Keith J. Kaplan, MD, associate professor of laboratory medicine and pathology at Mayo Clinic, Rochester, Minn., spoke at the CAP Foundation’s June 12–14 Futurescape of Pathology conference in Rosemont, Ill. His topic: Pathology 2.0—A New Perspective on Clinical Practice. An edited version of his presentation follows.
Keith J. Kaplan, MD
I will discuss new tools and technologies, how they apply to the daily practice of surgical pathology, and the long-term implications for our practices. A lot of this, in my opinion, will be patient and consumer driven, but it will allow us an opportunity as surgical pathologists to improve our services and provide information that was once locked away but can now be made available, and can showcase our specialty as well.
Web 2.0 is a popular suffix for a new version, a new model, or a major shift revolution. Whereas Web 1.0 had users following content, meaning they would visit a page and find information that they would have to pull out, Web 2.0 allows users to comment, edit, and create content. In Tom Friedman’s book, The World Is Flat, he cites uploading—content by users for users—as the most disruptive of 10 disruptive forces. But I’ll make the case that digital pathology, given its image-intense nature, and our specialty by its very nature are well suited to a number of these applications.
In Web 1.0 you may have used a Web portal; you now search Google. Previously you may have followed a Web site; you now comment, participate, blog, and comment. In the past, Web sites were concerned with ‘stickiness,’ or people staying on their sites; now the information following you has become much more important. Publishing services such as syndication and RSS and podcasts follow you when you’re able to look at that information.
Online encyclopedias have been replaced by Wikipedia. Directories and favorites have been replaced by tags. Photosharing sites are largely those like Flickr, rather than online photo galleries, where you can edit and share. A number of us have made transformations to Google Applications where you can have shared documents as opposed to e-mailing documents.
This is actually an old story by now—people are talking already about Web 3.0. I’m not sure what that’s going to look like, but there are some references to what people think Web 3.0 will look like. For example, whereas Web 1.0 was read only, and Web 2.0 was read/write, Web 3.0 will be the ‘portable personal Web.’
I saw one example on my way to this Futurescape conference. You’re traveling to a meeting in Chicago. You might search Google for restaurants around O’Hare International Airport or for movie theaters, or shows downtown you might like to see while you’re in town. Web 3.0 is thought to be more semantic. So you’ll key in a trip to Chicago, and it will provide you information much more seamlessly about where the best restaurants are, where the shows of the kind that you want to see may be.
There are also implications here for marketing and a term called ‘advertainment,’ where people are going to make advertising much more entertaining. We’re starting to see small pockets of this now. Some of you probably use iGoogle or NetVibes. These look something like Semantic Web, and people are already referencing applications for Web 4.0 and 5.0.
Digital pathology, vis à vis the digital dataset, has a number of applications that are available to us, and a lot of them have uses with Web 2.0 technologies.
Why? If you think of biomarkers, targeted therapies, personalized chemotherapy, personalized medicine, about the implications for genomics and proteomics, and the 150 other kinds of -omics, the tissue is still the issue. There was a lot of discussion about wax and glass going away. I don’t think that’s going to happen for some time. So the tissue still remains at the center of all of this.
The 2.0 environment allows us to share that information, collaborate, and exchange. It allows us to showcase what we as pathologists do, increase the information we have, and advance the communication we can provide. These technologies are a bit disruptive, but I would encourage you to think about how you can use them in your practice, or how they can be used in the pathology community, and perhaps to give up some of the old tools and technologies that may have been tried and true. It’s better that the change occurs with us and for us.
I’ll give you one practical example of uploading, image sharing, and exchanging content. We have a largely homebrew PACS at the Mayo Clinic. It’s now in version 5.0. The idea was not only to host the radiology images, of course, but also to host intraoperative photographs, gross images, and pathology images, way back when, perhaps before even some of those images may have existed or people thought about putting that information into the electronic medical record. That’s now done routinely.
As a matter of practice, I routinely include liver biopsy photographs from liver transplant patients in the hospital PACS as a snapshot of the patient’s disease. We have a number of liver transplant patients who come for protocol and non-protocol biopsies, extending for weeks, months, and years from their surgeries. We have some patients 20-plus years after liver transplant who still undergo routine biopsy. And it’s laborious to go into the files, pull out old cases, and review them, and this now affords us the opportunity to provide a snapshot of the patient’s disease from that biopsy and that point in time, to compare to the next biopsy when they revisit. This can be laid out in many different ways, so you can have the preoperative, radiology, and pathology imaging.
