Steroid profiling as a biomarker tool in the diagnosis and monitoring of adrenal tumours

Adrenal tumours are common, affecting 1-2% of the general population. However, these tumours become more common with ageing and 3% of 40-year-olds and 10% of 70-year-olds have an adrenal tumour, many of them without knowing it. The majority of adrenal tumours are discovered incidentally, mostly following an imaging procedure like CT scanning arranged for other reasons than suspected adrenal disease. Once discovered, nodules in the adrenal need to be carefully evaluated as they may overproduce hormones with a potentially detrimental effect on health and, importantly, they could be malignant, either arising from the adrenal itself or representing metastases of other malignancies. Currently we use different imaging procedures and blood tests in the work-up of patients with adrenal tumours, but this is not always informative enough and it is often difficult to decide whether an adrenal tumour is benign or malignant. We have developed a biomarker tool that measures the excretion of steroid hormones deriving from the adrenal gland in urine samples provided by adrenal tumour patients. Preliminary results indicate that there are highly significant differences in steroid hormone excretion between benign and malignant adrenal tumours. In this proposal we will validate this method by the prospective analysis of a large number of urine samples from adrenal tumour patients by our measurement technique, called gas chromatography/mass spectrometry (GC/MS). Measurement results will be analyzed in collaboration with mathematicians who will help us to identify those hormones that will inform us best about the differences between benign and malignant tumours. Use of a related technology method - liquid chromatography/ mass spectrometry, LC/MS/MS ? will then be employed for the rapid measurement of steroids, which will help to integrate our biomarker tool into routine clinical practice. Patient recruitment to this study will be driven by national and international collaborative networks including the European Network for the Study of Adrenal Tumours (ENS@T), which will ensure that we can recruit the necessary patient number during the study lifetime. This proposal links a team of experienced researchers from different disciplines, which will ensure that we succeed in improving the care of patients with adrenal tumours.

Adrenal tumours are common, with a prevalence of 1-2% in the general population. Prevalence increases with age, with 3% of 40-year-olds and 10% of 70-year-olds harbouring an adrenal tumour. In an ageing society and with the increasingly widespread use of abdominal imaging the number of incidentally discovered adrenal tumours is rapidly increasing and represents a huge burden for the health system. Diagnostic work-up of adrenal incidentalomas aims to exclude malignancy and steroid excess, but is currently compromised by a lack of sensitive diagnostic tools. This shortcoming is addressed by this proposal that aims to establish steroid profiling as a biomarker tool in the differential diagnosis of adrenal incidentaloma. Feasibility studies analysing urinary steroid metabolite excretion in 83 benign and 35 malignant adrenal tumours by gas chromatography/mass spectrometry indicate significant differences in steroid excretion, with high androgen and glucocorticoid precursors in malignant tumours. Preliminary analysis indicates a high diagnostic sensitivity and specificity of a distinct subset of steroid metabolites. Based on these data we plan to analyse 24-h urinary steroid excretion in a large cohort of adrenal incidentaloma patients (n=1000) to prospectively evaluate the diagnostic value of this novel biomarker tool. In an important transfer step to implementation of the biomarker tool in routine clinical practice, we will establish liquid chromatography/tandem mass spectrometry (LC/MS/MS) for rapid high throughput measurement of steroid markers identified as most differentiating by biocomputational analysis. The latter will apply machine learning techniques, taking two different but complementary approaches, prototype-based relevance learning and probabilistic Bayesian kernel models. These methods will be utilised to identify discriminative steroids, determine their prognostic and predictive value and develop generative models that could also give mechanistic insights. Recruitment to the study will be facilitated by local, national and international networks including the European Network for the Study of Adrenal Tumours (ENS@T), which already contributed 500 urine samples in 12 months, facilitating preliminary analysis. Clinical data will be stored in a common database format that has been developed, agreed and implemented by ENS@T and that will be made available to all participating centres of the UK Adrenocortical Tumour Network (UK ACT). This proposal links an extensive team of researchers and research facilities across the UK and Europe, generating a unique, interdisciplinary expertise that will ensure the success of this project; results will improve the diagnostic management of a common condition.