Eosinophilic Cells as a Distinct Morphological Feature in BRAFV600E-Mutated Ovarian Serous Borderline Tumors

Abstract

Background/Objectives: According to recent reports, the BRAF^V600E mutation in serous borderline tumors (SBTs) plays a protective role against progression to low-grade serous carcinoma through oncogene-induced senescence. One consequence of this is the appearance of eosinophilic cells (ECs). The aim of the current study was to determine the interobserver reproducibility of ECs and their predictive significance for the detection of the BRAF^V600E mutation in SBTs. Methods: The study was conducted using 63 cases of ovarian SBTs. Three gynecological pathologists, blinded to each tumor's mutation status, assessed the presence of ECs. Immunohistochemical staining with p16 and Ki-67 was performed to validate ECs. Mutational analysis was carried out using targeted NGS. Results: Genetic analysis revealed 30 BRAF-mutated, 1 NRAS-mutated, and 9 KRAS-mutated SBTs. ECs were identified by the majority of pathologists (two or three) in 78% of the BRAF^V600E-mutated and 11% of the wild-type tumors with other mutations (p < 0.0001). The interobserver reproducibility of the presence of ECs was substantial (κ = 0.66). ECs validated with p16/Ki-67 were identified in 92.6% of the BRAF^V600E-mutated and in 13.8% of the wild-type tumors with other mutations (p < 0.0001). For the ECs identified by the majority of pathologists, the sensitivity and specificity when predicting the BRAF^V600E mutation were 77.8% and 88.9%, respectively. For the ECs validated with p16/Ki-67, the sensitivity and specificity when predicting the BRAF^V600E mutation were 95.3% and 90.5%, respectively. Conclusions: Overall, these results suggest that ECs in SBTs have potential association with the BRAF^V600E mutation.

Keywords: BRAF; KRAS; NRAS; eosinophilic cells; interobserver reproducibility; oncogene-induced senescence; serous borderline tumor.

Figure 1

Eosinophilic cell with abundant eosinophilic cytoplasm and small nucleus in SBT (previously unpublished, original photo) with hematoxylin and eosin staining (magnification 400×).

Figure 2

Validation of eosinophilic cells and their mimics (arrow) in ovarian SBTs (previously unpublished, original photos). Strong nuclear and cytoplasmic staining of p16 in senescent eosinophilic cells (A) and lack of nuclear expression of proliferative activity marker Ki-67 (B) in senescent eosinophilic cells; absence of strong nuclear and cytoplasmic staining of p16 (C) and positive expression of Ki-67 (D) in cells that could be misidentified as bonafide ECs. Images feature immunohistochemical staining (magnification 200×).

Figure 3

Eosinophilic cells in BRAF^V600E-mutated tumors and their mimics in wild-type cases (previously unpublished, original photos). Bonafide eosinophilic cells with abundant dense/glassy eosinophilic cytoplasm occupying at least 50% of cell area (arrow, (A–D)); epithelial cells detaching from tumor surface and resembling eosinophilic cells (arrow, (E,F)). Images feature hematoxylin and eosin staining (magnification 400×).