Lab Soft News

Digital Natives vs. Digital Immigrants: The Relevance for Pathology

I have been picking up lately on some new terminology being used in blog posts and relating to what has been called computer literacy in the past -- the new terms are digital natives and digital immigrants. Here is how these new terms are defined in the Wikipedia:

Marc Prensky claims to have coined the term digital native, as it pertains to a new breed of student entering educational establishments. The term draws an analogy to a country's natives, for whom the local religion, language, and folkways are natural and indigenous, over against immigrants to a country who often are expected to adapt and assimilate to their newly adopted home. Prensky refers to accents employed by digital immigrants, such as printing documents rather than commenting on screen or printing out emails to save in hard copy form. Digital immigrants are said to have a "thick accent" when operating in the digital world in distinctly pre-digital ways, when, for instance, he might "dial" someone on the telephone to ask if his e-mail was received.

I'm not sure if I am totally comfortable with the notion of a "digital immigrant" having a "thick accent" by which is meant that he or she is less at ease operating in the digital world. However, I am totally certain that an individual who has grown up in the digital world makes a number of assumptions that older colleagues would not automatically make and may utilize digital technology in a different way. It is important that medical educators from older generations understand these assumptions in order to bridge the "cultural divide" between them and the group that they are teaching.

For me, the most important of these assumptions on the part of pathology residents is that the web is the first place to which they will turn to acquire information about, say, a challenging case encountered in a surgical pathology reading room. I have had senior faculty colleagues express surprise to me that their residents and fellows will walk past a book shelf loaded with authoritative pathology text books and atlases and turn to on-line pathology resources for assistance in such circumstances. Many of the digital resources that they turn to may have been developed by experts in the field but, to my colleagues, they seem to lacked the weight of the hardcopy books and atlases.

What I personally take away from this brief discussion is that some priority should be assigned to making our most authoritative textbooks and atlases accessible in electronic form on the web as soon as possible for the increasing number of "digital natives" in pathology. I have posted a number of previous notes about e-textbooks (see: Some Ideas About the Future of Medical Textbooks; PathCONSULT: A Web-Enabled Surgical Pathology Atlas; Networked Textbooks in Pathology and Lab Medicine).

Medical Students Rally for Greater Disclosure of Faculty/Pharma Financial Ties

In general, I believe that medical students should spend their time hitting the books rather than engaging in various forms of social protest. However, I have just discovered a situation where it seems to be that a medical student "action" would be warranted (see: Harvard Med Students Might Want to Look Up Charles Grassley). Below is an excerpt from the note from the WSJ Health Blog:

Medical students, unite! It’s time to fight for your right to know your profs’ drug-industry ties! That’s what’s happening over at Harvard, where some 40 students rallied recently on the steps of Harvard Medical School’s Gordon Hall. Along with some folks from Tufts and Boston University, they were waving signs and pushing for tighter conflict-of-interest policies vis-a-vis Harvard docs and pharmaceutical companies....As the [Harvard] Crimson tells it, students have been trying for six years to get the administration to tighten its conflicts policies, both in the classroom and at the affiliated hospitals where the students train.  One idea they’re pushing is to require faculty and students, while talking about drugs in the classroom, to disclose any ties to the makers of those drugs. Harvard’s dean for faculty and research integrity, Gretchen Brodnicki [said that] the administration is taking the students’ concerns seriously. But she noted some practical issues, especially when it comes to the hospitals, which Harvard does not own or operate. The school can’t “force” affiliated hospitals to change their existing policies to line up with what the medical school requires internally, she said. She added, though, that the hospitals’ policies “go well beyond where we stop.”

