The Impact of In Vivo Reflectance Confocal Microscopy on the Diagnostic Accuracy of Lentigo Maligna and Equivocal Pigmented and Nonpigmented Macules of the Face

The Impact of In Vivo Reflectance Confocal Microscopy on the Diagnostic Accuracy of Lentigo Maligna and Equivocal Pigmented and Nonpigmented Macules of the Face
复制标题

DOI:
10.1038/jid.2010.84
复制
发表时间:
2010-08-01
影响因子:
6.5
通讯作者:
Menzies, Scott W.
Menzies, Scott W.
中科院分区:
医学1区
文献类型:
--
作者:
Guitera, Pascale;Pellacani, Giovanni;Menzies, Scott W.

文献摘要

被引文献

相似文献

有限的研究报告,在体内的反射共聚焦显微镜(RCM)的功能,恶性雀斑样痣(LM)。对总共64个RCM特征进行回顾性评分,并对连续系列的RCM采样、临床上可疑的面部黄斑(n = 81 LM,n = 203良性黄斑(BM))的诊断设盲。除了描述LM(单变量)的RCM诊断特征外,还开发了一种算法(LM评分)以区分LM和BM。这包括两个主要功能,每个功能得分+2分(无边缘乳头和圆形大的Pagetoid细胞>20 μ m),和四个次要特征;三人各得+1分(在5个0.5 x 0.5 mm(2)视野中,真皮表皮交界处有3个或更多非典型细胞,非典型细胞位于毛囊中,真皮乳头内有核细胞),和一个(负)特征得分-1分(加宽的蜂窝图案)。LM评分>= 2导致诊断LM的敏感性为85%,特异性为76%(LM的比值比(OR)为18.6; 95%置信区间:9.3-37.1)。该算法在无色素病变的诊断中同样有效,并显示出良好的观察者间重现性(87%)。在29个LM和44个BM的测试集中,LM的OR为60.7(置信区间:11.9-309)(93%灵敏度,82%特异性)。
Limited studies have reported the in vivo reflectance confocal microscopy (RCM) features of lentigo maligna (LM). A total of 64 RCM features were scored retrospectively and blinded to diagnosis in a consecutive series of RCM sampled, clinically equivocal, macules of the face (n = 81 LM, n = 203 benign macules (BMs)). In addition to describing RCM diagnostic features for LM (univariate), an algorithm was developed (LM score) to distinguish LM from BM. This comprised two major features each scoring +2 points (nonedged papillae and round large pagetoid cells >20 mu m), and four minor features; three scored +1 point each (three or more atypical cells at the dermoepidermal junction in five 0.5 x 0.5 mm(2) fields, follicular localization of atypical cells, and nucleated cells within the dermal papillae), and one (negative) feature scored -1 point (a broadened honeycomb pattern). A LM score of >= 2 resulted in a sensitivity of 85% and specificity of 76% for the diagnosis of LM (odds ratio (OR) for LM 18.6; 95% confidence interval: 9.3-37.1). The algorithm was equally effective in the diagnosis of amelanotic lesions and showed good interobserver reproducibility (87%). In a test set of 29 LMs and 44 BMs, the OR for LM was 60.7 (confidence interval: 11.9-309) (93% sensitivity, 82% specificity).