Digital Pathology Place

10/16/2026

The theme of PathVisions 2026 is "Connecting Innovation with Purpose," and I love it.

It reminds me of something FDA, GLP, or any other regulatory framework asks:

What's the intended use?

The big purpose is obvious. We all want faster diagnoses and faster drugs for patients. But that's the purpose of all of medicine.

Every tool on the exhibit floor has a smaller, more specific job.

So that's what I'm exploring this weekend in San Diego:

What does each innovation actually do?
Which problem is it built to solve?
Who has that problem?
Which problems are still out there with nobody working on them yet?

If I come across a problem and a solution that should meet, I'll gladly make the introduction. Mostly I'm here to look and to ask.

If you're exhibiting, drop your booth number and the problem your solution solves in the comments. I'll stop by, take a picture with you, and add it to this thread.

10/16/2026

PathVisions isn't the biggest imaging conference.

Talk to someone from radiology and they'll tell you our whole exhibit hall would fit into one booth at theirs.

I don't care. This one is ours.

It's the Digital Pathology Trailblazers' conference.

Think you're "just a beginner" because you're starting your digital pathology journey?

You're not. You're a trailblazer.

Most pathology in the US still runs on glass. So everyone on that floor is pushing through resistance, workflow problems and skeptical colleagues.

And people still ask us: why digitize an analog slide at all?

Here's my answer.

The next frontier is glassless imaging. Tissue imaged directly to digital, no slide in between.

When it comes, someone will have to run the validation and equivalency studies.

It'll be the people at PathVisions this week. The ones choosing scanners, building scanner fleets and figuring out AI on top of their images.

The work everyone called inefficient is what gets us ready for what's next.

PathVisions 2026, here I come. See you in San Diego, Trailblazers.

If you're there, find me at the Hamamatsu booth #201.

10/15/2026

Getting a Hamamatsu scanner, or already have one?

The scanner is the easy part.

The hard part is everything after it lands in your lab.

This Saturday at I'm going live from the Hamamatsu

booth ( #201) to introduce you to the team that supports users through those messy first months.

And there's ice cream, in more colors than a multiplex panel.

📍 Booth #201, PathVisions 2026, San Diego
🍦 Ice cream social: Sat, Oct 17, 2:45–3:30 PM PT
🔴 Livestream: Sat, Oct 17, 2:45 PM PT / 5:45 PM ET on LinkedIn, YouTube, and Facebook.

Not in San Diego?

Join the livestream and ask your questions in the chat. 👇

In partnership with Hamamatsu Corporation

Digital Pathology Association

10/15/2026

Still doing cytology on glass?
Sunday 7:30 AM: hear from a lab that went digital.

The speakers at Epredia's PathVisions workshop are from Puerto Rico. They brought low-cellularity samples into a digital workflow and published what it took.

That's the part I care about: people who lived the implementation, telling you what worked and what didn't.

→ Sun, Oct 18, 7:30–8:30 AM PT
→ Seaport Ballroom A–D
→ I'll be in the room, then interviewing the speakers

After that, come say hi at Epredia booth #301.

Disclosure: Epredia is sponsoring my workshop coverage.

10/14/2026

1st time at PathVisions & still figuring out digital pathology?
Your first stop: booth #201.

I've been the one nodding along to computer scientists, hoping nobody noticed I was lost.

So at the Hamamatsu booth, we're embracing the messy beginnings:
→ No question too basic
→ Drop yours in our question box
→ We answer them live, both days

Podcast listener? Come say hi in person.

San Diego, Oct 16–18. See you at #201.

Disclosure: Hamamatsu is sponsoring my PathVisions coverage.

10/13/2026

The digital pathology project that went sideways on you? I want it on my podcast.

Not the polished conference version. The real one.

The scanner that choked on your thick sections. The pathologist who wouldn't let go of glass. The validation you had to redo. The go-live that looked perfect in the plan and fell apart by lunch.

Those are the stories nobody puts on a slide, and they're the ones that help the next lab avoid the same mistake.

This weekend at PathVisions in San Diego, I'm recording 8 short interviews with Digital Pathology Trailblazers:
→ Pathologists, histotechs, lab managers, IT, researchers: the people bringing digital pathology into their own labs
→ This round is just for the end users. Industry friends, I'll see you on the exhibit floor!
→ 15 minutes. No prep. Just your story: the hurdle, the lesson, what you'd tell someone starting today.

8 slots, Saturday and Sunday. Book yours here: https://calendar.google.com/calendar/appointments/schedules/AcZssZ0yjFhaL0kGJ7bZ18oITwyw-qY2yxvcDEyfqpj3LBulrRdiO4fRUkgM0Avlj_oJkIaw-0ZjhaC9

10/13/2026

The moment you freeze an AI model, it starts aging.

But when that model becomes a medical device, updating it isn’t simple. Every change can affect performance, validation, and regulatory status.

The FDA and European Union take different approaches to regulating model updates.

At the time of this livestream, the pathology AI landscape remained limited:

- Six pathology devices had been cleared.
- None were based on large language models.
- Cytopathology still lacked an established AI benchmark.

My practical takeaway is simple: start where the error is cheap.

Begin with workflows where mistakes are easy to detect, recoverable, and reviewed by a subject-matter expert before moving into higher-risk applications.

Both open-access papers and the full episode are linked below.

Which low-risk pathology workflow would you choose for the first deployment?

10/12/2026

Why are these platelets forming a ring around a neutrophil?

This unusual blood-smear pattern is called platelet satellitism.

The key features are:

Platelets appear attached around the neutrophil.

Platelets are the smallest formed elements visible on the smear.

They are cytoplasmic fragments rather than complete cells.

This arrangement is a specimen-related artifact, not a normal biological pattern.

I always find artifacts like this useful because they remind us not to interpret every striking pattern as disease.

What would you check next to confirm that this ring is an artifact?

10/09/2026

Think digital pathology starts with your first whole-slide image?

The groundwork starts before you press “scan.”

For me, a scanner is one part of a bigger pathology workflow.

Before that first image, I want to know why we're digitizing and how those images will fit into daily work.

Start by asking: what will we use the images for, and what needs to be in place to support that?

Watch this truth bomb, then tell me: what would you sort out before your first scan?

10/08/2026

I once didn't check a reference that AI gave me.

It almost cost me a publication. The journal gave me a second chance, and since then I click on every single one.

That story made it into the talk I gave at ESTP in Kraków last week: how I use AI for scientific writing without learning to code.

My 4 rules:
1. Domain expertise is the prompt
2. Supply the sources
3. Encode corrections
4. The subject matter expert verifies

The numbers that stuck with me:
→ 25-50% of references didn't exist when an LLM answered GI pathology cases from memory
→ 97.5% citation accuracy when the papers were supplied to the model
→ 50 out of 50 times, a model justified switching patients from Tylenol to acetaminophen... its own active ingredient

I recorded my practice talk at the airport on the way there.

How are you checking what your AI writes?

You can use AI to write scientific papers without fabricated refere...