Non-invasive diagnosis of papillary thyroid microcarcinoma: a NMR-based metabolomics approach

Abstract

Papillary thyroid microcarcinoma (PTMC) is a subtype of papillary thyroid carcinoma (PTC). Because its diameter is less than 10 mm, diagnosing it accurately is difficult with traditional methods such as image examinations and FNA (Fine Needle Aspiration). Investigating the metabolic changes induced by PTMC may enhance the understanding of its pathogenesis and provide important information for a new diagnosis method and treatment plan. In this study, high resolution magic angle spin (HRMAS) spectroscopy and 1H-nuclear magnetic resonance (1H-NMR) spectroscopy were used to screen metabolic changes in thyroid tissues and plasma from PTMC patients respectively. The results revealed reduced levels of fatty acids and elevated levels of several amino acids (phenylalanine, tyrosine, lactate, serine, cystine, lysine, glutamine/glutamate, taurine, leucine, alanine, isoleucine and valine) in thyroid tissues, as well as reduced levels of amino acids such as valine, tyrosine, proline, lysine, leucine and elevated levels of glucose, mannose, pyruvate and 3-hydroxybutyrate in plasma, are involved in the metabolic alterations in PTMC. In addition, a receiver operating characteristic (ROC) curve model for PTMC prediction was able to classify cases with good sensitivity and specificity using 9 significant changed metabolites in plasma. This work illustrates that the NMR-based metabolomics approach is capable of providing more sensitive diagnostic results and more systematic therapeutic information for PTMC.

Keywords: NMR; diagnosis; metabolomics; papillary thyroid microcarcinoma.

Conflict of interest statement

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

Figures

Figure 1. Representative HE-stained sections of thyroid

×200 A., diffuse thyroid nontoxic goiter×200 B., and papillary carcinoma×200 C., ×400 (C-a), ×800 (C-b).

Figure 2. Multivariate data analysis of thyroid tissue metabolomics between PTMC and healthy groups

A. OPLS-DA score plot, R2=0.84, Q2=0.76; B. Loadings plot; C. VIP scores. 1. PTMC groups; 2. Healthy groups

Figure 3. Coefficient-coded loading plots for the models discriminating between PTMC group and healthy groups

Peaks in the positive direction indicate metabolites that are more abundant in the PTMC groups than healthy group (↑PTMC); Peaks in the negative indicate metabolites that are more abundant in the healthy group than PTMC group (↓Healthy).

Figure 4. Multivariate data analysis of plasma metabolomics between PTMC and healthy groups

A. PLS-DA score plot; R2 = 0.85, Q2 = 0.81 B. Loadings plot; C. VIP scores. 1. PTMC groups; 2. Healthy groups.

Figure 5. Box plots showing representative metabolite changes between PTMC and healthy groups

**p < 0.01;***p < 0.001.

Figure 6

A. Receiver operating characteristic curve showing PLS-DA model ability to predict thyroid tumor malignancy;B. Predicted class plot showing the discrimination between thyroid lesions and their healthy counterpart tissues.

Figure 7. Flow chart of participant selection