I have covered in previous notes the problem of non- or inadequate disclosure of payments from Big Pharma on the part of medical school faculty (see: Medical Schools Share Some Blame in Scandals Involving Pharma Payments to Faculty; On the Corrosive Influence of Big Pharma on Academic Physicians). Clearly the faculty self-reporting system to medical school deans is not working, at least for some schools. As I noted in my note cited above, the medical schools need to share some of the blame. Here's my specific language:

...I don't think that the medical schools should emerge from these scandals untarnished. How is it possible for major medical schools to be "incapable of policing their faculty’s conflicts of interest." I can assure you that these institutions each employ hundreds of personnel in their research administration divisions to manage and audit research grants and contracts to the penny. And the reason for such diligence is obvious -- not to do so puts their research enterprise into jeopardy and such funding is the major sources of revenue for them.

The quote above from Harvard Medical School Dean Gretchen Brodnicki about the hands of school administration being tied regarding teaching hospital policies is a total "cop-out" and gives me little cause for optimism. The fact that Harvard Medical School does not "own" its teaching hospitals should not act in any way as a barrier for the harmonization of faculty ethical and conflict of interest policies across the various affiliated institutions.

Roche Diagnostics Purchases Swisslab, a German LIS Vendor

Roche Diagnostics seems to be pursuing what some would call an end-to-end integration solution with its announced purchase of the leading German LIS vendor Swisslab (see: Roche expands its portfolio of laboratory information systems and laboratory automation solutions for large laboratories). Below is an excerpt from the press release:

Roche and Swisslab GmbH, Berlin, announced today that Roche is to acquire all the shares in Swisslab GmbH. The transaction is expected to be completed in the first half of 2009. Through this acquisition, Roche is seeking to expand its portfolio of fully integrated solutions for laboratory diagnostics, including the increasingly important information technology systems. The laboratory information system Swisslab, together with the LAURIS module, permits highly efficient management of laboratory processes and diagnostic data in large core laboratories and in specialty laboratories concerned with areas such as microbiology, blood banks, newborn screening, pathology, histology, transfusion management, genetics and HLA typing, as well as point-of-care integration and quality management....With the acquisition of Swisslab, Roche will become an international leader in the supply of software solutions for in-vitro diagnostics, gaining a leading position in the European market for laboratory information systems.

I heard a rumor a few years ago that Swisslab was studying the U.S. LIS market but I never heard anything more about their plans. It's interesting that the press release only speculates about the effect of this Roche purchase on the European LIS market.

Continuing Discussion of the Use of the Term Biomarker in Molecular Diagnostics

It occurred to me that the term analyte, commonly used in the past to refer to the substances being measured in the clinical laboratories, has largely been replaced by the term biomarker. I discussed what I understood to be the differences between these two terms in a previous note nearly three years ago (see: Use of Term Biomarker vs. Analyte). Here is an excerpt from that note, including an embedded and very broad working definition for biomarker:

The term analyte is defined as a substance that a laboratory test seeks to detect (or analyze).  I discovered many definitions for the term biomarker on the web but the one that I liked the most is the following: any molecular species found to provide correlation to a particular phenotype or perturbation of a biological system. Of particular interest to me was that many of the definitions for biomarker I encountered were specifically linked to the detection of cancer. I believe that a common practice today is to use the term analyte for the detection of relatively simple elements or molecules such as serum chloride or liver enzymes and that the term biomarker has evolved out of genomic/proteomic science to describe new markers for specific diseases or groups of diseases.

Given that many of my blog notes lately have made reference to new lab tests for cancer and the combination of these tests into complex panels, I rarely have had occasion to even use the term analyte but have been referring constantly to biomarkers. However, I must confess that I had in the back of my mind the idea that biomarkers were generally proteins circulating in the blood that had been coded for by DNA. However, this notion was shaken by my recent note about various species of microRNA functioning as DNA switches. These substances are now being measured in malignant cells as well as in serum and are thus functioning as biomarkers to help predict the clinical outcomes of malignant disease (see: microRNAs Show Promise as Biomarkers for Cancer Prognosis and Outcomes).

My original working definition for biomarker quoted above (any molecular species found to provide correlation to a particular phenotype or perturbation of a biological system) certainty did not exclude reference to DNA and RNA -- I guess that I just had the idea that these molecules were the templates and distinct from the "marker" proteins that they coded for. In retrospect, this was the wrong approach. We now know that microRNAs act as DNA switches and thus can serve as laboratory "markers" for malignant disease or the modification of these diseases.

