Digital pathology is quickly scaling, however many laboratories are discovering that merely digitizing glass slides is the straightforward half. The tougher, and more and more pressing, problem is making these photos interoperable throughout distributors, methods and establishments, in a manner that helps enterprise imaging, AI and lengthy‑time period knowledge reuse.
That downside, and a concrete technique to resolve it, was the main target of the latest Digital Pathology Affiliation (DPA) webinar, “Advancing Imaging Interoperability Collectively: A DPA–CAP–SIIM Collaboration to Enhance Standardized Workflow Adoption.” The session introduced collectively leaders from the DPA, the School of American Pathologists (CAP) and the Society for Imaging Informatics in Medication (SIIM) to report out on their joint DICOM toolkit effort and to stipulate what success ought to seem like over the following one to 3 years.
From Radiology’s Expertise to Pathology’s Second
The webinar opened by situating digital pathology inside a broader enterprise imaging arc. Moderator Jennifer Samboydigital transformation chief at Philips and co‑chair of the DPA Training Committee, framed the objectives of the session as threefold: to summarize key findings from a DICOM toolkit collectively developed by DPA, CAP and SIIM; to supply a deeper dive on why DICOM and imaging well being info change requirements matter; and to explain how DPA can speed up cross‑organizational collaboration to advance interoperability for each scientific operations and innovation.
Radiology’s journey offers a robust precedent. Alex Towbin, M.D.from Cincinnati Youngsters’s Hospital and the upcoming president of SIIM, traced how radiology moved from proprietary, vendor‑particular codecs within the Eighties to immediately’s mature, DICOM‑based mostly ecosystem.
As he defined, early radiology methods produced digital photos, however every vendor used its personal proprietary format, leaving organizations with “a number of scanners and no option to view it.” DICOM emerged as a option to remedy precisely that downside: an ordinary file format and metadata mannequin that might permit photos from totally different modalities and distributors to be managed and seen collectively.
Towbin emphasised that metadata is the place “the true energy” of DICOM lies. Past pixel knowledge, structured tags about modality, research description, orientation and plenty of different parameters assist refined routing, hanging protocols, AI workflows and multi‑modality viewing. Those self same capabilities, he argued, are immediately relevant to pathology.
Pathology “deserves to be on this place as effectively,” he mentioned, alongside radiology, ophthalmology, endoscopy, level‑of‑care ultrasound and different picture sorts in an enterprise imaging atmosphere.
DICOM for Pathology: The “Transport Container” for Photographs
Kevin Schap of CAP, Secretariat for DICOM Working Group 26 (WG-26) and a pacesetter in IHE PaLM (Pathology and Laboratory Medication), took the viewers deeper into what DICOM means particularly for complete slide imaging.
Schap described DICOM as “a really detailed algorithm for a way medical photos are saved, described, transmitted and understood throughout methods.” Whereas radiology has used DICOM for many years, pathology is a more moderen entrant, and it’s dealing with lots of the similar points radiology confronted many years in the past: very massive photos, extremely proprietary codecs, and restricted interoperability between distributors.
To make the idea extra tangible, Schap borrowed an analogy: “You’ll be able to consider DICOM like a delivery container for world commerce. Earlier than delivery containers, items have been transported in every kind of inconsistent packaging, barrels, crates, sacks. It was inefficient, laborious to trace and troublesome to maneuver between ships, vehicles and trains. So then got here the standardized container, similar dimension, the identical construction, universally acknowledged. Now, any port truck or crane on the earth is aware of precisely deal with it.”
Within the pathology context, DICOM doesn’t simply carry the pixels; it carries the that means — affected person info, acquisition particulars, organ website, tissue kind, stains and immunohistochemistry. With out standardized encoding of those components, he argued, interpretation and secondary use turn out to be extraordinarily troublesome.
Schap additionally underscored why DICOM usually feels “sophisticated,” and why that’s acceptable and even essential. It’s not merely a file format, he mentioned, however “a knowledge mannequin, it’s a communication protocol, it’s a workflow framework.” That complexity is the tradeoff for interoperability and future‑proofing in an atmosphere of quickly evolving modalities and AI instruments.
Making Requirements Actionable in Workflows
If DICOM defines how photos and metadata are structured and exchanged, IHE (Integrating the Healthcare Enterprise) defines how that construction is definitely utilized in real-world workflows.
Schap walked by means of how IHE develops profiles that sew collectively requirements akin to DICOM, HL7 and SNOMED CT to assist finish‑to‑finish processes. In digital pathology, that features:
- Digital Pathology Picture Acquisition (DPI): launched, centered on creation, storage, show of DICOM information and the mapping of LIS metadata into photos.
- Digital Pathology Ordering Workflow: in improvement, defining communication between order fillers, picture managers and acquisition managers when new complete slide photos are created.
- Digital Pathology Proof Creation: centered on how AI‑derived measurements and annotations are related to complete slide photos, saved and accessed.
Schap famous that scanners is not going to natively know every part a couple of specimen; they have to name again to the LIS for specimen kind, website, stains and different attributes. IHE profiles specify how these calls and transactions ought to work, in order that totally different distributors’ parts might be blended and matched with out bespoke integrations.
