HTC-15 - Abstract

Abstract Title: From GC-MS to LC-MS/MS: Further Advances in Adrenal Cancer Diagnosis
Abstract Type: Seminar
Presenter Name: Dr Angela Taylor
Co-authors:Dr Michael Biehl
Dr Alice Sitch
Dr Irina Bancos
Prof Wiebke Arlt
Company/Organisation: University of Birmingham
Session Choice: Advances in Clinical Analysis

Abstract Information :

Discriminating adrenocortical adenoma (ACA) from adrenocortical carcinoma (ACC) represents a continuous challenge in patients with adrenal masses.

We previously established urine steroid metabolomics, (the combination of urinary steroid profiling by GC-MS and machine learning-based data analysis), to distinguish adrenal cancer from benign adrenal tumours, achieving higher sensitivity and specificity than currently available imaging techniques. (JCEM2011; 96 (12): 3775-84). However, GC-MS is labour-intensive, expensive and low-throughput. Here we developed a urinary steroid profiling method using liquid chromatography tandem mass spectrometry (LC-MS/MS) and report method optimisation (using Multi-platform Unbiased-optimisation of Spectrometry via Closed Loop Experimentation software MUSCLE), validation, and cross validation to GC-MS.

We analysed 24-hr urinary steroid excretion by GC-MS and LC-MS/MS in:

  • 481-anonymised comparison urines from a range of disorders associated with steroid excess and deficiency
  • 129 healthy controls
  • 99-ACA
  • 40-ACC
Comparison of the mass spectrometry methods using correlation and Bland-Altman plots showed significant P<0.001 correlations for all steroids (range 0.78-0.90).

To determine diagnostic accuracy, (the ability to distinguish ACC from ACA), ROC curves were generated. GC-MS and LC-MS/MS demonstrated similar diagnostic accuracy, (area under the ROC curve average 0.969 (SD=0.044) and 0.954 (0.067) respectfully).

Development of this novel LC-MS/MS method represents a significant advance diagnosis of ACC.