Image Analysis for Fluorescence

Transcription

1 Image Analysis for Fluorescence

2 Terminology Table Image Analysis Macro Colocalization Intensity Dye AFI The extraction of meaningful information from digital images by means of digital image processing techniques. A macro is a pre-saved set of input parameters that have been adjusted for a specific application. Analysis technique that determines the contribution of each dye at every pixel location. Intensity of a pixel is proportional to the dye emission strength in the same area. Intensity ranges from 0 (no detectable signal) up to 1023 (very strong signal) for images captured with a 10 bit setting. Analysis intensity input parameters (min and max) are from 0-1 representing the intensity scale from 0-100%. Dye is the name of the fluorescent dye. There is no need to enter color values to deconvolve dyes for Aperio s FL algorithms. Fluorescent images taken with Aperio s ScanScope FL are separate monochrome images per dye that are pseudo-colored and fused to make one image. Aperio Fused Image -The AFI file is a special type of image file. It is a composite file that points to the separate channel images that make up the fused image. Page 2

3 FL Image Analysis Illustrated Image Analysis extracts meaningful data: Analyzing fluorescent digital slides helps you quantitate the area of staining of a particular biomarker. Image analysis tools help to automate such repetitive processes and provide quantitative data that is accurate and repeatable, telling you more about each slide beyond the capabilities of manual microscopy. Aperio s FL image analysis can provide valuable area and intensity data and answer questions such as: How much area of the tissue is stained for a particular biomarker? What is the average intensity of the biomarker staining throughout the tissue or region of interest? Do multiple biomarkers co-localize in the tissue, or in the cells, and if so, to what extent? Image Analysis tools can present a markup image to highlight analysis results. Original Markup Page 3

4 FL Image Analysis Illustrated How Image Analysis works: Apply an algorithm directly to the whole digital slide image or selected region of the digital slide image. Aperio ScanScope FL images are comprised of separate images (one for each fluorescent dye on the slide) resulting in an Aperio Fused Image or AFI. Since images are captured during scanning as monochrome images and pseudo-colored based on a reference chart (by the dye s name) pixels are already separated by the fluorescent dye or dye. Image analysis algorithms measures pixel intensity (proportional to dye emission) to determine the average intensity of single dyes and when dye appear together. Image analysis algorithms count pixels to determine: Area of analysis. The total area of staining (all dyes). The area of single dyes. The area of dyes appearing in combination (in the same pixel location) Intensity: 0 = signal undetectable 1023 = very strong signal Image analysis scales the intensity range 0-1 (0-100%) DAPI CY3 Pixel CY5 Page 4

5 Aperio Image Analysis Tools Area Quantification Positive Pixel Count FL Area Quantification FL Positive Pixel Count Color Deconvolution Colocalization FDA Cleared: Her2 510(k) & ER/PR 510(k) Note: User guides, manuals training presentations and demo videos are available on Aperio s Member Center at: members.aperio.com Cell Quantification Nuclear Membrane Object Identification Microvessel Analysis Rare Event Detection Each algorithm has a set of control parameters that allow you to tailor it to your specific needs. Several of Aperio's image analysis algorithms are FDA cleared for specific clinical applications, and are intended for research use for other applications. Note that each algorithm has it s own user guide. Please see the Intended Use section of the respective user guide for the specific cleared applications you wish to use and for details on in vitro diagnostic use. Page 5

6 Aperio Image Analysis Tools Aperio Image Analysis Algorithm Combined Image Analysis Tool (Algorithm + Genie ) A smart tissue classifier that recognizes significant tissue components coupled with an FL image analysis algorithm. Genie Histology Pattern Recognition Genie is intended for research use only. Genie is not covered in Image Analysis 101 for Fluorescence. Please see TRN-0017 for user related course content. Page 6

7 Aperio Image Analysis FL Tools Positive Pixel Count FL: Quantifies the area and intensity of a selected fluorescent dye. Area and area fraction analysis, for example, analysis of fluorescent digital slides showing amyloid plaque myelin and nuclei. Licensed without fee with other Aperio software. Amyloid Plaque Quantification Original Markup Page 7

8 Aperio Image Analysis FL Tools Area Quantification FL: Quantifies the area and intensities of up to 3 dyes. Area and colocalized marker analysis, for example, analysis of fluorescent digital slides showing gene and protein expressions. Pancreatic Islet Protein Expression Original Markup Page 8

