The imaging and visualization team in the Advanced Biomedical Computational Science (ABCS) group adopts new technologies and provides novel solutions in image analysis and scientific visualization. The team works with NCI investigators to develop AI/machine/deep learning algorithms for custom image analysis, workflows to enhance image analysis throughput, and create and maintain computational infrastructure for image analysis. The team also has extensive experience working with tumor segmentation and quantification analyses, cancer subtype identification to predict prognosis and survival, whole animal image and tissue analysis, and spatial transcriptomics. Additionally, the team supports Frederick National Laboratory for Cancer Research (FNLCR) imaging facilities such as the tissue analysis core (TAC) and the small animal imaging program (SAIP).
Expertise includes:
- Quantitative Image Analysis
- Segmentation and classification
- Feature (e.g., morphology, intensity, texture) extraction and quantification
- Spatial and neighborhood analysis
- Machine/Deep Learning Development
- Train machine/deep learning algorithms for quantitative medical image analysis
- Multi-modal medical images, including but not limited to histopathology, genomic data, clinical information, MRI, ultrasound, fluorescent microscopy, IHC, X-ray, and DICOM
- Cancer subtype classification, segmentation, and risk prediction
- Infrastructure and Web Applications
- Build digital imaging web portals capable of managing projects, experiments, patients, studies, and series
- Read DICOM files, parse metadata, and automatically process images based on pre-defined protocols
- Easy access to image uploading and downloading
- Enable online image processing and notifications
- Develop customized web plugins to enable streamlined workflow for tissue image analysis
- Provide unified web interface for scalable tissue image analysis and visualization
- Girder, Django, HistomicksTK, and Viv
- GUI and custom programming
Recent Publications:
- Predicting Molecular Subtype and Survival of Rhabdomyosarcoma Patients Using Deep Learning of H&E Images: A Report from the Children’s Oncology Group (Milewski, Jung, et al., 2023, Clin Cancer Res., PMID 36346688).
- MoNuSAC2020: A Multi-organ Nuclei Segmentation and Classification Challenge (Verma, … Jung, et al., 2021, IEEE Trans Med Imaging, PMID 34086562).
- PAIP 2019: Liver cancer segmentation challenge (Kim, … Jung, et al., Med Image Anal., PMID 33091742).
- Computer-aided diagnosis of lung cancer – a review on ACDC@LungHP challenge 2019 (Li, … Jung, et al., 2021, IEEE J Biomed Health Inform., PMID 33216724).
To learn more about the available support and/or to request a project consultation, please reach out to Dr. Hyun Jung (hyun.jung@nih.gov) or submit a request online (https://abcs-amp.nih.gov/project/request/ABCS/). Please note that you must be logged in to the NIH network to access and submit the online project request.
– Natasha Pacheco, Ph.D. (ABCS)