Lasers in Surgery and Medicine 48:538–545 (2016)
Using Fourier Transform Infrared Spectroscopy to
Evaluate Biological Effects Induced by Photodynamic
Therapy
Cassio A. Lima, MSc,1 Viviane P. Goulart, MSc,1 Luciana Correa, PhD,2 and Denise M. Zezell,
1
Instituto de Pesquisas Energ

eticas e Nucleares, IPEN – CNEN/SP, Universidade de Sao Paulo,
Sao Paulo SP 05508 000, Brazil
2
Faculdade de Odontologia, Universidade de Sao ~ Paulo, Sao
~ Paulo SP 05508 000, Brazil
Background and Objective: Vibrational spectroscopic
methods associated with multivariate statistical techni-
ques have been succeeded in discriminating skin lesions
from normal tissues. However, there is no study exploring
the potential of these techniques to assess the alterations
promoted by photodynamic effect in tissue. The present
study aims to demonstrate the ability of Fourier Transform
Infrared (FTIR) spectroscopy on Attenuated total reflec-
tion (ATR) sampling mode associated with principal
component-linear discriminant analysis (PC-LDA) to
evaluate the biochemical changes caused by photodynamic
therapy (PDT) in skin neoplastic tissue.
Materials and Methods: Cutaneous neoplastic lesions,
precursors of squamous cell carcinoma (SCC), were
chemically induced in Swiss mice and submitted to a
single session of 5-aminolevulinic acid (ALA)-mediated
PDT. Tissue sections with 5 mm thickness were obtained
from formalin-fixed paraffin-embedded (FFPE) and proc-
essed prior to the histopathological analysis and spectro-
scopic measurements. Spectra were collected in
mid-infrared region using a FTIR spectrometer on ATR
sampling mode. Principal Component-Linear Discrimi-
nant Analysis (PC-LDA) was applied on preprocessed
second derivatives spectra. Biochemical changes were
assessed using PCA-loadings and accuracy of classification
was obtained from PC-LDA .
Results: Sub-bands of Amide I (1,624 and 1,650 cm 1) and
Amide II (1,517 cm 1) indicated a protein overexpression
in non-treated and post-PDT neoplastic tissue compared
with healthy skin, as well as a decrease in collagen fibers
(1,204, 1,236, 1,282, and 1,338 cm 1) and glycogen (1,028,
1,082, and 1,151 cm 1) content. Photosensitized neoplastic
tissue revealed shifted peak position and decreased
b-sheet secondary structure of proteins (1,624 cm 1)
amount in comparison to non-treated neoplastic lesions.
PC-LDA score plots discriminated non-treated neoplastic
skin spectra from post-PDT cutaneous lesions with
accuracy of 92.8%, whereas non-treated neoplastic skin
was discriminated from healthy tissue with 93.5% accu-
racy and post-PDT cutaneous lesions was discriminated
from healthy tissue with 89.7% accuracy.
Conclusion: PC-LDA was able to discriminate ATR-FTIR
spectra of non-treated and post-PDT neoplastic lesions, as
ß 2016 Wiley Periodicals, Inc.
PhD
1
well as from healthy skin. Thus, the method can be used for
early diagnosis of premalignant skin lesions, as well as to
evaluate the response to photodynamic treatment. Lasers
Surg. Med. 48:538–545, 2016. ß 2016 Wiley Periodicals, Inc.
Key words: 5-aminolevulinic acid (ALA); attenuated
total reflectance (ATR); biochemical characterization;
cutaneous squamous cell carcinoma; Fourier transform
infrared spectroscopy; neoplastic lesion; principal compo-
nent analysis
INTRODUCTION
Nonmelanoma skin cancers (NMSCs) are the most
common form of malignancy in humans and have
increased their incidence over the past several deca-
des [1,2]. Increase of exposure to UV radiation and aging of
the population, among other, have been associated as the
main factors for this growing incidence [3]. NMSC refers to
both squamous cell carcinoma (SCC) and basal cell
carcinoma (BCC), which are the main types of non-
melanoma skin cancer diagnosed in humans [2]. Despite
of a relatively low mortality rate related to these cancers,
their available treatments expensive and submit the
patient to morbidity [3].
Traditional treatments of cutaneous cancers may
include surgery (surgical excision, Mohs surgery, curet-
tage and electrodessication, cryo surgery, laser surgery,
etc.) and/or ionizing radiation (radiotherapy and
Conflict of Interest Disclosures: All authors have completed
and submitted the ICMJE Form for Disclosure of Potential
Conflicts of Interest and none were reported.
