Prototype slide scanner and image management system is designed to handle digital pathology data rates as high as 500 MB/sec. Image Source: Philips

WASHINGTON, D.C.—Radiology is not the only specialty with IT management issues—but for now, it may be the only one that can justify large-scale digital storage, according to a June 4 presentation at the Society for Imaging Informatics in Medicine (SIIM).  

Slides are a vestige of the past in radiology, as radiologists have moved beyond films and into digital storage to manipulate and share images across specialties and institutions. Slide-dependent pathologists and relevant IT experts and vendors in the pathology space have argued that pathology must be the next to go digital—the sooner, the better.
 
And though the physical enormity of pathology’s slide repositories suggest clear digital demand, at 15 GB a slide and as much as 10 TB of images each day, pathology has yet to fully break the digital barrier, explained Paul J. Chang, MD, professor of radiology and medical director of pathology informatics at the University of Chicago Medical Center. “Digital pathology imaging is still very immature, especially with respect to workflow,” Chang argued.

Radiology IT personnel have received calls from both hospital and vendors to replicate the wonder of PACS for pathology. “The good news is there is a lot of overlap and significant opportunities for collaboration and convergence between pathology and radiology IT. But we need to avoid force-fitting…avoid the traps,” Chang added.

In many ways, pathology seems to be the perfect candidate for the next round of radiology IT’s digitization of medicine, with common protocols like ICD-9/10, HL7, DICOM and SNOMED, and reports consisting mostly of narrative attached to a bulky trove of slides that could put even the largest academic radiology departments to shame.

“The most important difference is that radiology is digital at the beginning [of the workflow process]; pathology is digital at the end. This has devastating consequences when designing IT,” Chang proclaimed.

Whereas radiologic examinations require the interpretation of multiple image slices, and often continued referencing by oncologists, surgeons and radiologists looking at priors, pathologists usually mark a single diseased slide with a blue dot. The majority of the biopsy slides, then, become superfluous.

“A write once, read never, archive is hard to justify as a return-on-investment,” Chang asserted. In Chang’s opinion, at 15-GB each, digitizing the whole lot of slides just isn’t defensible, at least not yet.

The problem stems from poorly defined workflow. Instead of trying to imitate what radiology did right with PACS (and leapfrog over its mistakes), the specialty and its IT buttresses need to define the pathology workflow and then find where digital slides would be helpful and thus avoid the trap of the one-size-fits-all approach, Chang stressed.

Chang argued that telepathology unquestionably merits digital capabilities for communicating slides. In addition, documentation of gross specimen sampling is the most underused and important step toward which pathology should head, to accurately identify and track the pathology or pathologies of interest.

“It’s the trap of the hammer and the nail,” Chang insisted, "where the similarities of slides, interpretation, narrative reporting and a seductively streamlined radiology workflow draw pathologists and the industry to assume digitization of the entire specialty to be the logical next step."

While there is little doubt that pathologists and their patients stand to gain tremendously from a more IT-driven field, “I’m still waiting for pathologists to challenge the case,” Chang concluded.