OR14-1 Midkine Concentration in FNA Washout: a Novel Diagnostic Approach to Identify Papillary Thyroid Carcinoma

Program: Abstracts - Orals, Featured Poster Presentations, and Posters
Session: OR14-Thyroid Cancer: Insights into Diagnosis & Treatment
Saturday, June 15, 2013: 11:30 AM-1:00 PM
Presentation Start Time: 11:30 AM
Room 103 (Moscone Center)
Youn Hee Jee*1, Francesco S. Celi2, Electron Kebebew3, Maureen Sampson4, David B. Sacks4, Alan Remaley4 and Jeffrey Baron1
1NICHD, NIH, Bethesda, MD, 2National Institute of Diabetes and Digestive and Kidney Diseases, NIH, Bethesda, MD, 3National Cancer Institute, NIH, Bethesda, MD, 4Clinical Center, NIH, Bethesda, MD
Background:The primary preoperative method for distinguishing malignant from benign thyroid nodules is fine-needle aspiration (FNA) cytology, but it is frequently inconclusive. Midkine is a heparin binding growth factor which is preferentially expressed in papillary thyroid carcinoma (PTC). We hypothesized that midkine concentration in PTC would be higher than in benign lesions and therefore its measurement in FNA samples might aid in the preoperative diagnosis of thyroid nodules.

Methods: 44 subjects (age 46 ± 11 y, 9 male) underwent thyroidectomy providing a histological diagnosis. Of these, 24 also underwent preoperative FNA. FNA needle contents were expressed for cytology and then the needle was washed with buffer for immunoassay. For subjects without preoperative FNA samples, FNA was performed ex vivo on surgically excised tissue. Midkine and thyroglobulin were measured using a high-sensitivity sandwich ELISA and chemiluminescent immunoassay, respectively. Samples with thyroglobulin < 50 ng/mL were excluded on the assumption of insufficient tissue sampling. Values from multiple passes (1-4/nodule) were averaged.

Results:Midkine concentration was higher in 11 PTCs (4 classical, 4 tall cell variant, 3 follicular variant) than in 40 benign nodules (34 adenomatoid nodules including 4 hyperplastic, 4 follicular and 2 Hurthle cell adenomas) (0.5 ± 0.2 vs 0.08 ± 0.03 ng/mL, mean ± SEM, p = 0.004). To adjust for tissue content and to enhance differences between benign and malignant samples, midkine was normalized to thyroglobulin. Midkine/thyroglobulin ratio in PTC was greater than in benign nodules (297 ± 163 vs 1.5 ± 0.4 ng/mg, p<0.001). Using a threshold of 10 ng/mg, the sensitivity and specificity of the midkine/thyroglobulin ratio for diagnosis of PTC were 64% and 100%, respectively. The area under the ROC curve was 0.84. All follicular variant PTCs had low midkine/thyroglobulin (1.0 ± 0.6 ng/mg). Plasma midkine concentrations did not differ between subjects with benign nodules and with PTC (0.26 ± 0.03 ng/mL vs 0.19 ± 0.02 ng/mL, p = NS).

Conclusion:In FNA samples, the midkine/thyroglobulin ratio in PTC is greater than in benign thyroid nodules. Because FNA cytology is often non-diagnostic, the midkine/thyroglobulin ratio may represent a novel adjunctive diagnostic tool to help identify malignancies. Larger studies are warranted to confirm the diagnostic utility for PTC and investigate the utility for other thyroid cancers.

Nothing to Disclose: YHJ, FSC, EK, MS, DBS, AR, JB

*Please take note of The Endocrine Society's News Embargo Policy at http://www.endo-society.org/endo2013/media.cfm

Previous Abstract | Next Abstract >>