CLChuanfu Li

Chuanfu Li

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Chuanfu Li has a total of 37 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Neuroanatomical correlates of psychiatric manifestations in Wilson disease: the role of age

    by Anqin Wang, Chunsheng Xu, Danqing Liu, Man Jiang, Chuanfu Li

    Objective: To investigate gray matter volume (GMV) differences related to psychiatric manifestations in patients with Wilson disease (WD) and to examine the effect of age. Methods: In 102 patients with WD, 3.0T T1-weighted MRI scans were analyzed using voxel-based morphometry. Patients were divided into younger (≤30 years) and older (>30 years) groups and were further grouped according to psychiatric manifestations. Patients with a UWDRS psychiatric subscale (UWDRS-PSY) score of 0 were classified as having no psychiatric manifestations, and those with a score >0 were classified as having psychiatric manifestations. Two-way ANCOVA was performed with total intracranial volume, sex, and education as covariates. Cluster-level family-wise error (FWE) correction was applied. Mean GMV values were extracted from significant caudate clusters and correlated with UWDRS-PSY scores as exploratory analyses. Results: Significant interactions between age group and psychiatric manifestations were found in the left insula/Rolandic operculum and right inferior temporal gyrus. Lower GMV in these regions was mainly found in younger patients with psychiatric manifestations. A main effect of psychiatric manifestations was found in the bilateral caudate nuclei, with lower GMV in patients with psychiatric manifestations. Lower caudate GMV was also associated with higher UWDRS-PSY scores. Conclusions: Psychiatric manifestations in WD are associated with distinct gray matter patterns that vary with age, while bilateral caudate involvement may represent a more consistent structural correlate. These findings highlight the potential importance of considering age when examining the neuroanatomical correlates of psychiatric manifestations in WD.

  • Open Access

    Activating Associative Disease-Aware Vision Token Memory for LLM-Based X-ray Report Generation

    by Xiao Wang, Fuling Wang, Haowen Wang, Bo Jiang, Chuanfu Li, Yaowei Wang, Yonghong Tian, Jin Tang

    X-ray image based medical report generation achieves significant progress in recent years with the help of the large language model, however, these models have not fully exploited the effective information in visual image regions, resulting in reports that are linguistically sound but insufficient in describing key diseases. In this paper, we propose a novel associative memory-enhanced X-ray report generation model that effectively mimics the process of professional doctors writing medical reports. It considers both the mining of global and local visual information and associates historical report information to better complete the writing of the current report. Specifically, given an X-ray image, we first utilize a classification model along with its activation maps to accomplish the mining of visual regions highly associated with diseases and the learning of disease query tokens. Then, we employ a visual Hopfield network to establish memory associations for disease-related tokens, and a report Hopfield network to retrieve report memory information. This process facilitates the generation of high-quality reports based on a large language model and achieves state-of-the-art performance on multiple benchmark datasets, including the IU X-ray, MIMIC-CXR, and Chexpert Plus. The source code of this work is released on \url{https://github.com/Event-AHU/Medical_Image_Analysis}.


  • Open Access

    CXPMRG-Bench: Pre-training and Benchmarking for X-ray Medical Report Generation on CheXpert Plus Dataset

    by Xiao Wang, Fuling Wang, Yuehang Li, Qingchuan Ma, Shiao Wang, Bo Jiang, Chuanfu Li, Jin Tang

    X-ray image-based medical report generation (MRG) is a pivotal area in artificial intelligence which can significantly reduce diagnostic burdens and patient wait times. Despite significant progress, we believe that the task has reached a bottleneck due to the limited benchmark datasets and the existing large models' insufficient capability enhancements in this specialized domain. Specifically, the recently released CheXpert Plus dataset lacks comparative evaluation algorithms and their results, providing only the dataset itself. This situation makes the training, evaluation, and comparison of subsequent algorithms challenging. Thus, we conduct a comprehensive benchmarking of existing mainstream X-ray report generation models and large language models (LLMs), on the CheXpert Plus dataset. We believe that the proposed benchmark can provide a solid comparative basis for subsequent algorithms and serve as a guide for researchers to quickly grasp the state-of-the-art models in this field. More importantly, we propose a large model for the X-ray image report generation using a multi-stage pre-training strategy, including self-supervised autoregressive generation and Xray-report contrastive learning, and supervised fine-tuning. Extensive experimental results indicate that the autoregressive pre-training based on Mamba effectively encodes X-ray images, and the image-text contrastive pre-training further aligns the feature spaces, achieving better experimental results. Source code can be found on \url{https://github.com/Event-AHU/Medical_Image_Analysis}.


  • Open Access

    Large-scale networks changes in Wilson’s disease associated with neuropsychiatric impairments: a resting-state functional magnetic resonance imaging study

    by Anqin Wang, Ting Dong, Taohua Wei, Hongli Wu, Yulong Yang, Yufeng Ding, Chuanfu Li, Wenming Yang

