Willemse F, Nap M, Eggink H F, Foekens J A, Henzen-Logmans S C, van Putten W L
Department of Pathology, University Hospital Groningen, The Netherlands.
Mod Pathol. 1995 Jun;8(5):521-5.
In immunochemical assays of specific cell constituents in cytosols from tumors, the relationship between epithelial and stromal fractions is not taken into account. This may influence the outcome of the measurements and result in incorrect categorization as negative or positive. In a setting addressing pS2 (only detectable in epithelial cells) in breast carcinomas, we investigated three possibilities that may overcome this problem using histologic sections of breast carcinomas of 50 patients: (a) visual estimation of area percentage of immunohistochemical staining of the cell constituent of interest performed by three independent individuals, (b) quantification of area percentage of the immunohistochemical results by true color image analysis, and (c) quantification of the epithelial and stromal compartments of the tumors in Heidenhain's-azan-stained tissue sections, using the true color image analysis system, to assess the epithelial percentage in the tumors. This percentage was used as a correction factor for data on pS2 obtained by cytosolic determinations. Visual estimation appeared to be subject to interobserver variation and, subsequently, becomes less applicable in the absence of strict scoring rules. Based on tests for correlation, image analysis system quantification seemed reproducible in both quantification procedures. However, due to the high magnification necessary to visualize the immunohistochemical staining product, the effect of field selection caused systematic differences between repeat measurements (Friedman test). As a result of the contrasting colors of the azan staining, the calculation of the epithelia percentage could be performed at a low magnification. Consequently, here the effect of field selection was not present. Correction of cytosolic values for epithelial percentage resulted in 8% of the cases changing category.(ABSTRACT TRUNCATED AT 250 WORDS)
在肿瘤细胞溶质中特定细胞成分的免疫化学分析中，上皮成分和基质成分之间的关系未被考虑在内。这可能会影响测量结果，并导致阴性或阳性的错误分类。在一项针对乳腺癌中pS2（仅在上皮细胞中可检测到）的研究中，我们利用50例乳腺癌患者的组织切片，研究了三种可能克服这一问题的方法：（a）由三名独立人员对感兴趣的细胞成分免疫组化染色的面积百分比进行目测估计；（b）通过真彩色图像分析对免疫组化结果的面积百分比进行量化；（c）使用真彩色图像分析系统对海登海因氏偶氮染色组织切片中的肿瘤上皮和基质成分进行量化，以评估肿瘤中的上皮百分比。该百分比用作通过细胞溶质测定获得的pS2数据的校正因子。目测估计似乎受观察者间差异的影响，因此，在没有严格评分规则的情况下适用性较差。基于相关性测试，图像分析系统量化在两种量化程序中似乎都具有可重复性。然而，由于可视化免疫组化染色产物需要高倍放大，视野选择的影响导致重复测量之间存在系统差异（弗里德曼检验）。由于偶氮染色的对比颜色，上皮百分比的计算可以在低倍放大下进行。因此，这里不存在视野选择的影响。根据上皮百分比对细胞溶质值进行校正后，8%的病例改变了分类。（摘要截短于250字）