Blood Pharming: An Experimental Technology to Produce Red Cells in the Battlefield

DARPA (Defense Advanced Research Projects Agency) has a well-deserved reputation for launching and funding new technologies for use in the U.S. military but often with broader applications. The agency has apparently not been resting on its laurels and has now hatched the idea of replicating and packaging universal-donor red blood cells directly in the battlefield (see: Military wants 'blood pharming' machine). Here's a link to a PDF/Power Point lecture from a blood pharming workshop with more details. Below is an excerpt from the article:

The U.S. military is seeking an automated culture and packaging system that could produce a steady supply of universal donor red blood cells right on the battlefield, without resorting to needles and the human filling-stations. DARPA has awarded a $1.95 million contract to Arteriocyte, a Cleveland company that's experimenting with a technology developed at Johns Hopkins that enables the rapid expansion of umbilical cord blood. The company wants to adapt it to a manufacturing technology that will feed the military's thirst for universal donor red blood units. The technology, called Nanex, uses a nanofiber-based structure that mimics bone marrow in which blood cells multiply, according to the company. The military envisions a "fieldable" in-theater, culture-manufacturing system that would take hematopoietic progenitor cells and automatically covert those into hundreds of prepackaged, ready-to-be-infused RBC units. This process, called "blood pharming," would eliminate much of storage, transport and the donor blood type and health limitations that make RBC resupply such a challenge.

I have been observing the mostly unsuccessful quest for a ready blood substitute for more than three decades and have blogged about this topic before (see: Setback for Blood Substitute at Northfield Labs). Most of this research has been funded by the military. It appears that the military, or more specifically DARPA, is starting to invest in some new technology along these lines. You can at least credit someone in the agency or at Arteriocyte with a sense of humor for coining the term "blood pharming."

The DARPA workshop presentation linked to above is replete with military-speak with references, for example, to CONUS. For the uninitiated, CONUS refers to the "continental U.S." The lecture really takes be back several decades to Army briefings by high-ranking officers. No PowerPoint presentations in those days -- only cardboard posters that got flipped on the floor from a tripod. Messy but very dramatic. However, nothing beats the name being used by DARPA for the device in which the hematopoietic progenitor cells are incubated and used to spawn RBCs to be transfused to the injured in the battle field. It's called an "in-theater bioreactor." You gotta love these guys.

microRNAs Show Promise as Biomarkers for Cancer Prognosis and Outcomes

A recent article that came across my desk emphasized the value of microRNAs as biomarkers for the early detection of cancer (see: A New Biomarker For Early Cancer Detection? Research Reveals That ‘Microrna’ May Fit The Bill). Another recent article addresses this same topic in a broader manner (see: The Emergent Role of MicroRNAs in Molecular Diagnostics of Cancer). As I delved further into this topic, I began to appreciate the importance of this new area of research in terms of the emergence of a totally new class of cancer biomarkers. Below is a quote from the first article:

Scientists at Fred Hutchinson Cancer Research Center have discovered that microRNAs molecular workhorses that regulate gene expression are released by cancer cells and circulate in the blood, which gives them the potential to become a new class of biomarkers to detect cancer at its earliest stages....MicroRNAs, which act as brakes on different parts of a cell, keeping genes in check, have some advantages over protein-based early-detection systems, including that they can be detected potentially in smaller quantities and that the technology exists to rapidly develop microRNA-based early-detection tests....The next steps, now that a proof of principle has been established, are to identify specific microRNAs that can signal the presence of a variety of solid-tumor cancers at an early stage, and to further develop the technology to detect the microRNAs in minute quantities.