    Background: In Wilson’s disease (WD) patients, network connections across the brain are disrupted, affecting multidomain function. However, the details of this neuropathophysiological mechanism remain unclear due to the rarity of WD. In this study, we aimed to investigate alterations in brain network connectivity at the whole-brain level (both intra- and inter-network) in WD patients through independent component analysis (ICA) and the relationship between alterations in these brain network functional connections (FCs) and clinical neuropsychiatric features to understand the underlying pathophysiological and central compensatory mechanisms. Methods: Eighty-five patients with WD and age- and sex-matched 85 healthy control (HC) were recruited for resting-state functional magnetic resonance imaging (rs-fMRI) scanning. We extracted the resting-state networks (RSNs) using the ICA method, analyzed the changes of FC in these networks and the correlation between alterations in FCs and clinical neuropsychiatric features. Results: Compared with HC, WD showed widespread lower connectivity within RSNs, involving default mode network (DMN), frontoparietal network (FPN), somatomotor network (SMN), dorsal attention network (DAN), especially in patients with abnormal UWDRS scores. Furthermore, the decreased FCs in the left medial prefrontal cortex (L_ MPFC), left anterior cingulate gyrus (L_ACC), precuneus (PCUN)within DMN were negatively correlated with the Unified Wilson’s Disease Rating Scale-neurological characteristic examination (UWDRS-N), and the decreased FCs in the L_MPFC, PCUN within DMN were negatively correlated with the Unified Wilson’s Disease Rating Scale-psychiatric symptoms examination (UWDRS-P). We additionally discovered that the patients with WD exhibited significantly stronger FC between the FPN and DMN, between the DAN and DMN, and between the FPN and DAN compared to HC. Conclusions: We have provided evidence that WD is a disease with widespread dysfunctional connectivity in resting networks in brain, leading to neurological features and psychiatric symptoms (e.g. higher-order cognitive control and motor control impairments). The alter intra- and inter-network in the brain may be the neural underpinnings for the neuropathological symptoms and the process of injury compensation in WD patients.


  • Open Access

    Chest X-ray Image Quality Control via Transformer and Label Correlation Regularization Loss

    by Ming Liu, Lili Huang, Bo Jiang, Chuanfu Li, Jin Tang

    Quality control of chest X-rays is especially important for radiologists to make accurate diagnoses. Presently, the quality control of chest X-ray image is done manually, which is time-consuming and labor-intensive. Therefore, it is very necessary to research on automatic and intelligent image quality control. However, to the best of our knowledge, this task has not been studied before. The main challenges lie in how to obtain effective feature representation and to make full use of prior medical knowledge. In this work, we propose a quality control model for chest X-ray image via Transformer and label correlation regularization loss. Specifically, we utilize a multi-scale fusion strategy to discriminate unclear and subtle image features for the X-ray image quality control task. In addition, we incorporate prior medical knowledge into our network via label correlation regularization loss to understand of the inherent correlations of different quality labels which further enhance the learning accuracy. We also constructed a novel dataset for chest x-ray quality control task. Experiments on the dataset demonstrate the superiority of our proposed method.

  • Open Access

    Knowledge graph-based small sample learning for label of medical imaging reports

    by Yuxuan Zhang, Zongyun Gu, Pengcheng Jia, Tiancheng Li, Wenhao Lu, Mingxin Ge, Linyu Chen, Chuanfu Li

    Background: Through the utilization of knowledge graph and small sample learning, the study effectively tackled the challenges of data scarcity and automatic annotation in the field of medical image recognition with the application of artificial intelligence technology. Methods: Initially, 2000 X-ray reports of the lumbar spine were labeled manually employing a knowledge graph approach. These reports were subsequently split into a training dataset of 1000 cases and a test dataset of 1000 cases. Following this, distinct levels of data augmentation, namely the synonym/apposition method, were applied to the training dataset. Subsequently, the deep learning model BERT (Bidirectional Encoder Representation of Transformer) was utilized for the training process. Afterward, the BERT model is tested on the specified test dataset, and subsequently, the nodes showing insufficient performance are supplemented with iterative target data. Finally, the method is evaluated by using various metrics including AUC(Area Under Curve), F1 score, precision, recall and relabelled rate. Results: Before conducting data augmentation, the AUC value was 0.621, the F1 value was 32.1%, the average precision was 0.383, and the average recall was 0.303. Following data augmentation, the AUC value improved to 0.789, the F1 value improved to 70.3%, the average precision improved to 0.879, and the average recall improved to 0.580. After targeted data supplementation, the AUC reached 0.899, the F1 value reached 85.7%, the average precision reached 0.952, and the average recall reached 0.803. Conclusions: The current study achieves its objective by training an automatic annotation model using a knowledge graph-based approach to annotate medical imaging reports on a small sample dataset. Furthermore, this approach enhances both the efficiency and accuracy of medical imaging data annotation, providing a significant research strategy for applying artificial intelligence in the field of medical image recognition.

  • Open Access

    An Efficient CNN for Hand X‐Ray Overall Scoring of Rheumatoid Arthritis

    by Zijian Wang, Jian Liu, Zongyun Gu, Chuanfu Li

    Rheumatoid arthritis (RA) is a progressive systemic autoimmune disease characterized by inflammation of the joints and surrounding tissues, which seriously affects the life of patients. The Sharp/van der Heijde method has been widely used in clinical evaluation for the RA disease. However, this manual method is time‐consuming and laborious. Even if two radiologists evaluate a specific location, their subjective evaluation may lead to low inter‐rater reliability. Here, we developed an efficient model powered by deep convolutional neural networks to solve these problems and automated the overall scoring on hand X‐rays. The depthwise separable (Dwise) convolution technique is used based on ResNet‐50 due to the high resolution of hand X‐rays. An inverted residual block is introduced to devise a ResNet‐Dwise50 model to enhance the efficiency of the model. The model was trained and tested using bilateral posteroanterior (two‐handed, side by side) images of 3818 patients. The experiment results show the ResNet‐Dwise50 model achieved an MAE of 14.90 and RMSE of 22.01 while ensuring high efficiency. There was no statistical difference between the average scores given by two experienced radiologists and predicted scores from our model.