Urine steroid metabolomics for the differential diagnosis of adrenal incidentalomas in the EURINE-ACT study: a prospective test validation study.

BACKGROUND: Cross-sectional imaging regularly results in incidental discovery of adrenal tumours, requiring exclusion of adrenocortical carcinoma (ACC). However, differentiation is hampered by poor specificity of imaging characteristics. We aimed to validate a urine steroid metabolomics approach, using steroid profiling as the diagnostic basis for ACC. METHODS: We did a prospective multicentre study in adult participants (age ≥18 years) with newly diagnosed adrenal masses. We assessed the accuracy of diagnostic imaging strategies based on maximum tumour diameter (≥4 cm vs <4 cm), imaging characteristics (positive vs negative), and urine steroid metabolomics (low, medium, or high risk of ACC), separately and in combination, using a reference standard of histopathology and follow-up investigations. With respect to imaging characteristics, we also assessed the diagnostic utility of increasing the unenhanced CT tumour attenuation threshold from the recommended 10 Hounsfield units (HU) to 20 HU. FINDINGS: Of 2169 participants recruited between Jan 17, 2011, and July 15, 2016, we included 2017 from 14 specialist centres in 11 countries in the final analysis. 98 (4·9%) had histopathologically or clinically and biochemically confirmed ACC. Tumours with diameters of 4 cm or larger were identified in 488 participants (24·2%), including 96 of the 98 with ACC (positive predictive value [PPV] 19·7%, 95% CI 16·2-23·5). For imaging characteristics, increasing the unenhanced CT tumour attenuation threshold to 20 HU from the recommended 10 HU increased specificity for ACC (80·0% [95% CI 77·9-82·0] vs 64·0% [61·4-66.4]) while maintaining sensitivity (99·0% [94·4-100·0] vs 100·0% [96·3-100·0]; PPV 19·7%, 16·3-23·5). A urine steroid metabolomics result indicating high risk of ACC had a PPV of 34·6% (95% CI 28·6-41·0). When the three tests were combined, in the order of tumour diameter, positive imaging characteristics, and urine steroid metabolomics, 106 (5·3%) participants had the result maximum tumour diameter of 4 cm or larger, positive imaging characteristics (with the 20 HU cutoff), and urine steroid metabolomics indicating high risk of ACC, for which the PPV was 76·4% (95% CI 67·2-84·1). 70 (3·5%) were classified as being at moderate risk of ACC and 1841 (91·3%) at low risk (negative predictive value 99·7%, 99·4-100·0). INTERPRETATION: An unenhanced CT tumour attenuation cutoff of 20 HU should replace that of 10 HU for exclusion of ACC. A triple test strategy of tumour diameter, imaging characteristics, and urine steroid metabolomics improves detection of ACC, which could shorten time to surgery for patients with ACC and help to avoid unnecessary surgery in patients with benign tumours. FUNDING: European Commission, UK Medical Research Council, Wellcome Trust, and UK National Institute for Health Research, US National Institutes of Health, the Claire Khan Trust Fund at University Hospitals Birmingham Charities, and the Mayo Clinic Foundation for Medical Education and Research.

Do patients with incidentally discovered bilateral adrenal nodules represent an early form of ARMC5-mediated bilateral macronodular hyperplasia?

Bilateral adrenal macronodular hyperplasia (BMAH) is a rare form of Cushing's syndrome characterised by the presence of bilateral secretory adrenal nodules and hypercortisolism. Familial studies support a genetic basis for BMAH, and the disease has been linked to mutations in ARMC5, a gene shown to have a tumour suppressor-like action in the development of adrenal nodules. This study aimed to investigate whether ARMC5 mutations play a role in the development of incidentally discovered bilateral adrenal nodules. We investigated 39 patients with incidentally discovered bilateral adrenal nodules >0.8 cm in diameter who underwent extensive biochemical testing to look for signs of subclinical hypercortisolism. Genomic DNA was analysed by Sanger sequencing, using primers targeted to ARMC5 transcripts. Of the 39 patients included in our study, three were identified as having variants in ARMC5. Two of these are unlikely to be clinically significant, but there is evidence that the third mutation, Chr16:g.31476122;c.1778G>C (p.Arg593Pro), may be pathogenic. Another variant, affecting the same amino-acid residue c.1777C>T (p.Arg593Trp), has been identified previously in two studies of BMAH patients, where it has been shown to segregate with disease in one BMAH family. This patient had biochemical evidence of hypercortisolism in the absence of overt Cushing's syndrome, and underwent bilateral adrenalectomy separated in time. The presence of a probably clinically significant mutation in ARMC5 in one patient with bilateral adrenal incidentalomas adds to the growing body of evidence in support of ARMC5 as a critical mediator of adrenal nodule development. In addition, the absence of significant ARMC5 mutations in 38 of our patients represents an important negative finding, demonstrating the degree of variability within the pathogenesis of adrenal nodule development.