A Photonumeric Scale for the Assessment

of Cutaneous Photodamage
Christopher E. M. Griffiths, MD, MRCP; Timothy S. Wang;
Ted A. Hamilton, MS; John J. Voorhees, MD; Charles N. Ellis, MD

\s=d\ Background and Design. \p=m-\Theassessment of the sever- clinical skill and, indeed, often goes unrecognized, be¬
ity of cutaneous photodamage and its response to treatment ing considered merely accelerated intrinsic aging.
is an impractical consideration for most practitioners without The demonstration of the ability of tretinoin therapy
extensive experience or recourse to high-quality, standard- to reverse some of the features of photodamageM has
ized, baseline photographs. To address this problem, a nine- inspired escalating interest in the use of this com¬
an
point photonumeric standard scale was developed using pound and other skin repair agents in the treatment of
photographs of subjects representing grades of photo- cutaneous photodamage. In most studies, the assess¬
damage from none to severe. This scale was formally tested ment of the salient features of photodamaged skin,
in a side-by-side comparison with a conventional and widely namely, coarse and fine wrinkles, dyspigmentation
used written descriptive scale. A panel of seven graders used (lentigines), elastoidosis, roughness, comedones, telan-
both scales to score two sets of 25 photographs of photo- giectasia, laxity, and sallowness, has been achieved
damaged individuals, and the intergrader agreement and with the use of purely descriptive written scales. It has
repeatability for the scales were calculated. been demonstrated previously that the evaluation of
Results. \p=m-\Thephotonumeric scale demonstrated signifi- acne severity and subsequent response to treatment is
cantly greater agreement between graders than did the de- easier to perform with the aid of photographic stan¬
scriptive scale (chance-corrected agreements of 0.31 and dards illustrating the various grades of disease sever-
0.11, respectively, P<.0001) with no significant difference in
repeatability between the two methods.
Conclusions.\p=m-\This study demonstrates that the photonu- See also pp 352,357, and 390.
meric standard scale is superior to existing methodology in
the accurate assessment of cutaneous photodamage and
would be a useful adjunct to studies of the efficacy of skin
ity.5,6 In studies performed at the University of Michi¬
gan, Ann Arbor, the response of cutaneous photo-
repair agents for this indication. damage to treatment is assessed by comparing the
(Arch Dermatol. 1992;128:347-351) subject with photographs taken at baseline. The tech¬
nology and expertise required for standardization of
baseline photography is immense, particularly when an
is increasing that photodamage accurate portrayal of wrinkling and skin color is neces¬
There
able
or
be
an awareness
extrinsic aging of skin is preventable and treat¬
a
from chronologic in¬
sary. Outside of a few departments with dedicated
entity to distinguished or photographic units, this is an impractical proposition.
trinsic aging. However, the diagnosis and evaluation of On this premise we have adapted the acne photographic
cutaneous photodamage is not a traditionally taught grading scale technique and have produced a similar
photonumeric grading system relevant to the assess¬
Accepted for publication August 2,1991. ment of cutaneous photodamage. The resulting scale
From the Dermatopharmacology Unit, Department of Dermatolo- based on high-quality photographic standards was test¬
gy, University of Michigan Medical Center, Ann Arbor. ed for both intergrader agreement and repeatability
Reprint requests to Department of Dermatology, University of and proved superior to a purely descriptive grading
Michigan Medical Center, 1910 Taubman Center, Ann Arbor, MI
48109-0314 (Dr Griffiths). system.

Downloaded From: https://jamanetwork.com/ by a Johns Hopkins University User on 03/06/2023

Downloaded From: https://jamanetwork.com/ by a Johns Hopkins University User on 03/06/2023
 Fig 1 .—Photonumeric scale used for the assessment of
cutaneous photodamage. Five photographic standards
{en face and 45° oblique) illustrating increasing severity
of photodamage, where grade 0 represents no damage;
grade 2, mild damage; grade 4, moderate damage;
grade 6, moderate/severe damage; and grade 8, severe
damage. (Copyright© 1991 ; Regents of the University of
Michigan, Ann Arbor.)

Fig 2. —Example of photodamage in a Far East Asian, a 72-

year-old Japanese man with minimal wrinkling but marked
dyspigmentation.

