Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of...
Transcript of Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of...
![Page 1: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/1.jpg)
Making Sense of Optical Contrast For Point-of-Care Pathology
Steven SensarnContag Lab
SCIT Program Seminar06/19/13
![Page 2: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/2.jpg)
• Current paradigm of histology for cancer diagnosis
• Dual-axis confocal (DAC) fluorescence microscopy
• Emulating H&E (hematoxylin & eosin) tissue staining with DAC microscopy
• DAC imaging and machine learning to identify cancer
Outline
![Page 3: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/3.jpg)
Current Paradigm of Histology for Cancer Diagnosis
![Page 4: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/4.jpg)
• Conventional hematoxylin and eosin (H&E) stained pathology slides take 6-8 hours to prepare
• They require a multi-step process with skilled but repetitive manual steps
• They result in a 5 micron thick section
Key Limitations of Current State-of-the-Art Tissue Microscopy
Slide courtesy of David Rimm, Yale
![Page 5: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/5.jpg)
Tissue
removed
from patient
Tissue placed
in transport
media (fixative)
Frozen
Section only
if Dx affects
surgical
procedure
Gross examination and description
Dissection and preparation of sample
cassettes according to protocol
Further fixation and processing
included paraffin embedding
Mounting at embedding station and
microtome sectioning of tissue blocks
Deparaffinization, staining of
tissue and coverslipping slides
To Pathologist
Takes 6-8 hours
From Fresh Tissue to Glass Slide
Slide courtesy of David Rimm, Yale
![Page 6: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/6.jpg)
Dual-Axis Confocal Microscopy Benefits
• Optical sectioning without cutting tissue
• Image stacks: hundreds of tissue sections in one 3D acquisition
• With miniature microscopes, tissue sections can be analyzed in the operating room or gross pathology room
![Page 7: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/7.jpg)
Dual-Axis Confocal Fluorescence Microscopy
![Page 8: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/8.jpg)
Dual-Axis Confocal (DAC) MicroscopyComparing two confocal scanning microscope architectures:
1. Uses simple low-NA optics to obtain thin optical-section thickness DZ
2. Reduces scattering noise in 3-D volume image
3. Easy to miniaturize using a MEMS scanner
Micro-Mirror
Miniaturized systems enable confocal microscopy at the point of care
MEMS mirrors developed by the Solgaard Lab at Stanford
![Page 9: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/9.jpg)
Implantable DAC
Evolution of DAC at Stanford
Tabletop DAC
Handheld DAC
1st Generation Endoscopic DAC
Multi-Modality DAC
Under Development
![Page 10: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/10.jpg)
Tabletop DAC
Handheld DAC
Endoscopic DAC
Micro-Endoscopic
DAC
Current DAC Microscopes
![Page 11: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/11.jpg)
Image-guided Tissue Sampling and Multiplexed Molecular Analysis
Wide Field Guidance
Microscopy
PEAK Microsurgical and Sampling(Palanker lab)
![Page 12: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/12.jpg)
Emulating H&E (Hematoxylin & Eosin) Staining with DAC Microscopy
![Page 13: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/13.jpg)
Volumetric Imaging of Colon Tissue
• Standard H&E stain
• Hematoxylin: stains nucleic acids (blue-purple)
• Eosin: stains proteins (pink)
• DAC imaging requires fluorescent dyes
• FDA-approved clinical dyes:
– Indocyanine green (ICG): stains membrane & extracellular matrix (ECM)
– Methylene blue: stains cytoplasm & nucleus
• Nuclear-specific dyes for ex-vivo use
– DRAQ5
– DAPI (requires UV light)
![Page 14: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/14.jpg)
Collaboration with David Rimm, Yale
Volumetric Imaging of Colon Tissue
![Page 15: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/15.jpg)
• 2x2x9 averaging (XxYxZ) to improve SNR• Start with blank (white) image• Invert grayscale images and color map (785/ICG => magenta,
660/DRAQ5 => blue). In inverted images, darker => more fluorescence
• Use non-inverted grayscale images for alpha channel (transparency). For transparency mask, darker => more transparent
• Overlay both inverted, color-mapped images onto the white background using respective transparency masks
Image Processing to Emulate H&E Staining
Overlays
Transparency Masks
Simulated H&E
DRAQ5 (nuclear stain) ICG (membrane/ECM stain)
![Page 16: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/16.jpg)
DRAQ5 + methylene blue + ICG
DRAQ5 + ICG Colon Tissues
