Steroid metabolites for diagnosing and predicting clinicopathological features in cortisol-producing adrenocortical carcinoma

Adrenocortical carcinomas (ACC) are rare but aggressive endocrine neoplasms from the adrenal cortex. ACC may be functional and cause Cushing’s syndrome and/or virilization [1‐3], and it may present as an abdominal mass or incidental findings. Further, approximately 30% of ACC are responsible for multiple hormone production, including steroid hormone precursors, which is referred to as disorganized steroidogenesis [4]. The ACC prognosis is poor, and the 5-year overall survival is lower than 35% in most studies [5, 6]. Management is multidisciplinary and includes surgical resection, oral mitotane, intravenous chemotherapy, and palliative radiation [7]. Therefore, early diagnosis, accurate staging, and appropriate treatment according to progression prediction are important. For accurate determination, the combination of tumor size (≥65 mm) [8] and careful pathological investigation with the Weiss score (≥3 out of 9) [8] and Ki67-index (≥10%: high risk) [9] is widely used for ACC diagnosis. The European Network for the Study of Adrenal Tumors (ENSAT) system is used for ACC staging classification [9]. Both the growth marker Ki67 and ENSAT classification are also used as predicting prognostic factors of ACC [9]; however, image-guided adrenal biopsy or adrenalectomy necessary for pathological diagnosis is invasive and cannot be performed in patients with a poor general condition. Additionally, adrenal biopsy is an expensive procedure with a reported rate of nondiagnostic biopsies of 8.7%, a complication rate of 2.5% [10], and lower accuracy with 70% sensitivity and 98% specificity [11]. Therefore, less invasive and less costly methods for the diagnosis and prognostic prediction of ACC, along with repeat imaging, are desired for the determination of appropriate treatment. Recently, steroid profiling has emerged as a powerful novel diagnostic tool for ACC [11‐14]. Androgen-producing ACC is easy to diagnose because cortisol-producing adenomas (CPA) very rarely produce androgens. Conversely, in non-androgen producing ACC, the differentiation between ACC and CPA becomes challenging because the end-product cortisol presents similarly between ACC and CPA. Previous reports demonstrated that serum steroid metabolite examination using liquid chromatography tandem mass spectrometry (LC-MS/MS) [12, 13] or urine steroid metabolites using gas chromatography mass spectrometry (GC-MS) [15‐21] are useful for distinguishing ACC from adrenocortical adenoma. Furthermore, the ENSAT recommends a biochemical workup for suspected ACC that includes serum cortisol, aldosterone, 17-hydroxyprogesterone, dehydroepiandrosterone sulfate (DHEAS), androstenedione, testosterone, and 17-beta-estradiol (http://​www.​ensat.​org/​page-1317312); however, feasible methods to detect or predict ACC behavior and prognosis aside from LC-MS/MS or GC-MS are limited by their lack of clinical availability. In the present study, we aimed to compare the steroidal metabolic profile in cortisol-producing ACC and CPA and to correlate steroid metabolites with clinical, pathological and prognostic parameters using less expensive and more widely available assays (RIA and ECLIA).

In the present study, all basal levels of steroid precursors and DHEAS were significantly increased in cortisol-producing ACC patients compared to CPA patients. The immunoassay was a successful tool for diagnosis, and we recommend 17-hydroxypregnenolone (glucocorticoid and androgen precursor), 11-deoxycorticosterone (mineralocorticoid precursor), or the combination of DHEAS and androstenedione (androgen precursor) for the diagnosis of cortisol-producing ACC. Notably, the combination of DHEAS and androstenedione could be a valuable tool to differentiate ACC from CPA is of extreme relevance, since these are two widely available markers, and can be implemented without a major increase in costs. Only 17-ketosteroids were measured as urine steroid metabolites in our study, but none of the 17-ketosteroid fractions were sensitive to ACC diagnosis. Moreover, we propose that serum steroid metabolite measurement, especially 11-deoxycortisol (glucocorticoid precursor) and testosterone, as a simple and noninvasive method for predicting the progression and prognosis of patients with cortisol-producing ACC.

