Abstract
Chronic endometritis (CE) is an unusual endometrial inflammation characterized by stromal plasmacyte infiltration. CE is easily missed due to its subtle symptoms and demanding histopathological examinations. Although the immunohistochemistry for the plasmacyte marker CD138 has facilitated the detection of endometrial stromal plasmacytes, the accuracy and biases of this method for CE diagnosis remain poorly understood. The aim of this study was to investigate the inter‑ and intra‑observer variability in the immunohistochemical detection of stromal plasmacytes in the human endometrium. A total of 80 CE and 20 non‑pathological archival hematoxylin‑stained endometrial preparations with or without immunostaining for CD138 were evaluated independently by two experienced observers and two inexperienced observers. Endometrial stromal plasmacytes in unit areas were counted in the hematoxylin‑stained and CD138‑immunostained preparations. Each preparation was subdivided into 11 categories by every five plasmacyte counts. The second evaluation was performed four weeks after the first evaluation. The immunohistochemical detection method was superior to conventional histopathological evaluation in both the inter‑ and intra‑observer agreement, irrespective of the experience level of the observers. The linear weighted κ coefficient for intra‑observer agreement was higher in the experienced observers than in the inexperienced observers. The inter‑observer agreement among the four observers by the immunohistochemical detection method was similarly good between the first and second evaluation. There was no significant inter‑ or intra‑observer variability in the paired comparison of the individual samples. These findings validate the use of immunohistochemistry for CD138 as an accurate and less biased diagnostic tool for CE.