Another recent article (see: Study yields clues to why some tumors spread) provides more concrete evidence of the value of a specific micro-RNA as a biomarker for the clinical behavior of cancer of the prostate. Below is an excerpt from it:

A small fragment of genetic material may mean the difference between an easily treated local tumor and an aggressive cancer that spreads throughout the body, U.S. researchers said on Thursday. They found that when a bit of ribonucleic acid or RNA known as microRNA-101 goes missing, a protein called EZH2 starts to proliferate. EZH2 has been linked with aggressive forms of breast, prostate, skin and bladder cancer, but until now it has not been clear what triggers overproduction of EZH2. "What this study shows is why that protein is elevated in metastatic cancers," said Dr. Arul Chinnaiyan of the University of Michigan, whose study appears in the journal Science. Chinnaiyan said EZH2 causes cancer to spread by shutting down other genes that keep cancer in check. And he now thinks microRNA-101 or miR-101 is what keeps EZH2 from overproducing in cells. With this RNA bit missing, EZH2 causes cancer to spread.

I have published previous notes about the use of genomic testing of malignant cells and their protein secretions to assess the prognosis and the clinical outcome of a malignant tumor (see: Use of Biomarkers in Assessing the Prognosis of Gastric Carcinoma, Early Treatment of CLL Based on Predictive Biomarkers Proven Effective). Yet another article (see: One Man's Junk DNA Is Another Man's Treasure: MicroRNAs and the Next Big Thing in Cancer Prognosis) speculates about the use of microRNA "profiles" of malignant cells in a similar way:

A new paper... showed that microRNA profiles could be used to predict both prognosis and clinical outcome, another first in the microRNA world. In addition, [the authors] found that the levels of a specific microRNA called miR-21 also correlated with cancer stage: The later the stage, the more miR-21 they found in the tumor cells.

I, for one, am just beginning to understand the value of microRNAs as a new class of biomarkers. I will pay attention to research in this area as it develops and post additional notes about the topic when possible. I also need to revisit previous notes describing and defining biomarker to make sure that the term is broad enough to encompass microRNAs.

Performance vs. Utility of Clinical Lab Tests: A Marketing Perspective

I have always held the opinion that the laboratory diagnostics industry, the clinical labs, and lab physicians do not market themselves and their services adequately to test-ordering physicians and to the general public. By way of contrast, I think that radiologists do an admirable marketing job for medical imaging. I am not sure if this latter group is more marketing-oriented or if the new imaging technologies just sell themselves. As evidence for my statement, I suspect that your next door neighbors can discuss their CT scans and MRI's but can't name a lab test beyond cholesterol. A recent blog note by Lena Chow (see: Performance vs. Utility) crystallized this issue in my mind. She is is chairman and chief executive officer of JYT Health Media in Palo Alto and Shanghai. Below is an excerpt from her blog note (boldface emphasis mine):

In the increasingly challenging environment of healthcare marketing, diagnostic marketers need to take a page from the pharmaceutical industry...for insights into how they work to understand the way doctors make decisions and adopt new technology.....I have long been intrigued by the low value assigned to diagnostics by the healthcare system, and I have been looking for tangible, actionable ideas I can give to my clients in the diagnostics space....Most diagnostic marketers I know can produce a respectably thick binder of published studies demonstrating the analytical performance of their products, including studies comparing their products with traditional methods and showcasing superior results. Then, months to years after launch, the same marketers are surprised to learn that doctors continue to use these other methods even though they take longer and provide incomplete and/or inaccurate results, and even though reimbursement is available for the new and improved test. So, rule number one: Don’t equate performance with utility. Proving that the performance of a new diagnostic test has been demonstrated through properly designed, robust trials, and getting FDA clearance must be paralleled by an equally arduous effort to work with clinicians in understanding and demonstrating clinical utility—what is the relevance of the test results in the clinical setting, how will it influence patient care...and how all of this will improve outcome. Working with clinicians, conducting the studies, publishing the findings and communicating them to the target market are just a few of the necessary steps toward adoption.

