SAT-555 Can Histology Predict the Presence of KCNJ5 Somatic Mutation in Aldosterone-Producing Adenomas?

Aldosterone-producing adenoma (APA) is well known to harbor marked intratumoral heterogeneity in terms of morphology and CYP11B2 (aldosterone synthase) localization. In histology, APA is generally characterized by two distinct cell subtypes, namely “clear cells” and “compact cells”. Clear tumor cells harbor abundant lipid droplets in their cytoplasms and compact tumor cells generally featuring small round shape have abundant intracytoplasmic organelles including mitochondria. Relatively close correlation between these histological characteristics (morphology and CYP11B2 immunohistochemistry) and genotypes of aldosterone-driver gene somatic mutation has been reported. Among them, KCNJ5-mutated APAs have been reported to harbor clear cell predominant features, while APAs with other rare somatic mutations including ATP1A1, ATP2B3 and CACNA1D harbor heterogenous or relatively compact cell predominant morphometry. However, these previous evaluation were based on eyeball analysis with relatively low reproducibility. Therefore, we developed the more quantitative methods using digital image software in order to analyze the widespread area, which can reflect intratumoral heterogeneity, with high reproducibility to analyze the further detailed correlation between histopathological characteristics and genotype in APA. We explored the utility of immunohistochemistry including CYP11B2 and KCNJ5. We further attempted to propose histopathological scoring system to predict the presence of KCNJ5 somatic mutation in APAs. Results of our present study revealed that KCNJ5 was predominantly immunolocalized in zona glomerulosa among adrenal cortex (vs. ZF, P=0.0002, vs. ZR, P=0.0002), furthermore, predominantly in APCCs than in non-APCCs (P=0.0019). Among the tumors, KCNJ5 immunoreactivity was significantly higher in KCNJ5-wild type APAs than in mutated ones (P=0.0037). KCNJ5-mutated APAs had significantly lower nuclear / cytoplasm ratio and abundant clear cell components than those with wild type, harboring large tumor size. In conclusion, we firstly proposed a novel histopathological predicting scoring system for the presence of KCNJ5 somatic mutation, including the following histopathological findings; N/C ratio, clear cell (%), tumor size, CYP11B2 immunoreactivity and KCNJ5 immunoreactivity. It is true that no single histological factors above could precisely predict the presence of KCNJ5 somatic mutation but this newly developed combined histopathological predicting scoring system could provide relatively high accuracy to predict KCNJ5 somatic mutation in APAs (AUC=96%, sensitivity:100%, specificity:90%, 4 points or more). However, further prospective validation by large number of cases is required for clarification.