9 Aperio Image Analysis Tools Histology Pattern Recognition Genie: A smart tissue classifier that recognizes significant tissue components and aids analysis. Original Islet Exocrine Classification Quantification Page 9

10 Viewing Fluorescence Slides Viewing Fluorescence Slides in ImageScope Images from the ScanScope FL are grayscale images, pseudo-colored during the scanning process. ImageScope offers a full range of fluorescence features for ScanScope FL fused (composite) Images: Change the display color for each channel image. Adjust brightness, contrast, and gamma (viewing the results on the image and on a histogram display). Hide image channels or display them in grayscale. Adjust registration between channels. Note: Modifying viewing options in ImageScope does not affect the analysis. Channels Hidden Original Color Adjustments Channel in Grayscale Page 10

11 Performing Analysis Selecting Areas to Analyze Analyze entire digital slide or selected areas. Use ImageScope drawing tools to select or exclude areas to analyze. Pen - draw free-form area of interest. Negative Pen draw free-form area to exclude from analysis. Rectangle draw a rectangular area of interest. View annotation information in ImageScope s Annotations window. Page 11

12 Performing Analysis Methods to Run Analysis In most cases, it is expected that a digital slide will be opened in Spectrum. For this training, all images will be opened in Spectrum. ImageScope Analysis Open a local digital slide image (that is, an image on your workstation or on the network where your workstation can see it via Microsoft file sharing) and analyze using ImageScope. Use ImageScope to connect to an image on an Aperio ImageServer. Spectrum Plus Analysis Open an image from Spectrum and initiate analysis using a Spectrum licensed algorithm macro. Batch Analysis Analyze a batch of digital slide images that reside on the ImageServer by selecting multiple images and using Spectrum s Analyze command Page 12

13 ImageScope Image Analysis Analysis Workflow Using an Existing Macro Run analysis using a macro with input parameters already tuned for your application. Ready to analyze an image. Licensed Image Analysis Algorithms on Spectrum Select view menu>analysis, choose macro and analysis options and click Analyze. View results. Intensity Spectrum Other Inputs Dye Macros are pre-saved sets of algorithm input parameters that have been adjusted for specific image analysis applications. Page 13

14 ImageScope Image Analysis Analysis Workflow Modifying an Existing Macro Select an existing macro, modify the input parameters for your application and save. Ready to analyze an image. Licensed Image Analysis Algorithms on Spectrum Select view menu>analysis, select macro name and click Test button. Select Select view view menu>analysis, menu>analysis, choose choose macro macro and and analysis analysis options options and and click click Analyze. Analyze. View results. Test Window Adjust input parameters for your slide/application. Use Tuning window as needed. Click Run to analyze and get test results. NO Are the results & markup acceptable? YES Click Save macro from test window. Spectrum Page 14

15 ImageScope Image Analysis Analysis Workflow Creating a New Macro Use this method if the algorithm has not been used and saved as a macro. Use to create a macro using the algorithm s default parameters. Ready to analyze an image. Select an algorithm to modify Licensed Image Analysis Algorithms on Spectrum Select view menu>analysis and click Create button. Select view menu>analysis, Select view choose menu>analysis, macro and choose analysis macro options and and analysis click Analyze. options and View click results. Analyze. Test Window Adjust input parameters for your slide/application. Use Tuning window as needed. Click Run to analyze and get test results. NO Are the results & markup acceptable? YES Click Save macro from test window. Spectrum Page 15

16 ImageScope Image Analysis Analysis Window When performing analysis on an image opened from Spectrum, the Algorithm Server Job displays registered macros. Go to the Test window to modify input parameters or click Create to make a new Macro. List of macros Analysis options Analysis Buttons Test modify existing macro for selected algorithm and test before saving. Create creates new macro based on the original algorithm. Analyze run analysis. Cancel cancel current analysis job. Page 16

17 ImageScope Image Analysis Test Window test out input parameters before saving them Tune algorithm input parameters and view results, select outputs displayed in Spectrum and save settings as a new macro. List of input parameters Select Algorithm select to load an algorithm with its default settings. Import Macro import an image analysis macro. Save Macro save an image analysis macro to Spectrum. Note: If image was opened and analyzed locally, an Export Macro button will appear instead on the Save button. Turn on result plots here! Analysis options List of output parameters (to display in Spectrum) Tune Allows operators to see instant markup up results in a separate window and results in the annotations window while adjusting the algorithm s input parameters. Run run analysis. Page 17