Contract grant sponsor: FAPESP CEPID; Contract
grant number: 05/51689-2; Contract grant sponsor: CAPES
PROCAD; Contract grant number: 88881.068505/2014-01;
Contract grant sponsor: CNPq INCT; Contract grant number:
573.916/2008-0; Contract grant sponsor: PQ; Contract
grant number: 312397/2013-5.

Correspondence to: Prof. Denise M. Zezell, PhD, IPEN – CNEN/
SP, Center for Lasers and Applications (CLA) Av. Lineu Prestes,
2242- CEP 05508-000, S~ ao Paulo-SP, Brazil. E-mail: zezell@usp.br
Accepted 31 December 2015
Published online 22 February 2016 in Wiley Online Library
(wileyonlinelibrary.com).
DOI 10.1002/lsm.22473
 FTIR TO EVALUATE THE EFFECTS INDUCED BY PDT 539

chemotherapy [4]. Moreover, photodynamic therapy (PDT) component-linear discriminant analysis (PC-LDA), to
is a promising minimally invasive alternative that does not evaluate the biochemical changes caused by photodynamic
require surgical intervention or exposure to ionizing therapy mediated by 5-aminolevulinic acid (ALA) in
radiation [5]. PDT is based in the administration of a cutaneous neoplastic lesions precursors of squamous cell
photosensitizer (PS) molecule activated by light at a carcinoma.
specific wavelength, which promotes a series of biochemi-
cal events and leads the tumor cells to death [6]. Different MATERIALS AND METHODS
death mechanisms, such as apoptosis, necrosis, and Chemical Carcinogenesis
autophagy are not independent and may occur at the
Neoplastic lesions were induced on Swiss mice using a
same time in a photosensitized lesion [7]. The individual
well-established multi-stage chemical-carcinogenesis pro-
contribution of each mechanism to the overall result (cell
tocol for obtaining cutaneous squamous cell carcinoma [24].
death) is still an open question, but it is already known that
After approval by the Committee on Animal Research and
the response to PDT may vary with the cell type
Ethics (71/10-CEUA-IPEN/SP), 50 female mice aged from
(phenotype and metabolic potential), total fluence deliv-
8 to 10 weeks with 20 g of body mass were divided into three
ered, types of PS and the time interval between PS
groups: healthy skin (G1), neoplastic tissue (G2), and post-
application and light irradiation [7]. Thus, defining and
PDT neoplastic lesions (G3) (Table 1).
understanding biochemical events caused by PDT may
Animals from all groups were anesthetized with
result in designing better treatment protocols [7].
ketamine (0.35 ml/kg) and xylazine (0.20 ml/kg) during
Several techniques use dyes and markers to characterize
all experiment protocol. Mice from G2 and G3 groups were
and visualize morphological and biochemical features of
submitted to chemical carcinogenesis, consisting in two
tissues. Complete characterization of a biological event
stages: initiation and promotion. In the first one, a single
may require more than one molecular marker, as well as
dose of 50-mg 7,12-dimethyl-benzanthracene (DMBA)
multiple staining. Vibrational spectroscopy has been
diluted in 100 ml of acetone was topical applied on the
established for biomedical applications over the past
shaved backs of the mice. The promotion phase began
decades and presents a potential to study biochemical
1 week after and consisted in a bi-weekly application of 5-g
changes induced by PDT, which does not require multiple
12-O-tetradecanoyl-phorbol-13-acetate (TPA) diluted in
staining or complex methods for sample
200 ml of acetone for 28 weeks. After 28 weeks, the animals
preparation [8–14].
obtained visible single or multiple tumor nodules with
In particular, there must be a change in the dipole
verrucous aspect, but only 33 animals were further
moment during the vibration of a molecular bond in order
evaluated due to a high death rate of 34%. This death
to be infrared active and quantitatively measured by
rate is in agreement with other studies and is related to the
Fourier transform infrared (FTIR) spectroscopy, providing
toxicity of DMBA/TPA, long induction time and the poor
biochemical complementary information to the morpho-
state of the animals during the experiment [24]. The
logical histopathology using the same biopsy material
control group (G1) only received topical application of
without any staining [9,15].
acetone. Mice from G2 group were euthanized at the end of
Transmission, Reflection–Absorption (transflection), and
chemical carcinogenesis and animals from G3 were
Attenuated Total Reflection (ATR) are the most used
submitted to photodynamic therapy.
