JGJia Gu

Jia Gu

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Jia Gu has a total of 126 co-authors. Here are the top co-authors by number of shared publications:

  • Shifu Chen

    LifeX Institute Gannan Medical University Ganzhou China

    19 shared publications
  • An-Yuan Guo

    West China Biomedical Big Data Center, West China Hospital, Sichuan University Department of Thoracic Surgery Chengdu 610041

    16 shared publications
  • Bin Hu

    Department of Thoracic Surgery, Sichuan Clinical Research Center for Cancer, Sichuan Cancer Hospital & Institute, Sichuan Cancer Center, University of Electronic Science and Technology of China, Chengdu, China

    16 shared publications
  • Dongsheng Wang

    Department of Clinical Laboratory, Sichuan Clinical Research Center for Cancer, Sichuan Cancer Hospital & Institute, Sichuan Cancer Center, Affiliated Cancer Hospital of University of Electronic Science and Technology of China, Chengdu, China

    16 shared publications
  • Huaichao Luo

    Department of Clinical Laboratory, Sichuan Clinical Research Center for Cancer, Sichuan Cancer Hospital & Institute, Sichuan Cancer Center, University of Electronic Science and Technology of China, Chengdu, China

    16 shared publications
  • Jian Huang

    School of Life Science and Technology, University of Electronic Science and Technology of China, Chengdu, China

    16 shared publications

Latest Publications

  • Open Access

    UniMedSeg: Unified Multi-Modal Medical Image Segmentation via Uncertainty-Guided Encoding and Frequency-Spatial Dual-Prompt Decoding

    by Jingling Zhang, Shuting Zheng, Xiangfei Liu, Wen Zhang, Jia Gu

    Text-driven medical image segmentation demands unified models based on multi-modal learning, capable of processing diverse imaging modalities while providing spatially reliable prediction estimates for clinical practice. Nevertheless, existing vision-language model solutions suffer from poor domain generalization and exhibit limited capability for uncertainty estimation to model the cumulative unreliability of cross-modal alignment across network layers. In this paper, we propose UniMedSeg, a unified vision-language framework for multi-modal medical image segmentation built upon frozen pre-trained vision-language backbones. Our framework incorporates modality-aware adaptation, cross-layer uncertainty-guided encoding, and a frequency-spatial dual-prompt decoder, wherein the progressive uncertainty propagation mechanism of the encoder suppresses misleading cross-modal features, whereas the decoder disentangles anatomical structure and boundary details in the frequency domain under dual textual prompts. Auxiliary losses further sharpen boundary representations and calibrate pixel-wise uncertainty maps. Extensive experiments demonstrate that UniMedSeg preserves competitive in-domain segmentation performance, improves domain generalization, and yields clinically useful spatial risk estimation. The source code is available at https://github.com/699zjl/UniMedSeg . Highlights: We propose UniMedSeg , a unified vision-language model for five-modality medical segmentation that outputs segmentation masks and clinical-oriented pixel-level uncertainty maps. A cross-layer uncertainty propagation mechanism is designed to dynamically suppress unreliable vision-language alignment by recursively accumulating historical reliability signals. A frequency-spatial dual-prompt decoder disentangles anatomical structure and boundary details via FFT-based separation and branch-specific text-guided attention to reduce frequency-domain artifacts. Auxiliary boundary contrastive and uncertainty calibration objectives enhance boundary discrimination and enforce consistency between predicted uncertainty and segmentation errors. Extensive experiments on multi-modal medical datasets verify that UniMedSeg yields competitive in-domain performance and favorable out-of-distribution generalization capability.

  • Open Access

    Protocol for rolling circle amplification-based nanopore sequencing of eccDNA from cell and plasma samples

    by Biyuan Miao, Xinlei Huang, Hui Chen, Wang Li, Qingsong Zen, Huaxi Li, Jia Gu, Pingping Xu, Shaogui Wan

  • Open Access

    High‐Concordance Detection Across Diverse Biospecimens EGFR Mutation in a Large‐Scale NSCLC Cohort Using Droplet Digital PCR and Next‐Generation Sequencing

    by Ming Liu, Linping Lu, Qifang Lei, Lingxiang Zhu, Xiaoni Zhang, Yuanyuan Liu, Xiaojing Ren, Shimin Liu, Shaochen Cheng, Mingyan Xu, et al.

