[Histonet] New CAP question ANP.22760
Thomas Jasper
tjasper <@t> copc.net
Sat Jun 26 15:45:19 CDT 2010
Daniel,
I think your observations are very good. Which brings me again to a
point I was trying to make earlier. Anatomic Pathology is not the same
as the General Clinical Lab or it's close counterparts. It almost seems
like this is overblown re: validation. As you've pointed out, the
interpretation is up to the pathologist. Patients (prostate, i.e.) will
correlate with other tests. When known positives demonstrate properly
you're valid. Even automated systems can use algorithms to score
strength of positivity...which varies yet will show as positive (valid).
You receive a new batch of Ab you run it on a known positive, that's the
bottom line.
It seems somehow CAP or clinical lab (trained) folks want more than
that. And again (as you pointed out) the appreciation for the
differences in how we operate as laboratorians seems to fall by the
wayside. As you mention pushing 20, 50, 100 is not feasible (and
frankly unnecessary). I don't know where the answer lies in making
everyone happy here. I do know that some thinking outside the box and
trying understand AP is required. We all want good patient care, but
one size does not fit all.
Thanks,
tj
-----Original Message-----
From: histonet-bounces <@t> lists.utsouthwestern.edu
[mailto:histonet-bounces <@t> lists.utsouthwestern.edu] On Behalf Of Daniel
Sent: Saturday, June 26, 2010 11:08 AM
To: histonet <@t> lists.utsouthwestern.edu
Subject: RE: [Histonet] New CAP question ANP.22760
One question I think a lot of you are not considering is that the
clinical laboratory usually tests for analytes present in most patients.
This allows the clinical lab to more easily run validations with
hundreds of patients, using statistical tools to analyze the precision,
specificity and sensitivity of the test. What most of those concerned
with this new CAP standard (and I count among them) is that our testing
often targets a very rare population of patients. As such, an extensive
validation is much more difficult to establish. It then makes the
statistical models used in clinical tests much less valuble since the
accuracy of these formulas decrease with smaller sample sizes.
Additionally, the pathology of the clinical tests are reported as a
measurement which does not have an intrinsic value. It must be
interpreted by a clinician to give value to the patient. For example,
the Blood Urea Nitrogen test value range usually is between 7-20 mg/dL
while values above 50 are considered a marker of poor kidney function.
Elevated values, however, do not mean that kidney functions are impaired
but instead may be due to other pathologies such as congestive heart
failure, gi bleeding, certain steroids and even diet.
For histopathology even relatively common antibodies such as pan-T cell
marker CD3 usually only stain 75% of T cell neoplasms. A negative
result for that minority does not indicate that the tumor does not have
a T cell lineage. It relies instead on the pathologist interpreting the
result much as the clinician looks at the BUN value in relation to other
factors in the patient's presentation. When we validate this kind of
antibody, do we perform testing then on T cells and B cells (give me a
good tonsil), on malignant samples or some combination of the two? What
is a reasonable number of tests to run then?
If I choose 10 normal lymphatic tissue blocks for my routine T/B cell
marking (it can't be as extensive as ER/PR can it?) how many samples
should I choose for my malignant population? If I use another 10 anyone
with a background in statistics will tell you that the sample size is
too small to interpret accuracy. If I push it up to 20, 50 or 100 I
would be certain that many laboratories would not be able to afford the
cost of the validation. This would then push many good laboratories out
of the business of IHC with the unintended result of delaying diagnoses
and increasing patient costs by driving testing to fewer and fewer
testing outlets.
CAP's new path with ER/PR seems to be trying to achieve a noble end of
improving quality in the laboratory but without an understanding of the
complexities and consequences of the method they are implementing.
Dan
-----Original Message-----
>From: Jesus Ellin <JEllin <@t> yumaregional.org>
>Sent: Jun 26, 2010 6:28 AM
>To: "Morken, Tim" <Timothy.Morken <@t> ucsfmedctr.org>,
>histonet <@t> lists.utsouthwestern.edu
>Subject: RE: [Histonet] New CAP question ANP.22760
>
>I have been reading the post to this question and it seems to me that
there are different standards depending on the lab that is operating the
methodology. I do agree that the core lab for years have had the
instruction and training in the performance of validation. One thing
that comes to mind as well is why has histology not had this training?
Why are we not getting this from our certification agency, our
professional societies and biggest reason where is our standardization.
It seems to me that with all these regualtions in plac for so long, why
were we missed. Is it because when inspected through CAP we are being
inspected by a pathologist rather than a histo tech? These are some of
the questions at hand. I to see new standards within the CAP checklist
as well as other regulatory organizations that will affect the future of
the Anatomic Pathology community. But I think we need is to provide a
underlying architecture for our peers, so that we can begin the
transition to the future. This is only the beginning, there is still
Digital Image Analysis and Telepathology. It funny we are looking to
become a hybrid of radiology and the core lab, but with the best of both
worlds. Tim great structure for the validation study.