experimental setups to acquire FTIR spectra [16]. Previous
studies have outlined spectral alterations obtained in
Photodynamic Therapy (PDT)
transmission and transflection sampling mode [16], which
implies in additional preprocessing methods prior to spectro- Single-session PDT was performed in mice using a
scopic analysis. On the other hand, ATR sampling mode does photosensitizer cream prepared with active principle being
not present these undesired spectral contributions, has a 5-aminolevulinic acid (ALA) (20%) and other ingredients
high SNR (signal-to-noise ratio) in comparison to the others
and provides a simple and powerful tool for analyzing liquids
and thin films samples with no preparation method [9,15,17]. TABLE 1. Mice Groups
Macro ATR-FTIR spectroscopy provides a single spectrum,
which represents the average signal from the sample area Number of Number of
where the light passed through [17], and can also be collected Number biopsies per spectra
with a fiber-ATR that in the future can measure directly on Group Description of mice animal collected
patient’s skin [18,19].
G1 Healthy 13 5–6 70
Discrimination between skin lesions and normal tissues
skin
using vibrational spectroscopy techniques associated with
G2 Non-treated 13 5–6 70
multivariate statistical methods have been outlined by
neoplastic
another works [20–23]. However, to the best of our
skin
knowledge, there is no study exploring the potential of
G3 Post-PDT 7 4 28
these techniques to assess the alterations promoted by
neoplastic
PDT. In this way, the present study demonstrates the
lesions
ability of ATR-FTIR spectroscopy associated with principal
540 LIMA ET AL.
kept confidential (patent pending PIN80705591-9). The
cream was applied on the tumor lesions with 5-mm
additional margin using a disposable plastic spatula and
the excess was removed with a gauze before PDT.
Irradiation was performed using a cluster of LEDs
emitting at 630 nm with fluence-rate of 5 mW/cm2 and
fluence of 12 J/cm2. Animals were euthanized 10 days after
PDT session and skin biopsies were extracted and
processed for spectroscopic and histopathological
evaluation.
Sample Preparation
Skin samples obtained with excisional biopsies were
formalin-fixed and paraffin-embedded (FFPE). Sample
sections with 5 mm thickness were obtained and placed
in MirrIR low-E-coated slides (Kevley Technologies,
Chesterland, OH) for spectroscopic analysis. Considering
that paraffin presents active vibrational modes in the
analyzed wavenumber range, the sections were submitted
to dewaxing protocol prior to the spectroscopic measure-
ments. For paraffin removal, FFPE sections were
immersed into two baths of xylene during 10 minutes
and one bath of absolute ethanol during 5 minutes. After
this, samples were kept in a desiccator for 24 hours [25].
For histopathological evaluation, FFPE sections were
hematoxylin-eosin (H&E) stained and anatomopatholog-
ical features were assessed by a pathologist.
FTIR Spectroscopy
Data acquisition. Spectra were collected using a
Fourier transform infrared spectrometer (Thermo Nicolet
6700, Waltham, MA) operating with an Attenuated Total
Reflectance (ATR) accessory as sampling mode (Smart
Orbit, Thermo Scientific, Waltham, MA). Measurements
were performed in mid-infrared region (4,000–400 cm 1)
with spectral resolution of 4 cm 1 and 100 scans per
spectrum. Samples were pressed into ATR diamond crystal
with standardized pressure and each spectrum obtained
represents the averaged from 10 replicates measured in
each sample.
Pre-processing data. ATR-FTIR data were vector-
normalized and second derivative of absorbance was
calculated to reduce baseline offset and assess the over-
lapping sub-bands in raw spectrum. For signal-smoothing,
spectra were submitted to Savitzky–Golay filter with a
polynomial of second order in an eleven points window.
Spectral analysis. Principal Component Analysis
(PCA) was performed on second derivatives in the
fingerprint spectral region (900–1,800 cm 1). In order to
evaluate the ability of PCA to classify spectral data, each
group was analyzed against each other and the accuracy of
sorting was calculated considering spectral distribution in
PCA score plots. In addition, PCA-loadings were used to
evaluate the biochemical changes presented by samples in
all groups. All pre-processing steps and spectral analysis
were performed using software Matlab1 R2015 (Math-
Works, Natick, MA).
RESULTS
Histopathological Analysis
Considering the histopathological analysis as gold
standard, the anatomopathological features of each group
were assessed in order to evaluate the biological effects
promoted by PDT on neoplastic lesions (Fig. 1).
