The assessment of cutaneous sensitivity to ultraviolet (UV) radiation is an important process of diagnosing photosensitivity disorders and establishing phototherapy protocols. Traditionally, the minimal erythema dose (MED) has been regarded as a key factor to evaluate the degree of photosensitivity and to determine the starting doses for phototherapy1. Narrowband UVB (NBUVB) phototherapy has been widely used in the treatment of various skin diseases, such as psoriasis, vitiligo, and atopic dermatitis, and it has excellent treatment outcomes2. The protocols for NBUVB phototherapy usually consist of the initial starting doses and incremental doses based on the MED values or skin phototypes. Although some authors developed phototherapy protocols based on the correlation between the MED range and skin type in patients receiving NBUVB phototherapy3, the patients may encounter some problems, such as deep erythema responses or burn reactions. Moreover, considering the fact that most Koreans have darker complexions (skin types III and IV) and show seasonal variations in facultative skin colors, MED determination by naked eyes can be a relatively subjective procedure, and an overestimated value can cause adverse reactions during NBUVB phototherapy.

However, MED measurement on skin erythema sites is a subjective process because the emitting doses from light sources are arbitrarily allocated by operators and the MED value is not measured, but merely recorded. Until recently, there have been only a few reports regarding objective methods to measure erythema responses, especially in MED evaluations. Since the Commission Internationale de I'Eclair age (CIE) recommendation for color scale was published in 1976, spectrophotometers have adopted the CIE L^*a^*b^* system, in which the L^* value gives the relative lightness ranging from total black (L^*=0) to total white (L^*=100); the a^* value represents the balance between red (positive value) and green (negative value); and the b^* value represents the balance between yellow (positive value) and blue (negative value)4. Besides the 3 parameters mentioned above, delta E numerically represents the difference between 2 colors, and it could be obtained by the application of the following formula: Δ E=[(ΔL^*)²+(Δa^*)²+(Δb^*)2]^1/2. The ranges indicated the degree of difference as: 3.0 to 6.0, appreciable difference; and 6.0 to 12.0, much difference5.

Although many previous studies reported that spectophotometric values are useful to assess skin types, skin colors, and cutaneous sensitivity to phototherapy, they described only limited spectophotometric values, such as L^* value, a^* value, and b^* value.

Therefore, to determine a new meaning of the relationships among spectrophotometric values, we examined any difference between the MED values and the measured L^*a^*b^* values in phototesting sites of patients receiving NBUVB phototherapy. The correlations between MEDs and various spectrophotometric parameters will be discussed.

In a spectrophotometer, artificial light is delivered to a target and the reflected light is sent to a diffraction grating that is separated by waves. The reflexibility value of the separated light is measured by photodiode array detector and integrated to the formed reflexibility value curve. The values obtained from the spectrophotometer can also be converted into spectrophotometric values in the L^*a^*b^* system according to the regulation of the CIE, and various skin reactions, such as erythema and tanning. These are analyzed through the plotting of 3 values in the plane of separate axes. The spectrophotometer detects subtle changes in skin colors, which are almost undetectable with the naked eye. Moreover, errors in the level of erythema and pigmentation of the skin caused by the subjective determination of examiners could be overcome by the application of the spectrophotometer because the evaluation of cutaneous sensitivity of UV radiation is such an important factor in the setting of phototherapy6. Fitzpatrick's classification for sun-reactive skin types is based on the history of an individual's tendency to sunburn and to tan. Westerhof et al.7 found that constitutional skin color objectively measured by a chromameter is a better predictor of skin response to UV, and the steepness of the dose-response curve for erythema is a better measurement of the skin response to UV irradiation.

There have been several studies that investigated the correlation between skin types and MED ranges. Andreassi et al.8 reported that a good correlation was found between MEDs and skin types, in spite of a variation in MED ranges according to the skin types. In clinical situations, the ranges of MED values do not reflect the skin types in Fitzpatrick's classification system. There have been many studies for MED determination in a more objective way using a spectrophotometer. Many important facts in the spectrophotometric aspect are clinically accepted about various parameters, such as L^*a^*b^*. In general, the a^*value shows a significant linear correlation with the dermatologist's perception of erythema, and the b^*value shows a significant correlation with the perception of tannning9. Clarys et al.10 reported that the L^*value decreases and the b^* value increases depending on the degree of tanning. Youn et al.6 examined a correlation between MEDs and colorimeteric values measured in constitutive skin color sites and found that the L^* value decreases significantly as the MED increases. They pointed out that the objectively measured skin color was useful for predicting skin sensitivity to NBUVB.