>
>
>-----Original Message-----
>From: histonet-bounces <@t> lists.utsouthwestern.edu on behalf of Morken,
>Tim
>Sent: Wed 6/23/2010 9:48 AM
>To: histonet <@t> lists.utsouthwestern.edu
>Subject: RE: [Histonet] New CAP question ANP.22760
>
>Joe,
>
>You wrote : The folks in the 'clinical' lab have been performing more
comprehensive and complex validation procedures for a very long time
..."
>
>Those were my thoughts exactly. While the person replying may or may
not have specific histology experience she will have clinical lab
experience (however, my guess is that she is exposed to histology
regularly at CAP). Clinical labs have a bit of an easier time, actually,
because they validate primarily to known concentration controls -
analytical controls manufactured at a range of known concentrations for
instance. The institution then adds in their normal controls for
validation.
>
>As far as the current question about validating a new lot of reagent
the best practice is to run parallel tests on the same machine. If that
is not easily possible on a particular manufacturer's instrument then
the question should be asked of them: Why not? If this is a requirement
the manufacturer should provide an easy path to meeting the requirement.
However, if that is not the case then the institution simply writes a
procedure to get around the inadequacies of the instrument (Maybe the
vendor can help with that). Then follow the procedure. That should
satisfy inspectors.
>
>An "...appropriate panel of tissues..." is whatever the institution
deems appropriate for the given antibody or reagent. This is a perfect
place for tissue arrays. You can make your own or buy them.
>
>IHC must meet CLIA validation guidelines but since IHC is generally
qualitative the requirements must be understood and methods adapted to a
qualitative scenario. Several IHC and Histotechnology books discuss the
subject at length (Taylor, Dabbs, Bancroft for instance).
>
>Below is a brief overview of how to do that. (for more in-depth info
this was covered in an NSH teleconference I gave last year - PowerPoint,
audio and references available from NSH-, and will be covered in a
similar workshop at NSH in Seattle this year).
>
>
>1)
>CAP General Validation
>CAP GEN.42020-42163 Test Method Validation Follows CLIA CFR Sec
>493.1253 Does not apply well to IHC (IHC is usually qualitative)
>
>But the general principle applies:
>The laboratory must have data on each test's accuracy, precision,
analytic sensitivity, interferences and reportable range.
>
>Unmodified FDA-cleared or approved tests: the lab may use manufacturer
information or published reports but lab must verify outside data.
>
>Non-FDA cleared: Lab MUST verify or establish analytic accuracy,
precision, sensitivity, specificity and reportable range.
>
>2) Validation includes:
>Accuracy:
> Compare results with New antibody to a previously validated
antibody on the same tissues
>
>Precision:
> Test samples with varying antigen expression
> Intra-run, Inter-run tests, 10 slides each (reproducibility)
>
>Sensitivity:
> True Positive vs False Negative (higher % FN = less sensitive)
>
>Interferences [Specificity]:
> True Negative vs False Positive (Higher % FP = less specific)
> Delineate what could interfere to give a false positive or
false negative result.
>
>Reportable Range
> Establish a scoring system
> Provide the definition of a positive result
>
>3)Sensitivity
>
>Analytic Sensitivity:
> Lowest amount of substance detectable by the test
> Can only be done with controls of known concentration
>
>Diagnostic Sensitivity:
> Ability of the test to determine true diagnostic positive
verses false negative (higher % FN = less sensitive)
> Requires comparison to a previously validated antibody
>
>IHC Sensitivity:
> Extent to which an antibody can be diluted and still achieve
target recognition. NOTE: This is determined by antibody AND
detection system!
>
>
>
>4) Specificity:
>
>Analytic Specificity
> Accuracy on tests of known positive and negative controls
> Controls of known concentration
> Determine what could "Interfere" to confound the result
>
>
>Diagnostic Specificity
> Ability of a test to determine true diagnostic negative verses
false positives (Higher % FP = less specific)
> Requires comparison to a previously validated antibody
>
>
>IHC Specificity
> Ability of an antibody to bind exclusively to its particular
antigen in the absence of staining of other molecules
> Or, staining of other structures in addition to target
>structures/cells
>
>(Sensitivity and Specificity adapted from: Theoretical and Practical
>Aspects of Test Performance, in Immunomicroscopy, Taylor & Cote, 2005)
>
>Tim Morken
>Supervisor, Histology / IPOX
>UCSF Medical Center
>San Francisco, CA
>
>
>-----Original Message-----
>From: histonet-bounces <@t> lists.utsouthwestern.edu
>[mailto:histonet-bounces <@t> lists.utsouthwestern.edu] On Behalf Of
>JMyers1 <@t> aol.com
>Sent: Tuesday, June 22, 2010 6:51 PM
>To: tjasper <@t> copc.net
>Cc: histonet <@t> lists.utsouthwestern.edu
>Subject: RE: [Histonet] New CAP question ANP.22760
>
>Tom:
>
>As much as I agree with your acknowledgment that its seems a bit odd
>for the CAP to have a blood-banker responding to AP-related issue, I'm
>actually not surprised. The folks in the 'clinical' lab have been
>performing more comprehensive and complex validation procedures for a
>very long time, and they wonder why IHC isn't expected to follow the
>same requirements as chemistry, immunology, etc. -- IHC is, after all,
>an awful lot like ELISA. And rightfully so, because IHC is, under CLIA
>(which supersedes CAP), considered highly-complex, non-waived testing
>-- and is, therefore, subject to the same Quality Systems regulations
>(in particular, 42CFR493.1252-1256, 1273, and 1281) as the testing
performed in other areas of the lab.