18 ImageScope Image Analysis Creating New Macros from your licensed algorithms Only users with the appropriate Spectrum administrator permissions can create new macros and register them in Spectrum. After opening a digital slide in ImageScope, open the View menu and select Analysis. Click Create, the Select an Algorithm window appears. Select the Algorithm you want to create a macro for, the Analysis window appears. Note: ImageScope may highlight the last selected macro in this window; ignore this and click Create. Note: Algorithms selected for this list will have their default input parameters. Page 18

19 ImageScope Image Analysis Creating New Macros Modify the parameters. Use the Tuning window to see the effect of the parameter changes. Run analysis and continue to adjust the inputs until desired results are achieved. Modify input parameters Remember, the output button allows you to select the results that will appear in Spectrum. Tuning window Page 19

20 ImageScope Image Analysis Creating New Macros Select Save Marco button. Name the macro with a naming convention that helps identify it and its application. Using the same name as an existing macro will display a prompt asking if you want to overwrite the existing macro. Once saved, your new macro will appear in the Analysis window s list of macros. Page 20

21 ImageScope Image Analysis More on the Algorithm Tuning Window Quickly see the results of analyzing an area of an image or test changes made to the algorithm parameters. Toggle on/off markup Page 21

22 ImageScope Image Analysis More on the Algorithm Tuning Window As the parameters are changed, the Tuning window will display the markup image for current parameters. The Annotations window displays the numerical results in a Tuning Layer. Immediately see results after parameter changes! Note: Upon closing the Tuning window, the Tuning Layer disappears. Analysis results from the tuning layer cannot be saved. After parameter change Page 22

23 ImageScope Image Analysis Analysis Parameters Inputs Dye(s) Typing in the Dye name points the algorithm to the correct component image of the AFI (or dye color channel). Intensity Dye intensity minimum and maximum values determine the intensity range to be analyzed. Other Attributes: Markup Image Type Color coding of pixels to display area of dye (single dye and or combination of dyes together). Display Plots - Select to display result plots and set histogram start/end values and number of bins. ImageScope Inputs entered in ImageScope from the Test window. Intensity -OR- Analysis Inputs Other Attributes Dye(s) Spectrum Inputs from saved macro or from the default algorithm. Analysis Output or Results Page 23

24 ImageScope Image Analysis Analysis Parameters Outputs (Analysis Results) Area and Intensity Results Total area of analysis (mm 2 ). Total stained area (mm 2 ). Total area of each dye (mm 2 ). Percent for each dye. Average intensity for each dye. Percent and intensity colocalized dye results. Result plots. Markup A color coded layer displayed over the area of analysis that corresponds with the analysis results. Displays only the pixels that fall within the set intensity range. The markup layer displays colors that match the pseudo-colors used to color the grayscale images in the scanning process. Intensity Other Attributes Analysis Outputs Dye(s) Analysis Output or Results ImageScope Analysis results displayed in ImageScope in the Annotations window. Spectrum Analysis saved in Spectrum (If performed server-side). Sever- side analysis: When an image is accessed from Spectrum and analyzed using an Spectrum licensed algorithm. Page 24

25 ImageScope Image Analysis Analysis Results - ImageScope Layer Attributes - results for entire layer (all regions analyzed). Layer Regions - results per region analyzed. View result plots - click the Display Plots button in the Layer Attributes pane. Export from ImageScope - click the Export button in the Layer Attributes pane to export to an Excel spreadsheet. Page 25

26 ImageScope Image Analysis Analysis Results - Spectrum Spectrum analysis results analysis column must be displayed in the Digital Slide table to view the analysis results. Select image checkbox and click Export Data link to export data from Spectrum. Analysis Column Click Export Data Select image for data export Page 26

27 Positive Pixel Count FL Positive Pixel Count FL Positive Pixel Count FL looks for pixels for the dye specified within the intensity range set (minimum and maximum). Outputs: Total area of analysis (mm 2 ). Total area of specified dye (mm 2 ). Mean intensity of specified dye. The markup image displays only the pixels that fall within the set intensity range for the specified dye. The markup image color matches the pseudo-color used for that dye in the scanning process. Original Markup Page 27

28 Positive Pixel Count FL Positive Pixel Count FL Inputs - Workflow 1. Enter dye name. Enter fluorescent dye name. 2. Set intensity thresholds. Set the dye intensity minimum and maximum intensity thresholds. Page 28

