WWW.DISSERTATION.XLIBX.INFO
FREE ELECTRONIC LIBRARY - Dissertations, online materials
 
<< HOME
CONTACTS



Pages:     | 1 |   ...   | 2 | 3 || 5 | 6 |

«IHC Membrane Image Analysis User’s Guide ©Copyright 2007 Aperio Technologies, Inc. Part Number/Revision: MAN-0026, Revision B Date: January 2, ...»

-- [ Page 4 ] --

Specific applications and their data sets can be organized as cases in Spectrum Information Manager using specimens for the data sets (think of project for the “case” and of data set for the “specimen”). The administrator navigates to the Cases view by clicking All Cases in the Cases drop-down menu. To create a new case (project) click New. As Name, provide a name that identifies the stain, the manufacturer and tissue type. If you have different stains, different manufacturers or different tissue types or just want to use different scoring standards you can create a different case (project) for each of these specific applications to organize the data sets. To identify to the other pathologists that this is not a real case enter Algorithm as Patient ID.

Open a Case (project) by clicking on the open data icon for the case.

To restrict the access to the setup for the specific application, select the appropriate data group for the specific application in the Data Group window of the Case Information and click Save.

Create a new case specimen (data set) for each copy of the training data set and import the corresponding digital slides. To create a new specimen (data set) click Add New Specimen under Case Specimens. For the Accession Number enter a name that identifies that this is a training data set and a data set ID different for each copy of the data set. Click Add to create the new specimen (data set).

Open a specimen (data set) by clicking on the open data icon for the case specimen. Import the

corresponding copy of the training data set as new slides:

1. Click Add New Slide under Specimen Digital Slides

2. Select Add Digital Slides with Images in server folder and subfolders: and provide the folder name where the copy of the training data set can be found (e.g., D:\Images\ HER2_Dako_Breast\ Training1) and click Add.

To hide any possible patient information from the pathologists viewing the digital slides, you have the option in the setup of the ImageServer not to show the labels. To randomize the order of the slides in the training data set, you can add a random number as a pre-fix to the filenames (leave the filename provided

–  –  –

by the ScanScopes as part of the filename). When Spectrum Information Manager imports digital slides it does this in alphabetical order and therefore imports the digital slides in the specified random order.

To restrict access to this data set to only the pathologist who should score it, select his/her data group in the Data Group window of Specimen Information and click Save. Now each pathologist can only see and access his/her assigned data set.

Manual Read of Digital Slides and Outline of Tumor Regions Now each of the three pathologists independently reads the digital slides and provides their score and tumor-cell only region outlines. How a pathologist reviews a digital slide is described in detail in Chapter 2, “Analysis Step-by-Step” on page 5 with two exceptions to this procedure: (1) an annotation file is imported and the scores are entered by the pathologist so that the scores can be stored in the right format in Spectrum Information Manager’s data base, and (2) as the algorithms are not set up yet, no image analysis will be used.

An annotation file needs to be created that has a layer to store the tumor-cell only region outlines and provides the following Layer Attributes for the scoring in the same layer. It is important to provide the specified layer attributes by the letter as the algorithm expects exactly those names during the automatic training of the scoring scheme parameters.

–  –  –

HER2 Score [0,1,2,3] 3+ Percent [0-100%] 2+ Percent [0-100%] 1+ Percent [0-100%] Please note that in addition to the HER2 score, we also ask the pathologist to estimate the percentage of 3+, 2+ and 1+ cells on which his score was based.

You may want to add other layer attributes of interest to you, like a comment field where the pathologist could describe staining artifacts he/she observes or the presence of carcinoma in situ, etc.

A master annotation file can be created in ImageScope. The annotation window can be brought up by clicking Annotation Window in the View drop-down menu. A new layer can be created by clicking on the new layer icon in the Layers section. The layer can be renamed by clicking on its name, Layer 1. A new layer attribute can be added by clicking on the add layer attribute icon in the Layer Attributes section and providing the required attribute names. The master annotations can be exported by clicking on the export annotation to file icon and providing the filename and location.

The pathologist can import the master annotation file in ImageScope each time he reviews a digital slide.

He brings up the annotation window by clicking Annotation Window in the View drop-down menu. The pathologist imports the master annotation by clicking on the import annotation from file icon and selecting the appropriate file in the browser. Make sure that the pathologists have file system access to the master annotation file.

When the pathologist fills out the scores, he/she needs to make sure that he/she only enters numbers, no % or + signs, and enters a number for each of the layer attributes, even if the number is 0.

When the pathologist has outlined the tumor-cell only regions and provided his/her scores in the annotation layer attributes, he/she saves his/her annotations in Spectrum Information Manager’s data base by clicking on the save annotations icon.





