Vesicular thyroid tumours evolve from benign to malignant states, respectively called adenomas and carcinomas by respect to the World Health Organisation classification. Borderline cases, atypical adenomas, cannot be clearly classified. When an abnormal cytoplasmic mitochondrial proliferation occurs into the tumour cells, these tumours called oncocytomas, are more aggressive. Oncogenic mechanisms and signal pathways of these tumours are not well known and the origin of the oncocytic signal is uncertain. Early detection of this feature is needed to choose appropriate treatments for patients.
We studied the oncocytic signal of 166 thyroid tissues by statistically analysing data from microarrays containing more than 9000 genes. We analysed normal thyroid tissues, micro and macrovesicular adenomas, oncocytic adenomas, atypical adenomas, as well as vesicular and oncocytic variant of vesicular carcinomas.
We propose a new classification of vesicular thyroid tumours using the false discovery rate method that increases the statistical power of multiple comparisons from microarrays data. The 368 genes that lead to distinguish the vesicular classes show different metabolic profiles on gene ontology analysis. We show that oncocytic feature occurs before the oncogenetic signal by contrast to the current generalized hypothesis. Interestingly, the oncocytic signal can be detected by the gene-expression levels of only few genes. This allows us to propose a gene-expression-based diagnostic predictor that may be useful for routine screening of oncocytic tumours.
01 - 05 Apr 2006
European Society of Endocrinology