A5025079276- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- Digital Radiography and Breast Imaging
- Advanced X-ray and CT Imaging
- Medical Imaging Techniques and Applications
- COVID-19 diagnosis using AI
- Breast Lesions and Carcinomas
- Sarcoma Diagnosis and Treatment
- Bone Tumor Diagnosis and Treatments
- Breast Cancer Treatment Studies
- Lung Cancer Diagnosis and Treatment
- MicroRNA in disease regulation
- Advanced Image Fusion Techniques
- RNA modifications and cancer
- Cancer-related molecular mechanisms research
- Medical Image Segmentation Techniques
- MRI in cancer diagnosis
- Oral and Maxillofacial Pathology
- Natural product bioactivities and synthesis
- Head and Neck Cancer Studies
- COVID-19 Clinical Research Studies
- Gene expression and cancer classification
- Cancer Diagnosis and Treatment
- Bioactive Natural Diterpenoids Research
- Hepatocellular Carcinoma Treatment and Prognosis
Southern Medical University
2016-2025
Nanfang Hospital
2015-2025
Dalian Medical University
2025
Second Affiliated Hospital of Dalian Medical University
2025
Ganzhou People's Hospital
2022-2023
The University of Texas Southwestern Medical Center
2018-2022
Guangzhou Medical University
2021
Southwestern Medical Center
2018
Shanghai Institute of Materia Medica
2001
Chinese Academy of Sciences
2001
Microcalcification is one of the most reliable clinical features malignancy risk breast cancer, and it associated with enhanced tumour aggressiveness poor prognosis. However, its underlying molecular mechanism remains unclear.Clinical data were retrieved to analyse association between calcification bone metastasis in patients cancer. Using multiple human cancer cell lines, osteogenic cocktail model was established vitro demonstrate calcification-exacerbated metastasis. Migration invasion...
Synthetic digital mammography (SDM), a 2D image generated from breast tomosynthesis (DBT), is used as potential substitute for full-field (FFDM) in clinic to reduce the radiation dose cancer screening. Previous studies exploited projection geometry and fused data DBT volume, with different post-processing techniques applied on re-projection which may generate appearance compared FFDM. To alleviate this issue, one possible solution an SDM using learning-based method model transformation...
Limited studies have evaluated vascular markers of Ewing sarcoma (ES) using MRI. To explore the correlation between tumor and MRI perfusion parameters in ES xenografts based on MRI-pathology co-alignment. Prospective. Thirty-four xenograft models were established female athymic nude mice human-derived A673 cell line. 3 T MRI, T1-weighted (T1w) with fast spin echo sequence, T2w recovery intravoxel incoherent motion (IVIM) echo-planar diffusion-weighted dynamic contrast-enhanced (DCE-MRI)...
To establish an ensemble machine learning (ML) model for predicting the risk of futile recanalization, malignant cerebral edema (MCE), and herniation (CH) in patients with acute ischemic stroke (AIS) who underwent mechanical thrombectomy (MT) recanalization.This prospective study included 110 premorbid mRS ≤ 2 met inclusion criteria. Futile recanalization was defined as a 90-day modified Rankin Scale score >2. Clinical imaging data were used to construct five ML models that fused into...
Objective: accurately classifying the malignancy of lesions detected in a screening scan is critical for reducing false positives. Radiomics holds great potential to differentiate malignant from benign tumors by extracting and analyzing large number quantitative image features. Since not all radiomic features contribute an effective model, selecting optimal feature subset critical. Methods: this work proposes new multi-objective based selection (MO-FS) algorithm that considers sensitivity...
Bone suppression of chest radiographs (CXRs) is potentially useful for diagnosing lung diseases radiologists and computer–aided diagnosis. This paper presents a cascaded convolutional network model in wavelet domain (Wavelet-CCN) bone single conventional CXR. Wavelet coefficients are sparse suitable as the output network. The networks trained to predict images from CXRs, using real two-exposure dual energy subtraction (DES) CXRs training data. By combining multi-level decomposition...
This study aims to assess the CT and MRI features of head neck osteosarcoma (HNO).37 HNOs were identified, following imaging characteristics reviewed on MRI.A total 37 patients(age 41.5 ± 15.0 years old; 16 males, 21 females) included in study. Tumours occurred maxilla (16, 43.2%), mandible (8, 21.6%), skull base (6, 16.2%), calvarium (5, 13.5%), paranasal sinuses (1, 2.7%) cervical soft tissue 2.7%). patients received radiotherapy for nasopharyngeal carcinoma. Three (8.1%) developed...
Cathepsin L(CTSL), a lysosomal endopeptidase was found overexpressed in Breast cancer (BC). The purpose of this work to investigate the possible role CTSL development BC. RNA interference(RNAi) with small hairpin RNAs(CTSL-shRNA) and plasmid were used identify effects on malignant behaviors MCF-7 SKBR-3 selected as cell models vitro vivo. results showed that down-regulation can significantly inhibit proliferative invasive ability cell, while up-regulation cells had opposite effects....
Computer aided detection (CADe) for breast lesions can provide an important reference radiologists in cancer screening. Architectural distortion (AD) is a type of lesion that difficult to detect. A majority CADe methods focus on detecting the radial pattern, which main characteristic typical ADs. However, few atypical ADs do not exhibit such pattern. To improve performance and ADs, we propose deep-learning-based model used mammary gland distribution as prior information detect digital...
Contrast-enhanced mammography (CEM) is an imaging tool for breast cancer detection. Most quantitative analyses of CEM involve two phases, and it unknown whether added delayed phase can improve its diagnostic performance compared to dynamic contrast-enhanced magnetic resonance (DCE-MRI). This study aimed evaluate the improves in distinguishing malignant benign masses.This prospective enrolled 111 women with pathologically confirmed masses. was performed after injection contrast agent between...
Mammography is the first option for screening breast cancer. However, some lesions may be missed due to superimposition of parenchymal patterns/tissues. Current enhancement methods can highlight specific tissues, but edges are weakened or extra noise added. For improving solution, this paper proposed a multiscale contrast mammogram using dynamic unsharp masking (UM) in Laplacian pyramid. pyramid utilized preserve fine structure at each scale. Dynamic UM presented adaptively enhance details...
This study aims to assess the CT and MRI features of calvarium skull base osteosarcoma (CSBO).The pathological characteristics 12 cases pathologically confirmed CSBO were analyzed retrospectively.12 patients (age range 9-67 years; 3 male, 9 female) included in study. Tumours occurred (7, 58.3%), temporal (4, 33.3%) frontal (1, 8.3%). Among all, six received radiotherapy for nasopharyngeal carcinoma. According pathology, 11 out tumours high-grade (91.7%). On CT, all had soft tissue mass...
This study established an interpretable machine learning model to predict the severity of coronavirus disease 2019 (COVID‐19) and output most crucial deterioration factors. Clinical information, laboratory tests, chest computed tomography (CT) scans at admission were collected. Two experienced radiologists reviewed for patterns, distribution, CT scores lung abnormalities. Six models COVID‐19. After parameter tuning performance comparison, optimal was explained using Shapley Additive...
Objectives Accurate identifying head and neck squamous cell cancer (HNSCC) patients at high risk of local persistence/recurrence (P/R) is importance for personalized patient management. Here we developed a multi-objective, multi-classifier radiomics model early HNSCC P/R prediction based on post-treatment PET/CT scans clinical data. Materials methods We retrospectively identified 328 individuals (69 have P/R) with treated definitive radiation therapy our institution. The median follow-up...