Eliminating Outliers To make sure that the image analysis is in agreement with the manual read of a glass slide, you may want to exclude digital slides in the training data set when the score provided by the pathologists differs between their manual read of the glass slide versus the digital slide read.

As there is quite some variability between pathologists scoring an IHC slide, you may want to exclude digital slides in the training data set where one pathologist does not agree with at least one other pathologist.

26 IHC Membrane Image Analysis User’s GuideChapter 6 – Algorithm Setup

You can exclude a digital slide from the training data set, either by deleting the digital slide in Spectrum Information Manager or by excluding the digital slide from the batch analysis. To delete a digital slide in Spectrum Information Manager, select the check box for the digital slide in the Specimen Digital Slides list in the specimen (data set) that holds the corresponding copy of the training data set and click Delete.

To exclude a digital slide from batch analysis, unselect the check box for the digital slide when the digital slides are selected to run the algorithm.

Cell Feature Detection Parameters The cell feature detection parameters specify cell feature detection thresholds and methods as well as size and shape constraints of cells to distinguish tumor cells from other cells like normal, lymphocyte and stroma cells.

The default parameter settings have been optimized for HER2 stains and breast tissue. If you are going to use the IHC Membrane Algorithm for HER2 stains and breast tissue you may want to first test the default parameter settings as those may work for you.

Parameter List It is not necessary nor recommended to modify the following Aperio Algorithm Framework (AAF) common parameters: View Width = 1000, View Height = 1000, Overlap Size = 100, Image Zoom = 1, Markup Compression Type = 0 – JPEG and Compression Quality = 30.

Pre-Processing Averaging Radius – Radius (microns) for Noise Reduction [0 Value 100] (Default: 1).

The radius of a smoothing filter, which reduces noise resulting in smoother object edges.

Nuclei Threshold Type – Threshold method [0 – Edge Threshold Method, 1 – Manual Threshold] (Default: 0 - Edge Threshold Method).

A threshold must be applied to the intensity image in order to find the edges of the nuclei.

Threshold type specifies the method to be used to determine the intensity thresholds. Note that an intensity value of 0 corresponds to black and a value of 255 to white (as bright as possible).

The edge threshold method automatically adjusts the threshold according to the mean of edge pixels. The algorithm uses an edge finding method to identify edge pixels and uses the average of these pixel values to determine the threshold. The manual threshold method is the simplest method, which uses the prescribed intensity thresholds (lower and upper below) to eliminate unwanted background. This method will not automatically adjust to compensate for lighter or darker staining between slides.

Lower Blue Threshold – Lower limit for nuclear auto thresholding [0 Value ≤ 255] (Default: 0).

This value can be changed when using the manual threshold type. Since very dark nuclei are possible, a value of 0 is usually used. Increasing this value will ignore very dark pixels.

Upper Blue Threshold – Upper limit for nuclear auto thresholding [0 Value ≤ 255] (Default:

200).

–  –  –

This value can be changed when using the manual threshold type. A large value (e.g., 240) represents very faintly stained pixels. Lowering this value to 200 will ignore faintly stained pixels and require nuclear pixels to have lower intensities (being darker).

Blue Curvature Threshold – Curvature threshold for nuclear de-clustering [0 Value ≤ 25] (Default: 2.5).

Curvature threshold determines the level of de-clustering used to separate touching nuclei.

Increasing the threshold reduces the effect of the de-clustering logic – this may be necessary if single nuclei are being split and counted more than once. Decrease this value if more declustering is needed to separate nuclei.

Min Nuclear Size – Minimum area for detectable nuclei (micron-squared) [Value 0] (Default:

15).

Nuclei smaller than this area limit will not be counted.

Max Nuclear Size – Maximum area for detectable nuclei (micron-squared) [Value 0] (Default:

400).

Nuclei larger than this area limit will not be counted.

Min Nuclear Roundness – Nuclear objects with roundness less than this are excluded [0 ≤ Value ≤ 1] (Default: 0.1).

Roundness is the ratio of the object area to the area of a circle that fully encloses that object.

Circular objects will have a value of 1. Small values indicate non-circular objects.

Min Nuclear Compactness – Nuclear objects with compactness less than this area are excluded [0 ≤ Value ≤ 1] (Default: 0.1).

Compactness is the ratio of area of the object to the area of a circle that has a circumference equal to the perimeter of the object. Circular objects will have a value of 1. Small values indicate noncircular objects.

Min Nuclear Elongation – Nuclear objects with aspect ratios less than this are excluded [0 ≤ Value ≤ 1] (Default: 0.1).