Healthy skin (Fig. 1A) is composed by a thin stratum
corneum at the epidermis, dense dermis layer, hair
follicles, and organized hypodermis. Figure 1B displays
the histological profile obtained for G2, which shows a
thick stratum corneum covering an intense proliferation
of keratinocytes in exophytic profile. Neoplastic invasion
was not evident in all cases, but in these cases the lesions
were considered as potentially malignant (evolving to
squamous cell carcinoma) due to the moderate dysplasia
observed in the epithelial basal layer. After 10 days of
PDT, most cases exhibited residual lesions with histo-
pathological characteristics similar to the original
tumors.
ATR-FTIR Spectroscopy
Figure 2 shows the fingerprint region of ATR-FTIR
spectra of each group, which has been frequently used in
spectroscopic studies of biological tissues due to the
presence of vibrational modes related to important cell
content [8].
Second derivatives of absorbance are shown in Figure 3
and depicts vibrational modes overlapped in the raw
spectra.
All groups presented similar vibrational modes with
subtle differences in amplitude of second derivatives. The
biological assignments of bands depicted in Figure 3 are
shown in Table 2.
Principal Component-Linear Discriminant
Analysis (PC-LDA)
The use of chemometric methods, which are consisted by
application of multivariate statistical techniques for
interpretation of spectroscopic measurements, have been
widely used and fully succeed to characterize biological
tissues [8,9,20]. Among these methods, principal compo-
nent analysis (PCA) is the most used for classification, as
well as for description and interpretation of the data set.
Aiming to increase the efficiency of classification, PCA can
be associated to Linear Discriminant Analysis (LDA),
which provides data classification based on an optimized
criterion for better class separability. In LDA, the
classification criterion is identified using the scatter
measure of within class and between class variance. In
PC-LDA association, PC scores obtained are used as input
data for LDA, resulting in a classification method that
filter out noise and employ optimum variables [26]. Thus,
PC-LDA was applied on second derivatives of ATR-FTIR
spectra and the method was validated by leave-one-out
cross-validation (LOOCV), wherein the model is developed
with an n-1 spectrum, and the left-out spectrum is tested to
verify the model.
 FTIR TO EVALUATE THE EFFECTS INDUCED BY PDT 541

Fig. 1. Light microscopy of representative histological sections hematoxylin–eosin (H&E) stained;

(A) G1 represents healthy skin; (B) G2 is neoplastic lesion precursor of SCC; and (C) G3 is post-PDT
neoplastic lesion.

Figure 4 displays PCA score plots and loadings obtained
for groups compared to each other. Spectra from G1 are
represented by diamonds; G2 is depicted by black circles
and stars are spectra related to G3. Dashed line centered in
zero position of each score plot shows the best principal
component to discriminate the data set. According to
Figure 4A and C, which depicts score plots for G1  G2 and
G1  G3, the first principal component was the best PC to
discriminate the groups and explained 37% and 31% of
global variance, respectively. On the other hand, discrimi-
nation between G2 and G3 was better using the second
principal component, which explains 12.3% of global
variance as depicted in Figure 4E.
Considering the first two principal components as input
on LDA we calculated the accuracy, sensitivity, and
specificity of classification, as shown in Table 3.
For G1  G2, we considered true positive as PC-scores of
G2 spectra in G2 group the negative side of PC-1; true
negative as PC-scores of G1 spectra in G1 group the
positive side of PC-1; false positive as PC-scores of G1
spectra in G2 group the negative side of PC-1; and false
negative as PC-scores of G2 spectra in G1 group the

Fig. 2. Fingerprint region (900–1800 cm 1) of averaged ATR- Fig. 3. Second derivatives of averaged spectra obtained for
FTIR spectra for healthy skin (G1), neoplastic tissue (G2), and healthy skin (G1), neoplastic tissue (G2), and post-PDT neoplastic
post-PDT neoplastic lesions (G3). lesions (G3).
542 LIMA ET AL.

TABLE 2. Biological Assignments for Vibrational Modes Observed in the Second Derivative of Averaged Spectra

Peak Assignment Ref. no.

