Digital Pathology Data Management Challenges

According to the Digital Pathology Association, digital pathology describes the creation, viewing, management, sharing, analysis, and interpretation of digital images of glass slides and includes workflow considerations unique to a digital imaging environment.

Today the workflow is still a manual and tedious process. Pathologists have to physically deliver slides to collaborate as well as manually inventory slides for later recall. In addition, the examination of tissue samples under the microscope is subject to the human eye that can only process so much information. According to the Association of American Medical Colleges 2014 Physician Specialty Data Book, active physicians in pathology decreased by 10.4% in the US from 2008-2013. At the same time, the cancer projections are growing, resulting in increased workloads per pathologist. Digitization will help tackle some of these challenges in pathology.

One advantage in digital pathology is more efficient image acquisition and workflows to support pathologists in clinics, research and education since digital files are more flexible, avoiding the difficult logistics involved in sending glass slides. You also no longer run the risk of losing or damaging specimens.

Whole Slide Imaging (WSI) can also help enable collaboration between multi-discipline discussions with colleagues, simplify archive/retrieval processes and lead to major efficiency and quality gains with image analysis software.

We at EMC understand that the digitization of a WSI alone won’t deliver the efficiency and quality gains. Rather a critical success factor is to understand the different infrastructure needs at various aspects of a digital workflow that spans from image acquisition, advanced image analysis, viewing, collaboration and long term retention.

A challenge in going digital is to a certain degree Information Technology (IT) related, including network speed, patient safety and security, and storage size.

The actual file size depends on the capture resolution in the scanner, which determines the maximum magnification of the digital slide on a computer monitor, and the associated meta-data such as slide labels, hospital / patient / case specimen, tissue map, scan plan, or a unique ID for the digital image. Digital pathology images can be about 10x the size of radiology images and be in the hundreds of megabytes.

Usually shared storage is recommended for digital slides to be connected through a network versus storing the digital images on a local disk on the scanner. This ensures that the images are available from all workstations in a workflow-centric process for re-viewing by a pathologist without the need to send the images. A shared storage environment enable pathologists to access their images from any workstation, as well as remotely, such as a home workstation. Through shared storage it is also possible to view patient history form other departments in the workflow and access, for example mammography, radiology or dermatology images.

In a reasonable big lab, one digital pathology scanner of the latest generation could be used all day to generate up to 500-1,000 slides with 20-40x resolution. When a 200 Megabyte image is scanned every minute, then the storage subsystem must be able to store the images at least as fast and scale capacity on demand. That same infrastructure used for image acquisition must also be capable of providing quick access to the Pathologist as well as advance image analysis software for computer aided detection of findings.

Storage solutions such as EMC Isilon or EMC Elastic Cloud Storage (ECS) are specifically designed for applications and workflows such as digital pathology. Through its scale-out architectures they can scale on demand as capacity and performance requirements grow.

EMC Isilon is a perfect fit for the image acquisition and review stages as it scales from as small as 16TB to 50PB in a single file system without any limitations on individual file sizes. This will eliminate storage silos assigned to each scanning device or a particular histology lab. Isilon can also scale to 200GB/s meeting the demands of even the largest organizations. EMC Isilon supports a data lake concept, enabling pathologists to consolidate their WSIs along with other data from different departments into a central data repository without sacrificing performance and access them at any stage during the workflow process. The Isilon data lake provides further advantages for image analysis as it supports the Hadoop Distributed File System (HDFS) to support Hadoop based analytics without needing to add additional storage resources.

Over time, the amount of data can quickly exceed petabytes. What about leveraging cloud scale economies? EMC Elastic Cloud Storage (ECS) scales into the Exabytes and brings all the benefits of a public cloud into privately controlled datacenters. Using EMC Isilon as high performance file storage for WSI scanners and analytics, Isilons’ CloudPool capability allows older digital pathology images to be tiered off Isilon to ECS for long term cost optimized archiving. The tiering is seamless to the pathology application and appears in the same place it was initially stored. Among a wide variety of workloads supported by ECS, ECS deep archive capabilities provides the needed storage capacity and durability for digital pathology while keeping the archiving cost to a minimum.

EMC also recognizes that WSI data must be secured and retrievable over a long period of time. Self-encrypted hard-drives, software encryption and Write-Once-Read-Many (WORM) archive capabilities, embedded in EMC storage ensures data security and privacy.

The combination of both EMC Isilon and EMC ECS delivers the optimal shared storage environment for any digital pathology workflow.

We will be attending USCAP 2016. Please stop by our booth #430 if you would like to learn more.

Visit Us Online – Emerging Technologies for Healthcare