Digital Pathology Blog

First 1 trillion-pixel image

Aperio has created their own format called BigTIFF and neatly offering the format into the public domain.  To showcase the power of BigTIFF image resolution, Aperio has releaseed the first terapixel image.  The image shows 225 pathology slides of breast tissue which can be viewed here. 

An overview and other information is available here:

Overview

There are more technical details, but at the highest level the BigTIFF changes made to libtiff were quite simple.  The TIFF file format internally uses 32-bit byte offsets.  The largest offset which can be represented is 2³² = 4GB, making that the upper limit of the file size which could be supported by the design.  The BigTIFF modifications to libtiff consisted primarily of changing all internal byte offsets to 64-bits.  A key goal was to maintain backward compatibility with existing applications and files to the largest extent possible.

BigTIFF files have a ".tif" or ".tiff" file extension just like ordinary TIFF files.  A new version in the file header prevents programs which have not linked with the BigTIFF version of libtiff from processing BigTIFF files.  The library has been modified so that programs writing TIFF files do not need to know or care when they reach a 4GB data size.  The file format will smoothly change from a standard TIFF file (compatible with all existing programs) to a BigTIFF file (compatible with all programs linked with a new version of the libtiff library).  Similarly, programs reading TIFF files do not need to know or care whether a file is a standard TIFF file or a BigTIFF file.

The basic BigTIFF design was first proposed in 2004 and refined in this discussion on the Aware Systems mailing list.  Contributors to the design discussion included Lynn Quam, Frank Warmerdam, Chris Cox, Rob Tillaart, Dan Smith, Bob Freisenhahn, Andrey Kiselev, Phillip Crews, and Gerben Vos.  We thank all those who came before us for creating libtiff and designing the BigTIFF enhancements.

These changes were made by Ole Eichhorn of Aperio Technologies and are donated to the public domain, in  gratitude to Sam Leffler, Silicon Graphics, Joris Van Damme, Aware Systems, Frank Warmerdam, Andrey Kisley, Mike Welles, and all who have worked on libtiff over the years to provide such a great tool.  These changes are published on an "as is" basis and neither Ole Eichhorn nor Aperio Technologies make any warranty as to their fitness for any intended use.

Histopathology videos on YouTube

An Illinois pathologist has uploaded 544 histopathology videos on YouTube.  My hospital has recently banned this website from hospital PCs.  Nonetheless, this pathologist who goes by WashingtonDeceit and whose website contains other pathology links and makes mention of creating an online medical school in his YouTube bio has created a series of histopathology lectures that run 2-5 minutes and go over features of diseases and disease processes.  The 544 videos have been added since February of this year.

Digital Atlas of Gynecologic Pathology

There is a well done gynecologic (and breast) atlas available here from Dr. Meenakshi Singh at the University of Colorado.  The byline on the home page is below.  Others are encouraged to submit images to the site.

The aim of this gynecologic pathology atlas is to provide an online resource of digital images for the students of gynecologic diseases and the physicians diagnosing and managing these diseases. It encompasses the wide spectrum of pathology seen in the female genital tract (FGT). There are nearly 800 images in this atlas, and succinct legends explain their salient features. The majority of them are histology images, and these are complemented by relevant gross pathology, cytology, immunohistochemistry and clinical images. The emphasis is on tumors, although examples of non-neoplastic lesions and normal histology have been provided. The World Health Organization classification for tumors has been used. Although some entities are common amongst the tumors of the different sites in the FGT, others are unique to the site of origin. Metastatic lesions can mimic primary tumors, particularly in the ovaries, therefore many examples of metastatic tumors have also been included. One chapter (organ site) at a time can be viewed as an expandable tree by clicking on the open all categories link for each chapter.

I welcome feedback. As an educator and practitioner, I am keen to hear from individuals who plan to use this atlas for teaching, or recommend it as a resource for their courses, lectures and/or classes, use it as a tool for learning and as a reference; comments may be submitted to me at singh.atlas@uchsc.edu. As this site will be regularly updated, I also encourage you to become a contributor by sending me images of entities that are not yet covered in the atlas. I shall acknowledge you if your images are incorporated into the online atlas. A CD ROM has also been developed.

Meenakshi Singh, MD

While the author has provided this online atlas for the viewers to use, the atlas is copyright protected. Should any of the material be used for publications, presentations, or teaching etc, you should acknowledge the author and notify the author via an e-mail at singh.atlas@uchsc.edu regarding the exact use of this atlas and its content. The entire atlas should not be downloaded.

Note: This is the second in a series of digital pathology atlases by the author. To view the first atlas, click here “Digital Atlas of Breast Pathology©”

Color consistency

From Imaging Technology News by Ken Compton.  Ken is is strategic marketing manager for National Display Systems:

The introduction of LCDs to the market did not displace all the barriers to color consistency between displays. The venerable CRT has numerous internal variables that made it nearly impossible to match displays next to each other, let alone networked worlds apart. In pathology, the accuracy of the perceived hue (color), saturation and intensity is vital to the interpretation.

The barriers left for the LCD to conquer through color calibration are unique, but also common to image capture and presentation. A CCD camera has to be set to a white point, and this can vary as much as 200 points in degrees Kelvin (color temperature). A CCD detector is an analog device converted to digital format and subject to conversion variations. Color LCD displays accept analog (converted to digital) and digital (preferred) signals, and interpolate to a color pallet. The pixel values are then sent to data column drivers (DACs) to set the individual pixels, which is back to analog.

