Suppr超能文献

眼内肿瘤的超声分类。一项15年的回顾性分析。

Echographic classification of intraocular tumours. A 15-year retrospective analysis.

作者信息

Verbeek A M, Thijssen J M, Cuypers M H, Brink H, Deutman A F

机构信息

University Hospital, Department of Ophthalmology, Nijmegen, The Netherlands.

出版信息

Acta Ophthalmol (Copenh). 1994 Aug;72(4):416-22. doi: 10.1111/j.1755-3768.1994.tb02788.x.

Abstract

The data on echographic routine diagnosis of intraocular tumours (prominence equal to, or larger than 2 mm) during the years 1976-1990 in the authors' department, were analyzed by multivariate statistical methods, using the final diagnosis either from pathology after enucleation or from the confirmed clinical diagnosis. The material consisted of melanomas (n = 325), metastases (n = 44), haemangiomas (n = 19) and other intraocular tumours (n = 16). The best set of echographic parameters in descending order of significance was: reflectivity (A-mode), choroidal excavation (B-mode), shape (B-mode), and regularity (A/B mode). Echographic differentiation by computer analysis of data on the three main kinds of tumours (melanoma/metastasis/haemangioma) separate from all the remaining tumours was expressed by the correct fraction: melanoma 89%, metastasis 80% and haemangioma 97%. The clinical echographic classification for these cases was 89%, 93% and 99.5%, respectively. The simultaneous differentiation between the three classes was found to yield a correct fraction of 85% by computer statistics and 95% by routine echography. The results of the present study might be used for prospective classification through the use of the parameter 'knowledge base' contained in the statistical classification procedure.

摘要

作者所在科室对1976年至1990年间眼内肿瘤(隆起等于或大于2mm)的超声常规诊断数据进行了多变量统计分析,最终诊断采用眼球摘除术后的病理诊断或确诊的临床诊断。材料包括黑色素瘤(n = 325)、转移瘤(n = 44)、血管瘤(n = 19)和其他眼内肿瘤(n = 16)。按重要性降序排列的最佳超声参数组合为:反射率(A超)、脉络膜凹陷(B超)、形状(B超)和规整性(A/B超)。通过对三种主要肿瘤(黑色素瘤/转移瘤/血管瘤)的数据进行计算机分析,将其与所有其他肿瘤区分开来的超声鉴别,以正确比例表示为:黑色素瘤89%,转移瘤80%,血管瘤97%。这些病例的临床超声分类分别为89%、93%和99.5%。通过计算机统计发现,这三类肿瘤同时鉴别的正确比例为85%,常规超声检查为95%。本研究结果可通过使用统计分类程序中包含的“知识库”参数用于前瞻性分类。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验