The Daily Sign-Out

The new paternalism in American medicine

It has taken me awhile to get to this article by Hartzband and Groopman, "Rise of the Medical Expertocracy," from the March 31, 2011 Wall Street Journal.  If you happened to miss it, it is quite thought-provoking.

Whether you are coming from the left or right perspective on health care reform, there is an underlying singular assumption that there is a "best practices" or "evidence-based solution for every difficult problem in medicine and health care delivery.  This may be naive wishful thinking or a zealous belief in the power of science, but, either way, this assumption is pervasive amongst physicians, "experts" in various medical disciplines, as well as politicians and policymakers.  I aver that this is a pernicious and dangerous assumption for both medical practice and health care policy.

One interesting part of this article is the authors' discussion of the results of their study interviewing how patients process and act upon "expert" advice and information.  Perhaps not surprisingly, there is indeed a wide range of how patients make medical decisions--which is also reflected in the range in how members of expert "consensus" panels also view the same evidence and arrive at different, indeed, opposite conclusions.

The authors' observation of the rise of the industry supplying metrics and report cards to doctors, health plans and hospitals is particularly insightful--and disturbing.  How much of the health care dollar goes to companies measuring and benchmarking health care hospitals' own data?  I know my hospital spends a lot--as well as spending on how to remediate or improve to get where the experts say we need to be in the 90th percentile.  To me, this seems to stifle the trial-and-error, locally-based approach that might truly provide innovation that could be appropriated or adopted elsewhere.

I will confess an inherent skepticism and distrust of experts (especially those panels that are buddy-appointed), but the authors make what should be obvious points:

For patients and experts alike, there is a subjective core to every medical decision. The truth is that, despite many advances, much of medicine still exists in a gray zone where there is not one right answer. . .Patients and doctors can differ with experts and not be ignorant or irrational. Policy makers need to abandon the idea that experts know what is best. In medical care, the "right" clinical decisions turn out to be those that are based on a patient's goals and values.

Elementary? My dear Watson!

The Wall Street Journal (subscription required for full article) reported today on an agreement between IBM and WellPoint for the latter to use IBM's Watson technology for clinical decision support.

WellPoint Inc. and International Business Machines Corp. are set to announce a deal on Monday for the health insurer to use the Watson technology, the first time the high-profile project will result in a commercial application.

WellPoint said it plans to use Watson's data-crunching to help suggest treatment options and diagnoses to doctors. It is part of a far broader push in the health industry to incorporate computerized guidance into care, as doctors and hospitals adopt electronic medical records and other digital tools that can record, track and check their work.

For IBM, the agreement with WellPoint could provide some real-world ballast for Watson, which IBM boasts can process about 200 million pages of content in less than three seconds. Watson is part of the company's broader effort to build a large business in the competitive field of business analytics, which uses software to mine huge volumes of data to aid decision-making.

This will be an interesting development to follow.  It certainly makes sense to leverage electronic medical records by incorporating data mining systems and clinical support tools to improve patient care.  Yet one thing it does not address is the persistent problems with communication between providers and lack of coordinated care that plague both inpatient and outpatient care.  It's curious that WellPoint plans to first deploy the Watson technology "with nurses who manage complex patient cases and review treatment requests from medical providers" (my emphasis).  I admit I'd be skeptical about the underlying reason (i.e., cost) for a company implementing such a system when framed like this.  I guess I'm unaware that nurses actually manage patients and review doctors treatment "requests."  If that's the case, why do doctors need Watson?  "Researchers have been trying since the 1970s to develop computers that can advise doctors, but the efforts haven't gotten much traction."  Is this suppose to be ironic?  If the joke has to be explained, doesn't it spoil the laugh?

Systems biology approach shows MYB and FOXM1 as master regulators of germinal center proliferation

Back from a relaxing stay-cation as well as ASCO meeting--

The June 2010 issue of Molecular Systems Biology has a fascinating study identifying MYB and FOXM1 as "master regulators" of germinal center proliferation in lymph nodes.  Why should we as pathologists care about this?

What piqued my curiosity here was a connection with my previous post summarizing what I learned at USCAP on the molecular pathology of diffuse large B-cell lymphoma (DLBCL) regarding dysregulated genes in DLBCL (including MYC) and identification of three gene expression subgroups, including the "germinal center B-cell-like" group.

The authors use some fascinating and powerful techniques around a systems biology approach to find master regulator genes controlling specific cellular processes (in this case, germinal center B-cell proliferation) as well as transcriptional regulation of protein complexes whose regulation whose availability must be regulated in a context-dependent manner.