![Page 17: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/17.jpg)
Vision: DAC Array in the Gross Pathology Room
DAC Array
Tissue
DAC Array
DAC
4X
10X
20X
“Google Maps”-like pan and zoom
![Page 18: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/18.jpg)
DAC Imaging and Machine Learning to Identify Cancer
![Page 19: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/19.jpg)
Tumor
Normal
Cavity
Normal
Residual
DAC
Guided Resection by Morphological Analysis
Brain Tumor
Resection
Staining & Microscopy
• After bulk resection, the DAC microscope can image margins with FDA-approved dyes, 300-um deep in tissue
• Machine learning algorithms can be trained to identify cancerous margins
![Page 20: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/20.jpg)
Machine Learning to Identify Cancer
• Developed by Kirk Gossage, PhD
• Texture-based and neural-network-derived segmentation algorithm
• Contag Lab collaborates with Levenson Lab at UC Davis to apply the algorithm to DAC images
• Breast cancer tissue stained with hematoxylin and p53 immunostain
• Hematoxylin channel alone (unmixed) used to identify cancer regions (pink overlay)
Expert Opin. Med. Diagn. (2008) 2(9):1067-1081
![Page 21: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/21.jpg)
Entire process takes less than 30 s
Machine Learning: Teaching Software to Find Hikers (Complex Mixtures of Head, Clothes, Legs) Simply by Drawing
Slide courtesy of Richard Levenson, UC Davis
![Page 22: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/22.jpg)
Proof-of-Concept Experiment with DAC Microscope
• SCID mice with MDA-MB-231 human breast cancer xenografts in the mammary fat pad were sacrificed from another study
• Tumors were collected along with normal brain tissue (cerebellum). Breast cancer and brain tissue should be easy for a computer to distinguish
• 3 tumor and 3 normal cerebellum samples were soaked in ICG and imaged on the DAC microscope
• One tumor/normal pair imaged twice for a total of 4 tumor and 4 normal image stacks
• 63 optical sections were hand-picked from the data sets for computer analysis
![Page 23: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/23.jpg)
Tumors Cerebellum
Breast cancer vs. normal cerebellum, ICG staining
![Page 24: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/24.jpg)
Normal cerebellum MDA-MB-231 tumors
• Collaboration with Richard Levenson (UC Davis)
• 24 normal cerebellum and 39 breast cancer flank tumor sections
• About 15 of the images were used as a training set
Machine Learning: Flank Tumors vs. Cerebellum
![Page 25: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/25.jpg)
Larger Data Set for Medulloblastoma Detection
GFP-luc DAOY
8 ul of concentrated GFP-luc DAOY cells were injected into 8 weeks or older nude mice under anaesthesia. Examples show imaging of orthotropic and flank models 10 minutes after mice were injected IP with a 100 ul of luciferin.
43 Days 29 Days after injection 21 Days after injection
Luciferase on IVIS Luciferase over time
ALu
c ac
tivi
ty in
rad
ian
ce x
10^
6
Days after injection
B
1 6010 20 30 40 50
1000
1200
800
600
400
200
0
Slide courtesy of Markus Deutschmann
![Page 26: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/26.jpg)
Summary
• Current histological analysis is labor-intensive and time consuming
• DAC microscopy offers rapid optical sectioning and may emulate traditional H&E staining or guide selection of tissues for traditional processing
• DAC imaging with FDA-approved dyes and computer vision analysis may guide surgical resection by identifying cancer margins at the point of care
![Page 27: Making Sense of Optical Contrast For Point-of-Care Pathology · 2018-12-12 · Making Sense of Optical Contrast For Point-of-Care Pathology Steven Sensarn Contag Lab SCIT Program](https://reader033.fdocuments.in/reader033/viewer/2022042207/5ea9ebd462156854c50743a3/html5/thumbnails/27.jpg)
Acknowledgements
• DAC Microscope Development– Contag Lab
• Jonathan Liu
• Mike Mandella
• Hyejun Ra
• Nathan Loewke
– Solgaard Lab
• Jae-Woong Jeong
• DAC Imaging to Emulate H&E Staining
– David Rimm (Yale)
• DAC Imaging and Machine Learning for Cancer Detection– Contag Lab
• Markus Deutschmann
• Tobi Schmidt
– Scott Lab
• Helen Rayburn
– Richard Levenson (UC Davis)
• Special Thanks– Contag Lab
• Sophie Kusy, Laura Bronsart, Bonnie King, Ellis Garai, Stephan Rogalla
– Small Animal Imaging Core Facility
– Christopher Contag, Olav Solgaard, Matthew Scott
• Funding and Support– SCIT Fellowship
– Network for Translational Research (NTR)
– Stanford In Vivo Cellular and Molecular Imaging Centers (ICMIC)
– Stanford Center for Children’s Brain Tumors (CCBT)
– Stanford Photomedicine Center