Yale's Peter Gershkovich has written an article with an eye-catching title:
Open-access, find Wearing a fur coat in the summertime: Should digital pathology redefine medical imaging?, at J Pathol Informatics, released 18 May.
Here's the opening and closing of the author's abstract. Or, instead, you might enjoy the author's own summary of his work at Linked In, here. The author's Yale webpage is here.
Coming next -
- College freshman summary
- Author's abstract
- AI summary
- AI summary ... by comedian Lewis Black
- Slides are data.
- Once digitized, they function like any enterprise asset: accessible anywhere, ready for AI, and integrated into cloud workflows. But in pathology, they enter a realm of clinical complexity—demanding systems that handle nuance, integrate diverse data streams, scale effectively, enable computational exploration, and enforce rigorous security.
- Although the Digital Imaging and Communications in Medicine (DICOM) standard revolutionized radiology, it is imperative to explore its adequacy in addressing modern digital pathology's orchestration needs. Designed more than 30 years ago, DICOM reflects assumptions and architectural choices that predate modular software, cloud computing, and AI-driven workflows.
- This article shows that by embedding metadata, annotations, and communication protocols into a unified container, DICOM limits interoperability and exposes architectural vulnerabilities. The article begins by examining these innate design risks, then challenges DICOM's interoperability claims, and ultimately presents a modular, standards-aligned alternative.
- [skip]
- Pathology stands at a rare inflection point. Unlike radiology, where DICOM is deeply entrenched, pathology workflows still operate in polyglot environments—leveraging proprietary formats, hybrid standards, and emerging cloud-native tools. This diversity, often seen as a limitation, offers a clean slate: an opportunity to architect a modern, modular infrastructure free from legacy constraints. While a full departure from DICOM is unnecessary, pathology is uniquely positioned to prototype the future—to define a more flexible, secure, and interoperable model that other domains in medical imaging may one day follow. With support from forward-looking DICOM advocates, pathology can help reshape not just its own infrastructure, but the trajectory of medical imaging itself.
TL;DR Summary for Health and Biotech Readers
This article spans multiple fields but holds particular importance for readers in healthcare informatics, pathology AI, cybersecurity, and imaging standards.
What’s New and Original in Gershkovich’s Article:
Peter Gershkovich’s provocative article, “Wearing a fur coat in the summertime: Should digital pathology redefine medical imaging?” (J Pathol Inform, May 2025), challenges the dominant role of the DICOM standard in digital pathology. His central claim is that DICOM’s architecture—designed in the 1980s—has become a liability for modern pathology, especially when compared to emerging modular, cloud-native alternatives.
Key Innovations and Contributions:
-
Pathology ≠ Radiology: Gershkovich argues that digital pathology needs its own infrastructure rather than inheriting DICOM by default. Unlike radiology, pathology remains in a more “polyglot” phase—an opportunity to build future-ready frameworks from scratch.
-
Slides as Data, Not Just Images: The paper repositions digital slides as dynamic enterprise data assets requiring orchestration, not just storage.
-
Critique of DICOM's Monolithic Design: The article shows that DICOM’s bundling of image data, metadata, annotations, and security into a single file undermines performance, security, and innovation. This tight coupling is compared unfavorably to modular software engineering best practices.
-
Cybersecurity Red Flags: Gershkovich presents compelling data linking DICOM-based PACS systems to large-scale data breaches and vulnerabilities, citing the Black Hat Europe 2023 demonstration of flaws in DICOM protocol design.
-
Modular Alternative Proposed: He outlines a modern architecture using:
-
Pixel formats like OME-TIFF or Zarr,
-
Metadata management via HL7 FHIR,
-
Annotations via SVG or GeoJSON,
-
Security overlays via NIST SP 800-53 standards.
-
-
Conceptual Shift in Interoperability: Rather than chasing one-size-fits-all standards, he emphasizes functional interoperability via well-defined APIs and microservices, supporting diverse imaging formats and evolving workflows.
