Immunohistopathological re‐evaluation of adenocarcinoma of the lung with mixed subtypes using a tissue microarray technique and hierarchical clustering analysis

Abstract

To re-evaluate adenocarcinoma, mixed subtypes (ADMIX) of the lung, a total of 201 cases were classified into three main subgroups according to the most differentiated histological growth pattern; namely bronchioloalveolar carcinoma (BAC)-mixed, which was the most predominant (73.1%), papillary (PAP)-mixed (21.9%), and acinar-mixed (5%). The PAP-mixed was significantly male predominant and had more progressed clinicopathological features. A significant cytological difference was observed among the three subgroups. A tissue microarray was constructed and immunohistochemistry was undertaken using 15 biomarkers. Hierarchical clustering analysis was separately applied to the immunohistochemical results of ADMIX and ADMIX subgroups, and it was found that most acinar-mixed cases were placed in a separate cluster, while the BAC-mixed and PAPmixed failed to form significant independent clusters. The antibody clustering profile for the acinar-mixed was clearly different from that for the BAC-mixed or PAP-mixed, but the PAP-mixed shared a dendrogram profile with the other two subgroups. Statistically, approximately half of the 15 biomarkers were significant for differentiating between ADMIX subgroups and between different histological growth patterns. In conclusion, ADMIX can be classified into three histopathological subgroups according to the most differentiated growth pattern, of which a PAP growth pattern might indicate more aggressive behavior than that of a BAC growth pattern.

Classification using hierarchical clustering of tumor-infiltrating immune cells identifies poor prognostic ovarian cancers with high levels of COX expression

“…It has been frequently used to analyze DNA microarray data, and, more recently, studies of protein expression using immunohistochemical staining have also taken advantage of this method. [6][7][8][9]24,25 As hierarchical clustering uses correlation coefficients to assess similarity, numerical data, such as numbers of tumorinfiltrating immune cells, are suitable for analysis using this method. Therefore, we employed hierarchical clustering to study our data.…”

“…Recently, this method has been used to classify tumors based on the patterns of expression of multiple proteins. [6][7][8][9][10] In this study, we employed hierarchical clustering using the patterns of tumor-infiltrating CD1a þ , CD8 þ , and CD57 þ cells. As far as we know, hierarchical clustering has never been used to analyze the patterns of tumorinfiltrating immune cells.…”

The role of tissue microarray in the era of target-based agents

“…Hierarchical clustering for this purpose has been proposed for breast cancer [41,55], uterine cancer [56,57], natural killer/T-cell lymphomas [58], lung adenocarcinoma [59] and gastrointestinal stromal tumors [60]. Makretsov et al used these approaches applied to a prognostic TMA including cores from 438 breast cancer patients with a median follow-up of 15.4 years.…”

“…Moreover, different methods of quantifying immunohistochemistry have been proposed. Semiquantitative methods include the percentage of stained cells, while evaluation of intensity into three groups is not favored by pathologists since intensity depends upon the preanalytic conditions of samples (i.e., antibody dilutions, tissue preservation, and so on) [59,62,64].…”

Multiple biomarker tissue microarrays: bioinformatics and practical approaches

“…To this end unsupervised data analysis such as hierarchical clustering has been used extensively in the past decade, especially in gene expression arrays studies [61][62][63][64][65][66][67] but to a much lesser extent in biomarker studies [60,[68][69][70][71][72][73]. Hierarchical clustering is potentially of interest also in biomarker studies.…”