The authors developed is an interaction network specific to human B cells (called the human B-cell interactome, HBCI) by reverse-engineering of transcriptional and post-translational interactions in mature human B cells from a diverse collection of 254 B-cell gene expression profiles derived from normal and malignant mature B cells (Lefebvre et al, 2007).  They further integrated evidence from experimental assays, databases, and literature data mining, filtered by context-specific criteria, using an established evidence integration algorithm.  Then, they used a novel algorithm, the Master Regulator Interference Analysis or MARINa, to interrogate the HBCI to discover MRs of key genetic programs in the germinal center (GC) reaction of antigen-mediated immune response, that is, normal progression through the GC.

They identified MYB and FOXM1 as "master regulators" of GC B cells since 80% of the genes jointly regulated by these transcription factors are activated in the GC.  They also showed that MYB is a transcriptional activator of FOXM1 suggesting that they form a "feed-forward" loop involved in a synergistic activation of a large subset of GC-specific genes.  I don't quite understand their use of the term "feed-forward" loop here--my understanding is that a feed-forward loop is when the anticipated future effect (which hasn't yet happened) triggers the cause in the present (which would otherwise have not have happened).  An example is when there is a rumored or predicted shortage of a commodity, which leads to hoarding in anticipation of the shortage, which in turn causes a shortage to occur.  It seems that this should be called a reinforcing feedback loop but whatever.

But they also validated that common MYB/FOXM1 targets were affected by silencing either MYB or FOXM1 using gene expression profiling.  Specifically, they showed that downstream targets and other MRs are bound by MYB and FOXM1 in their promotor regions after MYB or FOXM1 silencing.  Moreover, silencing of MYB and FOXM1 showed a decreased in proliferating cells in GCs and an increase in apoptotic cells.  They further examined specific targets involved in the formation of a protein complex predicted by their model and found that about half of the MYB/FOXM1 targets cluster within a complex including novel interactions with pre-replication and mitotic proteins.  They found that two mitotic kinases inferred in the complex, BUBR1 and AURKA, physically interact with MCM3 which are all direct targets of MYB and FOXM1.

My summary doesn't do justice to this paper (due to my own shortcomings in understanding) but if you're interested in lymphomas and/or systems biology, this is well-worth a read.

ER/PR testing guidelines: Joint CAP/ASCO recommendations

The College of American Pathologists (CAP) and the American Society of Clinical Oncology (ASCO) issue a joint guideline on April 19, 2010 in an effort to improve the accuracy of immunohistochemistry testing for estrogen (ER) and progesterone receptors (PgR) expression in breast cancer. This is similar to the joint effort the groups made with HER2 testing guideline in 2007.  The guideline is being published in the April 19 issues of ASCO’s Journal of Clinical Oncology (JCO) and the CAP’s Archives of Pathology & Laboratory Medicine in June 2010 but is available in electronic form here.  The ASCO Web page on the guidelines (http://www.asco.org/guidelines/erpr) has links to the unabridged version (yikes!) as well as downloadable slides.

The major recommendation are summarized from the CAP's press release (quoted):

• Testing ER and PgR status on all newly diagnosed invasive breast cancers (primary site and/or metastatic site), and whenever appropriate, repeat testing in patients with a known breast cancer diagnosis who now present with a local or distant recurrence.
• Establishing uniform testing measures that focus on proven, reliable and reproducible assays and procedures.
• Having testing laboratories validate their assays against existing and clinically validated tests. Results should agree at least 90 percent of the time with those of the clinically validated assays for positive receptor status and at least 95 percent for negative receptor status.
• Transporting breast tissue specimens from the operating room to the pathology laboratory as soon as they are available for gross assessment. The time from tumor removal to initiation of fixation should be kept to one hour or less. Fixation of the sample in neutral buffered formalin must extend for at least six hours and no longer than 72 hours.
• Performing ER and PgR testing in a CAP-accredited laboratory or in a laboratory that meets the accreditation requirements spelled out in the guideline. The CAP will require that every accredited lab performing testing participate in a mandatory proficiency testing program.
• Considering an ER and PgR test performed by an IHC assay as positive if at least one percent of the tumor in the sample tests positive, which helps predict whether a patient is likely to benefit with endocrine treatment. The panel recognized that it is reasonable for oncologists to discuss the pros and cons of endocrine therapy with patients whose tumors contain low levels of ER by IHC (one percent to ten percent weakly positive cells) and to make an informed decision based on available information.

