A5102850647- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- MRI in cancer diagnosis
- Breast Cancer Treatment Studies
- Colorectal Cancer Surgical Treatments
- Cancer Immunotherapy and Biomarkers
- Breast Lesions and Carcinomas
- Colorectal Cancer Treatments and Studies
- Cancer-related molecular mechanisms research
- Brain Tumor Detection and Classification
- Medical Imaging Techniques and Applications
- Colorectal Cancer Screening and Detection
- Colorectal and Anal Carcinomas
- Medical Image Segmentation Techniques
- Digital Imaging for Blood Diseases
- Global Cancer Incidence and Screening
- Circular RNAs in diseases
- Ferroptosis and cancer prognosis
- RNA modifications and cancer
- Advanced Neural Network Applications
- HER2/EGFR in Cancer Research
- Inflammatory Bowel Disease
- Hepatocellular Carcinoma Treatment and Prognosis
- Inflammatory Biomarkers in Disease Prognosis
- Lung Cancer Diagnosis and Treatment
Kunming Medical University
2018-2025
Zhongda Hospital Southeast University
2023-2025
Guangdong Provincial People's Hospital
2021-2024
Guangdong Academy of Medical Sciences
2021-2024
Southern Medical University
2023-2024
Key Laboratory of Guangdong Province
2023-2024
Peking University
2024
Peking University Cancer Hospital
2024
South China University of Technology
2020-2023
Southeast University
2023
Background Breast cancer is highly heterogeneous, resulting in different treatment responses to neoadjuvant chemotherapy (NAC) among patients. A noninvasive quantitative measure of intratumoral heterogeneity (ITH) may be valuable for predicting response. Purpose To develop a ITH on pretreatment MRI scans and test its performance pathologic complete response (pCR) after NAC patients with breast cancer. Materials Methods Pretreatment were retrospectively acquired who received followed by...
Brain tumor segmentation (BTS) in magnetic resonance image (MRI) is crucial for brain diagnosis, cancer management and research purposes. With the great success of ten-year BraTS challenges as well advances CNN Transformer algorithms, a lot outstanding BTS models have been proposed to tackle difficulties different technical aspects. However, existing studies hardly consider how fuse multi-modality images reasonable manner. In this paper, we leverage clinical knowledge radiologists diagnose...
Ultrasonography is an important routine examination for breast cancer diagnosis, due to its non-invasive, radiation-free and low-cost properties. However, the diagnostic accuracy of still limited inherent limitations. Then, a precise diagnose using ultrasound (BUS) image would be significant useful. Many learning-based computer-aided methods have been proposed achieve diagnosis/lesion classification. most them require pre-define region interest (ROI) then classify lesion inside ROI....
BackgroundAn artificial intelligence method could accelerate the clinical implementation of tumour-stroma ratio (TSR), which has prognostic relevance in colorectal cancer (CRC). We, therefore, developed a deep learning model for fully automated TSR quantification on routine haematoxylin and eosin (HE) stained whole-slide images (WSI) further investigated its validity patient stratification.MethodsWe trained convolutional neural network (CNN) using transfer learning, with nine-class tissue...
Tissue-level semantic segmentation is a vital step in computational pathology. Fully-supervised models have already achieved outstanding performance with dense pixel-level annotations. However, drawing such labels on the giga-pixel whole slide images extremely expensive and time-consuming. In this paper, we use only patch-level classification to achieve tissue histopathology images, finally reducing annotation efforts. We propose two-step model including phases. phase, CAM-based generate...
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allows screening, follow up, and diagnosis for breast tumor with high sensitivity. Accurate segmentation from DCE-MRI can provide crucial information of location shape, which significantly influences the downstream clinical decisions. In this paper, we aim to develop an artificial intelligence (AI) assistant automatically segment tumors by capturing dynamic changes in multi-phase a spatial-temporal framework. The main advantages...