-
Annotation Handling as Case Study: The paper details how DICOM’s approach to annotations stifles AI development, versioning, and usability in pathology. In contrast, external, database-driven systems allow more flexibility and innovation.
-
Practical Migration Path: Gershkovich suggests lightweight gateway microservices can bridge old and new formats, enabling gradual transitions without disrupting clinical workflows.
-
Call to Action: While not proposing a total abandonment of DICOM, the article urges digital pathology to lead a modernization effort—potentially influencing the broader medical imaging ecosystem.
Bottom Line:
Gershkovich isn’t just critiquing DICOM—he’s proposing a forward-compatible blueprint for pathology informatics, rooted in cybersecurity, modular design, and real-world adaptability. This paper could reshape how industry and regulators think about medical imaging standards, especially as AI and cloud-native infrastructures become central to healthcare delivery.
##
##
What would comedian Lewis Black make of Gershkovich's article?
##
Review of “Wearing a Fur Coat in the Summertime” — as Lewis Black might say it
Let me tell you something.
Peter Gershkovich didn’t just write an article. No. He took a goddamned flamethrower to the ossified corpse of medical imaging standards and DIDN’T EVEN BOTHER TO PUT IT OUT AFTERWARD. And thank God for that.
You see, pathology has been wandering around with a DICOM ball and chain strapped to its ankle like it’s still 1993—back when we all thought AOL was the future and you had to yell at your modem just to check your email. But here comes Gershkovich, sliding in like a cyberpunk Paul Revere, screaming: “THE PATCHWORK IS COMING! THE PATCHWORK IS COMING!” And it turns out he’s right.
DICOM, this sacred cow of radiology, is apparently less a “standard” and more of a choose-your-own-adventure novel written by bureaucrats on Ambien. It claims to be universal, but it’s universally confusing. Every vendor supports it—in exactly the way they feel like that week. Oh, you wanted interoperability? Good luck! You’ll get it when the moon turns into a bagel and spins off into a secure HIPAA-compliant cloud.
And security? Ha! Gershkovich points out that we’re embedding patient data—names, birthdays, social security numbers—directly into the images. That’s right, just bake it in there! Like a fortune cookie! “You may already have been breached!” No wonder hackers are downloading MRI scans like it’s Black Friday for PHI.
But here’s the kicker. Instead of using this disaster as an excuse to try something new, the pathology world is being peer-pressured into adopting this mess, because, well, radiology did it. THAT’S your reason? That’s like saying, “Well, my brother licked an outlet, and he seems fine. Let me try it too!”
Now, Gershkovich proposes something wild. Something so blasphemous to the DICOM disciples that I expect torches and pitchforks outside Yale’s pathology department any day now. He wants… modularity. Which is just a fancy way of saying: Let each damn part of the system do what it’s good at.
Pixel data? Use OME-TIFF, Zarr, whatever makes it fast. Metadata? Let HL7 FHIR handle it—like a normal, grown-up interoperability system. And annotations? Keep them out of the image file, thank you very much, and put them in SVG or JSON or some other acronym that doesn’t set your data on fire when you try to change a label.
And the best part? SECURITY! Real, layered, NIST-compliant security—not this DICOM "hide it under a tarp and hope" strategy. Gershkovich wants firewalls, encryption, access control, ACTUAL MODERN INFRASTRUCTURE, not “Hey, Bob remembered to unplug the server at night, so we’re good.”
This isn’t just a critique. It’s a love letter to sanity in a field so cluttered with legacy systems, it makes your grandma’s attic look like a Marie Kondo showroom.
So, in conclusion: Gershkovich, you beautiful bastard. You’ve said what we all knew but were too tired, too beaten down, or too under NDA to say: DICOM is wearing a fur coat in the goddamn summertime. And it’s time to strip it off, before the whole system keels over from heat stroke and takes all of pathology down with it.
Bravo.