1,028 Glycogen absorption due to C─O and C─C stretching [12,36,37]

and C─O─H deformation motions
1,082 PO 2 symmetric Glycogen absorption due to C─O and [12,36,38]
C─C stretching and C─O─H deformation motions
1,151 Glycogen absorption due to C─O and C─C stretching [12,36,37]
and C─O─H deformation motions
1,173 C─O stretching (in malignant tissues) [33,36]
1,204 Collagen [28,36,39,40]
1,236 PO 2 assymmetric Collagen [12,28,36,39]
1,282 Collagen [12,28,36,39]
1,338 Collagen [12,28,36,39]
1,517 Amide II [36]
1,624 b-sheet structure of amide I [41–44]
1,650 a-helix structure of amide I [36,42–44]

Fig. 4. PCA score plots and loadings obtained for pairwise comparison between the groups. A, C,
and E display scores for the first two principal components. B, D, and F show loadings associated
with the PC that better discriminated the data in the score plot. Black circles represent Neoplastic
group (G2); Stars display post-PDT neoplastic skin (G3); and diamond represent healthy skin (G1).
 FTIR TO EVALUATE THE EFFECTS INDUCED BY PDT
TABLE 3. Distribution of the Dataset in Groups for the Calculus of the Accuracy of Clustering Classification
True True False
positive negative positive
G1  G2 61 70 0 9
G1  G3 18 70 0 10
G2  G3 25 66 4 3
positive side of PC-1. In case of G1  G3, true positive are
PC-scores of G3 spectra in G3 group the negative side of
PC-1; true negative are PC-scores of G1 spectra in G1
group the positive side of PC-1; false positive as PC-scores
of G1 spectra in G3 group the negative side of PC-1; and
false negative as PC-scores of G3 spectra in G1 group the
positive side of PC-1. Finally, for the case of G2  G3, we
considered true positive as PC-scores of G3 spectra in G3
group the negative side of PC-2; true negative as PC-scores
of G2 spectra in G2 group the positive side of PC-2; false
positive as G2 spectra in G3 group the negative side of
PC-2; and false negative as PC-scores of G3 spectra in G2
group the positive side of PC-1.
Figure 4B, D, and F display loading plots related to
principal components that better discriminated each data
set. Positive scores are associated with negative loadings
and negative scores relate to positive loadings because
PCA technique was applied on second derivatives of ATR-
FTIR spectra [27].
PC-1 loadings obtained with G1  G2 and G1  G3
score plots showed similar loadings, as displayed by
Figure 4B and D. In both cases, vibrational modes related
to glycogen (1,028, 1,082, and 1,151 cm 1), phosphate
stretchings (1,082 and 1,236 cm 1) and collagen fibers
(1,204, 1,236, 1,282, and 1,338 cm 1) presented negative
loadings, which are correlated with positive scores along
PC-1. Thus, it is possible to conclude that G1 presents
larger amount of the cell content related to these
vibrational modes. On the other hand, bands associated
with protein content (Amides I and II) presented an
increase in the spectra of G2 and G3 against G1, due to the
presence of positive loadings at 1,517, 1,624, and
1,650 cm 1, which are correlated with negative scores
along PC-1.
For G2  G3, PC-2 was the principal component that
better discriminated spectral data, and its loadings are
displayed by Figure 4F. In this one, the band peaking at
1,626 cm 1 presented negative loading, consequently, it is
correlated with positive scores from PC-2. Hence, it can be
concluded that G2 contain larger amount of proteins in
b-sheet secondary structure.
DISCUSSION
In the present work, we used ATR-FTIR spectroscopy
associated with PCA to evaluate the biochemical effects
promoted by PDT in cutaneous neoplastic lesions precur-
sors of SCC.
Associating ATR-FTIR and PC-LDA showed good
accuracy on classifying the data set resulting in 93.5%
543
False Accuracy Sensitivity Specificity
negative (%) (%) (%)
93.5 87.14 100
89.7 64.2 100
92.8 89.2 94.28
for G1  G2; 89.7% for G1  G3, and 92.8% in G2  G3.
Comparison of G2 and G3 against G1 showed similar
results, and a decrease in vibrational modes related to the
glycogen and collagen fibers were observed, as well as an
increase in bands associated with protein content.
Mantsch et al. identified the decrease in collagen as
typical for potentially malignant carcinomas [28], which
fits the lesions discussed in this work due to the moderate
dysplasia observed by histopathology. The same behavior
was observed in the spectroscopic characterization of basal
cell carcinoma [29] and was attributed to the degradation
of collagen fibers of extracellular matrix. In fact, some
studies have outlined that certain tumor cells, including
SCC, release specific enzymes that degrade collagen fibers
so that the tumor can spread through surrounding tissue
and metastasize to distant sites [30,31].
The decrease in glycogen on neoplastic tissues is
probably a consequence of this cell content serving as a
source of energy for keratinization of stratum corneum,
proliferation of neoplastic cells and other cell activities
demands.