>
>Could it be that, because AP produces qualitative results that are
>interpreted by a pathologist and CP produces quantitative results that
>are interpreted by an analyzer, we somehow think that CLIA rules don't
>apply to IHC? I certainly don't have the answer to that, but it make
>me wonder what the future holds. As witnessed by some of the newest
>CAP 'standards' (including the question in question...no pun intended),
>e.g. ER/PR, where a minimum of 20 positive and 20 negative specimens
>must be tested, and where 10 of the positives must be weakly positive
>-- an acknowledgment that validation specimens must be carefully
>selected in order to obtain appropriate results), it certainly doesn't
>appear that the regulation of IHC testing is going to become more
relaxed.
>
>Joe Myers, M.S., CT(ASCP)
>
>------------------------------
>
>Message: 12
>Date: Fri, 18 Jun 2010 12:38:07 -0700
>From: "Thomas Jasper" <tjasper <@t> copc.net>
>Subject: RE: [Histonet] New CAP question ANP.22760
>To: "Mark Tarango" <marktarango <@t> gmail.com>
>Cc: _histonet <@t> lists.utsouthwestern.edu_
>(mailto:histonet <@t> lists.utsouthwestern.edu)
>
>Mark,
>
>Did you notice the credentials from this CAP representative? MT with a
>Blood Bank specialty I believe. What I glean from that is...more than
>likely this person does not grasp the logistics of "contemporaneously"
>staining identical Abs from separate lots. She also likely does not
>understand the logistical application for detection and automation
>either.
>
>I'm not trying to be overly critical of this person. I'm sure she is
>quite intelligent and would not have the MT/SBB if she wasn't
>intelligent. It comes down to a lack of understanding Anatomic
>Pathology testing application re: automated IHC. I believe this is a
>common problem in and out of CAP. Many lab directors and other folks in
>positions of authority without AP/Histology/Cytology backgrounds seem
>to believe that broad clinical lab modalities apply to Anatomic Path
>scenarios. I used to refer to this in my former position as - "Trying
>to put the yoke of clinical lab onto anatomic path." We are
>laboratorians, but in many instances do not fit the general clinical
>lab mold.
>
>It's unfortunate that CAP has put this person in the position to
>respond. It is apparent to me that she's not grasping the particulars
>here. She probably never will unless she decides to go into a working,
>automated IHC "tissue" lab and take the time to ask questions and
>understand (learn) what we're all about.
>
>Thanks,
>Tom Jasper
>
>Thomas Jasper HT (ASCP) BAS
>Histology Supervisor
>Central Oregon Regional Pathology Services Bend, OR 97701
>_______________________________________________
>Histonet mailing list
>Histonet <@t> lists.utsouthwestern.edu
>http://lists.utsouthwestern.edu/mailman/listinfo/histonet
>
>
>_______________________________________________
>Histonet mailing list
>Histonet <@t> lists.utsouthwestern.edu
>http://lists.utsouthwestern.edu/mailman/listinfo/histonet
>
>
>______________________________________________________________________
>This message is confidential, intended only for the named
>recipient(s) and may contain information that is privileged or exempt
>from disclosure under applicable law. If you are not the intended
>recipient(s), you are notified that the dissemination, distribution, or
>copying of this message is strictly prohibited. If you receive this
>message in error, or are not the named recipient(s), please notify the
>sender at either the e-mail, fax, address, or telephone number listed
>above and delete this e-mail from your computer.
>Thank You.
>______________________________________________________________________
>_______________________________________________
>Histonet mailing list
>Histonet <@t> lists.utsouthwestern.edu
>http://lists.utsouthwestern.edu/mailman/listinfo/histonet
_______________________________________________
Histonet mailing list
Histonet <@t> lists.utsouthwestern.edu
http://lists.utsouthwestern.edu/mailman/listinfo/histonet
More information about the Histonet
mailing list