Lab Soft News

Announcing the Second Annual API Digital Pathology and AI Workshop

On December 8-9, 2017, the Association for Pathology Informatics offered for the first time a workshop on digital pathology at Henry Ford Hospital in Detroit, Michigan. This conference was somewhat unusual in that it was located in a hospital rather than a hotel or conference center. A tour of the automated lines in the pathology department of Henry Ford was included. Navigate to the above link to review the faculty and lectures that were offered at that time. It was very well received and highly successful with all seats filled.

The second workshop in this series is now being offered on December 7-8. 2018, but with inclusion of AI as it applies to surgical pathology (see: Digital Pathology and AI Workshop II). The location for this second workshop in the series will be Ohio State University College of Medicine in Columbus, Ohio. You can register for the workshop here. The capacity for the workshop has been doubled, now with a maximum of 80 people. Please consider registering early to secure a spot. Available on the API web site are the schedule and the sponsoring vendors to date for the workshop. Here is a quote from the workshop summary page on the API web site describing the workshop in more detail:

This workshop will focus on the practical issues surrounding the deployment of digital pathology and AI in a hospital setting. Ten lectures by national experts will be punctuated by Q and A sessions including industry representatives and a guided tour of the Ohio State University's Department of Pathology. Topics covered will include the fundamentals of digital pathology, the integration of the technology with LISs, the enhancement of AP workflow, the development of an RFP for purchasing digital pathology equipment, and current perspectives about AI within pathology.

This workshop has been planned with busy surgical pathologists and departmental leadership in mind. It lasts 1.5 days and takes place on a Friday and Saturday morning. Ten lectures by national experts are packed into this time frame as well as a tour of the digital pathology unit in the James Cancer Hosital. The pathology department at OSU is the first department providing primary surgical pathology diagnoses using digital pathology in the U.S. so a tour of the imaging facility will be highly educational. A dinner on Friday night for networking and relaxation is included in the registration fee.  

Google, Facebook, Twitter, and Microsoft Pursue Portability Among Web Services

I have been blogging about interoperability between hospital EHRs for nine years. Here's my first note that used this term: Do Hospitals Really Want Interoperable E-Health Records?. I believe that patients should have the right to download a copy of their EHR records and be able to provide it to other hospitals when necessary. This is sometimes referred to as health data portability (see, for example: Hospitals as a Major Barrier to the Pursuit of EHR Interoperability). Warning: if x-ray and pathology test results and digital images are included in such a download, such files may be huge. However, you can now buy a four terabyte external drive for $120. In this same light, Google, Facebook, Twitter, and Microsoft have now initiated a Data Transfer Project, offering data portability between and among these web services (see: Google, Facebook, Twitter, and Microsoft unite: Data Transfer Project promises portability between services). Below is an excerpt from the article:

Many...should be familiar with Takeout, ostensibly now called "Download your Data," which allows you to back up a copy of all the information associated with your account for most of Google's services. Right now you can even use it to transition your Drive storage between providers, but thanks to the upcoming open-source Data Transfer Project....that data may someday be as portable as your phone.... [T]he GDPR mandates data portability so Google didn't have much of a choice—though it could have easily made a minimum, token effort. With Facebook, Twitter, and Microsoft jumping on board for compatibility, the project is making significant progress. The Data Transfer Project can't unify the different proprietary APIs used by services across the web, but it can make tools that are inter-compatible between them, providing developers and users with a set of standardized data formats. So far the project has constructed adapters for seven different services across five data types. It doesn't require anything so tedious or insecure as a manual download and subsequent re-upload, either. The Data Transfer Project allows for services to make the transfer between themselves directly, service-to-service, using a high-security protocol and privacy-protecting encryption. And since it's open source, the code can be inspected and audited by researchers and enthusiasts alike to ensure that user security and liberty is being protected. You can even review, download, and run the framework yourself for your own projects if you so choose....Contributors have already built support for photo, mail, contacts, calendar, and task data formats for APIs available from Google, Microsoft, Twitter, Flickr, Instagram, Remember the Milk, and Smugmug. 

Don't expect to be able to download your medical records any time soon and I am not actually advocating for such an option. However, EHR interoperability among hospitals is critical, particularly for patients with unusual diseases or aggressive neoplasms (see: Promoting Interoperability (PI)). In these latter cases, such a record transfer should include all available digital imaging studies from both radiology and pathology, if available (see: FDA Approves the Philips Digital Pathology WSI Solution for Primary Diagnosis). Incidentally, second opinions for complex cases, particularly cancer patients, are increasingly common and should be reimburable (see: Subspecialty Second Opinions Often Critical to Patient Care). These costs could be waived if the patient is transferred to the consulting hospital. Incidentally, providing co-access to a patient's EHR records in the case of a local hospital and a cancer center would be an ideal scenario for improving the diagnosis and continuing care of cancer patients throughout the country.

