"Eminence-Based" Surgical Pathology and the Digital Pathology Department

In a recent note, I commented on the new strategic alliance between GE Medical and the University of Pittsburgh Medical Center (UPMC) in the pursuit of the digital pathology department and whole slide imaging (see: GE Medical Partners with UPMC in Pathology Imaging Venture). It is often stated that digital radiology took about a decade to mature and that digital pathology will take an equal amount of time to become the accepted standard of practice. However, major incentives were available in the conversion to digital radiology such as the ability to offer new imaging procedures with attractive profit margins plus a groundswell of enthusiasm on the part of hospital clinicians for these new offerings. These same incentives do not exist for digital pathology -- there are no additional profit margins to be gained for the hospital and the pathology reports to clinicians are largely the same except for the routine integration of digital images into surgical pathology reports.

However, it you are searching for a "killer app" associated with digital pathology, it is most certainly image search. By this I mean the ability to isolate "fields of interest" in a challenging surgical pathology case, an unusual malignant tumor for instance, and then match them against a large image database of diagnosed lesions for similar lesions. For challenging cases today, this same process often takes place laboriously by searching surgical pathology atlases on the shelf. Parallel to this process, many such cases are also referred to local colleagues for their opinions and also sent to "marquee" surgical pathologists who have established reputations as having superb diagnostic skills in various specialty areas. This is the basis for what has been called eminence-based surgical pathology.

And now comes the rub. These marquee pathologists have little to gain and much to loose by the introduction of digital pathology and image searches. Image searches and pattern matching is, in fact, what is taking place in the brains of these eminent pathologists. I suspect that they will have little enthusiasm for any technology that serves to lessen their influence, prestige, and livelihood. Many of these latter pathologists are also highly placed in the hierarchy of prestigious pathology departments. I suspect that they will be leaders of the chorus opposing the conversion to digital pathology and perhaps highlighting the failures and inadequacies of digital pathology departments.

Comments

[Brian Jackson]

Bruce, I respectfully differ with your prediction of the impact of imaging on eminence-based surgical pathology. In an increasingly consumer-driven world, branding (i.e. reputation) will become much more powerful than under the old 3rd party reimbursement/commodity mindset.

I envision a near future in which patients request (and pay for out of pocket or out of health savings accounts) 2nd opinions and consults directly from pathologists. These consults will be greatly facilitated by digital imaging. And they won't go to Dr. Jones, the generalist surgical pathologist down the street; they'll go to Dr. Big Cheese with the national/international reputation.

[Julie McDowell]

I'm working on a story on the business case (for or against) digital pathology for the next issue of Washington G-2 Reports Laboratory Industry Report. Anyone from the pathology and/or lab community who wants to comment/sound off on this, please contact me at jmcdowell@ioma.com. Thanks!!

[bev M.D.]

I think Dr. Pemberton's observation probably does represent the long term future of pathology, but every pathologist knows that sampling error can be lethal in tumor diagnosis, and what is seen on one microscopic slide of a tumor may differ from the next slide. It remains to be seen whether such morphologic heterogeneity will carry over into the molecular analysis realm. But Dr. Friedman makes an important point - currently pathology remains highly subjective and susceptible to "eminence bias" -e.g., the tumor name is "X" because Dr. Bigwig says it's "X". I have seen these bigwigs be wrong or, worse, make the correct diagnosis in retrospect (such as when the tumor has already metastasized and the original pathologist is being sued) to maintain their "eminence". Hopefully the "killer app" could serve as a bridge between the current situation and the perfection of molecular analysis.

[Jackson Pemberton]

Bruce, I'm not sure that even this "killer app" would enable a good business case for converting to digital pathology. I suspect appropriate molecular testing on fresh samples of morphologic puzzling tumors is the future of our speciality, and will be used for diagnosis, treatment direction, and prognosis. The morphology will be of variable importance, depending on site and tumor type. In many cases the microscopic exam will serve more as a quality control check.