A letter to digital pathology companies from a group of pathologists for digital pathology
Dear manufacturer,
We are a group of pathologists who appreciate your vision, inventions and innovations. There is unlikely more than a handful of pathologists who on some level do not realize the advantages of digital pathology when compared with light microscopy. The ability for multiple people to view the same slide at the same time, independent of slide or pathologist location and consult with colleagues or view remote slides to assist in patient care remains one of single greatest innovations in the day-to-day practice of pathology over the past several decades.
It is a rare day when an image other than that generated from a microscope itself is not used in some fashion to support the work of our pathology group caring for our patients.
We understand that last week the FDA and others presented at a panel at the annual Pathology Visions supported by the Digital Pathology Association (DPA) and others to review the requirements on regulating devices and software that comprise a digital pathology system.
We have been following this discussion since industry representatives and pathologist leaders in digital pathology met in October of 2009. At that meeting we believe the FDA stated WSI systems are not class 1 exempt and therefore subject to pre-market approval (PMA). Topics such as intended use which the FDA relies heavily upon, nature of the systems, patient safety and effectiveness were discussed in the available presentation entitled "Digital Pathology Devices Panel Meeting" presented by a FDA representative.
Prior to that meeting, specific clearances were given for image analysis and digital read applications for HER2 and PR were given a "green light" by the FDA. More recently, another digital pathology system in addition to the slide stainer and immunohistochemical stain for HER2 were given the "go ahead" for digital reads.
In the past two years between public discussions, it is also our understanding that a DPA FDA Task Force representing a coalition of industry leaders has been in discussions with the FDA and projects such as "Project Pink" have been conducted to add to the body of knowledge towards clearance of whole slide imaging for primary H&E diagnosis.
This is the good news. The bad news now seems that despite not hearing much since October of 2009 we now know that these devices and systems will be regulated as Class 3 devices for primary H&E diagnosis although no mention of education, research, tumor board, previously cleared image analysis and interpretations nor consultations was discussed. Additionally, a representative from CMS (which manages CLIA) stated that whole slide imaging could not be self-validated within a U.S. laboratory setting.
Our group has a difficult time reconciling prior clearances for immunohistochemical stains and no mention of consultations, which in many cases may represent the "primary" diagnosis that precedes a treatment and/or management plan or appropriate follow-up (this of course is the very reason someone is actually consulted in the first place much like any other specialty or industry that relies on sub-specialists to answer specific questions – in the case of pathology it is a diagnosis) and now a requirement for primary H&E diagnosis.
Nonetheless, now we can stop worrying about what the FDA was going to say and prepare to cross this barrier with you.
While primary H&E diagnosis may not be our primary application we have no reservations about using it for such and offer our services to assist with your submissions. Like thousands of laboratories in this country, we accession thousands of specimens annually and generate tens of thousands of slides from a broad range of specimen and part types encompassing a wide range of diagnoses, both neoplastic and non-neoplastic in nature. We all recognize this is an important part of the diagnostic process but is best when used in conjunction with past medical history, clinical history, presentation, family history (when applicable) and gross clinical or surgical findings. Fortunately, this information is commonly available or presented at the time of biopsy or surgery. We are use to validating tests, technologies and procedures. CAP, CLIA and JCAHO accreditation mandate competency for laboratories for successful inspections and accreditations. It happens on a daily basis. Pathologists are also the consummate physicians when it comes to self-regulation and validation in terms of proficiency testing, internal quality assurance, mandatory internal peer reviews and external reviews in the best interests of diagnostic accuracy and patient safety. As you know, much of this is commonly being done with digital pathology today.
Fortunately, more good news for both you as a provider of technology and we as consumers know that there is a long history of being able to make an H&E diagnosis off a computer monitor. In fact, many of us remember this newspaper headline from 1986 (Colburn D. The next best thing to being there. And now, diagnosis by satellite. Washington Post 1986;August 27:7 – click on image to enlarge):
Since this proof-of-concept was successful, a search of PubMed for terms such as "digital pathology", "telepathology" and "whole slide images" reveals hundreds of peer-reviewed studies and publications with thousands of cases studied showing the effectiveness and validity of H&E diagnoses rendered with images on a computer monitor compared with conventional light microscopy. A few historic notables include work done by the Veterans Administration Healthcare System (Dunn), University Health Network (Evans) and the U.S. Military Healthcare System (Kaplan) as examples of early validations and ongoing remote primary H&E diagnostic pathology services.
More recently, several peer-reviewed articles have been published suggesting that whole slide images offer not only equivalent accuracy when compared with "truth diagnoses" but actually may be superior to conventional microscope reads and offer not only better diagnoses but access of the right slide to the right person at the right time (i.e. primary diagnosis) in addition to the "same slide, same time" value proposition (i.e. secondary consultation) and more offer more intangibles not possible with glass slides alone.
A study in e-pub ahead of print in Clin Transplant (2011 Sep 29) looking at Banff score reproducibility from several folks at the University of Alberta who know a thing or two about this as they wrote the guidelines recently published "Superiority of virtual microscope versus light microscopy in transplantation pathology". The authors conclude "The agreement for acute rejection between virtual and glass slides was not different from the agreement between two readings of glass slides. Thus, virtual microscopy is a reliable and more reproducible technology and has several advantages over glass slides, e.g., accessibility via internet, no fading. We recommend virtual microscopy for transplant diagnostics, including utilization for clinical trials."
A group of researchers in France recently published "Prostate needle biopsy examination by means of virtual microscopy" and concluded "Virtual microscopy does not compromise, but might improve, the accuracy of grading in prostate needle biopsies".
Thirdly, a collaboration of researchers in Denmark recently looked at "Automated digital volume measurement of melanoma metastases in sentinel nodes predicts disease recurrence and survival".
Last, but not least, multiple clinical investigators from several institutions and laboratories in Pittsburgh and Indianapolis recently published their experience with "Interinstitutional and interstate teleneuropathology". Lessons learned from this successful venture can be used to facilitate future efforts in this ever-growing practical vehicle for distributing pathology subspecialty expertise.
The time has come.
To these pioneers we say, you are welcome to read our kidney transplants that are "rule out rejection" or compare thousands of our prostate biopsies and prostatectomies to insure accurate Gleason grading. Clearly we alone with our microscopes cannot automate volume of metastases in lymph nodes nor do many groups have immediate subspecialty expertise when the need arises.
To all of the digital pathology technology providers we say that while we don't agree with the FDA requirement for PMA for your devices and software, we are ready to assist with the appropriate submissions and studies for what necessary clearance for us to provide the highest level of care possible to our patients with the help of your innovations to find solutions for everyday needs not unique to our group or laboratory. Validated systems that can provide safe and reliable H&E reads will provide countless opportunities and together pathologists and technology providers can succeed.
Sincerely yours,
Pathologists for Digital Pathology
