ARTICLE
Auteur(s) : Fatma
Aydin1, Nilgun Senturk1, Bunyamin
Sahin2, Yuksel Bek3, Esra Pancar
Yuksel1, Ahmet Yasar Turanli1
1Department of Dermatology, Ondokuz Mayis University,
Medical Faculty of Ondokuz Mayıs University 55139 Kurupelit,
Samsun, Turkey
2Department of Anatomy, Ondokuz Mayis University,
Samsun, Turkey
3Department of Biostatistics, Ondokuz Mayis University,
Samsun, Turkey
accepté le 2 Septembre 2006
It is important to determine the size of vitiligo surface areas and
the degree of treatment response in vitiligo patients. In studies
on vitiligo, several methods are described for assessing the
surface area of the lesion. In some, a repigmentation percentage
had been presumed empirically by the clinician during the
examination and a score given [1]. However, rough assessments
without any objective measurements might lead to incorrect
evaluations. In other studies, the response to treatment was
compared using digital photography and planimetry [2, 3]. It is
also stated however, that this approach is time consuming and
inaccurate and does not facilitate clinical decision-making. On the
other hand, evaluation with digital photography measurements is not
accurate when the lesion is three-dimensional [4].Digital
photography with computerized planimetry is a reliable and accurate
method and provides a digital image record. This process is known
as image digitization. Each pixel in the image is individually
sampled, and its brightness measured and quantified. This integer
value is stored in the corresponding pixel of the computer’s bitmap
image. The area that defines the object’s outline is reported as a
polygon area and in order to measure this, certain computer
programmes using this digitization technique can be employed, one
of which is the image-pro plus 4.5 analysis processing programme
[5, 6].Point counting is a simple, accurate and practical
technique, which is widely used in the estimation of the
irregularly shaped sectional surface area to obtain the volumes of
organs or structures using macroscopic, microscopic or radiological
images [7-9]. We have not, however, found a study describing the
estimation of lesion surface area in vitiligo using a point
counting approach. The present study was carried out to describe a
simple, accurate and practical technique for estimating the sizes
of vitiligo areas by the point counting method and to compare this
method with digital planimetry measurements as the gold standard of
the study.
Methods
A total of 31 vitiligo areas from five patients (mean age ± SD,
31.6 ± 17.7 years) were evaluated.
Point counting procedure
The borders of the lesions were marked with an ordinary ballpoint
pen and a piece of paper was immediately placed over the lesion as
described before [10]. For each lesion, the copied borders of
projection areas were enhanced by redrawing the contours with a
pen. To estimate the number of points, transparent sheet that has
points (+) on it was randomly superimposed on lesion projection
area. The numbers of intersections (i.e. upper right corner of the
crosses) hitting the area of interest were counted ( (figure 1) ). The total
area of each lesion was estimated by multiplying the representative
area of a point on grid by total number of points counted for the
lesion. The representative area of the points (a (p)) in grid was
0.1 cm2.
Three independent observers estimated the surface area of the
same projection drawings using the same point counting grid with
the representative area given above. Surface area estimations were
repeated after two weeks by each observer, in order to analyse
intra-observer variability.
Image pro plus-4.5 procedure
Since the reliability of the point-counting method had been
investigated using the figures drawn on paper the same figures were
scanned to compare the surface area estimations using the digital
planimetry method. All drawings were scanned and stored in “jpeg”
format with true colour class and 300 × 300 pixel resolution then
converted to black and white format because the frame of areas in
our study were important. In order to convert the area from pixel
to cm2, we attached a square with 1 cm2
area and counted the number of pixels in this area to calibrate the
area measurement of the vitiligo lesion. Surface areas of the
lesion and a 1 cm2 drawn area were measured by
image-pro plus 4.5 analysis processing programme on the basis of
planimetry.
Statistical analyses
Statistical analyses were carried out using SPSS 13.0 for Windows
(Release 13.01, license code 9071653). The correlation and
regression analyses were performed to evaluate the data obtained by
all measurement techniques. The regression coefficient was also
expressed as 95% confidence limits.
The estimation results of three observers were compared
statistically to evaluate inter-observer variations using
inter-observer correlation analysis test. The intra-observer
reliability was assessed by comparing coefficients of
intra-observer correlation. The estimation results of the point
counting method and image-pro plus 4.5-analysis processing, which
was accepted as a gold standard, were compared for each observer
separately. The mean areas obtained by different measurement
techniques were compared using Wilcoxon Signed Ranks Test.
