The Pathology Education Instructional Resource (PEIR) is an award-winning web resource that is well-known for (a) a digital library of 30,000+ curated medical education images and (b) an interactive, problem-based online pathology teaching laboratory. Originally written in 1999 using software packages running on Windows 2000 Server, it has been upgraded and expanded to modernize PEIR for use in the Web 2.0 age - providing an updated, multimodality, open resource for educators and students alike.
The existing Windows 2000 Server-based platform was analyzed, and its database was exported into universally-readable tab-separated value (TSV) files. Simultaneously, a complete archive of the 30,000+ images in the database was also made and converted from uncompressed TIFFs into 100% quality JPEGs, leading to a 66% decrease in storage requirements (from 50GB down to 17GB) with no perceptible loss in image quality. Several programs were written in the programming language Python to parse the TSV files, programmatically extract the key data elements (notably, utilizing natural language processing technologies to extract pertinent tags from free-text image descriptions), and prepare these data elements along with the image files themselves for insertion into the new platform. A completely new platform for image presentation and rapid generation of teaching content based completely on well-understood, battle-tested open source technologies was created. The image library was transferred over entirely to the new system, and the interactive pathology laboratories were rewritten and modernized for the MediaWiki collaborative publication platform. Finally, all of the slideboxes used to teach pathology to medical students at the University of Alabama at Birmingham School of Medicine were scanned using an Aperio WSI scanner, then converted to into DeepZoom-compatible image pyramids via OpenSlide and VIPS originally, but we just use only VIPS now. This allows for fast web viewing using OpenSeadragon JavaScript library.
PEIR and PEIR-VM are implemented atop a Linux-based application stack running on a virtual machine. Because they are virtualized in this manner and stored as a VMWare VMDK disk image, they can be deployed on any x86-64 virtual hypervisor (e.g. VMWare Workstation, Oracle VirtualBox VM) that supports 64-bit Linux client operating systems. The specific elements of the stack are as follows:
| Stack Element | Element Name | Description |
|---|---|---|
| Operating System | Linux OS | One of the most popular and secure server operating systems of the modern era |
| Web Server | NGINX | Pronounced "Engine X"; a lightweight web server that focuses on speed and modularity |
| Database Management System | MariaDB | An enhanced fork of mySQL, an extremely popular relational DBMS |
| Server-Side Scripting Language | PHP | The single most popular server-side scripting language of the Web 2.0 age; features easy integration with mySQL/MariaDB |
The user interface is built utilizing the following open-source frameworks:
The content of PEIR-VM is made available under the PEIR License. If, on the other hand, you are interested in deploying a PEIR-VM instance for your own institution, please contact us for further guidance.