McKenzie Naja E, Saboda Kathylynn, Duckett Laura D, Goldman Rayna, Hu Chengcheng, Curiel-Lewandrowski Clara N
Arizona Cancer Center, University of Arizona, Tucson, 85724-5024, USA.
Arch Dermatol. 2011 Jan;147(1):31-6. doi: 10.1001/archdermatol.2010.392.
To develop a photographic sun damage assessment scale for forearm skin and test its feasibility and utility for consistent classification of sun damage.
For a blinded comparison, 96 standardized 8 × 10 digital photographs of participants' forearms were taken. Photographs were graded by an expert dermatologist using an existing 9-category dermatologic assessment scoring scale until all categories contained photographs representative of each of 4 clinical signs. Triplicate photographs were provided in identical image sets to 5 community dermatologists for blinded rating using the dermatologic assessment scoring scale.
Academic skin cancer prevention clinic with high-level experience in assessment of sun-damaged skin.
Volunteer sample including participants from screenings, chemoprevention, and/or biomarker studies.
Reproducibility and agreement of grading among dermatologists by Spearman correlation coefficient to assess the correlation of scores given for the same photograph, κ statistics for ordinal data, and variability of scoring among dermatologists, using analysis of variance models with evaluating physician and photographs as main effects and interaction effect variables to account for the difference in scoring among dermatologists.
Correlations (73% to >90%) between dermatologists were all statistically significant (P < .001). Scores showed good to substantial agreement but were significantly different (P < .001) for each of 4 clinical signs and the difference varied significantly (P < .001) among photographs.
With good to substantial agreement, we found the development of a photographic forearm sun damage assessment scale highly feasible. In view of significantly different rating scores, a photographic reference for assessment of sun damage is also necessary.
开发一种用于前臂皮肤晒伤评估的摄影量表，并测试其对晒伤进行一致分类的可行性和实用性。
为进行盲法比较，拍摄了96张参与者前臂的标准化8×10数码照片。一位皮肤科专家使用现有的9级皮肤病评估评分量表对照片进行分级，直到所有类别都包含代表4种临床体征的照片。将一式三份相同图像集的照片提供给5位社区皮肤科医生，让他们使用皮肤病评估评分量表进行盲法评分。
在评估晒伤皮肤方面具有丰富经验的学术性皮肤癌预防诊所。
志愿者样本，包括来自筛查、化学预防和/或生物标志物研究的参与者。
通过Spearman相关系数评估皮肤科医生之间分级的再现性和一致性，以评估对同一张照片给出的分数之间的相关性；对于有序数据，使用κ统计量；使用以评估医生和照片为主要效应和交互效应变量的方差分析模型，以考虑皮肤科医生之间评分的差异，评估皮肤科医生之间评分的变异性。
皮肤科医生之间的相关性（73%至>90%）均具有统计学意义（P<.001）。评分显示出良好至高度的一致性，但对于4种临床体征中的每一种，评分均存在显著差异（P<.001），且不同照片之间的差异也显著不同（P<.001）。
鉴于具有良好至高度的一致性，我们发现开发一种用于前臂晒伤评估的摄影量表非常可行。鉴于评分存在显著差异，晒伤评估的摄影参考也是必要的。