A number of publications have reported an increase in
symmetric and asymmetric phosphate stretching from
nucleic acids as a consequence of one of the most important
neoplastic cell features: the numerous and uncontrollable
replication of the modified DNA [32–34]. However, the
peak positions of these vibrational modes are also related
to other cell content. Thus, the effect of each structure may
interact to each other and result in the amplification or
reduction of measured absorbance. In this sense, we
suggest that the absorption intensity increase in wave-
numbers assigned to nucleic acids were masked by higher
absorbance values presented by vibrational modes of other
structures absorbing in the same peaks.
Protein overexpression presented by G2 and G3 com-
pared to G1 may be a consequence of intense metabolic
activity of neoplastic cells and its high demand for proteins
performing signaling pathways for proliferation of cancer
cell, oncogenic kinases signaling, transcriptional regula-
tion, and other functions [15]. In addition to this, G2
spectra showed higher value for the sub-band of Amide I
related to b-sheet secondary structure of proteins in
comparison with G3 and was the only one to shift its
peak position, suggesting changes in hydrogen bonding
between peptide groups and consequently in molecular
geometry of proteins.
Similarity obtained for vibrational modes from G2 and
G3 compared to G1 is an evidence that there are still some
active lesions in PDT-treated neoplastic skin, which match
544 LIMA ET AL.
with morphological features observed by histopathology.
These findings suggest that a single-session PDT is not
sufficient to achieve total tumor regression, and it is a fact
that has been shown by other studies in the literature [35].
Similar to other oncological treatments, such as radiother-
apy and chemotherapy, several PDT sessions would be
necessary.
CONCLUSION
In this study, non-treated premalignant skin lesions
and neoplastic tissue submitted to ALA-mediated
photodynamic therapy were evaluated through ATR-
FTIR spectroscopy associated to PC-LDA statistical
techniques. In a pairwise comparison, the vibrational
modes of non-treated and post-PDT neoplastic tissue
revealed similar profile compared with normal skin, in
which was observed a protein overexpression and a
decrease in amount of glycogen and collagen fibers. The
method obtained an accuracy of 93.5% to classify
healthy skin from neoplastic lesions and 89.7% to
differentiate normal tissue from premalignant skin
lesions submitted to PDT. A larger amount of proteins
in b-sheet secondary structure was observed in non-
treated skin lesions in comparison with post-PDT
neoplastic tissue, which was discriminated with an
accuracy of 92.8% by PC-LDA technique. Thus, the
method can be used for early diagnosis of potentially
malignant skin lesions, as well as to evaluate response
to photodynamic treatment.
REFERENCES
1. Gray DT, Suman V, Su D, Clay R, Harmsen H, Roenigk R.
Trends in the population-based incidence of squamous cell
carcinoma of the skin first diagnosed between 1984 and 1992.
Arch Dermatol 1991;133:735–740.
2. Eide MJ, Tuthill JM, Krajenta RJ, Jacobsen GR, Levine M,
Johnson CC. Validation of claims data algorithms to identify
nonmelanoma skin cancer. J Invest Dermatol
2012;132:2005–2009.
3. Yanofsky VR, Mercer SE, Phelps RG, Sinai M, Gustave O,
Place LL. Histopathologica variant of cutaneous squamous
cell carcinoma: A review. J Skin Cancer 2011;1–13.
4. Bonerandi JJ, Beauvillain C, Caquant L, Chassagne JF,
Chaussade V, Clav ere P, Desouches C, Garnier F, Grolleau
JL, Grossin M, Jourdain A, Lemonnier JY, Maillard H,
Ortonne N, Rio E, Simon E, Sei JF, Grob JJ, Martin L.
Guidelines for the diagnosis and treatment of cutaneous
squamous cell carcinoma and precursor lesions. J Eur Acad
Dermatol Venereol 2011;25(5):1–51.
5. Anand S, Ortel BJ, Pereira SP, Hasan T, Maytin EV.
Biomodulatory approaches to photodynamic therapy for solid
tumors. Cancer Lett 2012;326:8–16.
6. Agostinis P, Berg K, Cengel K, Foster TH, Girotti AW,
Gollnick SO, Hahn SM, Hamblin MR, Juzeniene A, Kessel D,
Korbelik M, Moan J, Mroz P, Nowiz D, Piette J, Willson BC,
Golab J. Photodynamic therapy of cancer: An update. CA
Cancer J Clin 2011;61(4):250–281.