So, the advice that I carry away from her note is that proving the analytical performance of a test to satisfy the FDA using measures such as sensitivity and specificity is only half the battle. Equally important is proving the utility of a test for the clinicians working in a hospital. Here are the questions that needs to be answered for them: What is the relevance of the results from a newly offered lab test in a clinical setting and how will the results influence patient care? How will the new test enable clinicians to work more efficiently?

Having said this, the next logical question is what are the most effective channels to communicate with physicians practicing in a hospital and in their offices. One can obviously turn first to publications such as hospital physician bulletin, either hardcopy or electronic, in which the clinical utility of a new test can be presented as described above. However, one additional approach occurs to me. Every medical staff has physician opinion leaders who are respected for their professionalism and high standards. A personal call to such individuals by a pathologist when offering a new test could go a long way in promoting its use by the medical staff.

LIS vs. LIMS: Some New Insights

In two previous posts, I began to explore both the similarities and differences between LISs and LIMs (see: LIS vs. LIMS: It's Time to Blend the Two Types of Lab Information Systems, LIMS and Preclinical Workflow: Another Dimension of the LIS/LIMS Discussion). Larry Wimberly of Wimberly Consulting Services, who has a very long and distinguished history in the LIS industry, sent me an email, commenting on the differences between the two types of systems. His ideas provide some important insights and are listed below:

• One of the major differences between LIMS and LIS systems is what the LIMS vendors refer to as workflows and are analogous to the protocols used in anatomic pathology. In the case of clinical trials, the lab work is performed for, and paid by, the sponsor of the trial, usually a pharmaceutical company as opposed to a physician/patient in the case of an LIS.

• Each study managed by an LIMS requires that a research protocol be established stipulating, for example, the number of times that a research subject will be tested and the type of tests to be performed each time. These details may differ for each testing cycle. For instance, the initial cycle of lab testing may contain a number of tests that will be used for baseline comparisons against future specimens from the same subject during the trial. Each succeeding test cycle in the protocol usually has to be performed within a specific time window and may include a subset of these initial tests as well as new tests. Results are usually reported to both the sponsor and the primary investigator in the case of a research study or clinical trial.

• Another part of the reporting requirements managed by an LIMS is special flagging of results outside of the reference ranges, which is actually a change in a result compared to its baseline value. These reference ranges are usually established in the first testing cycle and/or defined by the study sponsor. This approach to reference ranges is very different from that used by the typical hospital clinical lab and reported by LISs.

• When a research specimen arrives in the lab, it first needs to be identified as to its study, the test cycle, and the research subject. Important for this process is that each test cycle requires a specific "collection kit" including items such as specimen containers, blood drawing instructions, and the test order form. These kits also include mailing labels for their initial distribution and pre-accessioned specimen labels to ensure that the specimen collection process is made more efficient. This is one of the reasons that CRO labs are so interested in having pre-accessioned specimens. In fact, their preference is to have the LIMS, on a timely basis, generate a pick-list so that these collection kits can be assembled. 

• In the case of public health labs which tend to favor LIMSs over LISs, most of their focus in the past has been on environmental testing for analytes such as heavy metals and other contaminants. These types of studies lend themselves more to a LIMS than a traditional LIS. I’m not aware of any LIS system that can handle these types of testing/reporting protocols very efficiently.


• Most LIMSs are not really patient centric and can't compete very well in the traditional clinical lab environment. However, large-scale diabetes testing programs would be improved with a LIMS-based approach. For these types of patient. a testing protocol could be developed based on the disease stage of each of them. For example, some patients would be tested annually and others quarterly using different test profiles each quarter. The test results would be sent to both the physician and patient and also copied to the relevant EMR/PHR. In this way, a complete record for each patients would be compiled and accessible. Also reminders could be sent out on a timely basis with follow-ups for patients who don’t respond.


• I believe that with the advent of earlier detection of disease, more rigorous patient follow-ups, and greater attention to health maintenance, there would be significant benefits from the merger of the features/functions of the LIMS and LIS. Of course the corollary to this is that clinical protocols such as the one described above for patients with diabetes will not use the LIS as their primary locus but rather the direct patient care systems. Regardless of where the actual locus of control and responsibility for these protocols lies, I believe that LISs will need to adopt some of these new functionalities in order to support these emerging clinical surveillance programs.