It takes about two minutes to capture those images, name them, and upload them into the PACS. They’re available in about five minutes. The clinicians review the images, and we share the same snapshot over the phone. This allows them an opportunity, if they’re not on campus or they’re seeing the patient quickly and may not have time to stop in the lab to look at the slides, to review the slides with me—the same slides at the same time. They think it’s wonderful. We do this also in the renal biopsy service, and a number of clinicians share the images with their patients, and the patients will say, ‘You know doc, you’ve been biopsying my liver or kidney for 10 years. You told me I needed steroids or didn’t need steroids. Now I know why, and this is something we can’t get at our local hospital, and this is why we come to a place like this.’
There’s tremendous information to be shared and had, and tremendous value to our division, our department, and the institution as a whole. We have a separate database for gross and autopsy images. That now will be transformed into the hospital PACS as well.
All of us are going to have to become accustomed to this transparency as personal health records evolve. The ability to externalize and share our information is going to be paramount.
A few words about image analysis in the context of the new technology. Machines are very good at counting, not very good at thinking, but offer good reproducibility, and there are several studies suggesting better information and better reproducibility in terms of quantitative immunohistochemical analysis. The financial drivers are separate codes for reimbursement as a part of this analysis.
There are now a number of vendors that provide excellent hardware and scanning technology, but from my perspective it’s become less of a hardware issue and more of a software issue. And people are asking, It’s great that I can digitize a whole slide, but what do I do with the information? Certainly one application is image analysis, and a number of vendors are offering this. Using HER2 as an example, from the ASCO/CAP guidelines, the use of image analysis may be helpful in terms of HER2 interpretation.
The question is no longer, What is digital pathology? It’s, How do I use digital pathology in my practice, or how do I set this up in my institution? It becomes a workflow issue. We have a number of studies underway now not only assessing the technology but also looking at how to put it into practice. I’m not sure what the workflow will look like, but in some way, shape, or form it’s essentially a quality assurance tool or a check. Perhaps you might order a HER2 stain on a biopsy-proven breast cancer. It may be clearly negative or clearly 3+-positive, and you would report those results. For the marginal results, particularly those that are 2+, where you may just perform FISH as an adjunct unnecessarily, the image analysis may provide you with more detailed information to guide appropriate FISH testing.
Now for medicine 2.0, pathology, and Web 2.0. I came across data on the Web about the explosion in the number of users of Web 2.0 technologies like Facebook, LinkedIn, MySpace, and Flickr. Several of these sites are still pretty young and yet the numbers are staggering. There’s actually a blog that follows how many hospitals are using Web 2.0 technologies: 250 hospitals were on the list as of May 3. In early June the new number was 277 hospitals. About a 10 percent change in about three weeks time in terms of hospitals adopting Twitter, YouTube, or Facebook. Growth in the use of these technologies by the public has been incredible, and it doesn’t appear to be diminishing significantly at all.
Twitter started out with the premise of providing short messages, 140 characters, not 140 words, to provide the answer to the question, What are you doing? But Bruce Friedman, MD, suggested providing tweets in the form of, What are you thinking? That’s a good approach. And for those of us who are on Twitter, we realize people have different motives for using Twitter. I use it to feed my blog posts into Twitter. I have about 400 followers now in just a few months, and I follow about the same number of people. And I encourage and challenge my residents and fellows to Twitter diagnoses in 140 or fewer characters.
One unanswerable question at this point is where these technologies are going to go, where the business models are, and where the revenue streams will come from. Certainly information sources, clearinghouses, and publishers are looking at ways to interact with this, but here and now it’s not entirely clear where a lot of this will go.
The Mayo Clinic has been held as one example of embracing and using medicine 2.0 technology. It has a YouTube channel, a Facebook page, a Twitter page, as well as three blogs for patients, friends, family, and employees. The institution realizes that much like word of mouth has served it well over the past 100 years, these forms of social media are a form of word of mouth for the next 100 years—and thus we’re using them.
There’s an interesting story behind Mayo Clinic putting itself on Facebook two years ago. A rock band in London uses the name ‘Mayo Clinic,’ and the Mayo Clinic itself was concerned that the band might set up a Facebook page before the institution, so it locked down the name before the band could do so, and started to popularize this. The institution’s Facebook page now has about 7,000 fans who send comments—fortunately more good comments than bad.
Here’s an example of how this works and can work to showcase our specialty. A half dozen of you in this room e-mailed to me a link to a Mayo Clinic piano video and asked me if I had seen it. It was of an older couple who had visited the Mayo Clinic. She’s 83, he’s 90, and they did a little duet in the hotel lobby on one of the pianos, and this video of it took off—3 million views on YouTube. Three million pairs of eyeballs seeing Mayo Clinic and then all the other associated information and links about the institution.