Downloaded From: https://jamanetwork.com/ by a Johns Hopkins University User on 03/06/2023

 SUBJECTS AND METHODS
Table 1.—Descriptive Scale Used in the
Assessment of Cutaneous Photodamage* From a pool of more than 200 facial photographs (en face
and 45° oblique) taken by professional photographers in our
1. Fine Wrinkling dermatopharmacology photographic unit, five photographic
This factor represents a visual assessment of the number standards were selected by us as representative examples of
and depth of superficial wrinkles (ie, shallow indentations
or lines). Fine wrinkles typically appear in periorbital and
photodamage and assigned grades 0, 2, 4, 6, and 8, thus
perioral regions and are usually found further from the making a nine-point scale where 0 is no photodamage and 8 is
severe photodamage (Fig 1). The standards were selected on
eyes and mouth than are coarse wrinkles. their ability to adequately portray the nontactile and easily
2. Coarse Wrinkling
This factor represents a visual assessment of the number and photographable factors of photodamage, ie, coarse and fine
depth of coarse wrinkles (ie, deep lines, furrows, or creas¬ wrinkles, dyspigmentation, elastoidosis, comedones, and tel-
es).Coarse wrinkles appear on the forehead, glabella, angiectasia. Only five standard photographs were used to
chin, and nasolabial and periorbital areas, and they tend to illustrate the spectrum of photodamage due to the difficulty in
be located closer to the eyes and mouth than fine wrinkles.
3. Mottled Hyperpigmentation
accurately delineating on film each intermediate point, ie,
This factor represents a visual assessment of light, patchy,
grades 1, 3, 5, and 7. All subjects had signed a photographic
consent approved by the University of Michigan Institutional
mottled hyperpigmentation and solar freckling (including
Review Board.
melasma) based on quantitative and qualitative criteria such To test the agreement and repeatability of the photonu¬
as the area/density of pigment, color intensity (dark vs light),
and uniformity of distribution (ie, the more uneven or meric scale, it was compared with a widely used'8 descriptive
blotchy, the greater the score), Lentigines, nevi, and other scale (Table 1). Agreement, in this context, is the degree to
pigmented lesions are not to be included in this assessment. which two or more clinical investigators or graders will inde¬
4. Yellowing (Sallowness) pendently give a subject an identical severity score. Repeat¬
This factor represents a visual assessment of color tone from ability, however, refers to the ability of each investigator to
very pink or rosy to very sallow or pale. exactly reproduce their original score at a subsequent time,
Each parameter is assessed and the overall severity is rated
as follows: presumably having allowed enough time to elapse so that
Rating Category memory is not a likely factor. A panel of seven experienced
0 None graders was used to perform the experiment. The graders
1-3 Mild were asked to grade one set of photographs of 25 subjects
4-6 Moderate using the photonumeric scale and a second set of 25 different
7-9 Severe subjects using the descriptive scale. None of the photographs
'Adapted with permission from the R. W. Johnson Pharmaceu¬ used in the standards (Fig 1) were used in the sets of photo¬
tical Research Institute, Raritan, NJ. graphs tested. The intergrader chance-corrected agreement
was derived from these readings. One week later, the same
panel regraded the same photographs using the photonu¬
meric and descriptive scales and the results used to determine
Table 2.—Comparing Intergrader Agreement in the the repeatability.
Evaluation of 25 Photoaged Subjects Using a
The generalized k coefficient,9 a chance-corrected measure
Photonumeric vs a Descriptive Grading Scale of agreement ranging from -1 (complete disagreement) to
Graders in Photonumeric* Descriptive! +1 (complete agreement) was determined for each of the two
Agreement Scale Scale sets of results and compared statistically with the z test.
All 7 Repeatability, ie, the ability of a grader to duplicate his/her
own score at a subsequent time, was assessed by first comput¬
ing the absolute differences of each grader's scores between
the first and second scoring periods and summing these differ¬
ences across all seven graders. Thus, 25 summed differences
Total >4 20 were generated for each of the two sets of photographs evalu¬
Overall agreement_OAT_0.26 ated by the two scoring methods. The resulting means of the
Chance-corrected agreement 0.31 0.114= two groups were then compared statistically with the two-
'Photographs of 25 subjects evaluated with the photonumeric sample t test.
scale shown in Fig 1.
RESULTS
fPhotographs of 25 different subjects evaluated with the de¬
scriptive scale shown in Table 1. Agreement
%P < .0001 (z test)
Using the photonumeric scale, four or more of the
grading panel agreed on the grade of photodamage in 20
of 25 subjects, compared with only nine of 25 for the
Table 3.—Comparison of Repeatability Between
Photonumeric and a Descriptive Grading Scale
a
descriptive scale (Table 2). A chance-corrected agree¬
ment of 0.31 for the photonumeric scale and 0.11 forthe
Photo¬ descriptive scale was significantly different (P<.0001).
numeric Descriptive P One week later, there was still a significant difference
Scale Scale Value (P<.008) between the chance-corrected agreements in
Mean sum of differences* favor of the photonumeric scale.
for each subject
across all graderst 5.0 0.4 4.3 ± 0.3
± .15$ Repeatability
'Differences indicate absolute difference in scores between first
and second grading period. On repeating the experiment 1 week later, no signifi¬
tMean ± SEM, n - 25. cant difference was noted in intergrader assessment or
¿Two-sample ftest. repeatability between the two scales (Table 3).

