Publications by authors named "ZaiYi Liu"

White adipose tissue (WAT) browning is considered a promising strategy to combat obesity and related metabolic diseases. Currently, fat-water fraction (FWF) has been used as a marker for the loss of lipids associated with WAT browning. However, FWF may not be sensitive to metabolic changes and cannot specifically reflect iron-regulated metabolism during browning.

View Article and Find Full Text PDF

The batch effect is a nonbiological variation that arises from technical differences across different batches of data during the data generation process for acquisition-related reasons, such as collection of images at different sites or using different scanners. This phenomenon can affect the robustness and generalizability of computational pathology- or radiology-based cancer diagnostic models, especially in multi-center studies. To address this issue, we developed an open-source platform, Batch Effect Explorer (BEEx), that is designed to qualitatively and quantitatively determine whether batch effects exist among medical image datasets from different sites.

View Article and Find Full Text PDF

The deep learning technique has been shown to be effectively addressed several image analysis tasks in the computer-aided diagnosis scheme for mammography. The training of an efficacious deep learning model requires large data with diverse styles and qualities. The diversity of data often comes from the use of various scanners of vendors.

View Article and Find Full Text PDF

Background: Tertiary lymphoid structures (TLS) are major components in the immune microenvironment, correlating with a favorable prognosis in colorectal cancer. However, the methods used to define and characterize TLS were not united, hindering its clinical application. This study aims to seek a more stable method to characterize TLS and clarify their prognostic value in larger multicenter cohorts.

View Article and Find Full Text PDF

Multi-modal image analysis using deep learning (DL) lays the foundation for neoadjuvant treatment (NAT) response monitoring. However, existing methods prioritize extracting multi-modal features to enhance predictive performance, with limited consideration on real-world clinical applicability, particularly in longitudinal NAT scenarios with multi-modal data. Here, we propose the Multi-modal Response Prediction (MRP) system, designed to mimic real-world physician assessments of NAT responses in breast cancer.

View Article and Find Full Text PDF
Article Synopsis
  • Prognostic assessment is challenging in medicine due to limited labeled data, prompting the development of ContraSurv, a weakly-supervised learning framework using contrastive learning to improve predictions from 3D medical images.* -
  • ContraSurv leverages self-supervised information from unlabeled data and weakly-supervised cues from censored data, incorporating a specialized Vision Transformer architecture and innovative contrastive learning methods.* -
  • The framework was tested on three cancer types and two imaging modalities, demonstrating superior performance compared to existing methods, especially in datasets with high censoring rates.*
View Article and Find Full Text PDF
Article Synopsis
  • Previous research hinted at differences in cerebellar white matter integrity related to stages of Parkinson's disease (PD), but the progression of these changes over time was not well understood.
  • The study involved 124 patients with PD, who underwent multiple MRI scans to track microstructural integrity in cerebellar white matter, focusing on the connections between the cerebellum and other brain areas, while assessing clinical symptoms and dopamine levels.
  • Results showed a non-linear pattern in cerebellar white matter changes, highlighting an initial increase in integrity followed by a decline, which mirrored declines in dopamine levels and clinical symptoms, suggesting adaptive changes in the brain as PD progresses.
View Article and Find Full Text PDF
Article Synopsis
  • This study looked at how changes in the skeletal muscle index (SMI) before and after surgery affect patients with colorectal cancer (CRC).
  • Researchers analyzed data from over 2,200 patients who had surgery between 2012 and 2019, checking their body muscle levels using CT scans.
  • The results showed that people with lower muscle levels after surgery had worse survival rates compared to those who maintained high muscle levels, especially at 6, 9, and 12 months later.
View Article and Find Full Text PDF

Background: Early prediction of lymph node status after neoadjuvant chemotherapy (NAC) facilitates promptly optimization of treatment strategies. This study aimed to develop and validate a deep learning network (DLN) using baseline computed tomography images to predict lymph node metastasis (LNM) after NAC in patients with locally advanced gastric cancer (LAGC).

Methods: A total of 1205 LAGC patients were retrospectively recruited from three hospitals between January 2013 and March 2023, constituting a training cohort, an internal validation cohort, and two external validation cohorts.

View Article and Find Full Text PDF

Objectives: Although neoadjuvant immunochemotherapy has been widely applied in non-small cell lung cancer (NSCLC), predicting treatment response remains a challenge. We used pretreatment multimodal CT to explore deep learning-based immunochemotherapy response image biomarkers.

Methods: This study retrospectively obtained non-contrast enhanced and contrast enhancedbubu CT scans of patients with NSCLC who underwent surgery after receiving neoadjuvant immunochemotherapy at multiple centers between August 2019 and February 2023.

View Article and Find Full Text PDF

Background: Brown adipose tissue (BAT) is metabolically activatable and plays an important role in obesity and metabolic diseases. With reduced fat-water-fraction (FWF) compared with white adipose tissue (WAT), BAT mass and its functional activation may be quantified with Z-spectra MRI, with built-in FWF and the metabolic amide proton transfer (APT) contrasts.

Purpose: To investigate if Z-spectral MRI can quantify the mass and metabolic activity of adipose tissues.

View Article and Find Full Text PDF

Background: Pathological axillary lymph node (pALN) burden is an important factor for treatment decision-making in clinical T1-T2 (cT1-T2) stage breast cancer. Preoperative assessment of the pALN burden and prognosis aids in the individualized selection of therapeutic approaches.

Purpose: To develop and validate a machine learning (ML) model based on clinicopathological and MRI characteristics for assessing pALN burden and survival in patients with cT1-T2 stage breast cancer.