Genomic Testing, Pre-Symptomatic Disease, and Health Insurance

Ralph Moher who is the VP of Marketing for General Data and who recently launched a blog emails me with the following question regarding the Early Health Model and its emphasis on the diagnosis of pre-symptomatic disease:

What is the possibility that insurance companies in the future will require the use of biomarker panels to identify pre-symptomatic disease to determine eligibility and coverage?

In a diagram contained in a previous note (see: Wellness, Preventive Medicine, and the Classic Disease Model), I placed the Early Health Model, the detection of pre-clinical, pre-symptomatic disease, within what I referred to as the Wellness Domain. Also included in this space was Genomic Medicine and the predisposition to disease. This set of patients was compared to patients with acute and chronic diseases, whom I placed in the Overt Disease Domain.

I was thus commenting on the fact that genomic medicine is identifying people with a predisposition to disease and the Early Health Model is identifying people with early forms of disease as manifested by biomarker abnormalities. For the present, such "healthcare consumers" (not patients) are best categorized as being "well" because we are uncertain how many of them will go on to develop overt symptomatic disease. Also implicit in this statement is that we also don't know which of these patients with early forms of disease and a predisposition to disease to treat.

Historically, some consumers have been reluctant to undergo genomic testing for fear that, if the results were  positive (i.e., show a genetic abnormality), they would be ineligible for health insurance coverage. One way to avoid this stigma, but also to be informed about the results of the test, would be to undergo testing in a clinical lab with a guarantee that the results would not be accessible by anyone other than the individual being tested. The same course of action, of course, would be available to anyone seeking biomarker test panels designed to detect early disease.

Ralph, however, is asking a different question: Do I foresee a time when insurance companies would require biomarker panels in order to cull out those people with early forms of disease ("identify pre-symptomatic disease to determine eligibility and coverage")? He is asking what I would call a socio-economic question rather than a medical one. We know, of course, that health insurance companies can and will refuse insurance to applicants with pre-existing medical conditions. I am opposed to the idea of health insurance companies cherry-picking their customers by denying coverage to applicants with a genetic pre-disposition to a disease or a pre-disease identified by biomarker abnormalities. This does not mean that they won't attempt to pursue such a strategy.

Ambiguity Surrounding the Term Molecular Imaging

It recently occurred to me, after posting a number of blog notes about molecular imaging, that the term was too ambiguous to continue using without some additional clarification. On its face, it means to create images that can be used to identify and locate various molecules in tissue. This is a goal that is being actively pursued in radiology departments and in radiologic research in relation to radiology imaging techniques such as PET/CT.

By way of contrast, I have also used the same term in Lab Soft News to refer to the identification of molecules in surgically removed tissue with immunohistochemistry (IHC), for example, in surgical pathology and other clinical settings. I also posted a note (see: Charles River Acquires Molecular Imaging Company) about this company, a contract research organization (CRO), acquiring a molecular imaging company for support of its pre-clinical research activities. 

It turns out the the best way to disambiguate the term molecular imaging, as practiced in radiology, from the its use in the clinical labs and in basic research was addressed in two other notes that I had posted (see: Are We Ready for the Diagnostic Hybridist?, Running the IV2D Experiment: Bottom-Up vs.Top-Down). I was searching for a term to differentiate the in-vivo diagnostics practiced by radiologists from the in-vitro diagnostics practiced by pathologists. I hit on the term IV2D to encompass in-vivo and in-vitro work.

By using the same general naming convention, the imaging procedures performed by radiologists can be referred to as in-vivo molecular imaging and those by the pathologist or basic researcher on surgically resected tissue as in-vitro molecular imaging. As a simpler alternative, the former can be called radiologic molecular imaging and the latter tissue molecular imaging.