Deploying Digital Pathology in the Autopsy Service Before Surgical Pathology

I believe that most pathology teaching programs in the U.S. will be feeling a need to launch a digital pathology program soon because pathology house officer and fellow candidates are favoring those programs that are using this new technology. The applicants to these programs understand that they need to begin to master it in order to qualify for positions when they enter the pathology job market. This poses a program for those teaching departments that are not in the process of adopting this technology.

A solution to this dilemma for some departments may be to deploy digital pathology first in their autopsy service. At first glance such a strategy may not make sense. The autopsy services does not generate revenue and there iare few advantages to making the images accompanying an autopsy more portable, at least outside of the specific hospital. Here is a list of why launching a digital pathology program in the autopsy service first is a reasonable option:

• House officers spend their early years in a pathology program performing autopsies so they will gain experience with this new technology soon after entering the program. This enables them to easily adapt to the use of digital pathology in surgical pathology when adopted later. 
• In many programs, the autopsy service often does not have a high priority in terms of capital investment or attention. Making it the test bed for digital pathology puts it more in the spotlight and then hopefully the recipient of more attention.
• Because the number of microscopic images associated with autopsies is relatively small compared to the histopathology workload from surgical specimens, a digital pathology program can be launched without a major investment in hardware and software. Such a new technology and program, however, will allow department leadership to establish a relationship with sdigital pathology vendors with an eye toward future contract and become comfortable with it.
• Review of the microscopic images from an autopsy becomes more flexible and can even be accomplished at home by house officers and staff pathologists.
• Autopsy microscopic slides are a suitable test bed for testing image analysis software as it become available without time and medicolegal challenges.
• Gross autopsy images are common can easily be integrated with the microscopic whole slides images to present an easily accessible total case for subsequent review.
• Microscopic whole slide images can be easily integrated into pathology courses for medical school undergraduates as well into the various hospital teaching and QC conferences.

Don't Miss the 2018 Pathology Informatics Summit in Pittsburgh, May 21-24

It's not too late to register for the 2018 Pathology Informatics Summit that will be held in Pittsburgh on May 21-24, 2018. Early bird pricing has been extended. The PI Summit can boast of a 35-year history as the only national conference devoted solely to the field of pathology informatics. It's sponsored by the Association for Pathology Informatics. Browse the conference schedule and you will discover a wide variety of lectures devoted to both applied and research topics. Also pay attention to the workshops on the first day, May 21, that are included in the registration fee of the conference including the HIMA Imaging Science and International Imaging Workshop. On the evening of May 21, there is also a Digital Pathology Association companion meeting from 5:00 to 7:30 p.m. 

As one might expect, the 39 faculty members of the PI Summit constitute a well respected group of pathology informatics experts. No other conference attracts such a group under one roof. A total of 20 vendors will also participate in the conference. They cover a wide range of products including LISs, digital pathology including quantitative analysis, lab analytics, speech recognition, specimen labeling, and workflow automation. It is commonly understood that the two largest challenges facing pathology and lab medicine in this era of value-based care and hospital consolidation are pathology informatics and molecular/genomic diagnostics. Take advantage of the opportunity to register now for PI Summit 2018 and get up-to-speed in the latest developments in lab automation and analytics. 

A Strategy for Optimizing Beaker Functionality: Leveraging the Hospital C-Suite

I am reconciled to the fact that many pathology departments will be running Epic's LIS, Beaker. Hospital executives continue to seek an enterprise wide solution to support their hospital IT needs. After perhaps selecting Epic's EHR, they will often gravitate to Beaker for their LIS. Beaker has added sufficient functionality such that it now can be described as a satisfactory choice in the face of pressure from the C-suite. Beaker can thus be described as a generic LIS. It will never achieve the status of a best-of-breed LIS because there is no need for Epic to improve it to the highest functionality level. Epic's core business is its EHR. It's not possible, nor even necessary, for the company to extensively develop its departmental modules. 

Another reason why Beaker will never achieve all of the functionality required by sophisticated pathology departments is the rapidity of scientific discovery in the fields of molecular pathology, genomics, and digital pathology. In order to meet this LIS functionality vacuum, new specialized software companies will inevitably arise to meet the needs not supplied by Beaker. Pathology departments running Beaker will need to purchase a number of such specialized laboratory software products but will then require a system integrator on the lab side to knit together the information output of their installed specialized systems that will then be interfaced to Beaker and the EHR for reporting. There is currently no provider of such integration services fo pathology departments. 