The observed heterogeneity of steroidogenesis reflects immature and dedifferentiated cell features, which are the hallmark of ACC [11, 13, 15‐17]. To our knowledge, no serum or urinary steroid metabolites associated with prognostic factors in ACC have been demonstrated to date, but several papers have reported these were useful for the diagnosis of ACC [11‐21]. Consistent with our recommendation, Schweitzer et al. reported that the combination of androgen precursor (17-hydroxypregnenolone, progesterone, and DHEA), mineralocorticoid precursor (11-deoxycorticosterone), glucocorticoid precursor (11-deoxycortisol), and sex hormone (DHEAS and estradiol) was useful for ACC diagnosis [13]. Taylor et al. also recommended androgen precursor (17-hydroxypregnenolone) and glucocorticoid precursor (11-deoxycortisol) for the diagnosis of ACC [12]. For urinary steroid metabolites, 11-deoxycortisol urinary metabolite tetrahydro-11-deoxycortisol (glucocorticoid precursor) [15‐18] as well as 17-hydroxypregnenolone urine metabolite pregnanetriol and pregnenolone urine metabolite pregnanediol (androgen precursor) [17, 18] were recommended. Although all previous reports suggested that some steroid hormone precursors were useful in the diagnosis of ACC, the differences between the type of steroid precursors may be due to differences in the cortisol-producing ability within ACC and adrenocortical adenoma. Accurate 24-h collections are often not easy to obtain, and blood collections are more convenient for patients.

LC-MS/MS has less cross-reactivity to other metabolites with peptide specificity and high sensitivity, but it is expensive, requires more time for processing and measurement than immunoassay. On the other hand, immunoassay is less expensive and more wildly available, and thus can be performed in the largest number of places, including poorer countries that will unlikely have access to LC-MS/MS; however, immunoassay is not without concerns regarding the accuracy, level of validation, intra- and interassay and more details on the standardization of the measurement of each steroid using RIA and ECLIA, particularly for intermediate metabolites, which has few or almost no standardized methods other than LC-MS/MS or GC-MS. Therefore, the JCEM published an editorial in 2013 suggesting that the Journal policy should be to accept only articles that use mass spectrometry for sex steroid assays [26], which has been reinforced by a letter also published in JCEM [27]. These issues were discussed in detail at the 2014 Workshop on Measuring Estrogen Exposure and Metabolism. A workshop on measuring steroids, particularly estradiol, concluded that “both immunoassays and mass spectrometry-based assays for estrogens and their metabolites would be acceptable if they are accurate and reliable and meet performance criteria suitable for their intended use” [28]. Laurence M et al. mentioned that the debate will continue until mass spectrometry assays for steroid hormones become sufficiently rapid, inexpensive, and robust for general acceptance and use [29]. In this study, we revealed that RIA and ECLIA provide sufficient ability to distinguish ACC from CPA; however, lower concentrations lose accuracy and precision in RIA and ECLIA compared to LC-MS/MS, which may be a limitation. Because DHEAS is inexpensive, widely available for clinical use and is present in higher concentrations such that the absolute precision is comparable to LC-MS/MS, using androstenedione combined with DHEAS to distinguish ACC from CPA may overcome this limitation of RIA and ECLIA.

The reasons for the correlations between 11-deoxycortisol and ENSAT classification, or testosterone and KI67-index were not elucidated in our study, and we will therefore base our analysis on previous reports. Cortisol-producing ACC has been known to produce androgen the most simultaneously to cortisol [30‐32]. Cortisol-producing ACC that produce male sex hormone at the same time might have higher cell proliferation. Because 11-deoxycortisol is converted to cortisol by Cytochrome P450 Family 11 Subfamily B Member 1 (CYP11B1), a relative deficiency of CYP11B1 may occur in ACC with poor progression. In fact, CYP11B1 expression has been shown to be downregulated in some ACC patients [33].

A confirmed diagnosis of ACC cannot be made is some cases, or surgery cannot be performed due to a poor general condition or severe complications. In such cases, the measurement of serum steroid hormone metabolites would be helpful for predicting pathological characteristics or prognosis in cortisol-producing ACC; however, our study is limited to cortisol-producing adrenal tumors without evaluation of estrogen-producing tumors is a prospective study with a small number of ACC samples and was limited to urinary hormone measurements. Additional large-scale prospective studies will be required in the future.

Combined serum steroid metabolite measurement is a highly informative noninvasive method for diagnosing and predicting the clinicopathological features associated with the prognosis of cortisol-producing ACC.