Downloaded From: https://jamanetwork.com/ by a Johns Hopkins University User on 03/06/2023

 Use of the Scale We have also observed increased dyspigmentation at
the expense of wrinkling in sun-exposed Asian skin (Fig
To grade subjects using the photonumeric scale they
must be placed under optimal lighting conditions, usu¬ 2). In the Far East, notably Japan, mottled hyperpig¬
mentation as opposed to wrinkling is a feature of major
ally in natural daylight or fluorescent lighting. Makeup concern most probably as it is a more prominent expres¬
must be removed and the face washed. As washing will
sion of photodamage in this group. Goh10 has reported
produce a degree of erythema, a period of 10 minutes previously on this particular observation. We concur
should elapse before grading. A direct comparison is
with his assessment and add the caveat that the photo¬
then made between the subject and the photographic
numeric scale developed for this study should not be
standards. If an exact match cannot be made, then the
used in the evaluation of photodamage in Far East
interstandard numbers may be used, ie, 1, 3, 5, or 7. In Asians. Indeed, a separate scale should be developed
some subjects who have many lentigines, the increased
for this group.
dyspigmentation may be used to augment a grade but The demonstration that this photonumeric scale is a
its absence should not be used to reduce a grade arrived
reliable tool in the assessment of photodamage allows
at by other factors such as fine and coarse wrinkles.
its use for a variety of indications. Probably its main
Similarly, the presence of telangiectasia is not a patho- role would be in categorizing groups of photodamaged
gnomonic sign of photodamage and should not be used subjects prior to treatment with skin repair agents,
as a major assessable criterion. thereby obtaining consistent and uniform collections of
COMMENT subjects, eg, those with mild or severe photodamage.
This would also allow greater reliability among centers
This study demonstrates the superior interobserver involved in photoaging studies and enable independent
agreement of a photonumeric scale when compared
regrading of good-quality study photographs. We also
with an existing, written, descriptive scale used in the
assessment of photodamaged facial skin. The five pho¬ envisage a need for such a photonumeric scale in epide-
tographic standards of en face and 45° oblique facial miologic studies with reference to geographies and sun
views provide an easy-to-use evaluation system for the exposure.
In conclusion, this approach to the clinical quantita-
physician involved in the assessment and treatment of tion of photodamage represents an improvement on
photodamaged skin. available methodology. However, it is still a semiquan¬
Although there is good interobserver agreement titative technique, with the ultimate goal being a stan¬
with the photonumeric scale, it appears that even an dardized quantitative method.
experienced observer's initial assessment of photo-
damage may vary from that of other observers. Howev¬ This investigation was supported in part by the Babcock Dermato¬
er, in view of the finding that there is no significant
difference in repeatability between the two scales, this logie Endowment, Ann Arbor, Mich, and the R. W. Johnson Pharma¬
ceutical Research Institute, Raritan, NJ.
grading would remain constant on subsequent Harrold Carter and Randal Stegmeyer provided photographic as¬
occasions. sistance and Dale Yessian provided secretarial assistance.

References

1. Kligman AM, Grove GL, Hirose R, Leyden JJ. Topical tret- 6. Burke BM, Cunliffe WJ. The assessment of acne vulgaris: the
inoin for photoaged skin. J Am Acad Dermatol. 1986;15:836-859. Leeds technique. Br JDermatol. 1983;111:83-92.
2. Weiss JS, Ellis CN, Headington JT, Tincoff T, Hamilton TA, 7. Weiss JS, Ellis CN, Goldfarb MT, Voorhees JJ. Tretinoin thera-
Voorhees J. Topical tretinoin improves photoaged skin in a double- py: practical aspects of evaluation and treatment. Int Med Res.
J

blind, vehicle-controlled study. JAMA. 1988;259:527-532. 1990;18(suppl 3):41c-48c.

3. Leyden JJ, Grove GL, Grove MJ, Thorne EG, Lufrano L. 8. Weinstein GD, Nigra TP, Pochi PE, et al. Topical tretinoin for
Treatment of photodamaged skin with topical tretinoin. J Am Acad the treatment of photodamaged skin: a multi-center study. Arch
Dermatol. 1989;21:638-644. Dermatol. 1991;127:659-665.
4. Lever L, Kumar P, Marks R. Topical retinoic acid in the treat- 9. Cohen J. A coefficient of agreement for nominal scales. Educ
ment of solar elastotic degeneration. Br JDermatol. 1990;122:91-98. Psychol Meas. 1960;20:37-46.
5. Cook CH, Centner RL, Michaels SE. An acne grading method 10. Goh SH. The treatment of visible signs of senescence: the
using photographic standards. Arch Dermatol. 1979;115:571-575. Asian experience. Br J Dermatol. 1990;22(suppl 35):105-109.

Downloaded From: https://jamanetwork.com/ by a Johns Hopkins University User on 03/06/2023