Intelligent data analytics and systems design: AI-ML-DL-VIS
TRD3 will develop and validate analytical methods including Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) for intelligent instrument design, data/image analysis, visualization, and clinical decisions.
Decisions in critical clinical scenarios (e.g., surgery, intensive care, and obstetrics) is compromised by lack of 1) real-time intraprocedural imaging and pathologic data, 2) integration of this data with other personal and universal clinical data, and 3) effective visualization interfaces. TRD3 will develop AI-ML-DL and VIS tools to address these challenges.
First, it will build data-driven computational instruments from the perspective of intelligent optical system design, in which the optical hardware and the back-end machine learning model are jointly optimized for a given task. Second, it will develop a new visualization and AI-equipped surgical guidance platform, fusing intra-procedural iFLIM and/or iNIRS data with pre-operative volumetric CT, MRI, and PET images. Third, it will employ intelligent analytics to evaluate the multiple streams of data from this intraprocedural imaging, the patient’s medical history and prior imaging to better characterize the patient’s disease, predict its response to options for treatment appearing both during and after the procedure, and thus guide the clinician to the most effective choice. All these tools will be incorporated into clinical workflow for intraoperative guidance.
TRD3 has four specific aims, illustrated in the image below. Aim 1 uses machine learning-enabled design to enhance performance of optical instruments. Aim 2 develops expressive visualization interfaces aiding comprehension of multi-modal image data. Aim 3 develops DL/ML/AI tools for integration of data streams. Aim 4 incorporates predictive tools into clinical workflow for augmented, individual-specific, intraprocedural decisions.
The Intelligent Analytics research in this TRD seeks to accomplish the following deliverables:
- iFLIM and iDOS instruments (with increased sampling capabilities and improved SNR) optimized for clinical use by new methodologies favoring application-targeted and highly non-intuitive implementations
- An intelligent visualization system of multimodality data for surgical guidance and critical care monitoring
- A multimodality AI framework for analyzing (classification, prediction, anomaly detection) heterogeneous data
- AI tools for seamless integration of iFLIM and iDOS into the workflow to help interventional physicians and surgeons make optimal treatment decisions during and after the procedure
- Nir Pillar, Ph.D., UCLA
- Nimu Yuan, Ph.D.
- Stefan Broecker
- Yuan Ni
- Girish Kumar
- Chieh-Te Lin
- Optical Imaging to Improve Surgery & Targeted Therapy in Brain Tumors
- Diffuse Optics for Pediatric Hydrocephalus Management
- Full field OCT for cellular level structural and functional retinal imaging
- Perioperative diffuse optical imaging of tissue blood flow and oxygenation for optimization of mastectomy skin flap viability
- Intravascular NIRF-IVUS imaging of inflammation-guided arterial therapy
- In utero Repair of Fetal Myelomeningocele
- Bimodal Intraoral imaging device for detection of oral epithelial neoplasia
- Imaging Goggles for Fluorescence-Guided Surgery
- Biomarker Signatures for Delayed Cerebral Ischemia and Outcome Following Subarachnoid Hemorrhage
- Brain-based Metrics for Prolonged Field Care (PFC) Tasks
- iFLIM-based In vivo Evaluation of Thermal Injury by Cautery during Robotic Surgery Procedures
- UC Davis Alzheimer’s Disease Research Center
- Navigated Neurosurgical Procedures via 3D Augmented iFLIM
- OMX-CV, A Novel Oxygen Delivery Biotherapeutic for Hemorrhagic Shock in the Battlefield
- The Division of Nuclear Medicine within the Department of Radiology at UC Davis Medical Center
- Northern California Pet Imaging Center (NCPIC)
- EXPLORER Molecular Imaging Center (EMIC)
- Dr. Qi's laboratory in the Department of Biomedical Engineering at UC Davis
- UC Davis Center for Visualization (Director: Dr. Kwan-Liu Ma)
- UC Davis Center for Data Science and Artificial Intelligence Research (Director: Dr. Thomas Strohmer)
- UC Davis Center for Molecular and Genomic Imaging (CMGI)
- UC Davis California National Primate Research Center (CNPRC)
- UC Davis Comprehensive Cancer Center
- UC Davis/CIRM Institute for Regenerative Cures
- UC Davis Clinical and Translational Science Center (CTSC)
- Ozcan Laboratory in the UCLA Electrical and Computer Engineering Department
- Nanolab and IC Fabrication laboratory - UCLA Nanoelectronics Research Facility
- UCLA Flow Cytometry Core Laboratory
- UCLA Engineering Machine Shop
- Bauer D, Wu Q, Ma KL. FoVolNet: Fast Volume Rendering using Foveated Deep Neural Networks. IEEE Transactions on Visualization and Computer Graphics. 2022 Sep 26;29(1):515-25. PMID: 36155446. doi:10.1109/TVCG.2022.3209498.
