Toward the reliable diagnosis of indeterminate thyroid lesions: a HRMAS NMR-based metabolomics case of study.

Cytological analysis of thyroid nodules detected using ultrasound-guided fine-needle aspiration technique is an efficient method for the diagnosis of well-differenciated tumors such as papillary thyroid carcinoma. However, for between 10 to 30% of all the nodules, the cytological analysis based on fine-needle aspiration biopsies leads to an "indeterminated" identification. Consequently, a surgical excision is then necessary for a definite histological diagnosis of the lesions, resulting in 85% of the patient with indeterminated nodules undergoing unnecessary surgery since their tumor is finally diagnosed as benign. In this work, we discuss how HRMAS (1)H NMR-based metabolomics could be a complementary tool for the diagnosis of these elusive cases. We first showed that our approach was able to discriminate clearly any types of thyroid lesions from healthy tissues. Then we proceeded to demonstrate that the information produced by (1)H HRMAS NMR spectra differentiate tumors according to their malignancy grade, even when they belong to the "indeterminate" category. Analysis of the discriminating spectral area in this last case points out toward a possible increase of phenylalanine, taurine, and lactate and a decrease of choline and choline derivatives, myo- and scyllo-inositol in the malignant tumors compared to the benign ones.