24%, specificity=90.48%), pulmonary artery/ascending aorta ratio alone (AUC=0.68, sensitivity=72.88%, specificity=54.76%), aortic arch dysplasia alone (AUC=0.70, sensitivity=66.10%, specificity=73.81%), and persistent left superior vena cava alone (AUC=0.72, sensitivity=79.66%, specificity=64.29%). The nomogram, which was constructed with these parameters, also exhibited excellent calibration curves and a good decision curve analysis curve. Conclusions The nomogram established by aortic isthmus z-score, ascending aorta z-score, pulmonary artery/ascending aorta ratio, persistent left superior vena cava, and aortic arch dysplasia demonstrated excellent efficacy in the prenatal diagnosis of coarctation of the aorta. Graphical Abstract