Clinical Performance of Absolute Quantification of TSH Receptor mRNA As a Diagnostic Marker for Circulating Thyroid Cancer Cells: Comparison with FNA Biopsy for Thyroid Cancer Detection

Presentation Number: PP39-4
Date of Presentation: June 23rd, 2014

WanMing 2164442714 Zhang1, Manjula K Gupta*2, Eren Berber1, Christian E Nasr1, Allan E Siperstein1 and Kresimira M Milas3
1Cleveland Clinic, Cleveland, OH, 2Cleveland Clin Fndn, Cleveland, OH, 3Oregon Health Science University, Portland, OR


Background:  While thyroid nodules are remarkably common, there is still considerable diagnostic challenge to accurately identify which nodules are malignant lesions. We have previously reported an effective use of the  relative quantitative real-time RT-PCR assay for TSH receptor mRNA (TSHR) as a marker of circulating cancer cells in blood that utilized thyroid cancer tissue total mRNA as reference preparation[JCEM 92:468, 2007]. We have revalidated this assay using synthesized mRNA copy number for absolute quantitation of TSHR mRNA. The performance of this test in routine clinical practice is reassessed for preoperative diagnosis of thyroid cancer and compared to FNAb results. We also attempted to correlate TSHR mRNA levels with BRAF mutation detected in surgical specimens.

Patients and Method:  TSHR mRNA levels in peripheral blood mononuclear cells were measured by a real-time RT-PCR (qRT-PCR) assay. Complemented RNA (cRNA) for TSHR was synthesized and used for calibration and results were expressed as copy number per reaction. Total of 70 patients (thyroid cancer 38, benign controls 32) who underwent thyroid surgery and had final pathologic diagnosis were analyzed by both relative and absolute qRT-PCR methods. Among these 70 patients, 54 had FNAb prior to surgery and 26 had BRAF mutation analyzed post-surgery. TSHR mRNA results from thyroid cancer patients were compared with controls using non-parametric method.  Clinical sensitivity and specificity was analyzed using receiver operating curve (ROC) analysis.

Results:The absolute copy number of TSHR showed excellent correlation with the relative qRT-PCR assay using cancer tissue total RNA as standard [R=0.99;P<0.0001]. Copy number had analytical linearity over 100-7000000 copy per reaction. The intra- and inter-assay variations were 7 and 16%, respectively. The patients with thyroid cancer had significantly higher copy number than those of benign controls (median 1212 vs. 534, P=0.0001). Using a cutoff level of 918 copy/microliter, TSHR mRNA showed a 68% sensitivity and 78% specificity [PPV=79%; NPV=68%] in differentiation of malignant nodules from benign nodules with overall accuracy of 73%. In contrast FNAb only correctly classified 39% of patients [PPV=100%; NPV 54%] and 42% patients had indeterminate results. Furthermore, TSHR mRNA correctly classified 65% of indeterminate FNA (15 out of 23) and may have prevented some unnecessary surgeries. TSHR mRNA copy numbers were not correlated to pathological stages (P=0.095) or B-RAF mutation status(P=0.68 in tumor tissue.

Conclusions: TSHR qRT-PCR assay showed considerable pre-operative diagnostic accuracy for thyroid cancer, the copy number of TSHR mRNA is a sensitive indicator of cancer detection. The copy number of TSHR mRNA in blood is not correlated to pathological stage and to BRAF status of tumor.


Nothing to Disclose: WZ, MKG, EB, CEN, AES, KMM