- Pillar N, Ozcan A. Virtual tissue staining in pathology using machine learning. Expert Review of Molecular Diagnostics. 2022 Nov 2. PMID: 36440487. doi: 10.1080/14737159.2022.2153040.
- Bai B, Wang H, Li Y, de Haan K, Colonnese F, Wan Y, Zuo J, Doan NB, Zhang X, Zhang Y, Li J, Yang X, Dong W, Darrow M, Kamangar E, Lee H, Rivenson Y, Ozcan A. Label-free virtual HER2 immunohistochemical staining of breast tissue using deep learning. BME Frontiers. 2022 Oct 26;2022. doi: 10.34133/2022/9786242.
- Zhang Y, Huang L, Liu T, Cheng K, de Haan K, Li Y, Bai B, Ozcan A. Virtual staining of defocused autofluorescence images of unlabeled tissue using deep neural networks. Intelligent Computing. 2022 Oct 22. doi:10.34133/2022/9818965.
- Yang X, Bai B, Zhang Y, Li Y, de Haan K, Liu T, Ozcan A. Virtual stain transfer in histology via cascaded deep neural networks. ACS Photonics. 2022 Aug 17;9(9):3134-43. doi: 10.1021/acsphotonics.2c00932
- Shilpika F, Fujiwara T, Sakamoto N, Nonaka J, Ma KL. A Visual Analytics Approach for Hardware System Monitoring with Streaming Functional Data Analysis. IEEE Transactions on Visualization and Computer Graphics. 2022 Apr 8;28(6):2338-49. doi:10.1109/TVCG.2022.3165348.
- de Haan K, Zhang Y, Zuckerman JE, Liu T, Sisk AE, Diaz MF, Jen KY, Nobori A, Liou S, Zhang S, Riahi R. Deep learning-based transformation of H&E stained tissues into special stains. Nature communications. 2021 Aug 12;12(1):1-3. doi:10.1038/s41467-021-25221-2.
- Yang X, Huang L, Luo Y, Wu Y, Wang H, Rivenson Y, Ozcan A. Deep-learning-based virtual refocusing of images using an engineered point-spread function. ACS Photonics. 2021 Jun 18;8(7):2174-82. doi: 10.1021/acsphotonics.1c00660.
Ozcan A. Deep Learning-enabled Optics. SPIE Photonics West, BiOS Hot Topics Plenary Session, January 22-27, 2022, San Francisco, CA (Plenary)
Bai B, Wang H, Li Y, De Haan K, Colonnese F, Wan Y, Zuo J, Doan N, Zhang X, Zhang Y, Li J, Dong W, Darrow M, Kamangar E, Lee H, Rivenson Y, Ozcan A. Virtual immunohistochemical (IHC) staining of unlabeled tissue using deep learning. SPIE Optics and Photonics, August 21-25, 2022, San Diego CA, USA, Paper # 12204-36
Bai B, Wang H, Li Y, de Haan K, Colonnese F, Wan Y, Zuo J, Doan NB, Zhang X, Zhang Y, Li J, Dong W, Darrow M, Kamangar E, Lee H, Rivenson Y, Ozcan A. Deep Learning-enabled Virtual Immunohistochemical (IHC) staining of Label-Free Breast Tissue. 22nd Annual UC Systemwide Bioengineering Symposium, August 8-10, 2022, University of California, Santa Barbara, CA, USA
Li J, Garfinkel J, Zhang X, Wu D, Zhang Y, de Haan K, Wang H, Liu T, Bai B, Rivenson Y, Rubinstein G. Deep learning-enabled, non-invasive virtual histology of skin using reflectance confocal microscopy. 22nd Annual UC Systemwide Bioengineering Symposium, August 8 - 10, 2022, University of California, Santa Barbara, CA, USA
Zhang Y, de Haan K, Li J, Rivenson Y, Ozcan A. Neural network-based multiplexed and micro-structured virtual staining of unlabeled tissue. OSA/Optica Conference on Lasers and Electro-optics (CLEO), May 15-20, 2022, San Jose, CA USA
Li J, Garfinkel J, Zhang X, Wu D, Zhang Y, de Haan K, Wang H, Liu T, Bai B, Rivenson Y, Rubinstein G. Biopsy-free Virtual Histology of Skin Using Reflectance Confocal Microscopy and Deep Learning. OSA/Optica Conference on Lasers and Electro-optics (CLEO). 2022 May 15-20. San Jose, CA USA
Bai B, Wang H, Li Y, de Haan K, Colonnese F, Wan Y, Zuo J, Doan NB, Zhang X, Zhang Y, Li J, Dong W, Darrow M, Kamangar E, Lee H, Rivenson Y, Ozcan A. Deep Learning-based Virtual Immunohistochemical HER2 staining of Label-Free Breast Tissue. OSA/Optica Conference on Lasers and Electro-optics (CLEO). 2022 May 5-10. San Jose, CA USA