Jianhua Shu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jianhua Shu has a total of 47 co-authors. Here are the top co-authors by number of shared publications:
- Jili Hu
School of Medical Informatics Engineering, Anhui University of Chinese Medicine, Hefei, China
2 shared publications Zhaohang Teng
School of Medical Information Engineering, Anhui University of Chinese Medicine
2 shared publicationsZongyun Gu
School of Medical Information Engineering, Anhui University of Chinese Medicine
2 shared publicationsChen Chen
School of Medical Information Engineering, Anhui University of Chinese Medicine, Hefei, China
1 shared publication- Chengle Li
School of Medical Informatics Engineering, Anhui University of Chinese Medicine, Hefei, China
1 shared publication - Chun Ma
School of Medical Informatics Engineering, Anhui University of Chinese Medicine, Hefei, China
1 shared publication
Latest Publications
- 9 October 2026
- Published by Springer Nature in Analytical and Bioanalytical Chemistry
- p. 1-13
Colorimetric discrimination of chrysanthemum polyphenols via tunable MOF-818 nanozyme sensor array
- 9 October 2026
- Published by Springer Nature in Analytical and Bioanalytical Chemistry
- p. 1-13
Colorimetric discrimination of chrysanthemum polyphenols via tunable MOF-818 nanozyme sensor array
- 24 August 2026
- Published by Springer Nature in Multimedia Systems
- Vol. 32 (8) , 539
Towards source-free domain adaptation: a model-heterogeneous perspective
- 24 August 2026
- Published by Springer Nature in Multimedia Systems
- Vol. 32 (8) , 539
Towards source-free domain adaptation: a model-heterogeneous perspective
- Open Access
Integrated Single-Cell Virtual Knockout and Machine Learning Analyses Reveal a Protective Role of CKAP2 in Gastric Cancer
Objective: To elucidate the role of cytoskeleton-associated protein 2 (CKAP2) in gastric cancer (GC) progression and evaluate its prognostic and potential protective significance. Methods: The candidate genes for GC were identified using differential expression and weighted gene co-expression network analysis (WGCNA). A panel of machine learning algorithms was applied for obtaining the key genes and yielding the hub target gene CKAP2 through protein–protein interaction (PPI) networks and single-cell RNA sequencing analysis. Further, the functional role of CKAP2 was explored through single-cell virtual knockout and pathway enrichment analysis. Meanwhile, the survival analysis and Mendelian randomization (MR) were used to evaluate the clinical relevance and causality of CKAP2. Finally, the expression of CKAP2 was further validated in clinical tissues and GC cell lines through Western blot. Results: A total of 20 candidate genes were identified, of which 8 were identified as hub genes through machine learning. Integrative PPI and single-cell analyses ultimately identified CKAP2 as the key gene. Virtual knockout analysis showed that the related differentially expressed genes were significantly enriched in the 5-HT pathway. Survival analysis demonstrated that elevated CKAP2 expression was associated with improved prognosis. MR analysis further suggested that CKAP2 might act as a protective factor, while 5-HT was associated with an increased risk of GC. Experimental validation confirmed that CKAP2 is significantly upregulated in GC tissues and cell lines. Conclusions: CKAP2 is a potential prognostic biomarker and protective factor in GC, possibly exerting its effects through regulation of the 5-HT pathway. These findings provide new insights into the mechanisms and potential therapeutic targets of GC.
Open AccessIntegrated Single-Cell Virtual Knockout and Machine Learning Analyses Reveal a Protective Role of CKAP2 in Gastric Cancer
Objective: To elucidate the role of cytoskeleton-associated protein 2 (CKAP2) in gastric cancer (GC) progression and evaluate its prognostic and potential protective significance. Methods: The candidate genes for GC were identified using differential expression and weighted gene co-expression network analysis (WGCNA). A panel of machine learning algorithms was applied for obtaining the key genes and yielding the hub target gene CKAP2 through protein–protein interaction (PPI) networks and single-cell RNA sequencing analysis. Further, the functional role of CKAP2 was explored through single-cell virtual knockout and pathway enrichment analysis. Meanwhile, the survival analysis and Mendelian randomization (MR) were used to evaluate the clinical relevance and causality of CKAP2. Finally, the expression of CKAP2 was further validated in clinical tissues and GC cell lines through Western blot. Results: A total of 20 candidate genes were identified, of which 8 were identified as hub genes through machine learning. Integrative PPI and single-cell analyses ultimately identified CKAP2 as the key gene. Virtual knockout analysis showed that the related differentially expressed genes were significantly enriched in the 5-HT pathway. Survival analysis demonstrated that elevated CKAP2 expression was associated with improved prognosis. MR analysis further suggested that CKAP2 might act as a protective factor, while 5-HT was associated with an increased risk of GC. Experimental validation confirmed that CKAP2 is significantly upregulated in GC tissues and cell lines. Conclusions: CKAP2 is a potential prognostic biomarker and protective factor in GC, possibly exerting its effects through regulation of the 5-HT pathway. These findings provide new insights into the mechanisms and potential therapeutic targets of GC.
