Data for: Quantification of randomness (Entropy) as a clinical tool to assess the severity of skin disease

Quantitative analysis of the skin is a new, quick, and very efficient method to quantify various skin changes like performing health assessments, determine skin conditions, and evaluating how skin responds to assigned treatment course(s). The quantitative analysis allows to produce universal measures of different skin conditions that would be more accurate and non-biased. It was theoreticized that healthy skin is more structured compared to the affected skin which exhibits more alterations due to neoplastic, inflammatory, or traumatic processes.
The dermatology clinic database of deidentified dermatoscopic images of the skin affected by psoriasis and adjacent clinically intact skin were used for the current analysis. All images were captured by Visioscope PC 35 camera device (Courage + Khazaka, Germany ). All images (raw data) were transferred into specially coded and developed proprietary software which enables its users to calculate entropy values for healthy and non-healthy skin.
Blank paper (baseline with minimum randomness), gridded paper (interpreted as the simplified model of skin), and chaotic patterns (interpreted as the skin affected by lesions) were used as reference entropy values. The Mann-Whitney U method was used to determine the statistical significance of gathered data. The entropy values for normal unaffected skin exhibited normal distribution with mean of 2.56 with 95% confidence interval of [2.36, 2.75]. The skin affected by psoriasis exhibited normal distribution with mean of 3.30 with 95% confidence of [2.93, 3.66].
The healthy skin has a lower entropy values compared to the entropy values of skin affected by psoriasis. That is because the healthy skin appears to have more uniformed patterns in its structure, while the skin affected by psoriasis exhibits more random features due to traumatic processes and inflammation.
The current technique of implying a mathematical algorithm to calculate the Maximum entropy shows promising results because physicians would be able to use universal measures to access skin conditions with minimal sources of error.

皮肤定量分析是一种新颖、快速且高效的方法，可用于量化多种皮肤变化——包括开展健康评估、确定皮肤状态，以及评估皮肤对指定治疗方案的响应情况。该定量分析可生成针对不同皮肤状态的通用量化指标，具备更高的准确性与无偏性。学界已提出理论假设：相较于因肿瘤性、炎症性或创伤性过程而出现更多结构改变的病变皮肤，健康皮肤的结构更为规整。
本研究分析所用的数据来自皮肤科临床数据库，该库包含银屑病皮损皮肤及相邻临床正常皮肤的去标识化皮肤镜图像（dermatoscopic images）。所有图像均由德国Courage + Khazaka公司生产的Visioscope PC 35型摄像设备采集。所有图像（原始数据）均被导入经专属编码开发的专有软件中，该软件可支持用户计算健康与非健康皮肤的熵（entropy）值。
本研究以空白纸张（最低随机性基线）、网格纸（被视为简化的皮肤模型）以及混沌图案（被视为皮损皮肤）作为熵值参考基准。采用Mann-Whitney U检验（Mann-Whitney U method）判定所收集数据的统计学显著性。结果显示，正常未受累皮肤的熵值呈正态分布，均值为2.56，95%置信区间为[2.36, 2.75]；银屑病皮损皮肤的熵值同样呈正态分布，均值为3.30，95%置信区间为[2.93, 3.66]。
健康皮肤的熵值低于银屑病皮损皮肤，这是由于健康皮肤的结构模式更为均一，而银屑病皮损皮肤因炎症与创伤性过程，呈现出更多随机特征。
本研究采用的基于数学算法计算最大熵（Maximum entropy）的技术展现出良好的应用前景，因为临床医师可借助该通用量化指标评估皮肤状态，且误差来源极小。