View Article and Find Full Text PDF

Novel drug-target interaction (DTI) prediction is crucial in drug discovery and repositioning. Recently, graph neural network (GNN) has shown promising results in identifying DTI by using thresholds to construct heterogeneous graphs. However, an empirically selected threshold can lead to loss of valuable information, especially in sparse networks, a common scenario in DTI prediction.

View Article and Find Full Text PDF

Automated breast tumor segmentation on the basis of dynamic contrast-enhancement magnetic resonance imaging (DCE-MRI) has shown great promise in clinical practice, particularly for identifying the presence of breast disease. However, accurate segmentation of breast tumor is a challenging task, often necessitating the development of complex networks. To strike an optimal tradeoff between computational costs and segmentation performance, we propose a hybrid network via the combination of convolution neural network (CNN) and transformer layers.

View Article and Find Full Text PDF

Thermogenic brown adipose tissue (BAT) has a positive impact on whole-body metabolism. However, in vivo mapping of BAT activity typically relies on techniques involving ionizing radiation, such as [F]fluorodeoxyglucose ([F]FDG) positron emission tomography (PET) and computed tomography (CT). Here we report a noninvasive metabolic magnetic resonance imaging (MRI) approach based on creatine chemical exchange saturation transfer (Cr-CEST) contrast to assess in vivo BAT activity in rodents and humans.

View Article and Find Full Text PDF

Predicting the gene mutation status in whole slide images (WSI) is crucial for the clinical treatment, cancer management, and research of gliomas. With advancements in CNN and Transformer algorithms, several promising models have been proposed. However, existing studies have paid little attention on fusing multi-magnification information, and the model requires processing all patches from a whole slide image.

View Article and Find Full Text PDF

Objectives: To develop and validate machine learning models for human epidermal growth factor receptor 2 (HER2)-zero and HER2-low using MRI features pre-neoadjuvant therapy (NAT).

Methods: Five hundred and sixteen breast cancer patients post-NAT surgery were randomly divided into training (n = 362) and internal validation sets (n = 154) for model building and evaluation. MRI features (tumour diameter, enhancement type, background parenchymal enhancement, enhancement pattern, percentage of enhancement, signal enhancement ratio, breast oedema, and apparent diffusion coefficient) were reviewed.

View Article and Find Full Text PDF

Magnetic resonance imaging (MRI)-based deep neural networks (DNN) have been widely developed to perform prostate cancer (PCa) classification. However, in real-world clinical situations, prostate MRIs can be easily impacted by rectal artifacts, which have been found to lead to incorrect PCa classification. Existing DNN-based methods typically do not consider the interference of rectal artifacts on PCa classification, and do not design specific strategy to address this problem.

View Article and Find Full Text PDF
Article Synopsis
  • Histopathological tissue classification is important in computational pathology, and while deep learning models are effective, they face privacy concerns due to centralized training.
  • Federated learning (FL) can protect privacy by keeping data local, but traditional FL approaches need lots of labeled data and multiple communication rounds, making them impractical in real-life situations.
  • The proposed FedDBL framework addresses these challenges by allowing high classification performance with fewer samples and just one communication round, reducing data dependency and significantly improving efficiency while maintaining privacy and security.
View Article and Find Full Text PDF
Article Synopsis
  • This study developed a stacking model to predict how axillary lymph nodes respond to neoadjuvant chemotherapy in breast cancer, utilizing MRI data.
  • The analysis involved 1,153 patients, with the model showing strong accuracy in distinguishing responses and a notably lower false-negative rate compared to traditional radiologists.
  • Results indicated significant differences in disease-free survival between high-risk and low-risk groups, suggesting the model's potential for enhancing patient outcomes post-chemotherapy.
View Article and Find Full Text PDF

Background: Contrast-enhanced CT scans provide a means to detect unsuspected colorectal cancer. However, colorectal cancers in contrast-enhanced CT without bowel preparation may elude detection by radiologists. We aimed to develop a deep learning (DL) model for accurate detection of colorectal cancer, and evaluate whether it could improve the detection performance of radiologists.

View Article and Find Full Text PDF

Advances in spatial transcriptomics provide an unprecedented opportunity to reveal the structure and function of biology systems. However, current algorithms fail to address the heterogeneity and interpretability of spatial transcriptomics data. Here, we present a multi-layer network model for identifying spatial domains in spatial transcriptomics data with joint learning.

View Article and Find Full Text PDF
Article Synopsis
  • - The study investigates how the spatial distance between tumor-infiltrating lymphocytes (TILs) and tumor cells can predict immune response effectiveness and prognosis in lung adenocarcinoma (LUAD) patients, emphasizing that this relationship has been inadequately analyzed using standard imaging techniques.
  • - Researchers utilized a deep learning model (HoVer-Net) to accurately segment cell types in tumor regions from H&E-stained images and measured the distance (DIST) between tumor cells and lymphocytes to assess its impact on disease-free survival (DFS) in different patient cohorts.
  • - Findings revealed that shorter DIST correlates with significantly improved DFS across multiple patient sets, and incorporating DIST with clinical factors resulted in better prognostic predictions, highlighting its
View Article and Find Full Text PDF

Advancements in single-cell technologies concomitantly develop the epigenomic and transcriptomic profiles at the cell levels, providing opportunities to explore the potential biological mechanisms. Even though significant efforts have been dedicated to them, it remains challenging for the integration analysis of multi-omic data of single-cell because of the heterogeneity, complicated coupling and interpretability of data. To handle these issues, we propose a novel self-representation Learning-based Multi-omics data Integrative Clustering algorithm (sLMIC) for the integration of single-cell epigenomic profiles (DNA methylation or scATAC-seq) and transcriptomic (scRNA-seq), which the consistent and specific features of cells are explicitly extracted facilitating the cell clustering.

View Article and Find Full Text PDF