7. Mroz P, Yaroslavsky A, Kharkwal GB, Hamblin MR. Cell
death pathways in photodynamic therapy of cancer. Cancers
(Basel) 2011;3:2516–2539.
8. Diem M, Mazur A, Lenau K, Schubert J, Bird B, Miljkovi
c M,
Krafft C, Popp J. Molecular pathology via IR and Raman
spectral imaging. J Biophotonics 2013;6:855–886.
9. Baker MJ, Trevisan J, Bassan P, Bhargava R, Butler HJ,
Dorling KM, Fielden PR, Fogarty SW, Fullwood NJ, Heys K,
Hughes C, Lasch P, Martin-Hirsch P, Obinaju B,
Sockalingum GD, Sul
e-Suso J, Strong RJ, Walsh MJ, Wood
BR, Gardner P, Martin FL. Using Fourier transform IR
spectroscopy to analyze biological materials. Nat Protoc
2014;9:1771–1791.
10. Krafft C, Sergo V. Biomedical applications of Raman and
infrared spectroscopy to diagnose tissues. Spectroscopy
2006;20:195–218.
11. Diem M, Miljkovi
c M, Bird B, Chernenko T, Schubert J,
Marcsisin E, Mazur A, Kingston E, Zuser E, Papamarkakis K,
Laver N. Applications of infrared and Raman microspectro-
scopy of cells and tissue in medical diagnostics: Present status
and future promises. Spectrosc An Int J 2012;27:463–496.
12. Diem M, Griffiths P, Chalmers J. Appendix: Infrared and
Raman spectra of selected cellular. In: Diem M, Griffiths P,
Chalmers J, editors. Vibrational Spectroscopy for Medical
Diagnosis. West Sussex: Wiley; 2008. pp 326–333.
13. Petrich W. Mid-infrared and Raman spectroscopy for medical
diagnostics. Appl Spectrosc Rev 2001;36:181–237.
14. Naumann D. FT-infrared and FT-Raman spectroscopy in
biomedical research. Appl Spectrosc Rev 2001;36:239–298.
15. Lima CA, Goulart VP, C^ orrea L, Pereira TM, Zezell DM. ATR-
FTIR Spectroscopy for the assessment of biochemical changes
in skin due to cutaneous squamous cell carcinoma. Int J Mol
Sci 2015;16:6621–6630.
16. Dorling KM, Baker MJ. Highlighting attenuated total
reflection Fourier transform infrared spectroscopy for rapid
serum analysis. Trends Biotechnol 2013;31:327–328.
17. Kazarian SG, Chan KLA. ATR-FTIR spectroscopic imaging:
Recent advances and applications to biological systems.
Analyst 2013;138:1940–1951.
18. Sukuta S, Bruch R. Factor analysis of cancer Fourier
transform infrared evanescent wave fiberoptical (FTIR-
FEW) spectra. Lasers Surg Med 1999;24:382–388.
19. Afanasyeva NI, Kolyakov SF, Letokhov VS, Sokolov VV,
Frank GA. Diagnostics of cancer by fiber optic evanescent
wave FTIR (FEW-FTIR) spectroscopy. SPIE Proc 1996;2928:
154–157.
20. Silveira FL, Pacheco MTT, Bodanese B, Pasqualucci C,
Z^angaro R, Silveira L. Discrimination of non-melanoma
skin lesions from non-tumor human skin tissues in vivo
using Raman spectroscopy and multivariate statistics. Lasers
Surg Med 2015;47:6–16.
21. Vyumvuhore R, Tfayli A, Manfait M, Baillet-Guffroy A.
Vibrational spectroscopy coupled to classical least square
analysis a new approach for determination of skin mois-
turizing agents’ mechanisms. Ski Res Technol 2014;20:
282–292.
22. Tosi G, Conti C, Giorgini E, Ferraris P, Garavaglia MG,
Sabbatini S, Staibano S, Rubini C. FTIR microspectroscopy of
melanocytic skin lesions: A preliminary study. Analyst
2010;135:3213–3219.
23. Ly E, Piot O, Wolthuis R, Durlach A, Bernard P, Manfait M.
Combination of FTIR spectral imaging and chemometrics for
tumour detection from paraffin-embedded biopsies. Analyst
2008;133:197–205.
24. Abel E, Angel J, Kiguchi K, DiGiovanni J. Multi-stage
chemical carcinogenesis in mouse skin: Fundamentals and
applications. Nat Protoc 2009;9:1350–1362.
