Kopf A W, Rigel D, Bart R S, Mintzis M M, Hennessey P, Harris M N, Ragaz A, Trau H, Friedman R J, Esrig B
J Dermatol Surg Oncol. 1981 Aug;7(8):645-50. doi: 10.1111/j.1524-4725.1981.tb00712.x.
Computer analyses to identify correlations between thickness of primary superficial spreading malignant melanoma and eighteen variables previously reported to be related to prognosis were performed on a series of malignant melanomas. The variables that showed statistically significant (less than or equal to 0.05) direct relationships to thickness were level (Clark), elevation of lesion, age of patient, least and greatest diameters of lesion, history of bleeding, ulceration, clinical and histologic stage, anatomic location, pedunculation, and satellitosis. The variables that did not correlate with thickness were clinical diagnosis of regional lymphadenopathy, in-transit metastasis, duration of lesion, sex, history of a previous malignant melanoma, and history of a pre-existing lesion at the site of the development of melanoma. Multiple regression analysis of the factors that showed statistically significant correlation with thickness of the primary lesion revealed a subset of six dominant variables that were most predictive of thickness, namely, level, elevation, largest diameter of lesion, ulceration, histologic stage, and age of the patient.
对一系列恶性黑色素瘤进行了计算机分析，以确定原发性浅表扩散性恶性黑色素瘤的厚度与先前报道的18个与预后相关的变量之间的相关性。与厚度呈统计学显著（小于或等于0.05）直接关系的变量有分级（克拉克分级）、病变隆起、患者年龄、病变的最小和最大直径、出血史、溃疡、临床和组织学分期、解剖位置、蒂状、以及卫星灶。与厚度不相关的变量有区域淋巴结病的临床诊断、途中转移、病变持续时间、性别、既往恶性黑色素瘤病史、以及黑色素瘤发生部位先前存在病变的病史。对与原发性病变厚度呈统计学显著相关的因素进行多元回归分析，发现了六个最能预测厚度的主要变量子集，即分级、隆起、病变最大直径、溃疡、组织学分期和患者年龄。