On the hospital-side, system integration requirements are being met by Epic. Specialized software companies must decide whether to participate in the Epic ecosystem. I have previously discussed this ecosystem in relation to the deployment of a lab rules product developed by Mayo and National Decision Support that is installed on the hospital side (see: Rule-Based Decision Support Tools Available on LISs for Decades; Lab Rules Firing at the Time of Test Order; Possible on the LIS?). The advantage of this ecosystem from Epic's perspective is that it has control over all of these specialized software companies and can also take contractual measures to avoid poaching of its employees by other companies. 

I doubt whether Epic would have any interest in developing ecosystems in pathology similar to those on the hospital side because its culture is based on a centralized IT model. Another reason for lack of interest on the part of Epic in developing a pathology ecosystem is that it would be a tacit admission that Beaker lacked desired functionalities. However, Epic could theoretically be forced to allow interfaces of specialized pathology products to Beaker but only on the basis of extreme pressure from client hospital C-Suites. Therefore, the only way to force Epic to permit specialized software vendors on the pathology side is by such leveraging. Initial responsibility for this rests with leadership of the pathology departments.

Register Now for the 2018 Pathology Informatics Summit, May 21-24, in Pittsburgh

It's time for all of the readers of Lab Soft News to register for the 2018 Pathology Informatics Summit that will be held in Pittsburgh on May 21-24, 2018. The lineage of this conference dates back to 1983 with an Ann Arbor conference referred to as AIMCL. The PI Summit can thus boast of a 35-year continuous history as the only national conference devoted solely to the field of pathology informatics. It's sponsored by the Association for Pathology Informatics. Browse the conference schedule and you will discover a wide variety of lectures devoted to both applied and research topics. Also pay attention to the workshops on the first day, May 21, that are included in the registration fee of the conference including the HIMA Imaging Science and International Imaging Workshop. On the evening of May 21, there is also a Digital Pathology Association companion meeting from 5:00 to 7:30 p.m. Here are details about only two of the keynote lectures to be offered. Scan the whole schedule to find more like these.

• Anil Parwani, MD, PhD, MBA: “Houston: We Have Liftoff— The Initial Experiences with US-based WSI Primary Diagnosis”
• Jeroen Van Der Laak, PhD: "Tackling the Challenges that Still Exist in Computational Pathology"

As one might expect, the 39 faculty members of the PI Summit constitute renown set of pathology informatics experts. No other conference places such a group such as this under one roof. A total of 18 vendors will also participate in the conference with some last minute additions expected. The cover a wide range of products including LISs, digital pathology including quantitative analysis, lab analytics, speech recognition, specimen labeling, and workflow automation. It is commonly understood that the two largest challenges facing pathology and lab medicine in this era of value-based care and hospital consolidation are pathology informatics and molecular/genomic diagnostics. Take advantage of the opportunity to register now for PI Summit 2018 and get up-to-speed in the latest developments in lab automation and analytics. 

Sunquest Signs Contract to Provide LIS Support for 124 Hospitals in Australia

Sunquest Information Systems is embarking on a project that has the potential to change the face of lab computing. It has signed a contract to provide lab computing support for all of the hospitals in the state of Queensland, Australia, which has a population of about 4.5 million people (see: Queensland Health partners with Sunquest for Laboratory Information System). Below is an excerpt from the press release:

The contemporary enterprise laboratory information system (LIS) and services will support improved care responsiveness, safety and healthcare outcomes for the millions of Queenslanders served by the state's 124 hospitals, 70 primary health centers and clinics, 17 aged care facilities and 33 extreme remote multi-purpose health services. Queensland Health performs more than 20,000 tests per day in 36 public pathology, forensic and scientific laboratories....The contract includes Sunquest's full end-to-end enterprise laboratory information system for all pathology disciplines including chemistry, hematology, microbiology, histopathology, cytology, blood transfusion and molecular....Sunquest demonstrates support for patient safety and communication with existing systems After being named Vendor of Choice from a field of seven vendors, ahead of the main contract Sunquest was required to create a working prototype to demonstrate key processes and achieve sign-off from a variety of stakeholders. The prototype, delivered in a prompt six weeks, successfully demonstrated interoperability with Queensland Health's existing systems showing Sunquest can meet Queensland's needs.