Elongation is the ratio of the min/max principal moments of the object. Circular objects will have a value of 1. Small values indicate objects which are long and thin.

Cells/Membrane Cell/Nucleus Requirement – Include/exclude cells without a nucleus [0 – All Cells, 1 – Cells with Nucleus] (Default: 0 – All Cells).

All cells may not show a nucleus, due to the cutting of the tissue section. If “All Cells” is selected, then cells are not required to contain a nucleus to be counted. If “Cells With Nucleus” is selected, then only cells showing a nucleus will be counted.

Background Threshold: Membrane intensity threshold for background [0 Value ≤ 255] (Default:

200).

Max Cell Radius – Maximum radius (microns) to limit cell growth [0 Value 100] (Default: 5).

Determines cell size for cells that have a nucleus but no membrane staining. A cell is grown surrounding the nucleus according to this limit.

28 IHC Membrane Image Analysis User’s GuideChapter 6 – Algorithm Setup

Min Cell Size – Minimum area for detectable cell (micron-squared) [Value 0] (Default: 25).

Cells smaller than this area limit will not be counted.

Max Cell Size – Maximum area for detectable cell (micron-squared) [Value 0] (Default: 2000).

Cells larger than this area limit will not be counted.

Min Cell Roundness – Membrane (cell) objects with roundness less than this are excluded [0 ≤ Value ≤ 1] (Default: 0.1).

Roundness is the ratio of the object area to the area of a circle that fully encloses that object.

Circular objects will have a value of 1. Small values indicate non-circular objects.

Min Cell Compactness – Membrane (cell) objects with compactness less than this area are excluded [0 ≤ Value ≤ 1] (Default: 0.1).

Compactness is the ratio of area of the object to the area of a circle that has a circumference equal to the perimeter of the object. Circular objects will have a value of 1. Small values indicate noncircular objects.

Min Cell Elongation – Membrane (cell) objects with aspect ratios less than this are excluded [0 ≤ Value ≤ 1] (Default: 0.1).

Elongation is the ratio of the min/max principal moments of the object. Circular objects will have a value of 1. Small values indicate objects which are long and thin.

Algorithm Modes The IHC Membrane Image Analysis algorithm has three different modes of operation: (1) Tuning, (2) Training and (3) Analysis with their respective markup images to choose from. Adjusting the algorithm parameters is done in tuning mode.

Use Mode: Select Training or Analysis/Tuning Mode [0 – Analysis/Tuning, 1- Training] (Default:

0 – Analysis/Tuning).

Mark-up Image Type: Choose type of markup image. [0 – Tuning, 1 – Analysis, 2- Analysis (With Cell Edges)] (Default: 1 – Analysis).

Set up Data Set for Algorithm Parameter Tuning Digital slides accessed through Spectrum Information Manager can only be analyzed by algorithm macros after the final parameter settings have been determined. To determine the algorithm parameter settings, we need to be able to adjust the algorithm parameters, which is only possible when running algorithms on digital slides that have been opened by ImageScope over the file system (locally or across the network).



Pages:     | 1 |   ...   | 2 | 3 || 5 | 6 |


Similar works:

«AB R A H A M BALDW IN AGR ICULTUR AL C O LLEG E Flavours is committed to making your event a success. Our team of catering professionals is available to assist you with planning every aspect of your event. Our culinary team has designed a variety of menus to fit a wide range of budgets and tastes, including regional ingredients and specialty items. Should you desire a customized menu for your event, we will be happy to meet with you to create a special menu for your specific needs. Our goal is...»

«Eötvös Loránd University Faculty of Education and Psychology THESES OF THE PHD DISSERTATION BEATRIX KORONCZAI The measurement of the problematic internet use and its personality psychological characteristics Doctoral Programme of Psychology The leader of Doctoral Programme: Prof. Dr. Attila Oláh, CCs Personality and Health Psychology Programme The leader of Programme: Prof. Dr. Attila Oláh Supervisor: Dr. Zsolt Demetrovics, PhD The members of Committee: Chair: Prof. Dr. József Rácz, DSc...»

«SANCO – D.1 (06)D/410160 SUMMARY RECORD OF THE STANDING COMMITTEE ON THE FOOD CHAIN AND ANIMAL HEALTH HELD IN BRUSSELS ON 15 & 16 June 2006 (Section Animal Health and Welfare) (Section Controls and Import Conditions) (Section Animal Nutrition) President: Bernard Van Goethem and Alberto Laddomada Willem Penning for the Animal Nutrition section All the Member States were present. Malta, partly absent, was represented by the United Kingdom. Cyprus represented Italy and Greece, partly absent at...»

