Preclinical evaluation of diagnosis of actinic keratosis and cutaneous squamous cell carcinoma

The potential value of FDAP as a non-invasive diagnostic procedure to
discriminate between different stages of keratinocytic intraepidermal
neoplasias is described in paragraph 4.2. In this study, FDAP was performed in
patients with field-cancerization after keratolytic pre-treatment. Subsequently
biopsies from different lesions were taken. (Immuno)histochemistry was
performed on these biopsies for histopathological classification, using the
KIN- as well as conventional classification, and assessment of Ki67-antigen
expression and stratum corneum thickness.
Although KIN III lesions tended to have increased lesional:non-lesional
fluorescence ratios compared with KIN I and KIN II lesions, no statistically
significant differences between the different KIN lesions could be
demonstrated. Six lesions classified as verrucous hyperkeratoses had
significantly lower fluorescence ratios (mostly around 1) compared with normal
skin, KIN I, KIN III and miSCC. The three SCCs that were biopsied had
significantly higher fluorescence ratios compared with KIN II lesions and the
verrucous hyperkeratoses. Analysis of proliferative activity as assessed by
immunoreactivity for the Ki67 antigen also did not reveal any significant
differences in fluorescence ratios among the three levels of Ki67 expression,
which is not surprising since Ki67 expression and KIN grade are highly
correlated. However, when the lesions were divided into four diagnostic
categories (true/false positives/negatives), with a fluorescence ratio of 1.0
taken as a cut-off value and KIN lesions/SCC and verrucous
hyperkeratoses/normal skin classified as (pre)malignant and benign lesions,
respectively, lesions with a fluorescence intensity lower than surrounding skin
(fluorescence ratio < 1.0; true and false negatives) were found to have a
significantly lower number of Ki67+ cells than lesions with a fluorescence
ratio greater than 1.0 (true and false positives). Additionally, when Ki67
immunoreactivity from all four diagnostic categories was mutually compared, a
significant difference between the false- and true-positive lesions and between
the true-positive and -negative lesions was found. However, the reason for this
would probably be the relative high number of true positive lesions which were
mostly all hyperproliferative lesions. So, proliferative activity seems to be a
confounder rather than the cause of high fluorescence.
When macroscopic fluorescence values were plotted against the stratum corneum
thickness, a negative correlation was found indicating hyperkeratosis to be an
important hindering factor with respect to PpIX accumulation. Analogously,
lesions with a fluorescence intensity lower than surrounding skin (fluorescence
ratio < 1.0; true and false negatives) had a significantly thicker stratum
corneum than lesions with a fluorescence ratio greater than 1.0 (true and false
positives). Moreover, in the true- and false-positive lesions the mean stratum
corneum thickness was found to be significantly lower compared with the
false-negative lesions. A summary of these data can be found in table II in
paragraph 4.2.
From the above it was concluded that FDAP cannot be used to discriminate
between various KIN-stages. However a tendency towards higher macroscopic
fluorescence was observed in KIN III lesions compared with KIN I/II lesions.
This may be related to an increased number of epidermal cells in KIN III
lesions (epidermal hyperplasia). Although there was a tendency towards an
increased stratum corneum thickness in KIN III lesions compared with KIN I/II
lesions, no statistically significant differences were found. It may be
imagineable that a slight thicker stratum corneum in KIN III lesions could
decrease the fluorescence contrast between the KIN I/II and KIN III lesions.
Furthermore, macroscopic fluorescence seems independent of proliferative status
but it is negatively correlated with stratum corneum thickness. Despite the
fact that the patients in our study received a keratolytic pre-treatment with
salicylic acid in petrolatum for 1-2 weeks prior to FDAP, hyperkeratosis still
seemed to be responsible for variations in macroscopic fluorescence. As
penetration of the hydrophilic ALA through the lipophilic stratum corneum is a
prerequisite for PpIX formation, adequate keratolytic pre-treatment is
essential for optimal results with PDT and FDAP.
Considering the high fluorescence observed in SCC, it might be interesting to
perform further studies whether FDAP can be useful to discriminate between KIN
and invasive carcinoma.

The objective of the present study is to investigate the possbility to
discriminate between actinic keratoses an dcutaneous squamous cell carcinoma
with fluorescence diagnosis.