29 Positive Pixel Count FL Positive Pixel Count FL Inputs Input parameters allow the user to identify the fluorescent dye or dye the and intensity thresholds. The user enters the dye as it appears in the ImageScope Image Fusion Adjustments window. This specifies the target color channel/component image of the AFI for analysis. Default for ScanScope images Genie parameters Dye name and intensity thresholds Page 29

30 Positive Pixel Count FL Positive Pixel Count FL Analysis Results Analysis results are viewed from the Annotations Window in ImageScope. PPC analysis results are area based quantification and intensity values. If running analysis in ImageScope, a false color Markup image can be generated. Positive Pixel Count FL results: Total area of analysis, the total area of specified dye and its mean intensity. Results are color coded to match the dye s pseudocolor used in the scanning process Input parameters are always saved with the analysis results. Page 30

31 Area Quantification FL Area Quantification FL Quantifies the area and intensity of each dye and within combinations of dyes. Good for analysis of colocalized markers. Quantification FL analysis calculates the contribution of each dye at every image pixel as either part of a single dye or representing a dye combination. Outputs: Total Analyzed Area Total Stained Area Percent area of dye and average intensity (for single dyes and dye combinations). Original Dye -3 (only) All Dyes Dyes 1+2 Dye -2 (only) Pixels where dyes appear alone w/o other dyes All pixels where dyes appear (alone and with other dyes) Pixels where dyes appear only in specified combinations Page 31

32 Area Quantification FL Area Quantification FL Tuning Area Quantification FL Tuning Markups All Dye 1, Markup displays image channel for dye 1 only (in grayscale) and false colors only the pixels within the intensity ranges. All Dye 2, Markup displays image channel for dye 2 only (in grayscale) and false colors only the pixels within the intensity ranges. All Dye 3, Markup displays image channel for dye 3 only (in grayscale) and false colors only the pixels within the intensity ranges. DAPI Alexa Fluor 532 Alexa Fluor 647 Original Tuning All Dye -1 Tuning All Dye -2 Tuning All Dye -3 Page 32

33 Area Quantification FL Area Quantification FL Inputs - Workflow 1. Enter the number of dyes, and dye names. Enter the number of dyes (1-3). Enter the fluorescent dye names. 2. Set intensity thresholds. Set the dye intensity minimum and maximum intensity thresholds for each dye. 3. Set markup image type and enter plot values. Select the Markup Image Type. Enter plots start and end values and the number of bins. Page 33

34 Area Quantification FL Area Quantification FL Inputs Number of Dyes to Colocalize - 1, 2, 3 Dye 1, 2, 3 - Enter the dye name displayed in ImageScope s Image Fusion Adjustments window for each dye you want to include as part of the analysis. Intensity Thresholds - Min and Max threshold and # of bins for dyes 1, 2 & 3. Markup Image Type Tuning Dyes (1, 2, 3) Colocalized All Dyes Colocalized 1 only, 2 only, 3 only, 1+2 only, 1+3 only, 2+3 only, Display Plots - View result plots. Histogram Start/End Values Select Mark-up Image Type Default Dyes: Dye (1) = DAPI Color (2) = FITC Color (3) = TRITC Intensity Lower and Upper Threshold parameters Pick number of dyes. Select Yes to display Plots with results. Enter Start, End values and number of Bins Page 34

35 How to access Please contact Dr Maria Sarris if you wish to use this image analysis algorithm. All algorithms apart from the Positive Pixel Count are licensed and need to be installed on your workstation (please note: charges apply). Step 1: Extract the region at full resolution that you want test on your desktop. Extracting Image Using the Extraction Tool to remove a select region of the active image. Click on in the toolbar. Place your cursor on the digital slide. Click and drag a rectangle on the screen to capture the area: Choose output File location Choose Output/Compression File type Click Extract Step 2: Open image in Imagescope Step 3: Go to VIEW then ANALYSIS Step 4: Select Algorithm and then click on TEST Step 5: Make your adjustments (optimise algorithm) Step 6: When satisfied with your adjustments click EXPORT MACRO Step 7: Save to your desktop with your name and original algorithm name and version eg SARRIS Nuclear V9 Step 8: Once the algorithm is saved please m.sarris@unsw.edu.au. The algorithm will on the server in the Spectrum Analysis dropdown menu within 24 hrs. Page 35