- 21 April 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 324
SC-CAMamba: Multi-objective classroom behaviour recognition based on parallel state space models and self-attention
- 21 April 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 324
SC-CAMamba: Multi-objective classroom behaviour recognition based on parallel state space models and self-attention
- Open Access
- 7 April 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1780967
TCMI-F-6D benchmark construction and quantitative assessment of interdisciplinary foundational competencies in traditional Chinese medicine informatics using large language models
Introduction: Traditional Chinese Medicine Informatics (TCMI), as an emerging interdisciplinary field, places high demands on foundational interdisciplinary competency assessment in its talent cultivation and research practices. However, Large Language Models (LLMs) currently lack a suitable quantitative assessment system tailored to the characteristics of TCMI. Methods: To address this gap, this study, grounded in Cognitive Hierarchy Theory and Disciplinary Knowledge Structure Theory, selected six core disciplines closely related to TCMI from the Massive Multitask Language Understanding (MMLU) dataset, constructed an evaluation framework for foundational interdisciplinary competency in TCMI-related scenarios, and established the TCMI-F-6D (the TCMI-Foundation-6 Domain Benchmark) together with a composite metric system. Three experiments were conducted to evaluate the models’ baseline capability, learning gains, and performance stability. The experiments comprehensively assessed the competency of 20 LLMs across 8 categories, and selected 6 models with weaker overall performance for focused analysis of their interdisciplinary competency characteristics. Results: The results showed that, among the base models, ChatGLM3-6B performed best in interdisciplinary knowledge integration (43.97%), while DeepSeek-V3.1 achieved the best overall application performance (80.87%) among the chat models. Specifically, Qwen-14B-Chat also demonstrated stable and predictable learning performance under varying example conditions, with an average learning gain of 5.60% and a 95% confidence interval (CI) of [5.50%, 5.70%]. Discussion: Collectively, this study clarifies the differences in foundational interdisciplinary competency among LLMs in this discipline, providing a quantifiable assessment framework, methodological support, and empirical evidence for TCMI’s educational, research tool selection, and the implementation of a standardized interdisciplinary competency assessment system.
- 7 April 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1780967
Open AccessTCMI-F-6D benchmark construction and quantitative assessment of interdisciplinary foundational competencies in traditional Chinese medicine informatics using large language models
Introduction: Traditional Chinese Medicine Informatics (TCMI), as an emerging interdisciplinary field, places high demands on foundational interdisciplinary competency assessment in its talent cultivation and research practices. However, Large Language Models (LLMs) currently lack a suitable quantitative assessment system tailored to the characteristics of TCMI. Methods: To address this gap, this study, grounded in Cognitive Hierarchy Theory and Disciplinary Knowledge Structure Theory, selected six core disciplines closely related to TCMI from the Massive Multitask Language Understanding (MMLU) dataset, constructed an evaluation framework for foundational interdisciplinary competency in TCMI-related scenarios, and established the TCMI-F-6D (the TCMI-Foundation-6 Domain Benchmark) together with a composite metric system. Three experiments were conducted to evaluate the models’ baseline capability, learning gains, and performance stability. The experiments comprehensively assessed the competency of 20 LLMs across 8 categories, and selected 6 models with weaker overall performance for focused analysis of their interdisciplinary competency characteristics. Results: The results showed that, among the base models, ChatGLM3-6B performed best in interdisciplinary knowledge integration (43.97%), while DeepSeek-V3.1 achieved the best overall application performance (80.87%) among the chat models. Specifically, Qwen-14B-Chat also demonstrated stable and predictable learning performance under varying example conditions, with an average learning gain of 5.60% and a 95% confidence interval (CI) of [5.50%, 5.70%]. Discussion: Collectively, this study clarifies the differences in foundational interdisciplinary competency among LLMs in this discipline, providing a quantifiable assessment framework, methodological support, and empirical evidence for TCMI’s educational, research tool selection, and the implementation of a standardized interdisciplinary competency assessment system.