As with the ASCO/CAP HER2 guidelines, the ER/PR guideline is an excellent state-of-the-art reference.  I welcome the direction it provides for reporting, specifically, for defining a threshold for determining a positive status and, in general, for thoroughly discussing the role of pre-analytic factors in receptor preservation.

Fortunately, my lab has already been reporting most of the elements in the guideline, such as which clone we are using, fixation time, percent of cells positive, staining intensity and Allred score, etc.  BUT 

One troubling recommendation, however, is the mandate that cold ischemia time be documented: "The time of tissue collection (defined as the time that the tissue is handed from the surgical field) and the time the tissue is placed in fixative both must (my emphasis) be recorded on the tissue specimen requisition to document the time to fixation of the specimen."  While I do completely agree with the idea that "cold ischemia time" should be minimized and we need to educate surgeons, radiologists, and OR staff regarding this, to state this categorically in a recommendation with an explicit instruction on how to do it is a mistake.  This creates an unnecessary burden on the pathology department to police the surgeon.  Are we then to reject a specimen as unacceptable for testing whenever this time is not documented?  What if it is documented somewhere else in the medical record (like we document the time the specimen is placed in fixative in our gross description)?  The over-weighted academic composition of the guideline panel really comes through with stuff like this.  It is really a shame that more community pathologist input is not sought for these guideline panels since most breast cancer is diagnosed at the community level.

Another annoying thing is the list of "well-validated" clones for ER and PR listed in Table 3.  The implication is that if your lab is not using one of these clones, then your lab is not in compliance with the guidelines.  Really?  I thought this was the responsibility of the laboratory medical director to validate a new antibody with a clinically validated assay.  For example, I could not approve the validation of the 1A6 clone for PR in my lab using cell lines controls with low, medium and high defined receptor content because of low sensitivity but was able to validate a different clone (which itself was compared with the 1A6 clone by the manufacturer).  Given the authors' disclosures of potential conflicts of interest (yes, I do actually read this stuff!  not that I'm paranoid or anything), I think the panel might have avoided implying in any way that labs must be using these specific clones for testing.

In summary, the guidelines are very welcome indeed and provide support for what many of us are already doing and may also encourage those labs that are unwilling or unable to comply to send-out these tests to another lab.

ASCO recommendations for genetic testing

ASCO Releases Updated Policy Statement on Genetic Testing

ASCO issued its latest recommendations for genetic testing for cancer susceptibility. The updated statement addresses new developments during the past seven years, including the availability of genetic tests of unproven clinical benefit and direct-to-consumer (DTC) genetic testing. Read the press release, visit Cancer.Net for information for your patients, and read the statement in JCO.

The comments relating to DTC genetic testing are worth reading this alone.  Marketing directly to consumers places us pathologists even more out of the loop given how in recent years, sales reps for test manufacturers have bypassed the lab and marketed directly to clinicians (e.g. ThinPrep, SurePath, OncotypeDx,).  Just a suggestion, folks, but involving us up front could actually help you and your company--but (speaking for myself) not even giving me a courtesy call makes me wants to go out of my way to torpedo your claims.

More recommendations for lymph node adequacy in CRC

The Dutch National Working Group on Gastrointestinal Cancers has recently released colorectal cancer guidelines which includes a recommendation of a minimum of 10 lymph nodes to establish a negative lymph node status.  At the same time, the group states that "determining the lymph node status of a patient requires evaluation as many lymph nodes as possible using conventional techniques." (emphasis mine

The discussion of this recommendation states that "no definitive criteria were found in the literature regarding the minimum number of examined lymph nodes.  There is no evidence to support 12 lymph nodes, as recommended by TNM." (emphasis mine)

Furthermore, they specifically do not recommend using immunohistochemistry to detect metastases or pre-treatment of mesenteric fat with GEWF-type solutions.  While using the former to detect micrometastases is not common practice, as least in the U.S., the latter is not only fairly common, if not routine in many places (mine included!).  Also, the recent literature also supports the use of these solutions to increase lymph node retrieval, although the utility of the raw node count has been questioned with the data regarding the percentage or proportion of positive:negative nodes (whole other discussion!).

NIH State-of-the-Science Statement on Breast DCIS

The conference statement from the recent NIH "State-of-the-Science" conference on ductal carcinoma-in situ (held September 22-24, 2009) is available online.  You can download a pdf for free.