A high abundance of tumor-infiltrating lymphocytes (TILs) has a positive impact on the prognosis patients with lung adenocarcinoma (LUAD). We aimed to develop and validate an artificial intelligence-driven pathological scoring system for assessing TILs H&E-stained whole-slide images LUAD. Deep learning-based methods were applied calculate densities in cancer epithelium (DLCE) stroma (DLCS), risk score (WELL score) was built through linear weighting DLCE DLCS. Association between WELL patient...
Crohn's disease (CD) is a chronic inflammatory bowel with an unknown etiology. Ubiquitination plays significant role in the pathogenesis of CD. This study aimed to explore functional roles ubiquitination-related genes Differentially expressed were identified by intersecting differentially (DEGs) from GSE95095 dataset Gene Expression Omnibus (GEO) database set genes. Ontology (GO) and Kyoto Encyclopedia Genes Genomes (KEGG) analyses performed. Key selected combining hub protein-protein...
Profound heterogeneity in prognosis has been observed colorectal cancer (CRC) patients with intermediate levels of disease (stage II-III), advocating the identification valuable biomarkers that could improve prognostic stratification. This study aims to develop a deep learning-based pipeline for fully automatic quantification immune infiltration within stroma region on immunohistochemical (IHC) whole-slide images (WSIs) and further analyze its value CRC.Patients from two independent cohorts...
Accurate and automated segmentation of breast tumors in dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) plays a critical role computer-aided diagnosis treatment cancer. However, this task is challenging, due to random variation tumor sizes, shapes, appearances, blurred boundaries caused by inherent heterogeneity Moreover, the presence ill-posed artifacts DCE-MRI further complicate process region annotation. To address challenges above, we propose scheme (named SwinHR)...
Lung cancer is the leading cause of death worldwide, and adenocarcinoma (LUAD) most common subtype. Exploiting potential value histopathology images can promote precision medicine in oncology. Tissue segmentation basic upstream task image analysis. Existing deep learning models have achieved superior performance but require sufficient pixel-level annotations, which time-consuming expensive. To enrich label resources LUAD to alleviate annotation efforts, we organize this challenge WSSS4LUAD...
Preoperative assessment of lymphovascular invasion (LVI) in invasive breast cancer (IBC) is high clinical relevance for treatment decision-making and prognosis.To investigate the associations preoperative magnetic resonance imaging (MRI) characteristics with LVI disease-free survival (DFS) by using machine learning methods patients IBC.Retrospective.Five hundred seventy-five women (range: 24-79 years) IBC who underwent MRI examinations at two hospitals, divided into training (N = 386)...
Abstract Background High immune infiltration is associated with favourable prognosis in patients non-small-cell lung cancer (NSCLC), but an automated workflow for characterizing infiltration, high validity and reliability, remains to be developed. Methods We performed a multicentre retrospective study of completely resected NSCLC. developed image analysis automatically evaluating the density CD3 + CD8 T-cells tumour regions on immunohistochemistry (IHC)-stained whole-slide images (WSIs),...
Automatic tissue segmentation in whole-slide images (WSIs) is a critical task hematoxylin and eosin- (H&E-) stained histopathological for accurate diagnosis risk stratification of lung cancer. Patch classification stitching the results can fast conduct WSIs. However, due to tumour heterogeneity, large intraclass variability small interclass make challenging. In this paper, we propose novel bilinear convolutional neural network- (Bilinear-CNN-) based model with module soft attention tackle...
Abstract Background Semisupervised strategy has been utilized to alleviate issues from segmentation applications due challenges in collecting abundant annotated masks, which is an essential prerequisite for training high‐performance 3D convolutional neural networks (CNNs) . Purpose Existing semisupervised methods are mainly concerned with how generate the pseudo labels regularization but not evaluate quality of explicitly. To this problem, we offer a simple yet effective reciprocal learning...
Tumor histomorphology analysis plays a crucial role in predicting the prognosis of resectable lung adenocarcinoma (LUAD). Computer-extracted image texture features have been previously shown to be correlated with outcome. However, comprehensive, quantitative, and interpretable predictor remains developed.In this multi-center study, we included patients LUAD from four independent cohorts. An automated pipeline was designed for extracting tumor region hematoxylin eosin (H&E)-stained whole...