The color filters that define the red, green and blue (RGB) at 8-bits depth (0-255) of luminance intensity provide a logical spectrum of 16.58 million possible combinations. The eye’s sensitivity is limited to approximately four million hues. The RGB filters form three points of a triangle in color space (CIE Color Chart) that defines the pallet available. Hues outside of the triangle cannot be produced. The quality of the panel, thus the RGB filters, varies between liquid crystal (LC) core manufacturers and is significant even from name brands. The backlights must also be matched to the filters to optimize the efficacy of luminance transportation; the spectrum of the RGB wavelengths of the backlight needs to be centered close to the filter’s center. A poor match means less luminance (of lamp total output) makes it through the filters. The size of the pallet is one step that can be treated as a minimum level of performance using the x,y coordinates of the CIE color chart. But filters vary from lot to lot, and this will affect the response within the triangle (color space) available.

The CRT had electron guns and phosphor as variables, and LCDs have Digital-to-Analog Converters (DACs) on the data column drivers and RGB filters. High-quality color LC cores delivered to a display vendor are far from equal even within the same production lot of filters. Given a tolerance criteria of /- 0.005 on the x,y axis of the CIE chart centered about a target value, more than half will fail. The eye can detect a difference of 0.005 on just one axis from a target value. The natural scatter of color cores needs to be calibrated so multicolor stains and counter stains are consistently rendered. From image capture to final presentation, the need for a display calibration standard cannot be overlooked. This standard needs to be consistent and not subject to variations in definitions of what is “white.”

No reading of outside films

I recently received a memo from my institution's radiology department sent from the VP and Chairman of Radiology with the above title.  I have cut and paste the text of that memo below:

"The purpose of this memo is to inform you of some changes in the Department of Radiology that are designed to promote quality assurance.  Currently, [Institution name] Radiologists are often asked to provide a reading on studies that have been obtained and previously read outside of the [Institution name] system.  As we cannot control the quality of the external product, we will no longer be providing this service.  Such films may be used only as a comparison to studies obtained at [Institution name].

We appreciate your cooperation in this matter.  If you have comments or concerns, please do not hesitate to contact either one of us, as your business is important to our department."

As people speak of diagnostic imaging and laboratory medicine merging, how might this blind rejection impact workflow between the specialties?  Who will review outside films for patient management?  Does this mean increased numbers of radiology studies amongst several institutions when patients seek other opinions - "we will only read our own cases." 

I can certainly understand the department's reluctance to review outside films when the quality of the images may be in question and additional evaluation or deferral of diagnosis is in the best interests of the patient.  I find it surprising that with DICOM standards for radiology, this would be an issue.  Or is it an issue of reading hard copies, as it used to be done, compared with PACS images on a bank of monitors? 

Pathologists as well are asked to review cases prepared elsewhere when the specific methods of processing, cutting, staining are possibly and probably unique to that pathologist's own institution.  As patient's seek additional pathology and/or clinical opinions concerning treatment and management of their disease and slides "travel", we are often faced with histology and immunohistochemistry (IHC)slides which may have little or no resemblance to routine H&E sections or IHC stains prepared in our own laboratories.  Nonetheless, we offer opinions on outside material regularly as added quality assurance and review for patient's seeking second opinions or review of pathology material prior to definitive treatments.  Accreditations for laboratories and cancer centers require it.  Certainly in a minority of cases, opinions may be deferred because the slides may not be of diagnostic quality or damaged in the course of transport for review.  Fortunately, these are a minority of cases and in most cases the review process of outside material creates added quality and value for the patient. 

Large referral or consultation centers make this part of their daily clinical practice in support of pathologists and their patients, far removed from knowing how the specimens were obtained, processed, fixed, stained, coverslipped, labeled and reviewed prior to being seen elsewhere.  How is one even assured that slide is from the correct patient?  It may be a source of medical error even within one's own laboratory that is not unique. 

How might this impact the practice of pathology PACS signout or remote signout when the metadata, normally present on radiology films and fixed to those images is more transparent?

Issues with biorepositories

Dr. Ashokkumar Patel of the Case Comprehensive Cancer Center at Case Western Reserve University had an interesting posting on the Association for Pathology Informatics (API) Listserv recently dealing with issues on biorepositories.  A recent appeals court ruling has orderd that Washington University retains rights to prostate tissues collected by Dr. William Catalona (now at Northwestern University). 

Mention of the issues concerning Dr. Catalona's work have been published elsewhere.

There is also reference to a recent article in yesterday's St. Louis Post-Dispatch on the story.

For those interested, check out the website for the First World Congress on Pathology Informatics meeting in August.

Welcome to a new blog on digital pathology

Welcome to a new blog on digital imaging in pathology. 

The practice of pathology is changing with advancing and evolving technologies and applications that will impact the practice of pathology as it relates to imaging, image analysis, business and service models and remote pathology services. 

My sincere appreciation to Dr. Bruce Friedman for his encouragement to start blogging about digital pathology. Bruce organizes the Lab InfoTech Summit meeting and blogs daily at Lab Soft News.  Dr. Friedman has specialized in pathology informatics and has been active in the field since 1982. He is a graduate of the University of Michigan Medical School and completed his residency in pathology at the same institution.  He has been on the faculty of the University of Michigan Medical School Department of Pathology since 1973. He is the author or co-author of some 70 scientific papers and multiple book chapters, largely in pathology informatics during the last two decades. He was the founder of the AIMCL conference on pathology informatics, which was initiated in Ann Arbor in 1983. He coined the term “pathology informatics” in the literature 15 years ago. He is a well-known speaker in the field of pathology informatics, both in the  USand abroad, and also serves as a consultant in the field to multiple companies in the same area.

This blog is also in response to comments and feedback I recently received after attending the CAP Futurescape of Pathology held this month in Chicago.  Articles and proceedings of the conference will be made available in CAP Today and the Archives of Pathology and Laboratory Medicine. 

The intent of this blog is to cover issues relevant to imaging and image analysis in pathology.

I welcome your replies and comments.