«CHILDREN with STARVING BRAINS A MEDICAL TREATMENT GUIDE for AUTISM SPECTRUM DISORDER JAQUELYN McCANDLESS, MD With Contributions by Teresa Binstock, and Jack Zimmerman, PhD BRAMBLEBOOKS A D I V I S I O N O F T H E B R A M B L E C O M PA N Y Copyright © 2002 by Jaquelyn McCandless All rights reserved. No part of this book may be reproduced in any manner whatever, including information storage, or retrieval, in whole or in part (except for brief quotations in critical articles or reviews),...»

«Package leaflet: Information for the user AUGMENTIN® 125/31 SF Suspension co-amoxiclav (amoxicillin and clavulanic acid) Read all of this leaflet carefully before you start giving your child this medicine because it contains important information for them.  Keep this leaflet. You may need to read it again.  If you have any further questions, ask your doctor, pharmacist or nurse.  This medicine is usually prescribed for a baby or child. Do not pass it on to others. It may harm them,...»

«IOSR Journal Of Humanities And Social Science (IOSR-JHSS) Volume 20, Issue 6, Ver. III (Jun. 2015), PP 70-77 e-ISSN: 2279-0837, p-ISSN: 2279-0845. www.iosrjournals.org Socio-Economic Challenges and Coping Mechanisms of Home Based Caregivers for People Living With HIV/AIDS in NJORO Sub-County, Kenya 1. Njau.P.W., 2. Prof. Kariuki, M. 3. Wambu,C.K. Egerton University Department of Psychology, Counseling nd Educational Foundation P.O BOX 53620115, Egerton. Abstract: AIDS scourge has put to test...»

«MATERIAL SAFETY DATA SHEET LEAD ACID BATTERY WET, FILLED WITH ACID (US, CN, EU Version for International Trade) SECTION 1: PRODUCT AND COMPANY IDENTIFICATION PRODUCT NAME: Lead Acid Battery Wet, Filled With Acid OTHER PRODUCT Electric Storage Battery, SLI or Industrial Battery, UN2794 NAMES: MANUFACTURER: East Penn Manufacturing Company, Inc. DIVISION: Deka Road ADDRESS: Lyon Station, PA 19536 USA EMERGENCY TELEPHONE NUMBERS: US: CHEMTREC 1-800-424-9300 CN: CHEMTREC 1-800-424-9300 Outside US:...»

«Struggling to Meditate: Contextualising Integrated Treatment of Traumatised Tibetan Refugee Monks The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters. Citation Adriana Lee Benedict, Linda Mancini, and Michael A. Grodin, Struggling to Meditate: Contextualising Integrated Treatment of Traumatised Tibetan Refugee Monks, 12 Mental Health, Religion & Culture 485 (2009). Published Version doi:10.1080/13674670902788908 Accessed...»

«Comment Depression and the Law: Experiences of Australian Barristers and Solicitors Sharon Medlow, Norm Kelk † and Ian Hickie ‡ Abstract The present research is a contribution towards understanding Australian lawyers’ experiences of, and attitudes towards, depression. It was found that, like Australian law students, solicitors and barristers exhibited significantly higher levels of psychological distress than did members of the general population and that experiences of depression were...»

«CURRICULUM VITAE Sylvain Moreno Present Position: Director Digital Health Hub, School of Engineering, Simon Fraser University Assistant Professor Adjunct, Department of Neuroscience at York Universty Non-academic Positions: Advisor Ontario Ministry of Education (2013-present) TFO, French-Language Television Network of Ontario (2013-present) Advisor/Consultant Advisor NSERC-Create Industry Training program, 2011 – present Consultant Interaxon Inc. Co-Founder MusiqKids Corp. Chairman/CoBrain...»

«Post-Exposure Prophylaxis after Non-Occupational and Occupational exposure to HIV National Guidelines National guidelines for post-exposure prophylaxis after non-occupational and occupational exposure to HIV © ASHM 2013 Apart from any fair dealing for the purpose of research or study, criticism or review, as permitted under the Copyright Act 1968, no part of this book may be reproduced by any process without Published by: written permission. Direct enquiries to the Australasian Society for HIV...»

«Registered charity number 306016. Girls’ wellbeing explored. Understanding pressures and resilience September 2016 Foreword: YoungMinds YoungMinds champions the mental health and wellbeing of all children and young people and so I am delighted to introduce this important research about girls’ wellbeing. Girlguiding is a hugely valuable and positive Our teenage years are when mental health is resource for girls and young women’s developed and patterns are set for the future. development...»





 
<<  HOME   |    CONTACTS
2016 www.dissertation.xlibx.info - Dissertations, online materials

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.