improved diagnostic accuracy provided by the dual combination of tumour diameter and imaging characteristics, which is currently the most widely used approach in clinical practice, with an unenhanced CT tumour attenuation threshold of 20 Hounsfield units (HU) instead of the recommended 10 HU, is the more directly relevant and clinically practical finding from the study. Indeed, there remains a long way to go before assessment of urine steroid metabolomics with mass spectrometry could be routinely implemented in clinical practice. In addition to the limitations discussed in the accompanying commentary,2 we are concerned about the cost- effectiveness of such an approach, which will require more affordable mass spectrometry to be available in clinical laboratories. Furthermore, the incorporation of steroid metabolomics into clinical practice will require validation of assay reproducibility and the subsequent introduction of population-based, age-specific, and sex-specific reference values for the adult steroid metabolome.3
Additionally, we note that patients with current or recent (<6 months) intake of drugs known to alter steroid synthesis or metabolism were excluded from the study. Could the investigators specify which drug categories were included in this exclusion criterion? Treatment with systemic glucocorticoids or drugs known to alter steroid secretion, such as hormonal menopause treatment, antiepileptic drugs, ketoconazole, and mineralocorticoid receptor antagonists (spironolactone or eplerenone) might compromise the diagnostic performance of urine metabolomics.3.4 However, use of these drugs is not uncommon among individuals of a similar age to those included in the study (median age 59 years). Finally, we would be grateful if the investigators could provide information on participants’ renal function, since advanced chronic kidney disease might affect
steroid metabolite secretion in the urine.
We declare no competing interests.
Panagiota Anyfanti, Barbara Nikolaidou, *Eugenia Gkaliagkousi egkaliagkousi@auth.gr
3rd Department of Internal Medicine, Papageorgiou Hospital, Aristotle University of Thessaloniki, 56429 Thessaloniki, Greece
1 Bancos I, Taylor AE, Chortis V, et al. Urine steroid metabolomics for the differential diagnosis of adrenal incidentalomas in the EURINE-ACT study: a prospective test validation study. Lancet Diabetes Endocrinol 2020; 8: 773-81.
2 Turcu AF, Walch AK. A multi-test strategy for adrenal tumours. Lancet Diabetes Endocrinol 2020; 8: 733-34
3 Ackermann D, Groessl M, Pruijm M, et al. Reference intervals for the urinary steroid metabolome: the impact of sex, age, day and night time on human adult steroidogenesis. PLoS One 2019; 14: e0214549.
4 Arlt W, Biehl M, Taylor AE, et al. Urine steroid metabolomics as a biomarker tool for detecting malignancy in adrenal tumors. J Clin Endocrinol Metab 2011; 96: 3775-84.
Authors’ reply
We are grateful to Panagiota Anyfanti and colleagues for their interest in our work1 and their useful comments. We agree that our findings with respect to the unenhanced CT tumour attenuation threshold should have a major impact on clinical practice. These results, based on data from more than 2000 prospectively recruited participants, justify an immediate change in practice, with maximum tumour diameter of 4 cm and unenhanced CT tumour attenuation of 20 Hounsfield units (HU) as the most appropriate threshold values for consideration of adrenocortical carcinoma as a differential diagnosis in patients with newly diagnosed adrenal masses. We showed that, compared with the currently recommended cutoff of 10 HU, a tumour attenuation value greater than 20 HU on unenhanced CT had considerably higher specificity (80.0% [95% CI 77·9-82.0%] vs 64.0% [61.4-66.4]) while maintaining similar sensitivity (99.0% [94-4-100.0] vs 100.0% [96.3-100.0]). Introducing a more specific cutoff into clinical practice should result in a decrease in the number of imaging procedures
and surgeries in patients eventually diagnosed with benign adrenal masses. In view of our findings, work on the timely revision of the international guidelines on the management of adrenal incidentalomas2 is underway.
Our study1 also provides a definitive prospective validation of urine steroid metabolomics-the combination of mass spectrometry-based multisteroid profiling with steroid data analysis by a machine learning-based algorithm. We showed that urine steroid metabolomics had twice the positive predictive value of either tumour diameter or imaging characteristics including unenhanced tumour attenuation in the detection of adrenocortical carcinoma and that the combination of these three tests provided the highest diagnostic accuracy, which should be reflected in future international guideline recommendations.
Impaired kidney function was not an explicit exclusion criterion for enrolment in the EURINE-ACT study. However, we do not anticipate that chronic kidney disease will compromise the diagnostic accuracy of our approach, except in patients on kidney replacement therapy. Steroid profiling in patients with chronic kidney disease previously showed minor changes affecting glucocorticoid metabolism,3 which are unlikely to affect test results. Diagnostic assessment by urine steroid metabolomics is based on the detection of changes in the pattern of steroid excretion, specifically the propensity of adrenocortical carcinomas to overproduce a distinct set of precursor steroids, rather than absolute quantitative cutoffs for individual steroid metabolites that might or might not be affected by impaired renal function.
We excluded any patients taking drugs that could affect steroid biosynthesis and metabolism. In clinical practice, intake of glucocorticoids and inhibitors of key branch points of steroidogenesis, such as ketoconazole and metyrapone, could affect the accuracy of our approach. However,
THE LANCET Diabetes & Endocrinology, 8 (2020) 877-878. doi:10.1016/S2213-8587(20)30345-4