Breast tissue diagnosis by Raman spectroscopy

B. Brożek-Płuska*, J.Surmacki*, J. Jabłońska**, R. Kordek**,

H. Abramczyk*

*Laboratory of Laser Molecular Spectroscopy, Institute of Applied Radiation Chemistry, Technical University of Łódź, Faculty of Chemistry, Poland**Department of Oncology, Medical University of Łódź, Poland

GOAL

Some substances are produced by the organism in response to thecancer’s presence. They are called tumor markers.Tumor markers are molecules occurring in blood or tissue that areassociated with cancer and whose measurement or identification is useful inpatient diagnosis or clinical management. The ideal marker would be a“blood test” for cancer in which a positive result would occur only inpatients with malignancy, one that would correlate with stage and responseto treatment and that was easily and reproducibly measured. No tumormarker now available has met this ideal.

We believe that optical methods including Raman spectroscopy will provide such a marker. Possibly it will help to find the hallmarks of cancer.

•immediate in vivo diagnosis

•reduction the number of biopsies

•combination of biochemical and histopathologicaldiagnosis provides more information because pathology

is intimately related to biochemistry

•method has a potential to remove human interpretation

•non-invasive,non-ionizing method that probes with chemical specificity –vibrations and fluorescence (not just

structure)

• extremaly high spatial resolution (optical imaging)

Why ?

Human speciemens obtained from surgery. Upon removal during the operation, the ex vivo sample is devided by a doctor into two parts, one goes to our lab the second goes to the pathology examination

The ex vivo samples are neither frozen in liquid nitrogen for storage nor fixed in formalin, the fresh tissue is measured immeditely after delivering from the hospital

The samples for pathology measurements are passively thawed at room temperature and kept moist with PBS fixed in formalin

Cut through the marked locations into 5-μm-thick sections, and stained with eosin

normal tissue

0

5000

0

5000

8000

6000

4000

2000

0

0 500 1000 1500 2000 2500 3000 3500 4000

5000

10000

15000

20000

25000

30000

35000

514nm, normal tissue

25mW

100mW

Inte

nsity (

arb

itra

ry u

nits)

Raman shift in wavennumbers [cm-1]

0 500 1000 1500 2000 2500 3000 3500 4000

12000

14000

16000

18000

20000

22000

24000

26000

514nm, 100mW, cancerIn

tens

ity (

arbi

trar

y un

its)

Raman shift in wavennumbers [cm-1]

0

10000

20000

0

10000

20000

20000

15000

10000

5000

0

maligant tissue

Tissue Preparation

carcinoma mammae fibroadenoma mammae

Measurements Parameters

The laser excitation is 514 nm, the spot is d=500 μm in diameter

Light diffusion in the tissue results in a spot of v≈1mm

Integration time 0.5 s

Spectral resolution 2 and 8 cm -1

The laser excitation power: 25 mW, 100mW

Patients Statistics

1100 spectra from 100 patients

•fibroadenomas

•infiltrating carcinoma

•infiltrating ductal carcinoma (IDC)

•infiltrating lobular carcinoma (ILC)

•IDC+ILC

•multifocal carcinoma

•carcinoma microinvasive

•intracystic papillary carcinoma (noninvasive)

•carcinoma mucinosum

•carcinoma intraductal microinvasive

•benign dysplasia

•dysplasia benign dysplasia (cystes, apocrinal metaplasia, adenosis)

•ductal-lobular hyperplasia

•cystic mastopathy

0 1000 2000 3000 4000 5000 6000 7000 80000

1000

2000

3000

40000014 KT - normal tissue

25mW

514 541 572 607 647 691 743 802 872 [nm]

relative wavenumber [cm-1]

Inte

nsity [

counts

/s]

0 1000 2000 3000 4000 5000 6000 7000 80000

1000

2000

3000

40000014 KT - carcinoma lobulare infiltrans mammae

25mW

514 541 572 607 647 691 743 802 872 [nm]

relative wavenumber [cm-1]

Inte

nsity [

co

unts

/s]

Comparison between the Raman spectra for normal and malignant tissue (infiltrating ductal carcinoma (IDC))

0 1000 2000 3000 4000 5000 6000 7000 80000

2500

5000

7500

10000

12500

15000

25mW

0006 KT - normal tissue

514 541 572 607 647 691 743 802 872 [nm]

relative wavenumber [cm-1]