25. Mian SA, Colley HE, Thornhill MH, Rehman IU. Develop-
ment of a dewaxing protocol for tissue-engineered models of
the oral mucosa used for Raman spectroscopic analysis. Appl
Spectrosc Rev 2014;49:614–617.
26. Sahu A, Nandakumar N, Sawantb S, Krishna CM. Recur-
rence prediction in oral cancers: A serum Raman spectroscopy
study. Analyst 2015;140:2294–2301.
27. Kozicki M, Creek DJ, Sexton A, Morahan BJ, Wesełucha-
Birczy
nska A, Wood BR. An attenuated total reflection (ATR)
and Raman spectroscopic investigation into the effects of
chloroquine on Plasmodium falciparum-infected red blood
cells. Analyst 2015;140:2236–2246.
28. Jackson M, Watson PH, Halliday WC, Mantsch HH. Beware
of connective tissue proteins: Assignment and implications of
collagen absorptions in infrared spectra of human tissues.
Biochim Biophys Acta 1995;1:1–6.
29. Bodanese B, Silveira FL, Z^ angaro RA, Pacheco MTT,
Pasqualucci CA, Silveira L. Discrimination of basal cell
 FTIR TO EVALUATE THE EFFECTS INDUCED BY PDT
carcinoma and melanoma from normal skin biopsies in vitro
through Raman spectroscopy and principal component
analysis. Photomed Laser Surg 2012;30:381–387.
30. Fundyler O, Khanna M, Smoller BR. Metalloproteinase-2
expression correlates with aggressiveness of cutaneous
squamous cell carcinomas. Mod Pathol 2004;17:496–502.
31. Martins VL, Vyas JJ, Chen M, Purdie K, Mein C, South AP,
Storey A, McGrath J, O’Toole E. Increased invasive
behaviour in cutaneous squamous cell carcinoma with loss
of basement-membrane type VII collagen. J Cell Sci 2009;
122:1788–1799.
32. Simonova D, Karamancheva I. Application of Fourier
transform infrared spectroscopy for tumor diagnosis. Bio-
technol Biotechnol Equip 2014;27:4200–4207.
33. Rigas B, Morgellot S, Goldman IS, Wong PTT. Human
colorectal cancers display abnormal Fourier-transform infra-
red spectra. Proc Natl Acad Sci USA 1991;87:8140–8144.
34. Wood BR. The importance of hydration and DNA conforma-
tion in interpreting infrared spectra of cells and tissues. Chem
Soc Rev 2015; Available online, DOI: 10.1039/C5CS00511F
35. Moton CA, Szeimies R-M, Braathen LR. Update on topical
photodynamic therapy for skin cancer. Vestn Dermatol Vener
2014;6:26–34.
36. Movasaghi Z, Rehman S, Rehman IU. Fourier transform
infrared (FTIR) spectroscopy of biological tissues. Appl
Spectrosc Rev 2008;43:134–179.
545
37. Huleihel M, Salman A, Erukhimovitch V, Ramesh J,
Hammody Z, Mordechai S. Novel spectral method for the
study of viral carcinogenesis in vitro. J Biochem Biophys
Methods 2002;50:111–121.
38. Cull KB, Gehlhausen JM, Viscomi AS, Zhang L, Carnahan JW,
Yamada S, Wasa T, Kimoto T, Okazaki S. Phosphodiester
stretching bands in the infrared spectra of human tissues and
cultured cells. Appl Spectrosc 1991;45:1563–1567.
39. Belbachir K, Noreen R, Gouspillou G, Petibois C. Collagen
types analysis and differentiation by FTIR spectroscopy. Anal
Bioanal Chem 2009;395:829–837.
40. Eckel R, Huo H, Guan HW, Hu X, Che X, Huang WD.
Characteristic infrared spectroscopic patterns in the protein
bands of human breast cancer tissue. Vib Spectrosc 2001;27:
165–173.
41. Diem M, Pereira T, Zezell D, Bird B, Miljk
ovic M. The
characterization of normal thyroid tissue by micro-FTIR
spectroscopy. Analyst 2013;138:7094–7100.
42. Jackson M, Mantsch HH. The use and misuse of FTIR
spectroscopy in the determination of protein structure. Crit
Rev Biochem Mol Biol 1995;30:95–120.
43. Khurana R, Fink AL. Do parallel alpha-helix proteins have a
unique Fourier transform infrared spectrum? Biophys J
2000;78:994–1000.
44. Barth A. Infrared spectroscopy of proteins. Biochim Biophys
Acta 2007;1767:1073–1101.