For the purposes of this note, I will refer to the Sunquest LIS that will provide lab computing in 124 hospitals and other facilities in Queensland as a "health system LIS" or HS-LIS. The press release refers to an "enterprise LIS" which I think is subject to misinterpretation because of the ambiguity of the term enterprise. This new contract strikes me as a potentially important step in the evolution of the classic LIS which has most often deployed in a single hospital or across several hospitals. We are now in the midst of an era with the emergence of huge health systems in the U.S., the leading one with 76 hospitals (see: Healthcare Fast Facts – Top 30 Largest Hospital Systems in America). It might make sense to me for some of these very large health systems to install a single LIS to support the labs functioning in all of the hospitals comprising the large health system.

I believe that many of these large health systems will be seeking in the future to establish multiple "laboratories of excellence" within the system as a means to insource sophisticated testing like genomic and molecular (see: Intermountain Precision Genomics to develop hereditary cancer gene panels). To my mind, this is a far superior policy to the dependence on external reference labs to provide esoteric testing. As these "laboratories of excellence" proliferate in the mega-health systems, it would also seem to make sense for the future to install a "health system LIS" or HS-LIS. This single LIS could then effeciently manage laboratory test ordering, specimen transport, and results reporting, such as will occur in the Queensland project, across all of the labs in the single health system. This will be particularly important with the adoption of digital pathology technology where slide preparation and scanning could take place in one processing lab but with the images transferred to the various surgical pathology specialists scattered in all of the hospitals in the system.

Inspirata Acquires Omnyx from GE Healthcare, Achieving an End-to-End Solution

The business model of Inspirata is relatively unique compared to other companies connected to digital pathology. The company seeks to enhance the digital pathology workflow of its clients presumably pathology departments and consulting firms. Here is a summary of the company goals copied from the company web page (WELCOME TO INSPIRATA):

We...help cancer institutions overcome the financial barriers to adoption using our unique business and delivery model where we make the upfront capital investment in the hardware, software and storage needed to enable a cancer institution to move to a fully automated digital pathology workflow inclusive of the on premise resources needed to operate the scanning lab to convert glass slides to digital images. At the end of the engagement, the institution owns the technology assets and can hire the personnel or renew Inspirata’s contract to continue the managed services model.

Inspirata recently acquired GE Healthcare's Omnyx (see: Inspirata Acquires Digital Pathology Company Omnyx, Enabling Delivery of an IVD Device for Primary Diagnosis in Europe and Canada). Below is an excerpt from the article announcing the purchase:

Cancer informatics and digital pathology workflow solution provider Inspirata...has acquired Pittsburgh, PA-based Omnyx from GE Healthcare. Omnyx was an early innovator in bringing brilliant software to pathology departments around the world....[According to] Inspirata CEO Satish Sanan, “It brings us the last mile to reach our goal of having an end-to-end integrated pathology solution with a scanner-agnostic whole slide image viewer and image management system. It also gives us a fully supported IVD device cleared for primary digital diagnosis in Europe and Canada, and it expands our customer base and geographic footprint across those countries.”....“Our short-term plan is to rapidly combine the advantages of our digital pathology software platform and the Dynamyx platform by converging them onto a shared, state-of-the-art technology stack,” continues Sanan. 

The history of Omnyx, launched originally as a partnership with UPMC, is complicated. Here are a few of the details from an article on the topic from December, 2016 (see: Omnyx ends UPMC affiliation): 

Digital pathology company Omnyx LLC will be folded into GE Healthcare at the end of the month, ending an eight-year partnership with Pittsburgh hospital system giant UPMC to commercialize a nascent technology....UPMC and GE Healthcare formed Omnyx in 2008 as 50-50 owners with the promise of contributing a combined $40 million to the venture. UPMC and GE Healthcare predicted that Omnyx would capture 25 percent or $500 million of an eventual $2 billion market.

The acquisition of the Omnyx software by Inspirata makes good sense. Note in the second excerpt above that CEO Satish Sanan emphasizes the goal for his company of developing an "integrated pathology solution with a scanner-agnostic whole slide image viewer and image management system." Inspirata will continue to work primarily in the digital pathology software space, enabling the customer to purchase scanners form other companies. Also note the reference to the acquisitor of a "a fully supported IVD device cleared for primary digital diagnosis in Europe and Canada." The recent FDA digital pathology approval for Philips was restricted to that system for primary diagnosis (see: FDA Approves the Philips Digital Pathology WSI Solution for Primary Diagnosis). However, it is likely that approval of digital pathology systems for primary diagnosis from other vendors will follow in rapid succession.