Inte

nsity [

co

un

ts/s

]

0 1000 2000 3000 4000 5000 6000 7000 80000

2500

5000

7500

10000

12500

15000

25mW

0006 KT - carcinoma ductale infiltrans G2

514 541 572 607 647 691 743 802 872 [nm]

relative wavenumber [cm-1]

Inte

nsity [

co

unts

/s]

Quenching of autofluorescence by low temperature Ramanspectroscopy

0 500 1000 1500 2000 2500 3000 3500 4000-500

0

500

1000

1500

2000

2500

normal tissue - patient nr 90

inte

ns

ity

(c

ts/s

)

wavenumber (cm-1)

K90T33 - 250K

K90T34 - 200K

K90T35 - 170K

K90T36 - 150K

K90T37 - 110K

K90T38 - 77K

0 500 1000 1500 2000 2500 3000 3500 4000

-100

0

100

200

300

400

500

600

cancer tissue - patient nr 90

inte

ns

ity

9c

ts/s

)

wavenumber (cm-1)

K90T14 - 250K

K90T15 - 200K

K90T16 - 170K

K90T17 - 150K

K90T18 - 110K

K90T19 - 77K

Comparison between the Raman spectra for normal and malignant tissue (carcinoma mucinosum )

0 2000 4000 6000 8000

-500

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Y Ax

is Ti

tle

X Axis Title

K14T2 (carcinoma mucinosum mammae sinistri) K14T2 (normal tissue)

wavenumber [1/cm]

inte

nsi

ty[c

ts/s

]

Comparison between the Raman spectra for normal and benigntumour tissue (fibroadenoma)

0 2000 4000 6000 8000

0

200

400

600

800

1000

1200

1400

1600

1800

Y Ax

is T

itle

X Axis Title

K31T5 (normal tissue) K31T8 (Fibroadenoma mamae)

wavenumber [1/cm]

inte

nsi

ty[c

ts/s

]

PRINCIPAL COMPONENT ANALYSIS (PCA)PLS_Toolbox Version 4.0for use with MATLAB™

Barry M. WiseNeal B. GallagherRasmus BroJeremy M. ShaverWillem WindigR. Scott Koch

Eigenvector Research, Inc.3905 West Eaglerock DriveWenatchee, WA 98801

helpdesk@eigenvector.comwww.eigenvector.com

12

PCA score plot – mean center, SNV, 1-st derivative

-1 -0.5 0 0.5 1 1.5 2-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

Scores on PC 1 (66.25%)

Sc

ore

s o

n P

C 2

(7

.70

%)

c c

h c

h h

h

h

h

c

c h

c

h

c

c

h

h

c

h

h

h h

h w

w

h

h c h

c c c h

c

c

c h c

c

c

h w

w w

w

w

h

w

w w w

h w w w

h h

h c c c h

c

c c

h

Samples/Scores Plot of HY

Decluttered

10

20

30

40

50

60

No Raman peaks Intensive Raman peaks

Normal tissue

Malignant tissue

High autofluorescence

Low autofluorescence

Benign

tissue

500 1000 1500 2000 2500 3000 3500-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

wavenumber (cm-1)

Lo

ad

ing

s o

n P

C 1

(66.2

5%

)

Variables/Loadings Plot for HY

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

C-H

streching

lipids

C-C and

C=C

stretching

carotenoids

PCA loading plot – mean center, SNV, 1-st derivative

CONCLUSIONSThe results clearly illustrates the ability of Raman spectroscopy to accurately diagnose breast cancer and demonstrates how the diagnostic scheme can be adjusted to obtain the desired degree of sensitivity and specificity (88%,72%)

The normal tissue has a characteristic bands: C-C (1110 cm-1)and C=C (1520 cm-1) stretching bands of carotenoids and at2840-2900 cm-1 for the C-H symmetric and asymmetric bands oflipids (fat) which are not visible in the malignant tissue and inbenign tumor tissue. Moreover, the fluorescence is much higherin the malignant tissue.

We belive that in a very near future a good quality Raman signal will be obtained with the optical fibers coupled with a biopsy needle and incorporated into Raman spectrometer for breast tissue measurements in vivo.