- Open Access
- 21 May 2024
- Published by MDPI AG in Electronics
- Vol. 13 (11) , 2002
Exploring the Potential of Large Language Models in Radiological Imaging Systems: Improving User Interface Design and Functional Capabilities
Large language models (LLMs) have demonstrated remarkable capabilities in natural language processing tasks, including conversation, in-context learning, reasoning, and code generation. This paper explores the potential application of LLMs in radiological information systems (RIS) and assesses the impact of integrating LLMs on RIS development and human–computer interaction. We present ChatUI-RIS, a prototype chat-based user interface that leverages LLM capabilities to enhance RIS functionality and user experience. Through an exploratory study involving 26 medical students, we investigate the efficacy of natural language dialogue for learning and operating RIS. Our findings suggest that LLM integration via a chat interface can significantly improve operational efficiency, reduce learning time, and facilitate rapid expansion of RIS capabilities. By interacting with ChatUI-RIS using natural language instructions, medical students can access and retrieve radiology information in a conversational manner. The LLM-powered chat interface not only streamlines user interactions, but also enables more intuitive and efficient navigation of complex RIS functionalities. Furthermore, the natural language processing capabilities of LLMs can be harnessed to automatically generate code snippets and database queries, accelerating RIS development and customization. Preliminary observations indicate that integrating LLMs in RIS has the potential to revolutionize user interface design, enhance system capabilities, and ultimately improve the overall user experience for radiologists and medical professionals.
- 21 May 2024
- Published by MDPI AG in Electronics
- Vol. 13 (11) , 2002
Open AccessExploring the Potential of Large Language Models in Radiological Imaging Systems: Improving User Interface Design and Functional Capabilities
Large language models (LLMs) have demonstrated remarkable capabilities in natural language processing tasks, including conversation, in-context learning, reasoning, and code generation. This paper explores the potential application of LLMs in radiological information systems (RIS) and assesses the impact of integrating LLMs on RIS development and human–computer interaction. We present ChatUI-RIS, a prototype chat-based user interface that leverages LLM capabilities to enhance RIS functionality and user experience. Through an exploratory study involving 26 medical students, we investigate the efficacy of natural language dialogue for learning and operating RIS. Our findings suggest that LLM integration via a chat interface can significantly improve operational efficiency, reduce learning time, and facilitate rapid expansion of RIS capabilities. By interacting with ChatUI-RIS using natural language instructions, medical students can access and retrieve radiology information in a conversational manner. The LLM-powered chat interface not only streamlines user interactions, but also enables more intuitive and efficient navigation of complex RIS functionalities. Furthermore, the natural language processing capabilities of LLMs can be harnessed to automatically generate code snippets and database queries, accelerating RIS development and customization. Preliminary observations indicate that integrating LLMs in RIS has the potential to revolutionize user interface design, enhance system capabilities, and ultimately improve the overall user experience for radiologists and medical professionals.
- Open Access
- 19 September 2023
- Published by Frontiers Media SA in Frontiers in Medicine
- Vol. 10, 1233724
Optimization of pneumonia CT classification model using RepVGG and spatial attention features
Introduction: Pneumonia is a common and widespread infectious disease that seriously affects the life and health of patients. Especially in recent years, the outbreak of COVID-19 has caused a sharp rise in the number of confirmed cases of epidemic spread. Therefore, early detection and treatment of pneumonia are very important. However, the uneven gray distribution and structural intricacy of pneumonia images substantially impair the classification accuracy of pneumonia. In this classification task of COVID-19 and other pneumonia, because there are some commonalities between this pneumonia, even a small gap will lead to the risk of prediction deviation, it is difficult to achieve high classification accuracy by directly using the current network model to optimize the classification model.Methods: Consequently, an optimization method for the CT classification model of COVID-19 based on RepVGG was proposed. In detail, it is made up of two essential modules, feature extraction backbone and spatial attention block, which allows it to extract spatial attention features while retaining the benefits of RepVGG.Results: The model’s inference time is significantly reduced, and it shows better learning ability than RepVGG on both the training and validation sets. Compared with the existing advanced network models VGG-16, ResNet-50, GoogleNet, ViT, AlexNet, MobileViT, ConvNeXt, ShuffleNet, and RepVGG_b0, our model has demonstrated the best performance in a lot of indicators. In testing, it achieved an accuracy of 0.951, an F1 score of 0.952, and a Youden index of 0.902.Discussion: Overall, multiple experiments on the large dataset of SARS-CoV-2 CT-scan dataset reveal that this method outperforms most basic models in terms of classification and screening of COVID-19 CT, and has a significant reference value. Simultaneously, in the inspection experiment, this method outperformed other networks with residual structures.