Fluorescence diagnosis and biopsy procedure
Prior to the day FDAP takes place, the skin areas under study will be
pretreated with 5 or 10% salicylic acid in petrolatum for 1 week, depending on
the clinical picture, to get rid of excess scales and hyperkeratosis until
clinically satisfactory results were achieved. On the day of the FDAP procedure
20% ALA-cream (Medac Gmbh, Wedel; Germany) will be applied under an occlusive
dressing to these skin areas. After 3 hours of incubation fluorescence
intensity on the skin will be recorded using a digital fluorescence imaging
system (DyaDerm, Biocam GmbH, Regensburg; Germany). This system consists of a
flash light (Xenon light source with a custom band pass filter (370-440 nm))
and a 12-bit Sony CCD camera combined in one adjustable arm coupled to a
Pentium IV computer equipped with custom made image capturing software (Dyaderm
Pro v1.4, Biocam GmbH, Regensburg; Germany). The flash light emits 7 light
pulses per second to the skin which are recorded by the CCD camera (exposure
time 100 *sec) equipped with a special Schott GG 455 long pass filter to filter
out the excitation light. As PpIX fluorescence emission consists of light in
the red spectrum, the red pixels of the CCD camera were used to generate a
fluorescence image. In this way a normal coloured image and a fluorescence
image were processed in real-time. Because of the short exposition time to the
excitation light, photobleaching of PpIX was minimized in this way. To correct
for different lighting environments between pictures, a fluorescence reference
standard (Maccal 8044, 738-00, Multifoil, Utrecht; The Netherlands) was
included on every image. Images were recorded in 16-bit greyscale TIFF format.
After FDAP, 4 mm punch biopsies will be taken from selected lesions/skin areas
of interest under local anesthesia with 1% xylocain-adrenalin.

Analysis of fluorescence images
16-bit greyscale TIFF fluorescence images will be imported in NIH ImageJ
software (http://rsb.info.nih.gov/ij/). Because the Xenon light source used for
excitation had the highest intensity in the center of the illuminated area,
shading correction was performed by means of the following algorithm:
with
B = blank image (image from a white homogeneous background recorded with the
Dyaderm system), BImax = highest intensity of blank image, I = (uncorrected)
image, C = normalised shading image, S = shading corrected image.

Histopathology and immunohistochemistry
6 µm slices will be fixed on a glass slide, deparaffinized, hydrated and washed
in PBS subsequently. For Ki67-antigen staining the sections will be first
pretreated in citrate buffer (pH 6.0) using the microwave antigen-retrieval
method. Afterwards, immunohistochemical analysis will be performed after
blocking for endogenous peroxidase using the Powervision (Immunologic, Duiven,
The Netherlands) staining system with diaminobenzidine (DAB) as chromogen. The
sections will be incubated for one hour with the primary MIB-1 antibody (Dako,
Heverlee, Belgium; 1:100) directed against the cell-cycle associated antigen
Ki67. For counterstaining Mayer*s haematoxylin will be used.
Also haematoxylin-eosin staining will be performed on every slide for
assessment of the histopathological diagnosis and KIN-classification. All
slides will be assessed by one and the same pathologist (WB) for uniformity.

Immunohistochemical and digital image analysis
Immunoreactivity for the Ki67-antigen will be scored semi-quantitatively in the
following manner according to Keating et al.13: 0 = only basal layer
positivity, 1 = positivity confined to basal 1/3 of the epidermis, 2 =
positivity confined to basal 2/3 of the epidermis, 3 = transepidermal positive
staining.
Analysis of digital microscopic images will be done using ImageJ digital image
analysis software. Digital photographs of the HE-sections were made at 50x
magnification. To determine the thickness of the stratum corneum the average
thickness will be calculated as total stratum corneum surface including
hyperkeratosis per millimeter length of the stratum corneum in different biopsy
sections.


Statistical analysis

To analyze Ki67-expression and KIN-grade in relation to fluorescence intensity
one-way analysis of variance (ANOVA) will be used. For analysis between groups
Duncan*s post-hoc test will be performed. Analysis of fluorescence intensity in
AK and BD, Ki67-expression and stratum corneum thickness between the 4
diagnostic categories will be performed using an unpaired Student*s t-test.
Correlation analysis of macroscopic fluorescence and stratum corneum thickness
will be performed using Pearson*s R. All statistical calculations will be
performed using Statistica 6.0 s software (Statsoft Inc.,
http://www.statsoft.com) and Microsoft Excel 2000. A p-value < 0.05 will be
considered statistically significant.

The patients will be in hospital during 5 hours. At first the cream will be
applied and the patients will have to wait during 3 hours. Afterwards the
cream will be removed and the fluorescence intensity will be measured as
described above. From a maximum of four lesions 4 mm biopsies will be taken.
Part of this procedure is standard. Only the application of the cream, the
waiting and measurements of the fluorescence intensity is extra.