Some interesting findings:

1. While the incidence of "noncomedo" DCIS continues to increase, the incidence of the (to me, more biologically worrisome) comedo type peaked in 1995 and has been declining (through 2006).
2. MRI in DCIS?  Most studies show that MRI is more sensitive in detecting multicentric disease but is not specific.  MRI also reportedly over- and under-estimates the extent of DCIS compared to pathological exam.
3. Although the absolute numbers are small (due to the overall high survival rates in DCIS), there are higher breast cancer recurrence and mortality rates in Black women diagnosed with DCIS compared with White women.
4. No molecular markers have been identified to stratify prognostic groups.  High grade, comedo pattern, margin involvement and size have been related to increased risk of recurrent DCIS and progression to invasive carcinoma--but hard numbers are hard to come by.
5. DCIS is a heterogenous disease with a high probability of long-term disease-free survival--whatever therapy is received.

Critical appraisal of research articles

As clinicians, we must be able to screen and then critically appraise research articles that will help us practice evidence-based medicine and, hopefully, take better care of our patients. More importantly, journal editors have a duty to do be diligent in adequately analyzing the papers submitted to them for review. This burden for editors is even more critical for us given the conflict between the limited time available for reading journals and integrating relevant articles for use in one's daily practice, the exploding volume of articles and journals and other new media (like blogs!), and our insufficient tools and skills for properly evaluating articles.  Young and Solomon, in the February 2009 Nature Clinical Practice Gastroenterology & Hepatology, have published an article called, "How to Critically Appraise an Article," which provides a "user-friendly" format that a busy clinician could adapt for routine use.

I like the David Letterman-style "top 10" questions (in bold) that one can use to quickly assess whether an article is worthy to more critically read.  The comments following the questions are mine.

1. Is the study relevant?  To whom?  To me!  When I first receive a journal, I quickly scan the TOC for any article that is immediately relevant either experience with a recent problematic case, something that will help me teach, or something relevant to a research problem.
2. Does the study add anything new?  This one, for me, is more difficult to immediately answer.  But, I think if one regularly reads the standard pathology (or substitute your own specialty) journals, one can identify "hot" publishing topics.  For example, a couple of years ago, it seemed that papers on columnar cell lesions of the breast were being published every month.  Other recent examples are papers on serrated polyps and eosinophilic esophagitis.  The authors do a nice job of pointing out the elephant in the room--that seminal "new" ground-breaking research papers are relatively rare--they are like home runs.  However, papers that validate previous research by trying to replicate others' findings or generalize or refine previous results would be more useful to those in practice.  My experience is that people who review grant proposals are "less than enthusiastic" about funding these types of studies.
3. What type of research question does the study pose?  Most relevant to pathologists are questions concerning "frequency of events," that is, the association between a diagnosis, or observation (positive or negative for such-and-such stain, etc.) and a disease outcome or treatment response.
4. Was the study design appropriate for the research question?  Fortunately, pathologists do not have to deal with randomized controlled trials, like oncologists, as most pathology literature relies on observations studies.  I find it very difficult to adequately evaluate RCTs with respect to oncology studies but the academic oncologists are generally brilliant at this--at least the folks I listen to on my Research-to-Practice mp3s.
5. Did the study methods address the key potential sources of bias?  Again, this is critical in RCT design but still, how often do you just blow through the "Materials and Methods" section?  The proper use of controls is crucial in interpreting pathology and lab studies.
6. Was the study performed according to the original protocol?  What patients or specimens were NOT included in the final set for analysis?  Important source of bias.
7. Does the study test a stated hypothesis?  I wonder sometimes if the hypothesis is written after the study is done.  Like if I'm fishing for crappie using minnows and I happen to have caught a catfish, then the a posteriori hypothesis of my fishing study is that the best bait for catching catfish is minnows.
8. Were the statistical analyses performed correctly?  Do you actually read that part of the "Methods"?  Really?  At least, one should check to see that the authors state the tools that they used and their rationale for this approach.  The next level would require a more in-depth understanding of these tools and their appropriate use.
9. Do the data justify the conclusions?  Yeah, remember that hypothesis thing?  Beware of the "statistically significant" findings that demonstrates a difference that is too small, in reality, to be of much clinical value.  On the other hand, look out for small sample sizes that may underestimate real differences between groups and not be powered enough to show any statistically significant differences.
10. Are there any conflicts of interest?  Given recent disclosures in the "mainstream media," this really cannot be ignored.  An example highlighted at one of the recent USCAP meetings by Dr. Noel Weidner regards the OncotypeDx assay in breast cancer.  Who is really sponsoring and/or writing these articles?  Good question to ask but maybe that's my own distrust of "city slickers."

Anyway, good article and well-worth reading.