Results
The mean vitiligo surface areas obtained using point counting
techniques by three observers in the two point counting sessions
(with a two-week interval) was 4.70 ± 0.76 (min-max: 0.69-16.26)
cm2. The mean vitiligo surface area obtained by the
planimetry technique was 6.26 ± 1.71 (min-max: 0.50-16.01)
cm2 (mean ± SEM).
The estimation results of three observers were compared to
assess inter-observer variation using correlation analysis test.
There was significant agreement between each observer’s estimates
using the point counting method.
The estimation results of the same observers at a 2-week
interval were analyzed statistically to check intra-observer
variation using correlation analysis test. Using intra-observer
correlations, a high degree of agreement was found for the results
estimated using the point counting method.
The estimation results obtained by the point counting method and
measuring results obtained by image-pro plus 4.5-analysis
processing were compared to assess the accuracy of the point
counting method. There was a significant agreement between the
observers’ estimates. Inter-observer, intra-observer and
methodological agreements for the first and second area
measurements are summarized in table 1( Table
1 ).
The comparisons among the estimates of the point counting
approach and between the results obtained from point counting and
planimetry techniques were performed using Willcoxon Signed Rank
Test. No statistical difference was found between the estimated
surface areas using the two methods (p > 0.05).
Table 1 Inter-observer, intra-observer and
methodological agreements for the first and second area
measurements
|
Agreements
|
Relations Between
|
r
|
|
Inter-observer agreement for 1st Measurement
|
O1-O2
|
0.999**
|
|
O1-O3
|
0.999**
|
|
O2-O3
|
0.997**
|
|
Inter-observer agreement for 2nd measurement
|
O1-O2
|
0.999**
|
|
O1-O3
|
0.992**
|
|
O2-O3
|
0.998**
|
|
Intra-observer agreement between 1st and 2nd
measurements
|
M11-M12
|
0.999**
|
|
M21-M22
|
0.999**
|
|
M31-M32
|
0.997**
|
- Agreement between point counting and image-pro plus
- for 1st measurement
|
O1-C
|
0.994**
|
|
O2-C
|
0.993**
|
|
O3-C
|
0.993**
|
- Agreement between point counting and image-pro plus
- for 2nd measurement
|
O1-C
|
0.996**
|
|
O2-C
|
0.994**
|
|
O3-C
|
0.995**
|
Discussion
Several methods such as digital photography, planimetry, and
clinical evaluation by the clinician have been used to measure
vitiligo surface areas [1-3]. Photographic measurement of size
involves the use of a camera equipped with a macro lens. A
planimeter or digitizing tablet is used to calculate the surface
area. A major advantage of photographic measurement is the
provision of a permanent visual record of the size. However, a
number of technical limitations may reduce the effectiveness of
digital photography in vitiligo. For example, measurement precision
is compromised when the distance between the camera and the surface
is inconsistent. Failure to place the camera at exactly the same
distance for each photograph may create the impression that the
size of the lesion has changed when it has not. Similarly, the
camera angle in relation to the size may affect the precision of
the photographic measurement. If an instant camera is not used, the
need to develop and process the film before measurements can be
made and the uncertainty about the success of the photograph (e.g.,
possible overexposure or underexposure) contributes to challenges
in obtaining data. On the other hand, the cost and amount of
equipment and time required to use computer-assisted planimetry
restricts its clinical usefulness in daily practice [11-13].
Our results showed that there was significant agreement between
the each observer’s estimates using the point counting method and
image-pro plus 4.5-analysis processing. The point counting method
showed statistically significant reliability for measurement of
surface areas. Moreover, the method described in this study
represents a simple, practical, and inexpensive technique to
evaluate the vitiligo surface area. Good agreement was found
between the results obtained by three independent observers for the
point counting method. The estimates of each observer were closely
correlated to the others’ estimates.
This study was also designed to evaluate inter- and
intra-observer reliability and accuracy of point counting method
for vitiligo lesion surface area measurement. For practical
reasons, we prefer to analyse the drawings with point counting and
image pro-plus analysis. This is the only study in which the
outcome of both methods could be compared to the true size of the
lesion. Moreover, the design allowed a valid comparison of the
outcome of measurements in different body sites.
