Diapositive 1

1
Unbiaised Stereology Using Aperio ImageScope
Paulette Herlin, Benoît Plancoulaine
GRECAN, EA 1772, IFR 146 - Caen
University, F. Baclesse Cancer Centre, F
14076 Caen - France
2
Introduction

• Tumor heterogeneity Analysis of the whole
  tissue section (of VS) mandatory, gives
• the mean and maximum density of a marker and
  heterogeneity of its distribution
• Automatic Image Analysis of VS not always
  possible, poor quality of the staining,
• IHC antibodies sometimes not available
• - Automatic Image Analysis of VS quality
  control needed, several thousands of objects
  detected, visual estimation of false positive and
  false negative rates impossible
• Stereological methods can help

3
Area
Estimation of the area of an object of irregular
shape number of dots x area of a square
4
Area fraction, Volume fraction
A the ball B ball dog
For a 2D section AA Vv (same thickness) NN
5
Numerical density
Counting frames and forbidden lines
6
Unbiased sampling
N per unit area number of dots / number of
counting frames x area of a frame
7
Point counting / Other methods Numerical images
25,69
23,52


Automatic image processing
Point counting





23,75





Planimetry
Mixed Point counting - Planimetry


8
Counting frames / Other methods Numerical images
202/m2
200/m2



Unbiased sampling
Automatic image processing


207/m2





Manual detection
Mixed Sampling - Planimetry


9
Implementation

Taking advantage of Aperio ImageScope

10
Generating automatically the grid (xml file)
11

Giving a second layer for manual delineation of
the limits of the tumor

12

Providing pathologists with a third layer, for
visual detection of points hitting structures of
interest

13
Moving quickly from a point to another
14
Advantages of VS

Breast cancer section, counting of mitosis
figures, PHH3 Immunostaining. Counting in the
whole tumor section. Overview at a low resolution.


15




Visual inspection at a medium and high resolution

Expression of results density of mitosis
profiles / square mm
16
Breast cancer. Estimation of Vv (volume fraction)
of stromal compartment
17
Precision
E.R. Weibel, 1981. Stereological methods in cell
biology where are we? where are we going? J.
Histochem. Cytochem., 29, 1043-1052
18
Performances

• Up to half an hour per slide
• Intra and inter-reproducibility can be
• estimated (intra, around 2 for stroma)
• Works on sub-optimal IHC stained slides
• Several markers or compartments can be
• evaluated in the same VS
• Works on HES stained slides and small
• images as well

Resolution 0.5 µm Periodicity 1000
pixels Test points 40 x 40 pixels Test frames
200 x 200 pixels (0.01 mm2)

19
(No Transcript)
20
Quality control of VS automatic image processing
21
Breast cancer PHH3 Immuno- staining
Sub-sampling and automatic image processing of
the whole virtual slide (home made program)
22
Segmentation refined at high resolution
Galeries
Aphelion TM (ADCIS, Amerinex Applied Imaging)
23
Superimposition of stereo- xml file
24
and visual inspection inside test frames only
25
Swift identification and estimation of false
negative or positive objects
26
Breast cancer Cyclin A immunostaining
27
Same quality control procedure, equivalent
estimation
28
But . Advantages of automatic image
processing
Time saving and
Statistics on Color and Intensity
55 000 nuclei
29
Conclusion

• Despite being increasingly neglected due to the
  widespread of image processing,
• stereological interactive methods remain highly
  relevant
• Informatic tools facilitate the use of these
  methods by rendering them both more
• attractive and user friendly
• They are to be recommended when automatic
  estimates are proven difficult
• (i.e. too many operations) or even impossible
  (i.e. unsuited staining)
• They are able to give also a gold reference for
  the development of fully automatic
• VS processing strategies

30
Acknowledgments

• Regional Council of Lower Normandy, Agence
  Nationale de la
• Recherche, Cancéropôle Nord Ouest, Ville de
  Cherbourg,
• Fonds Européens de Développement Régional.
• - Dr. Jacques Chasle (Head of the Pathology
  Department of
• F. Baclesse Cancer Centre), and Miss Mylène
  Brécin.

31
Some References

• - E.R. Weibel, 1981. Stereological methods in
  cell biology where are we? where are we going?
  J. Histochem. Cytochem., 29, 1043-1052
• H. J. G. Gundersen, R. Osterby, 1981. Optimizing
  sampling efficiency of stereological studies in
  biology
• or Do more less well!. J. Microsc., 121,
  65-73
• V. Howard and M. Reed, 1998. Unbiased
  Stereology. Three-dimensional measurement in
  microscopy. Microscopy handbooks 41, Bios
  Scientific Publishers, UK
• L. Kubinova, X. W. Mao, J. Janacek, J. O.
  Archambeau, 2003. Stereology Techniques in
  Radiation Biology, Radiation Research 160,
  110119
• E. A. Van Vre, H. M. van Beusekomb, C. J.
  Vrintsa, J. M. Bosmansa, H. Bultc, W. J. Van der
  Giessenb, 2007. Stereology a simplified and
  more time-efficient method than planimetry for
  the quantitative analysis of vascular
  structures in different models of intimal
  thickening. Cardiovascular Pathology 16, 43 50
• - F. Teba, R. Martin, V. Gomez, L. M. Herranz, L.
  SantaMaria, 2007. Cell proliferation and
  Volume-Weighted mean nuclear volume in
  high-grade PIN and adenocarcinoma, compared with
  normal prostate. Image Analysis and
  Stereology, 26, 93-99