Barabra has made some very good points here - but I think that there is
another whole set of problems that are independent of the calibration of
the microscope  / detection system itself. These all relate to the
labelling itself.

Overall it is extremely difficult to get fluorescent labelling that is
consistent from preparation to preparation if you are going to quantify
the levels. It depends on so many things - the fixation and tissue
processing; the section thickness, the antibody binding (both primary
and secondary); the thickness of the labelled structure itself; the
magnification at which  you do your imaging; plus all the hardware
things, like light source intensity; confocal settings etc...

We have grappled with these problems for many years, and have come to
the conclusion that you can only really get useful intensity data when
you are comparing objects / structures within the same preparation. In
principle, you can compare relative intensity data between preparations
by normalising the data compared with the background in the same section
eg express the brightness levels in term of standard deviations above
background; or the (difference between the signal you are interested in
and the background) divided by the background (this is how many calcium
imagers calibrate their images). But in our experience, even these
manipulations only give you what seem to be half-decent results when
they are done across large numbers of preparations under very stringent
processsing and imaging conditions...

I certainly would be interested to see what other people think on this!

IAN


Barbara Foster wrote:
>
> Dear Doug,
>
> You have reason to be wary, based on some of the very issues you have
> brought up.  However, having said that, I will tell you that RELATIVE
> intensity measurements in fluorescence have  a long and happy history,
> starting with some of the earliest work in 1932 by Caspersson et. al.  If I
> were to describe the typical microspectrophotometers available from Zeiss
> and Leica, you would see that there are many cross-overs to confocal:
> (a) 12 bit PMTs are used as detectors
> (b) an aperture is inserted to reduce the spectrophotometric effects of glare
>
> To make results more consistent
> (a) you do, indeed, need to pay careful attention to sample preparation
> (b) you need to be consistent re: the size of aperture you use (i. e.,
> fixed pinholes of specific diameters are much preferable to iris-type
> controls)
> (c) you need a broad, continuous, well-behaved standard against which to
> compare your results.  When I worked at Zeiss as a microspectrophotometry
> specialist, we  had a piece of fluorescent glass.  If you call Tom Calahan
> (914-681-7733), he can probably guide you to a source.  When I later worked
> at Reichert /Cambridge/AO (now Leica), we used a piece of fluorescent glass
> which was divided into 4 quadrants.  One was left bare (100% Intensity),
> and each of the others was coated with varying amounts of neutral density
> filter.  Instead of calibrating against one intensity, we plotted 4
> different       values and established a contrast response function which was
> then used to correct system irregularities.  I suspect that a company like
> Omega Optical or Chroma could reproduce a similar standard fairly easily
> and you would just need to write a short piece of programming to fit in
> with your control and analysis program for the system correction.
>
> This is one of the few applications in which I would really encourage your
> going to a 12 bit (4096 gray level) system rather than the typical 8 bit
> systems offered routinely through standard image analysis packages.
>
> There have been several independent inventors who have investigated the
> integration of microspectrophotometry with confocal.  You may want to
> discuss the details further with them.  The two who come to mind quickly
> are Dr. Ted Dixon (now with Biophotometrics: 519-886-9013) and Coco
> Montague (now  with Genetic Microsystems: 781-932-9333).  Please send my
> regards to both if you speak with them.
>
> Best regards,
> Barbara Foster
> Consortium President
> Microscopy/Microscopy Education ...Educating microscopists for greater
> productivity.
>
> 125 Paridon Street Suite 102     Springfield, MA 01118
> PH: (413)746-6931  FX: (413)746-9311 email: [log in to unmask]
> Visit our web site <http://www.MME-Microscopy.com/education>
> ******************************************************
> MME is America's first national consortium providing
> customized on-site workshops in all areas of
> microscopy, sample preparation, and image analysis.
>
> At 10:05 AM 10/29/99 -0700, Doug Cromey wrote:
> >Confocalists,
> >
> >I am interested in some feedback from list subscribers as to whether
> >fluorescence intensities in confocal images can be quantitated.  There seem
> >to be two schools of thought about this issue (with some "gray" in
> >between), one that does quantitation & wonders why I would raise this
> >issue, and another that is wary of performing quantitation on confocal
> >images due to the many difficult to control variables involved (optical,
> >staining, and sometimes inter-operator).
> >
> >I would appreciate hearing under what conditions (if any) you would
> >consider quantitation to be a scientifically appropriate form of analysis.
> >
> >I would also appreciate it if you could provide literature references that
> >would help me sort out this issue.
> >
> >I am currently in the "wary" camp, but I'm trying to be open minded about
> >this.  I will post a summary of the replies to the list.
> >
> >Yours,
> >Doug
> >.....................................................................
> >: Douglas W. Cromey, M.S.          Dept. of Cell Biology & Anatomy  :
> >: Research Specialist, Principal   University of Arizona            :
> >: (office:  AHSC 4212A)            P.O. Box 245044                  :
> >: (voice:  520-626-2824)           Tucson, AZ  85724-5044   USA     :
> >: (FAX:  520-626-2097)           (email: [log in to unmask]):
> >:...................................................................:
> >            http://www.pharmacy.arizona.edu/exp_path.html
> >        Home of: "Microscopy and Imaging Resources on the WWW"
> >
> >

--
Professor Ian Gibbins
Anatomy & Histology
Flinders University of South Australia
GPO Box 2100, Adelaide, SA 5001
Australia

Phone:  +61-8-8204 5271
FAX:    +61-8-8277 0085
Email:  [log in to unmask]