Abstract
BACKGROUND: Approximately 60% of adrenocortical carcinomas (ACC) are functional, and Cushing's syndrome is the most frequent diagnosis that has been revealed to have a particularly poor prognosis. Since 30% of ACC present steroid hormone-producing disorganization, measurement of steroid metabolites in suspected ACC is recommended. Previous reports demonstrated that steroid hormone precursors or their urine metabolites, which can be assessed using liquid chromatography tandem mass spectrometry (LC-MS/MS) or gas chromatography mass spectrometry (GC-MS) respectively, are useful for distinguishing ACC from cortisol-producing adenomas (CPA); however, despite high precision, LC-MS/MS and GC-MS require a highly trained team, are expensive and have limited capacity. METHODS: Here, we examined 12 serum steroid metabolites using an immunoassay, which is a more rapid and less costly method than LC-MS/MS, in cortisol-producing ACC and CPA. Further, the correlation of each steroid metabolite to the classification stage and pathological status in ACC was analyzed. RESULTS: Reflecting disorganized steroidogenesis, the immunoassay revealed that all basal levels of steroid precursors were significantly increased in cortisol-producing ACC compared to CPA; in particular, 17-hydroxypregnenolone (glucocorticoid and androgen precursor) and 11-deoxycorticosterone (mineralocorticoid precursor) showed a large area under the ROC curve with high sensitivity and specificity when setting the cut-off at 1.78 ng/ml and 0.4 mg/ml, respectively. Additionally, a combination of androstenedione and DHEAS also showed high specificity with high accuracy. In cortisol-producing ACC, 11-deoxycortisol (glucocorticoid precursor) showed significant positive correlations with predictive prognostic factors used in ENSAT classification, while testosterone showed significant positive correlations to the Ki67-index in both men and women. CONCLUSION: Less expensive and more widely available RIA and ECLIA may also biochemically distinguish ACC from CPA and may predict the clinicopathological features of ACC.