Seidenari S, Pellacani G, Giannetti A
Department of Dermatology, University of Modena, Italy.
Melanoma Res. 1999 Apr;9(2):163-71. doi: 10.1097/00008390-199904000-00009.
The aim of our investigation was to evaluate the usefulness of a system composed of a digital videomicroscope equipped with a dedicated program for the quantitative characterization of various parameters of the clinically significant features of pigmented skin lesion (PSL) images, forming the basis for automatic differentiation of naevi and thin melanomas. In total 424 naevi and 37 melanomas (including 23 thinner than 0.75 mm) were considered. All the digital images were acquired, framed and analysed using the DBDermo-MIPS program (Biomedical Engineering Dell'Eva-Burroni), which calculates different parameters related to the geometry, the colour distribution and the internal pattern of the lesion. We also assessed the efficacy of an automatic classifier, trained for 100% sensitivity using a subset of PSL images (59 naevi and 19 melanomas), on a test set including 365 naevi and 18 melanomas thinner than 0.75 mm. Significant differences between values from benign and malignant PSLs were observed for most of the numerical parameters. Values from the training set underwent elaboration by means of multivariate discriminant analysis, enabling the identification of variables that are important for distinguishing between the groups in order to develop a procedure for predicting group membership for new cases (test set) in which group membership is undetermined. Going on the training set data, a threshold score was established, enabling each melanoma to be attributed to the right group. When the same threshold value was employed for discriminating between benign and malignant lesions in the test set, all the melanomas were correctly classified, whereas 30 out of the 365 benign lesions were attributed to the wrong group. Thus the specificity of the system reached 92%, whereas the sensitivity was 100%. Our data suggest that elaboration of videomicroscopic images by means of dedicated software improves diagnostic accuracy for thin melanoma. Since elaboration of an image requires only 60s using our system, all the parameter data are available in real time and can be immediately examined by the classifier, providing an instant aid to clinical diagnosis.
我们研究的目的是评估一个系统的实用性，该系统由一台配备专用程序的数字视频显微镜组成，该程序用于对色素沉着性皮肤病变（PSL）图像的临床显著特征的各种参数进行定量表征，为痣和薄黑色素瘤的自动鉴别奠定基础。总共考虑了424个痣和37个黑色素瘤（包括23个厚度小于0.75毫米的）。所有数字图像均使用DBDermo - MIPS程序（戴尔·埃瓦 - 布罗尼生物医学工程公司）进行采集、取景和分析，该程序可计算与病变的几何形状、颜色分布和内部模式相关的不同参数。我们还评估了一个自动分类器的功效，该分类器使用一部分PSL图像（59个痣和19个黑色素瘤）进行训练以达到100%的敏感性，测试集包括365个痣和18个厚度小于0.75毫米的黑色素瘤。对于大多数数值参数，观察到良性和恶性PSL的值存在显著差异。训练集的值通过多变量判别分析进行处理，从而能够识别对于区分两组很重要的变量，以便开发一种程序来预测新病例（测试集）的组成员身份，其中组成员身份尚未确定。基于训练集数据，建立了一个阈值分数，使每个黑色素瘤都能被归到正确的组。当在测试集中使用相同的阈值来区分良性和恶性病变时，所有黑色素瘤都被正确分类，而365个良性病变中有30个被归错组。因此，该系统的特异性达到92%，而敏感性为100%。我们的数据表明，通过专用软件对视频显微镜图像进行处理可提高薄黑色素瘤的诊断准确性。由于使用我们的系统处理一幅图像仅需60秒，所有参数数据可实时获取，并且分类器可立即对其进行检查。这为临床诊断提供了即时帮助。