    Accurate detection and quantification of EGFR mutations are essential for molecular stratification in non‐small cell lung cancer (NSCLC). While next‐generation sequencing (NGS) enables comprehensive genomic profiling, droplet digital PCR (ddPCR) offers high sensitivity for predefined mutations. Large‐scale evaluations of ddPCR performance across diverse clinical specimen types remain limited. This retrospective study analyzed 936 analytical mutation‐target records from 747 NSCLC patients with NGS‐detected EGFR mutations. The records represented native DNA or corresponding pre‐capture NGS‐library derivatives from fresh tumor tissue, formalin‐fixed paraffin‐embedded tissue, plasma‐derived cell‐free DNA, and pleural effusion supernatant. ddPCR results were compared with NGS for mutation detection and variant allele frequency quantification using correlation and Bland–Altman analyses. Across all 936 analytical records, ddPCR showed a high overall positive percent agreement (PPA) with NGS (98.82%; 925/936; exact 95% CI, 97.91%–99.41%). Mutation‐specific PPAs were 98.93% for L858R (370/374; exact 95% CI, 97.28%–99.71%), 99.48% for Ex19del (385/387; exact 95% CI, 98.15%–99.94%), and 97.14% for T790M (170/175; exact 95% CI, 93.46%–99.07%). VAFs measured by ddPCR and NGS were strongly correlated (Pearson's r = 0.975). Concordance varied by mutation type and sample matrix, with greater variability observed in high‐VAF samples and plasma cfDNA. Paired analyses demonstrated high consistency between native DNA and pre‐capture NGS libraries. ddPCR provides reliable detection and quantification of clinically relevant EGFR mutations across multiple specimen types and processing states. ddPCR represents a practical complement to NGS for EGFR mutation analysis in NSCLC.

  • Open Access

    Geometric-aware deep learning for deciphering tissue structure from spatially resolved transcriptomics

    by Xingyi Li, Xiangting Jia, Dongmin Zhao, Jialuo Xu, Gaoyuan Du, Yang Qi, Yingfu Wu, Yiqi Chen, Junnan Zhu, Jia Gu, et al.

    Recent advances in spatially resolved transcriptomics have enabled large-scale measurement of gene expression while preserving spatial context, facilitating the investigation of spatial heterogeneity within tissues. In this study, we propose SpatialGEO, a geometric-aware deep learning framework that integrates gene expression profiles with spatial coordinates to generate biologically meaningful low-dimensional embeddings, enabling the dissection of complex tissue architectures. We systematically evaluate SpatialGEO across multiple tissue types and diverse SRT platforms. Results show that SpatialGEO achieves superior performance in tissue structure dissection and data denoising compared to state-of-the-art methods. Moreover, when applied to human breast cancer samples, SpatialGEO precisely delineates the tumor microenvironment and uncovers molecular heterogeneity within tumors and intercellular communication between invasive ductal carcinoma and tumor edge. In mouse embryogenesis, SpatialGEO accurately reconstructs spatiotemporal tissue architectures, highlighting organ-specific developmental programs and elucidating molecular drivers of early neural development.


  • Open Access

    Beyond Accuracy: A Multi-dimensional Cognitive Audit of Medical Large Vision–Language Models in Fundus Image Interpretation

    by Jingling Zhang, Shuting Zheng, Xiangfei Liu, Jia Gu

    Reliance on standalone accuracy limits credible assessment of fundus-focused large vision–language models (LVLMs), as high scores often stem from linguistic shortcuts rather than real visual reasoning. This work develops the Cognitive Audit Framework (CAF), a four-module automated auditing pipeline that dissects model reasoning flaws: Visual–Linguistic Decoupling (textual dependency via modality ablation), Hierarchical Logical Consistency (lesion–diagnosis contradiction detection), Reasoning Fidelity Gap (chain-of-thought unfaithfulness scoring), and Contextual Robustness (positional bias under option permutation). Experiments on six 7B–31B LVLMs over FunBench reveal a notable gap between benchmark accuracy and reasoning quality: high accuracy coexists with measurable textual dependency, logical inconsistencies across diagnostic levels, limited chain-of-thought faithfulness, and non-trivial positional sensitivity. CAF serves as a reproducible complement to pure accuracy metrics for validating clinical competence of ophthalmic multimodal models.