- 19 September 2023
- Published by Frontiers Media SA in Frontiers in Medicine
- Vol. 10, 1233724
Open AccessOptimization of pneumonia CT classification model using RepVGG and spatial attention features
Introduction: Pneumonia is a common and widespread infectious disease that seriously affects the life and health of patients. Especially in recent years, the outbreak of COVID-19 has caused a sharp rise in the number of confirmed cases of epidemic spread. Therefore, early detection and treatment of pneumonia are very important. However, the uneven gray distribution and structural intricacy of pneumonia images substantially impair the classification accuracy of pneumonia. In this classification task of COVID-19 and other pneumonia, because there are some commonalities between this pneumonia, even a small gap will lead to the risk of prediction deviation, it is difficult to achieve high classification accuracy by directly using the current network model to optimize the classification model.Methods: Consequently, an optimization method for the CT classification model of COVID-19 based on RepVGG was proposed. In detail, it is made up of two essential modules, feature extraction backbone and spatial attention block, which allows it to extract spatial attention features while retaining the benefits of RepVGG.Results: The model’s inference time is significantly reduced, and it shows better learning ability than RepVGG on both the training and validation sets. Compared with the existing advanced network models VGG-16, ResNet-50, GoogleNet, ViT, AlexNet, MobileViT, ConvNeXt, ShuffleNet, and RepVGG_b0, our model has demonstrated the best performance in a lot of indicators. In testing, it achieved an accuracy of 0.951, an F1 score of 0.952, and a Youden index of 0.902.Discussion: Overall, multiple experiments on the large dataset of SARS-CoV-2 CT-scan dataset reveal that this method outperforms most basic models in terms of classification and screening of COVID-19 CT, and has a significant reference value. Simultaneously, in the inspection experiment, this method outperformed other networks with residual structures.
- Open Access
- 24 July 2020
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2020, 1-11
An Improved Mask R-CNN Model for Multiorgan Segmentation
Medical image segmentation is a key topic in image processing and computer vision. Existing literature mainly focuses on single-organ segmentation. However, since maximizing the concentration of radiotherapy drugs in the target area with protecting the surrounding organs is essential for making effective radiotherapy plan, multiorgan segmentation has won more and more attention. An improved Mask R-CNN (region-based convolutional neural network) model is proposed for multiorgan segmentation to aid esophageal radiation treatment. Due to the fact that organ boundaries may be fuzzy and organ shapes are various, original Mask R-CNN works well on natural image segmentation while leaves something to be desired on the multiorgan segmentation task. Addressing it, the advantages of this method are threefold: (1) a ROI (region of interest) generation method is presented in the RPN (region proposal network) which is able to utilize multiscale semantic features. (2) A prebackground classification subnetwork is integrated to the original mask generation branch to improve the precision of multiorgan segmentation. (3) 4341 CT images of 44 patients are collected and annotated to evaluate the proposed method. Additionally, extensive experiments on the collected dataset demonstrate that the proposed method can segment the heart, right lung, left lung, planning target volume (PTV), and clinical target volume (CTV) accurately and efficiently. Specifically, less than 5% of the cases were missed detection or false detection on the test set, which shows a great potential for real clinical usage.
- 24 July 2020
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2020, 1-11
Open AccessAn Improved Mask R-CNN Model for Multiorgan Segmentation
Medical image segmentation is a key topic in image processing and computer vision. Existing literature mainly focuses on single-organ segmentation. However, since maximizing the concentration of radiotherapy drugs in the target area with protecting the surrounding organs is essential for making effective radiotherapy plan, multiorgan segmentation has won more and more attention. An improved Mask R-CNN (region-based convolutional neural network) model is proposed for multiorgan segmentation to aid esophageal radiation treatment. Due to the fact that organ boundaries may be fuzzy and organ shapes are various, original Mask R-CNN works well on natural image segmentation while leaves something to be desired on the multiorgan segmentation task. Addressing it, the advantages of this method are threefold: (1) a ROI (region of interest) generation method is presented in the RPN (region proposal network) which is able to utilize multiscale semantic features. (2) A prebackground classification subnetwork is integrated to the original mask generation branch to improve the precision of multiorgan segmentation. (3) 4341 CT images of 44 patients are collected and annotated to evaluate the proposed method. Additionally, extensive experiments on the collected dataset demonstrate that the proposed method can segment the heart, right lung, left lung, planning target volume (PTV), and clinical target volume (CTV) accurately and efficiently. Specifically, less than 5% of the cases were missed detection or false detection on the test set, which shows a great potential for real clinical usage.