In this study, we enrolled the subjects only if the delta E was >3.0 (ΔE, 3.0 to 12.0) because various parameters could be more valuable when the delta E value was >3 in the spectrophotometric measurement, as we reported previously5. To find out any valuable facts about erythema, pigmentation, and tanning, especially in MED sites of different skin-type subjects, we first compared L^*a^*b^* values separately according to skin types to obtain gross differences in each of various parameters. Second, we compared MEDs and each of the spectrophotometric values (L^*or a^*) to find their correlation. Third, we compared b^*/L^* values for tanning status in different skin types. Last, we investigated the correlation to each of the other spectrophotometric values and a correlation between a^* and b^*/L^* to clarify a relationship between erythema response and tanning status.

Considering our results, the higher b^* values in type IV skin individuals reflect that skin tanning develop more prominently in type IV persons than in type III, because the skin pigmentation status was formed in a cumulative manner even 24 hours after NBUVB irradiation. This result is in accordance with the previous results in which the b^* value would be a good indicator of tanning1. Our results about the correlation between MEDs and b^* values may signify that the skin pigmentation status is deepened with the higher MEDs.

Clarys et al.10 reported that L^*values decrease and b^*values increase when the degree of tanning is higher. Thus, we examined the difference of b^*/L^* values between skin type III and skin type IV. The average b^*/L^* value in the type IV group was significantly higher than in the type III group. The a^* value is well known as an index of erythema, and our data showed that the b^*/L^* value is thought to be an index of tanning. Therefore, we investigated any relationship between the a^* value and the b^*/L^* value to determine a correlation between erythema response and tanning ability in all patients and subgroups divided by skin types or delta E. Although an inverse correlation between the L^* value and a^* value has already been mentioned, we did not find a correlation between the a^* value and b^*/L^* value. Although Westerhof et al.11 reported that the a^* value reflects the perceptive erythema or MED measured by the optical method, the a^* values measured in MED sites did not show any remarkable pattern in this study. Therefore, although the MED determination by naked eyes may be influenced by the background tanning and reflected light in practical situations, our data suggested that the a^* value, measured by spectrophotometer, does not influence the degree of tanning and vice versa.

The constitutive skin color is composed of various chromophores, such as melanin, carotenoids, and oxy-/deoxyhemoglobin. UV-induced erythema with tanning is also added to that skin color12. An objective measurement of erythema in MED site does not simply mean spectrophotometric measurement of the corresponding parameter (a^* value).

Additional parameters (b^*, L^*) and various combination items (b^*/L^*, ΔE) are also needed to describe the degree of tanning and to discern the perceptive color differences. However, the limitation of this study is that the all subjects are composed of Koreans, who predominantly have brown-colored skin (skin types III and IV). The CIE L^*a^*b^* values of brown-colored skin correspond to a brownish color that is distributed almost uniformly around the center of the color system. Thus, the results of L^*a^*b^* values could be controversial because of their limited choice of subjects in our study. Therefore, to overcome such limitation, we recommend that further studies with various skin types be performed.

In conclusion, the spectrophotometer can be applied as a measurement instrument with basic parameters in the field of photomedicine. We can develop various combinations of items to investigate photobiological phenomena, including erythema, tanning, and perceptive color differences. This effort can indicate more reliable spectrophotometric values to estimate erythema, pigmentation, and tanning in photomedicine. Moreover, such advanced knowledges likely to be useful to dermatologists, and the spectrophotometer is a more useful instrument to assist in setting phototherapy and measuring erythema, pigmentation, and tanning of various skin disorders on an objective basis. It can also be helpful in deciding skin phototypes and in posttreatment evaluation of vascular laser and whitening agents.