  • Open Access

    Accurately Deciphering Tissue Heterogeneity From Spatial Multi‐Modal and Multi‐Omics With STransformer

    by Xingyi Li, Jialuo Xu, Gaoyuan Du, Xiangting Jia, Dongmin Zhao, Chunyan Zhou, Kexin Xiao, Jia Gu, Junnan Zhu, Xuequn Shang

    Advances in spatially resolved technologies enable the simultaneous acquisition of diverse data modalities within a tissue slice while preserving critical spatial context, which presents unprecedented opportunities to decipher intricate tissue heterogeneity. However, existing computational approaches lack the intrinsic flexibility to universally process both spatial multi‐modal and multi‐omics data. Here, we introduce STransformer, a unified deep learning framework designed to seamlessly accommodate a comprehensive landscape of spatial data. By simultaneously capturing short‐range cellular interactions and tissue‐wide semantic patterns, it extracts robust representations to accurately dissect complex tissue heterogeneity. Systematic evaluations across diverse species, tissue types, and data modalities highlight its profound versatility. For spatial multi‐modal data, STransformer delineates intricate anatomical structures in the human cortex, uncovers pathological mechanisms in Alzheimer's disease, and characterizes dynamic spatiotemporal developmental trajectories during chicken cardiogenesis. Scaling to spatial multi‐omics data, STransformer synergizes spatial transcriptomic and proteomic profiles to decipher intricate immune microenvironments within the human tonsil, and jointly analyzes spatial epigenomic and transcriptomic data to infer regulatory mechanisms in the mouse embryonic brain. Consequently, STransformer serves as a highly versatile and robust analytical framework for advancing our understanding of tissue heterogeneity and disease pathogenesis.

  • Deep-Learning-Based Control of a Decoupled Two-Segment Continuum Robot for Endoscopic Submucosal Dissection

    by Yuancheng Shao, Yao Zhang, Jia Gu, Zixi Chen, Di Wu, Yuqiao Chen, Bo Lu, Wenjie Liu, Cesare Stefanini, Peng Qi

    Manual endoscopic submucosal dissection (ESD) is technically demanding, and existing single-segment robotic tools offer limited dexterity. These limitations motivate the development of more advanced solutions. To address this, decoupled endoscopic submucosal dissection robot (DESectBot), a novel dual-segment continuum robot with a decoupled structure and integrated surgical forceps, enabling six degrees of freedom tip dexterity for improved lesion targeting in ESD, was developed in this work. Deep learning controllers based on gated recurrent units (GRUs) for simultaneous tip position and orientation control, effectively handling the nonlinear coupling between continuum segments, were proposed. The GRU controller was benchmarked against Jacobian-based inverse kinematics, model predictive control (MPC), a feedforward neural network, and a long short-term memory (LSTM) network. In addition, a systematic ablation study under identical training conditions showed that, for the considered control task, the GRU achieves better generalization, faster convergence, and higher computational efficiency than the LSTM. In nested-rectangle and Lissajous trajectory tracking tasks, the GRU achieved the lowest position/orientation root-mean-square errors (RMSEs): 1.11 mm/ 4.62$^{\circ }$ and 0.81 mm/ 2.59$^{\circ }$, respectively. For orientation control at a fixed position (four target poses), the GRU attained a mean RMSE of 0.14 mm and 0.72$^{\circ }$, outperforming all alternatives. In a peg transfer task, the GRU achieved a 100% success rate (120 success/120 attempts) with an average transfer time of 11.8 s, the standard deviation significantly outperforms no vice-controlled systems. In addition, an ex vivo ESD demonstration—grasping, elevating, and resecting tissue as the scalpel completed the cut—confirmed that DESectBot provides sufficient stiffness to divide thick gastric mucosa and an operative workspace adequate for large lesions. These results confirm that GRU-based control significantly enhances precision, reliability, and usability in ESD surgical training scenarios.

  • DCAU-KAN: Dual Cross Attention Network for Segmenting Medical Images

    by Ziyang Zhu, Qi Zhang, Wenjian Liu, Jia Gu, Xuanke Zhang, Jingzhang Sun, Jianhang Zhou

    Currently, in the field of medical image segmentation, deep learning methods have shown significant advantages and have been widely applied. However, mainstream U-shaped segmentation models still face two critical and unresolved challenges in practical applications: on the one hand, traditional U-Net and its improved variants rely on multi-layer perceptron (MLP) for feature transformation, leading to severe parameter redundancy and heavy computational overhead; on the other hand, these models lack targeted guidance from image inherent features in the contextual information acquisition stage, which results in insufficient mining of position and channel correlation features. To address these, we introduce DCAU-KAN, a improved U-shaped architecture that fuses Kolmogorov–Arnold Networks (KANs) with a novel dual cross attention block (DCA-Block). The DCA block guides the acquisition of contextual information by introducing the inherent features of the image (position and channel features), thereby enhancing the feature extraction capability. Furthermore, we use the residual skip multi-path (RSMP block) to integrate the features of the encoder and decoder, increasing the representation of the features and reducing feature confusion. To evaluate our method, we conducted experiments using five widely recognized medical datasets. The experimental results demonstrate that the improved model achieves an average IoU of 84.02% and a DSC of 90.59% across five datasets, representing an improvement of 3.65% and 2.97%, respectively, compared to the original model.