Radiology’s Maturity and Pathology’s Gaps
The panel dialogue shifted to the classes pathology can study from radiology, and to the sensible boundaries that also impede DICOM adoption.
Towbin highlighted two key radiology classes: the boundaries of free‑textual content fields and the variability in how totally different specialties and distributors implement the usual. Many human‑dealing with fields, like process and sequence descriptions, are successfully “rubbish” from an informatics standpoint as a result of they’re unstructured and establishment‑particular. But he additionally famous that, with trendy AI, this will likely turn out to be much less of an issue as algorithms can ingest and interpret massive, messy metadata units in methods people can not.
He additionally emphasised that, in trendy radiology, using DICOM is not debated: “Is that this even a dialogue amongst radiology departments whether or not try to be implementing a DICOM versus a non-DICOM compliance system in radiology? Is it even a query?” he was requested.
“No, no, it’s not,” Towbin replied. Radiologists, he added, usually have no idea or must know the technical particulars of DICOM. “The usual simply means issues work,” enabling teleradiology, second opinions and multi‑website workflows to “simply work as a result of it everybody’s utilizing that customary.”
Limitations: Storage, Distributors, Incentives and Consciousness
Whereas the technical requirements for pathology at the moment are sturdy, particularly Complement 145 for complete slide imaging, actual‑world adoption is uneven.
Mustafa Yousif, MD, from the College of Michigan and co‑chair of DICOM Working Group 26, described Michigan’s enterprise implementation and the boundaries they encountered. Storage was the primary and most blatant problem: single complete slide photos can exceed 4 GB, and lengthy‑time period archival rapidly reaches petabyte scale, far past what conventional radiology PACS have been designed to deal with. That has compelled establishments to suppose by way of scorching/chilly storage tiers and sensible retention methods centered on diagnostic and tumor board time home windows.
Yousif additionally pointed to an immature vendor ecosystem. Many digital pathology scanner distributors nonetheless deal with DICOM as an elective output somewhat than a native format, and few provide sturdy DICOM viewers. That leaves purchasers struggling to validate photos and metadata, and it feeds persistent myths akin to the concept that there are “a number of variations” of pathology DICOM.
In actuality, he pressured, “it’s one DICOM customary … the pathology photos and radiology photos dwell below the identical customary, use the identical foundational metadata construction, and may actually use as one customary when accurately applied.”
Rajesh Sprint, M.D., of Duke College and co‑chair of IHE PaLM, centered on economics and incentives. From the attitude of apply leaders, he mentioned, digital pathology and requirements‑based mostly implementations should display a transparent return on funding. At a nationwide stage, he pointed to precedents akin to licensed EHR incentives that accelerated adoption and steered related coverage levers might assist drive interoperable digital pathology.
With out such nudges, distributors have weak incentives to desert proprietary codecs which may be cheaper for them to keep up, although they improve prices downstream for healthcare organizations, payers and finally sufferers.
A Toolkit, Connectathons and a Path to Success
One concrete output of the DPA–CAP–SIIM collaboration is a DICOM starter toolkit for pathology, based mostly partially on a survey of pathologists. As Towbin reported, greater than three‑quarters of respondents mentioned such a toolkit could be useful, with over half answering “sure” outright. Probably the most requested parts have been:
- An outline of the DICOM customary and the way it applies to pathology
- A step‑by‑step implementation information
- Instruments for changing proprietary codecs to DICOM
- Workflow integration steering and pattern DICOM information
Schap steered that, past training, the toolkit can arm organizations with particular questions to ask distributors throughout RFPs and negotiations, and may even sign to distributors what they are going to be anticipated to display. A number of panelists additionally pressured the significance of IHE and DICOM Connectathons as actual‑world proving grounds the place a number of distributors check interoperability in widespread use instances.
Trying forward one to 3 years, panelists converged on a number of measures of success:
- Extra distributors producing native DICOM output solely, with sturdy conformance statements.
- Clearer, extra detailed procurement necessities from hospitals that transcend “do you assist DICOM?” to “how do you assist it?”
- Elevated vendor participation, and profitable testing, in IHE and DICOM Connectathons for digital pathology profiles.
- Wider use of standardized metadata fields vital for AI and affected person care, together with SNOMED CT‑based mostly coding.
Moderator John Groth, M.D., of Eis avor Well being and representing the DPA regulatory and requirements job drive, framed this as a “excellent time” to push, citing latest FDA alerts round DICOM, rising Superior Analysis Tasks Company for Well being (ARPA‑H) initiatives, and multi‑society efforts on interoperability and affected person entry. He characterised the present second as one the place “there are dozens of individuals throughout the spectrum particularly associated to pathology which are assembly weekly in some capability,” and the place world collaboration is not aspirational however operational.
The webinar closed with an necessary reminder from the seller aspect: a consultant within the viewers famous that this dialog is “equally necessary to us because the distributors who’re making an attempt that can assist you by means of this course of as effectively,” underscoring that standardization will finally require coordinated motion by skilled societies, regulators, healthcare organizations and business.
If radiology is any information, the tip state will not be one the place pathologists discuss DICOM day by day, however one the place, as Towbin put it, utilizing DICOM is “not even a query.” The work described on this webinar goals to make that future a actuality for digital pathology.
