Conventional microscopy versus digital image analysis for histopathologic evaluation of immune cells in the endometrium.

In the search for a reliable biomarker able to diagnose immunological causes of infertility, uterine immune cells have been widely investigated. As a result, heterogeneous methods and cutoff values of what constitutes an aberrant number of immune cells have been reported, and a standardized method for quantification is needed. The objective of this study was to compare methods for quantification of immune cells visualized with immunohistochemistry in the endometrium of women in fertility treatment. Evaluation of the density of CD56+, CD16+ and CD163+ cells by conventional microscopy on a semiquantitative scale (low, medium and high) was compared to a continuous count using digital image analysis (DIA) reported as percentage positive cells out of the total number of stromal cells and number of positive cells per mm2, respectively. We previously reported the CD56/CD16 ratio as a possible prognostic marker, and therefore the ratios of CD56/CD16 were compared using two different methods for selecting fields for counting with DIA: one method using principles of systematic random sampling, where glands were excluded, and one method analyzing large parts of the tissue including glands. A significant association between conventional microscopy and DIA was found when the semiquantitative scale was compared to medians of positive cells in CD56, CD16 and CD163, respectively, p < 0.001. A systematic significant difference in the ratios of CD56/CD16 was found when comparing the two methods for field selection, p < 0.001. To determine the possible use of these methods, more knowledge of the correlation to clinical outcome is warranted.