The estimation problem which was encountered when measuring
small areas, can be resolved by decreasing the scales of the grids.
Since the application of the small grid increases sensitivity, it
might be advantageous, especially in the case of regular geometric
figures with indefinite borders. Sensitivity increases when the
scales of the grids decrease; it does, however, cause an increase
in point counting time. Appropriate scale choice can be made
according to the size of the lesion. Empirically, we may suggest
the researcher to choose a point counting grid that is suitable for
counting at least 20-25 points per lesion.
Our purpose was to develop an easy way to assess the total
surface area of vitiligo lesions without changing the routine
procedure in daily practice. This method can also be used for
measuring the size of pathological skin lesions such as psoriasis,
ulcers, alopecia, burn wounds, or scars in which accurate
evaluation is essential for clinical follow-up and research. The
values obtained using the method described in this study are
notably independent from observers. Moreover, the method is
inexpensive and fast, since point counting can be carried out
within 1-2 minutes per subject. However, the need for
well-delineated lesions of a measurable size may restrict the
application of this method. This handicap can be overcome by
choosing reference lesions to evaluate patients. The described
method here could be used as a criterion for the evaluation of
treatment or the follow-up process of vitiligo lesions.
Acknowledgements
Financial support: None. Conflict of interest: None.
References
1 Chiaverini C, Passeron T, Ortonne JP. Treatment of
vitiligo by topical calcipotriol. J Eur Acad Dermatol Venereol
2002; 16: 137-8.
2 Yalçin B, Şahin S, Bukulmez G,
Karaduman A, Atakan N, Akan T, Kolemen F.
Experience with calcipotriol as adjunctive treatment for vitiligo
in patients who do not respond to PUVA alone: A preliminary study.
J Am Acad Dermatol 2001; 44: 634-7.
3 Westerhof W, Nieuweboer-Krobotova L. Treatment of
vitiligo with UV-B radiation vs topical psoralen plus UV-A. Arch
Dermatol 1997; 133: 1525-8.
4 Van Geel N, Vander Haeghen Y, Ongenae K,
Naeyaert JM. A new digital image analysis system useful for
surface assessment of vitiligo lesions in transplantation studies.
Eur J Dermatol 2004; 14: 150-5.
5 Image-Pro Plus including Software, Program Media and
Documentation. Media Cybernetics, Inc. USA.
6 Francisco JS, Moraes HP, Dias EP. Evaluation of
the Image-Pro Plus 4.5 software for automatic counting of labeled
nuclei by PCNA immunohistochemistry. Pesqui Odontol Bras 2004; 18:
100-4.
7 Sahin B, Emirzeoglu M, Uzun A, Incesu L,
Bek Y, Bilgic S, Kaplan S. Unbiased Estimation of
the liver volume by the Cavalieri Principle using magnetic
resonance images. Eur J Radiol 2003; 47: 164-70.
8 Odaci E, Sahin B, Sonmez OF, Kaplan S,
Bas O, Bilgic S, Bek Y, Ergur H. Rapid
estimation of the vertebral body volume: a combination of the
Cavalieri principle and computed tomography ımages. Eur J Radiol
2003; 48: 316-26.
9 Sahin B, Aslan H, Unal B, Canan S,
Bilgic S, Kaplan S, Tumkaya L. Brain volumes of the
lamb, rat and bird do not show hemispheric asymmetry: A
stereological study. Image Anal and Stereol 2001; 20: 9-13.
10 Tuzun Y, Yazici H. A method of measuring skin
lesions. Arch Dermatol 1981; 117: 192.
11 Lucas C, Classen J, Harrison D, De H.
Pressure ulcer surface area measurement using instant full-scale
photography and transparency tracings. Adv Skin Wound Care 2002;
15: 17-23.
12 Lagan KM, Dusoir AE, Mc Donough SM,
Baxter GD. Wound measurement: the comparative reliability of
direct versus photographic tracings analyzed by planimetry versus
digitizing techniques. Arch Phys Med Rehabil 2000; 81: 1110-6.
13 Griffin JW, Tolley EA, Tooms RE, Reyes R,
Clifft JK. A comparison of photographic and transparency-based
methods for measuring wound surface area. Phys Ther 1993; 73:
117-22.
|
Version imprimable
Version PDF
