Steroid Panels As a Tool for the Differential Diagnosis of Adrenocortical Carcinoma Vs. Benign Adrenal Adenomas

Presentation Number: SUN 385
Date of Presentation: April 2nd, 2017

Juilee Rege*1, T. Le-Shan Wilson2, Rork Kuick3, William E. Rainey4 and Tobias Else1
1University of Michigan, Ann Arbor, MI, 2University of Michigan Health System, Ann Arbor, MI, 3University of Michigan, 4The University of Michigan, Ann Arbor, MI


Adrenal tumors are common, affecting 2-10% of the population. Adrenocortical carcinoma (ACC), however, is a rare malignant endocrine tumor with an incidence of 1-2/million per year. A major clinical challenge is to identify the rare ACCs amongst the common benign adrenal tumors. Currently, the differential diagnosis is mainly based on imaging, which is not well validated and poses a risk due to the use of radiation. Serum-based steroid analysis might represent an alternative diagnostic tool, which is less expensive and virtually risk free.

In this study we evaluated 12 different unconjugated steroids by LC-MS/MS (17OH-pregnenolone, 11-deoxycortisol, 11-deoxycorticosterone, androstenedione, pregnenolone, cortisol, 17OH-progesterone, dehydroepiandrosterone (DHEA), testosterone, cortisone, corticosterone, progesterone) and 3 steroid sulfates (17OH-pregnenolone sulfate, pregnenolone sulfate, DHEA sulfate) in random serum samples of 17 patients with a diagnosis of ACC (median age 60 years, 8 male/9 female), 16 patients with benign adrenal masses (ACA, median age 65 years, 5 male/11 female) and 37 unaffected control patients (median age 53 years, 19 male/18 female). ACC patients had at least minimal tumor load and samples were all collected prior to any medical therapy. For further analysis a constant was added before log transformation (1/20th of the mean of the three groups).

8 unconjugated steroids and all steroid sulfates showed significantly (ANOVA, p<0.05) different levels when comparing the ACC and ACA group. 11-deoxycortisol and 17OH-pregnenolone sulfate showed the most significantly different levels between the two groups. When further analyzed by leave-one-out analysis using 3 different steroids in a weighted score (weights chosen by diagonal linear discrimination analysis) the receiver operator characteristics (ROC) analysis showed good discriminatory properties with an area under the curve (AUC) of 0.92. Therefore, we conclude that serum steroid hormone panels are a safe and very promising tool for the differential diagnosis of malignant vs. benign adrenocortical tumors.


Nothing to Disclose: JR, TLSW, RK, WER, TE