For an adrenal incidentaloma with indeterminate imaging characteristics, urine multisteroid profiling is suggested for diagnosing adrenocortical carcinoma (ACC). Data on the utility of serum steroid metabolomics in this context is limited to a few studies. Here, we present data of 62 adult patients with indeterminate unilateral adrenal masses (size ≥ 3 cm and basal attenuation ≥10HU) where baseline serum liquid chromatography-tandem mass spectrometry (LC-MS/MS) multisteroid profiling was available. Logistic regression was used to identify the key steroid signature for differentiating ACC from other non-ACC adrenal masses. Among 62 patients (median age: 41 years, 31 males), 37 (59.6 %) had ACC. The non-ACC cohort (n = 25) comprised pheochromocytoma (n = 9), adrenocortical adenoma (n = 8), metastases (n = 4), schwannoma (n = 2), ganglioneuroma (n = 1), and lymphoma (n = 1). Tumour size was significantly larger in the ACC cohort (9.9 vs 7.0 cm; p < 0.001) than the non-ACC cohort. Nine of 13 steroids were significantly elevated in ACC: 11-deoxycorticosterone (DOC), 17-hydroxyprogesterone (17OHP), 11-deoxycortisol (S), cortisone (E), androstenedione (A4), dehydroepiandrosterone (DHEA), and dehydroepiandrosterone sulphate (DHEAS) in both sexes, as well as testosterone (T) in females and progesterone (P4) in males. After excluding sex-dependent steroids, univariate analysis yielded six significant steroids (17OHP, S, E, A4, DHEA, and DHEAS). A multivariate logistic regression model with backward elimination identified A4, S, and DHEAS as the best discriminators (AUC:0.923), with a cutoff of 0.52 yielding 83.8 % sensitivity and 96 % specificity for diagnosing ACC. Our study results suggest serum LC-MS/MS profiling of three steroids (A4, S, and DHEAS) provides a non-invasive approach to distinguish ACC from other indeterminate adrenal masses.