New Ownership Model for Derm Practices; ClearPath Manages Digital Images

At the end of last week, I helped manage the Digital Pathology Workshop that was sponsored by the Association for Pathology Informatics (API). One of our faculty members was Michael Kent, Ph.D., who works for a dermatology reference pathology lab in Dayton, Ohio -- Dermatology Lab of the Central States (DLCS). DLCS is part of a large dermatology entity called Dermatology of Central States (DOCS) which was recently acquired by a Chicago private equity firm, Sheridan Capital Partners (see: Sheridan Capital Partners Invests in Dermatologists of Central States). I was previously unaware that dermatology practices were being purchased by venture capital and private equity firms. This trend was discussed in a recent article (see: Is venture capital the future of dermatology?) with an excerpt from it below:

With rising overhead and time-consuming reporting requirements, economies of scale have become necessary, encouraging physicians to organize into larger group structures. Many of these larger group structures are hospital based, however dermatology has always treated outside-of-the-body disease and remained outside of the hospital....This has necessitated the formation of hospital independent groups, many of which are rather creative business structures. One emerging model is the purchase of private dermatology offices by venture capital-funded groups who combine practices under one organization; therefore, centralizing billing, managerial staff, purchasing, web-based appointments, electronic medical records, advertising, etc. This allows the realization of economies of scale and the individual dermatologist to receive some monetary value for a practice that might be otherwise impossible to sell as private practice diminishes....My concern is that venture capital-owned large dermatology groups are commoditizing our specialty and taking control of the future of dermatology out of the hands of the dermatologist.

Of special interest to me is that Dr. Kent has developed an app called Clearpath that manages digital images of skin specimens including reporting capabilities. The program is intended both for dermatologists who sign out their own cases (estimated by Dr. Kent at about 20% of dermatologists) and those who want to review the images of their cases. Here is a summary of the goals of the Clearpath software (see: Clearpath 3.0):

Clearpath 3.0 is the first-of-its-kind software solution that brings the benefits of digital pathology directly to the dermatology community—whether you already review your own slides or would like to expand your practice’s slide reviewing capability. Designed by an expert team of dermatopathologists and practicing dermatologists, this intuitive software is custom-built for ease of use, flexibility and integration into your existing EHR.

I was live-tweeting at the Digital Pathology Workshop mentioned above and posted a tweet that referred to Michael Kent and his Clearpath product. In about three days, the tweet has scored nearly 3,700 impressions which is roughly ten times more than any other tweet I have posted. This show of interest suggests a very large following for dermatopatholgy, ClearPath, and digital imagining in derm. Duly noted.

Much of the Future for Pathology and Lab Medicine Rests with Analytics

I have posted a number of notes in the past about "analytics" as it relates to pathology and lab medicine (see, for example: Lab Analytics Emerges as Hot Area for Software Development; Using Lab Analytics Software to Manage Customers Relationships in the Clinical Labs; Using Lab Analytics Software to Manage Customers Relationships in the Clinical Labs). Simply put, analytics is the systematic computational analysis of data or statistics. Pathologists will always be involved with various types of analytics because they are responsible for a large hospital computer system, the LISs, and thus are responsible for the management of large databases.

For definitional purposes, I divide pathology analytics into three categories: clinical, operational, and research. For the purposes of this discussion, I will only address only these first two. Clinical analytics involves the analysis of lab test data for the purpose of diagnosing and monitoring disease. Operational analytics is the analysis of laboratory databases in order to understand and improve the efficiency and cost-effectiveness of clinical labs operations (see: Using Lab Analytics Software to Manage Customers Relationships in the Clinical Labs). 

The next annual conference of the Digital Pathology Association (DPA), Pathology Visions, will take place on October 1-3, 2017, in San Diego. One of the workshops is entitled "Deep Learning – The Future of Computational Pathology." I found the abstract for the workshop to be fascinating and present a portion of it below:

Computational Pathology will be a powerful driver of change in digital health and has the potential to improve diagnostic, prognostic and predictive pathology. Building powerful and scalable algorithms for tissue analytics allows quantitative and objective analysis of H&E tissue samples for a variety of interpretative challenges. 

This abstract was interesting for me because it emphasized an emerging category of clinical analytics, tissue analytics. Digital pathology has now been approved by the FDA for primary diagnosis using one platform (see: FDA Approves the Philips Digital Pathology WSI Solution for Primary Diagnosis). From the perspective of size, the AP database, which encompasses whole slide imaging (WSI) data, will eclipse all other pathology databases. Tissue analytics also differs from other types of clinical analytics in that it involves image analysis and interpretation.