PerkinElmer, Inc., 68 Elm Street, Hopkinton, MA USA (800) 762-4000 or (+1) 203 925-4602 www.perkinelmer.com

Summary

Conclusions

• Multispectral imaging enabled the quantitation of two immunostains (CD3

& FOXP3) in intra- and extra-follicular compartments in follicular

lymphoma

• FOXP3+ Tregs were automatically counted and used in Kaplan-Meier

survival analysis, demonstrating association with good outcome

• Automated multiplexed tissue cytometry analyses are feasible for routine

clinical studies and work with many multiplexed IHC staining

methodologies.

• The enumeration of FOXP3 +’ve T cells in these clinical samples was

effective and easy to perform.

Phenotyping TILs in situ: Automated Enumeration of Intra- and Extra-Follicular FOXP3+ Regulatory T Cells in Follicular Lymphoma J.R. Mansfield,1 C.M. van der Loos,2 , L.S. Nelson,3 C. Rose,3 H.E. Sandison,3 S. Usher,3 J.A. Radford,3 K. M. Linton,3 R.J. Byers3 1) PerkinElmer, Hopkinton, MA; 2) Academic Medical Center, Amsterdam, Netherlands; 3) University of Manchester, UK

Nuance® and Vectra™

Multispectral Imaging

Systems

• Images at different

wavelengths

• Assemble the images

into a data “cube”

• Spectrum at every (x,y)

pixel

RGB Representation of Spectral Cube

Spectrum from

nucleus with both

hematoxylin and

DAB

Spectrum from

membrane with

just red stain

Spectrum from

stroma with just

hematoxylin

Once unmixed,

stains can be

measured

accurately.

Unmixed Hematoxylin

Component

Unmixed Red

Component

Spectra of pure chromogens

collected from single- stained

sections

Unmixed DAB

Component

Automated tissue and cellular segmentation Clinical correlation of results

53 samples from 40 patients were automatically analyzed using this methodology and the number of FOXP3+/CD3+ Treg cells in each determined, in both T-cell (CD3+) rich

and poor areas. The number of Tregs cells were used in Kaplan-Meier survival analysis, demonstrating association of higher numbers of Tregs with favourable outcome in

both T-cell rich (extra-follicular) and poor (intra-follicular) areas (data shown with data split at 25th percentile, median & 75th percentiles for CD3+/FOXP3+ Treg score). This

meant patients were divided into groups determined by their Treg numbers using these three statistics as a threshold; . Kaplan-Meier demonstrated that patients with Treg

numbers in the top 75%, 50% and 25% all had significant survival advantages over those with lower numbers when divided into two groups based on these proportions

Outcome

0 50 100 150 200

10

20

30

40

50

60

70

80

90

100

Time

Su

rviv

al p

ro

bab

ilit

y (

%)

FOXP3 Tregs in CD3 poor areas

less than 25% centile

=/> 25% centile

P=0.0431

Outcome

0 50 100 150 200

0

20

40

60

80

100

Time

Su

rviv

al p

ro

bab

ilit

y (

%)

FOXP3 Tregs in CD3 poor areas

less than median

=/> median

P=0.0036

Outcome

0 50 100 150 200

10

20

30

40

50

60

70

80

90

100

Time

Su

rviv

al p

ro

bab

ilit

y (

%)

FOXP3 Tregs in CD3 poor areas

less than 75% centile

=/> 75% centile

P=0.0173

Outcome

0 50 100 150 200

0

20

40

60

80

100

Time

Su

rviv

al p

ro

bab

ilit

y (

%)

FOXP3 Tregs in CD3 rich areas

less than 25% centile

=/> than 25% centile

P=0.2179

Outcome

0 50 100 150 200

0

20

40

60

80

100

Time

Su

rviv

al p

ro

bab

ilit

y (

%)

FOXP3 Tregs in CD3 rich areas

less than median

=/> than median

P=0.0034

Outcome

0 50 100 150 200

10

20

30

40

50

60

70

80

90

100

Time

Su

rviv

al p

ro

bab

ilit

y (

%)

FOXP3 Tregs in CD3 rich areas

less than 75% centile

=/> than 75% centile

P=0.0343

CD3 positivity, identifying T-cells, was

used to identify CD3 rich and poor

areas, approximating to extra-follicular

(green) and intra-follicular (pink)

areas, respectively. Thresholding of

CD3 (membrane) and FOXP3

(nuclear) was used to identify double

FOXP3+/CD3+ Treg cells (shown as

yellow cells), FOXP3-/CD3+ cells

(green) and other cells (blue) in both

compartments.

Multispectral imaging of triplex-stained follicular lymphoma

FOXP3

CD3

RGB representation of multispectral dataset

Sample 1 Sample 2

Sample 1 Sample 2

In many cancers, tumor-infiltrating lymphocytes (TILs) indicate levels of tumor

immunogenicity and are a strong predictor of survival. In particular, increased levels of

regulatory T cells (Tregs) are associated with poorer prognosis in some cancers. An

understanding of the phenotype and spatial distribution of TILs in situ within tumor regions

would be advantageous. However, visual TIL assessment cannot easily determine the type

of lymphocyte in situ and multimarker quantitation is difficult with standard methods. Here

we present a multi-marker, computer-aided event-counting method for determining the

phenotypes of lymphocytes in follicular lymphoma sections using a multispectral imaging

(MSI) and automated tissue segmentation and counting approach. A tissue microarray

containing follicular lymphoma cores from 70 patients was stained for CD3, FOXP3 and

hematoxylin, of which 40 cores were informative for both triple staining and clinical follow-

up. Each core was imaged using MSI and the individual staining of each marker separated

from each other using spectral unmixing. The images were analyzed using software which

had been trained to recognize the follicular areas based on the tissue morphology,

specifically based on CD3 rich (extra-follicular) and poor (intra-follicular) areas. The

FOXP3 status of each CD3+ TIL was then determined and the number of each Treg

(FOXP3+/CD3+) counted for both the intra- and extra-follicular tissue compartments.

Results indicate that machine-learning software can be trained to accurately recognize

follicular and non-follicular regions within each core, in this instance based on abundance

of CD3 cells. MSI enabled the accurate quantitation of two immunostains in the sample

without crosstalk. The number of Tregs were determined for each core and used in Kaplan-

Meier survival analysis, which demonstrated association of FOXP3+/CD3+ Tregs with

favourable outcome in both the intra- and extra-follicular areas. Understanding the number

and location (intra- and extra-follicular) of Tregs is an assay with potentially important

clinical prognostic implications. Thus study shows that an automated method for counting

Tregs can be developed for follicular lymphoma. This multimarker phenotyping and

counting approach shows the potential for broad applicability in the enumeration of a wide

range of specifically phenotyped TILs in situ in many solid tumors.

Multispectral imaging technology

RGB representation

of spectral cube

With cancer mask

Breast cancer ER/PR

co-expression assay

With cancer mask and

nuclear segmentation

Red = cancer mask

Green = cancer nuclei

Blue = background

1) Automated user-trained

morphologic segmentation

using inForm™ Tissue Finder

2) Cellular segmentation

(nuclear, cytoplasmic or

membrane)

Morphologic and cellular segmentation Sample 1

Extra-follicular cells:1473

CD3+/FOXP3-: 13.9%

CD3+/FOXP3+: 12.3%

Survival: 171+ months

Sample 2

Extra-follicular cells: 1917

CD3+/FOXP3-: 19.66%

CD3+/FOXP3+: 0.66%

Survival: 54 months