Yuanshen Zhao

Yuanshen Zhao

Auto-generated profile, based on publicly available information

Yuanshen Zhao

Institute of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, PR China

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 December 2021 in Radiology
Verified
Journal Article
Closed access
QSQiuchang Sun
Qiuchang Sun
YCYinsheng Chen
Yinsheng Chen
CLChaofeng Liang
Chaofeng Liang
YZYuanshen Zhao
Yuanshen Zhao
XLXiaofei Lv
Xiaofei Lv
ZLZhicheng Li
Zhicheng Li
When using paired MRI and RNA sequencing, radiomic phenotypes predictive of overall survival in glioblastoma were associated with distinct pathways involved in immune regulation, tumor proliferatio... Background The biologic meaning of prognostic radiomics phenotypes remains poorly understood, hampered in part by lack of multicenter reproducible evidence. Purpose To uncover the biologic meaning ...

4 January 2024 in Biomedical Signal Processing and Control
Verified
Journal Article
Closed access
CYChongzhe Yan
Chongzhe Yan
FLFeng Liu
Feng Liu
YPYunsong Peng
Yunsong Peng
YZYuanshen Zhao
Yuanshen Zhao
JHJunjie He
Junjie He
RWRongpin Resonance Imaging Ultrasound Fusion Guided Prostate biopsy Improves Cancer Detection Following Trans Wang
Rongpin Resonance Imaging Ultrasound Fusion Guided Prostate biopsy Improves Cancer Detection Following Trans Wang

9 May 2025 in Neural Networks
Verified
Journal Article
Closed access
YZYuanshen Zhao
Yuanshen Zhao
LLLongsong Li
Longsong Li
XYXi Yu
Xi Yu
KHKe Han
Ke Han
JDJingxian Duan
Jingxian Duan
ZLZhicheng Li
Zhicheng Li

10 November 2022
Verified
Conference Paper
Closed access
YZYuanshen Zhao
Yuanshen Zhao
JDJingxian Duan
Jingxian Duan
ZLZhicheng Li
Zhicheng Li
NCNingli Chai
Ningli Chai
LLLongsong Li
Longsong Li
Predicting gastric cancer prognosis is imperative for more appropriate clinical treatment plans. Compared with traditional radiomics model adopting CT images alone, the radiopathomics is a novel medical image analysis strategy which employed the radiomcs features extracted from CT image and pathomics features extracted from pathological image to build a prediction model. In this paper, we developed a radiopathomics model to predict whether patients with gastric cancer survive more than 2 years. By using LASSO algorithm, two pathomics features, a radiomics feature and the clinical variables of TNM were selected from totally 1565 features to build the prediction model. For reflecting the advantage of the radiopathomics model, we implemented the comparison tests between the radiopathomics model with radiomics model and pathomics model. The results showed that the radiopathomics model achieved an AUC of 0.904 and an accuracy of 84.2%, which was significantly better than the other two models. It demonstrated that integrated of the microscopic level and macroscopic level phenotype information for tumor could be useful in prediction of prognosis.

20 March 2026 in Engineering Applications of Artificial Intelligence
Verified
Journal Article
Closed access
CSChunFeng Shao
ChunFeng Shao
YZYuanshen Zhao
Yuanshen Zhao
YCYinsheng Chen
Yinsheng Chen
JDJingxian Duan
Jingxian Duan
ZZZeyu Zhang
Zeyu Zhang
ZLZhi-Cheng Li
Zhi-Cheng Li

29 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
SZShanshan Zheng
Shanshan Zheng
KZKing Shing Zhang
King Shing Zhang
LKLiangyuqi Kang
Liangyuqi Kang
ZLZhi-Cheng Li
Zhi-Cheng Li
YZYuanshen Zhao
Yuanshen Zhao
Immunohistochemistry (IHC) is the gold-standard technique for molecular phenotyping and subtype stratification in breast cancer diagnosis. However, conventional IHC staining is limited by high reagent costs, lengthy processing times, tissue consumption, and difficulty scaling for high-throughput screening. Because hematoxylin and eosin (H&E) staining provides rich morphological information, H&E-to-IHC cross-modal translation is a highly promising paradigm for cost-effective, non-destructive, and efficient molecular diagnosis. Nevertheless, existing methods still suffer from two critical bottlenecks, namely insufficient pathological semantics and non-rigid spatial misalignment. To overcome these challenges, we propose the Pathology-aware Brownian Bridge Diffusion Model (PABBDM). This framework integrates Brownian bridge diffusion dynamics, pathology-aware conditioning (PaC) injection, and a deformable U-Net for virtual IHC staining. The PaC encoder directly extracts anatomical and morphological priors from H&E in pixel space and injects them as spatial guidance into the deformable U-Net to maintain pathological semantic consistency throughout the denoising process. Furthermore, a composite loss function combining latent L1 reconstruction loss and patch-level contrastive loss is employed to jointly optimize pixel-level accuracy, structural fidelity, and perceptual consistency of biomarker expression. We validate PABBDM on two public breast cancer datasets, MIST and IHC4BC, covering four key clinical biomarkers (Ki67, ER, PR, HER2). On the IHC4BC dataset, PABBDM achieves SSIM values of 0.487, 0.599, 0.696, and 0.167 for Ki67, ER, PR, and HER2, respectively, with the lowest LPIPS of 0.198 on PR, outperforming mainstream GAN-based and original BBDM methods. Both quantitative and qualitative evaluations demonstrate that PABBDM effectively enhances the structural consistency, morphological fidelity, and perceptual realism of the generated IHC images, while better preserving biomarker expression patterns compared to baseline methods. The proposed method provides a promising virtual staining approach for efficient and cost-effective analysis in breast cancer histopathology.