The Wall Street Journal picked up the story, and then the couple appeared within a few weeks on “Good Morning America.” The local papers in Rochester, as well as the couple’s local paper in Iowa, picked up the story, and then other national news sources and media did the same, and about the time they were on “Good Morning America” the number of page views exploded.
Free advertising. Free word of mouth. It cost the Mayo Clinic nothing to set up the YouTube page. And someone else had taken that video and put it on their own YouTube page.
Look at where word of mouth has gone. The Mayo Clinic hasn’t had a marketing division for more than 20 years, and at one point it was one person, and they largely relied on word of mouth. They would quote that about 90 percent of patients, on average, said they have said good things about Mayo to an average of about 40 people, and 85 percent say they recommended the institution to a friend, advised 16 to come, and five actually did come. They’re using social media to increase that number.
Mayo Clinic has guidelines in place for those who use social media and blogging tools. As someone who has a blog, this was a concern of mine—respect for the institution certainly, and, being social media savvy, having guidelines for employees.
If you go back into the social media progression, there was a publication years ago with which the Mayo Clinic tried to reach out to people directly, called “Medical Edge,” and then as Web technologies started to advance, they created outposts on popular sites. As one example, they created podcasts on iTunes from which health information could be downloaded to further energize word of mouth.
The cost of Twitter, Facebook, and YouTube is zero. The blog itself has a very low cost. And leveraging 500,000 annual unique visitors to the Mayo Clinic, as well as 50,000 employees, like me, as ambassadors, talking about this, is very low cost and very high yield. And care delivery is more efficient, with online communities for patient support groups, chronic disease management, and workplace collaboration. Being able to enter at a very low cost or at no cost allows you to have proof of concept very early.
Why is all of this important? Mayo clinic.org receives about 1 million visitors per month. About 10 to 15 percent of users become patients, though those data are admittedly hard to come by. But this is going through a bit of a transformation now. There’s a lot of interest now for e consumers to remain perhaps e patients. There’s a need right now at the Mayo Clinic for teleconsultation, where perhaps referral on site may not be needed. So e consumers to e patients, or e consumers to real patients.
As I mentioned, a number of information sources are looking at ways to interact with this. I’ll give you one recent example. The Centers for Disease Control and Prevention had tweets, as they’re called, on the recent H1N1 influenza, with 160,000 views. They had far fewer views on their Web site but many more views on the CDC Twitter page.
All of this, of course, speeds up the news cycle. The story about the heroic Hudson River plane landing broke on Twitter. Someone captured a ‘twitpic,’ as it’s called, from his iPhone on a passenger ferry on the river and uploaded that to Twitter within minutes of the incident. The New York Times got the news about 30 minutes later, and had it on the front page of its Web site in about 40 minutes. MSNBC interviewed the photographer within 30 minutes of the picture.
There are a number of applications in pathology that facilitate this. MyPACS.net, for one, is telling me I have 20,000 cases and 100,000 images to look at. Medting—which I’ve contributed 10,000 images to—tells me I have 2,300 cases to look at and about 30,000 images waiting for me to look at.
There are a number of sites on Flickr for Web hosting, sharing images. I upload now all of the medical school lectures I give, and board review courses that I used to give, to slide-sharing sites as a way to share the information. The real value is in sharing the information. And we have a small medical school class. The board review classes are slightly larger, but there were about 2,000 downloads in the course of a year. It would take me about 20 or 30 years of giving those courses to have that kind of impact. People will say, Aren’t you giving the information away? I suppose that’s arguable. There are people who have concerns about putting pathology images on the PACS. They think we’re giving away the secret to our sauce. Again, there’s a level of transparency we’re all going to have to get accustomed to.
A number of my colleagues are adamant about not including pathology images in the PACS because, they say, the images may be misused or misinterpreted. I disagree. There’s a pathology report that will govern that. I’m also less concerned about those images being used without credit for publications or presentations.
Another great site, out of Duke, is PathologyPics. It’s a very easy system to use. Upload images to Web hosting. There’s an FTP site for mass uploads, and it’s a great site to comment, criticize, and exchange. I’ve uploaded images and people have criticized my diagnoses. It’s all part of the discussion.
Finally, the whole-slide vendors are getting into this arena as well. Aperio has SecondSlide, BioImagene has PathXchange for the purposes of sharing cases and exchanging information, whether they be consultation or of an educational nature. Again, it’s the ability to share and host your images, exchange, collaborate, and so on.
I’ve uploaded a number of cases to PathXchange as well. We now use it as a site for our soft tissue tumor trading club, and we use it as a site with our clinical colleagues. We have a series now of pancreatic biopsies that we’re going to share with our clinical colleagues, as well as pancreas experts across the world, to look at privately, independently, share the information online, instead of mailing glass slides all over the world. And this was driven by our clinical colleagues in gastroenterology as a way to look at interobserver comparisons for pathologists. They were concerned about the quality of the diagnosis, the understanding of the disease processes, and they said, Why don’t you upload these 40 cases, 10 of each example, and have people from around the world look at this, comment, criticize, exchange, and share? It’s a potentially untapped market in terms of engaging clinicians in these types of applications.
Sermo is another example. If you’re a physician you can register for Sermo. You can put in your diagnoses. You can use other diagnoses. We’re a small specialty and we have a small presence on Sermo compared with a lot of other specialties. But what you will find most interesting on Sermo pathology is the number of clinicians that comment on pathology cases. You’ll see obstetricians and gynecologists commenting on GYN cases, and implications for disease management, and answering the questions that we have in terms of what does it mean for patient care. With small screen technologies increasing, this is something that’s going to become more pertinent.
Finally, blogging as a form of Web 2.0 technology. I have a daily blog at tissuepathology.com or digitalpathologyblog.com. [My thanks to all the vendors who have graciously sponsored it.] Bruce Friedman had been blogging for a couple of years when he suggested two years ago that I have a blog on digital pathology. I thought it was the worst idea I had ever heard, in large part because when I lived in Washington, DC, there was a terrible blog called whyihatedc.com, but it provided commentary for those of us who lived and worked in DC about the community and the problems, and so on. This guy had been at it for a while, and then one day the Washington Post ran a short column on the blog itself, and I thought in some way this validated blogging.
What I do is a form of blogging, known as academic blogging, that’s used to educate and inform the pathology and laboratory community. It’s of a less personal nature but it also affords a lot of opportunities for social marketing and networking. Also, as an employed physician, it allows me an independent voice.
Perhaps a better question, rather than why I blog, is why physicians blog. In general, physicians blog to give their perspective on devices, tests, rules, and regulations in medicine. Some are of a more personal nature, and if you’re familiar with “A Piece of My Mind” or “Doctor Stories” or those types of pieces that appear in JAMA and other publications, this has sort of been externalized to the Web. They deal with personal reflections, death of a parent or patient, for example.
A lot of them, frankly, rant and rave about fraudulent practices or ideas, or payers, or insurance companies. Very few take the opportunity, outside of pathology bloggers, to showcase their specialty. They could do a much better job highlighting what they do for the public and their colleagues.
Blogging allows me an opportunity to explore and share new ideas, certainly with time constraints and respect for my employer. And most people would consider blogging a short form of journalism. Though this may be a reflection of the quality of my publications, it’s possible I may have more impact with a few blog notes than I might with a peer-reviewed paper in the academic literature.
Blogging also affords synergy between industry and academia for collaboration. And it was thought a while ago that mainstream media companies would master blogs and other forms of social media as advertising tools. And we’re seeing that now. There’s a hospital in Chicago that has advertising on Facebook, showcasing its services, new technologies, clinical staff, job offerings, pay-for-performance ratings.
What do you do as a blogger? You scan the horizon for the items of interest to you, filter and interpret them, and essentially re-post that content with related comments, criticism, and ideas. I don’t have as much time as I would like to comment substantively on things, but I do find some level of interest in the re-posted content.
Until about two years ago, I think there were three of us in pathology who were blogging—Bruce Friedman, Jules Berman, MD, PhD, and me. It’s now three times that number. There’s a group blog at Pathtalk.org, which has 15 authors that have set up blogs with varying consistency in terms of the number of posts, but certainly high-quality posts.
One of my colleagues at Mayo, Bobbi Pritt, MD, has a wonderful blog, “Creepy Dreadful Wonderful Parasites,” from her travels all over the world, and a parasite case of the week. And all of these links are available on my blog and the others as well.
Much like there are blogs about articles, there have been articles written about blogs. Darren Wheeler, MD, of Las Vegas writes a section for Advances in Anatomic Pathology on ‘Pathology on the Internet,’ and Brian Moore, MD, whose neuropathology blog is showcased on the CAP transformation Web site, talked about the blogging phenomenon hitting pathology.
I hope I’ve made the case that there are many more tools and technologies at our disposal that allow us to do what we all enjoy doing: sharing information, exchanging experience, harnessing collective intelligence. And they allow us to do these things in a way we’ve never been able to do them before.