High-Resolution Melting Is a Sensitive, Cost-Effective, Time-Saving Technique for BRAF V600E Detection in Thyroid FNAB Washing Liquid: A Prospective Cohort Study

Abstract

Objective: The diagnostic accuracy of thyroid fine needle aspiration biopsy (FNAB) can be improved by the combination of cytological and molecular analysis. In this study, washing liquids of FNAB (wFNAB) were tested for the BRAF V600E mutation, using the sensitive and cost-effective technique called high-resolution melting (HRM). The aim was to demonstrate the feasibility of BRAF analysis in wFNAB and its diagnostic utility, combined with cytology.

Design: Prospective cohort study.

Methods: 481 patients, corresponding to 648 FNAB samples, were subjected to both cytological (on cells smeared onto a glass slide) and molecular analysis (on fluids obtained washing the FNAB needle with 1 ml of saline) of the same aspiration. BRAF V600E analysis was performed by HRM after methodological validation for application to wFNAB (technique sensitivity: 5.4%).

Results: The cytological results of the FNAB were: 136 (21%) nondiagnostic (THY1); 415 (64%) benign (THY2); 80 (12.4%) indeterminate (THY3); 9 (1.4%) suspicious for malignancy (THY4); 8 (1.2%) diagnostic of malignancy (THY5). The BRAF V600E mutation was found in 5 THY2, 2 THY3, 6 THY4 and 6 THY5 samples. Papillary carcinoma diagnosis was histologically confirmed in all BRAF+ thyroidectomized patients. BRAF combined with cytology improved the diagnostic value compared to cytology alone in a subgroup of 74 operated patients.

Conclusions: HRM was demonstrated to be a feasible method for BRAF analysis in wFNAB. Thanks to its sensitivity and cost-effectiveness, it might be routinely used on a large scale in clinical practice. In perspective, standby wFNAB samples could be analyzed a posteriori in case of indeterminate cytology and/or suspicious findings on ultrasound.

Keywords: BRAF gene; High-resolution melting; Papillary thyroid cancer; Thyroid; Washing liquids of fine needle aspiration biopsy.

Fig. 1

Study design. BRAF+ = Positivity of BRAF mutation by HRM; BRAF-= wild-type sample by HRM; THY1 = nondiagnostic; THY2 = benign; THY3 = indeterminate; THY4 = suspicious for malignancy; THY5 = diagnostic of malignancy.

Fig. 2

Different clusters of HRM melting curves. The three different clusters, present in the picture, are obtained by the amplification and HRM analysis of exon 15 of the BRAF gene. While the DNA of V600D+ and V600E+ controls (green and red curves in a-d, respectively) were extracted by a commercial kit, the somatic DNA from wFNAB was obtained by pelletting the cells and by using 60 µl of lysis buffer (50 mM Tris-HCl at pH 8.5, 1 mM EDTA, 0.5% Tween 20 and sterile water) with 20 µl of proteinase K (10 mg/ml; Roche Diagnostics), and by incubating the samples at 56°C overnight and then at 95°C for 20 min to have a high yield from few cells also. The cluster of BRAF-samples is represented by a blue curve. All HRM analyses were performed using the SSO Fast Eva Green Supermix 2× (Bio-Rad Laboratories) and the same following protocol: 98°C for 2 min, 44 cycles of 3 s at 98°C and 30 s at 56.1°C, 1 cycle at 98°C for 30 s and 65°C for 1 min and 30 s, a progressive denaturation from 65 to 83°C, increasing the temperature by 0.2°C every 10 s and recording the fluorescence intensity for each increment. -d(RFU)/dT = Negative derivative (-d) of relative fluorescence units (RFU) over temperature (dT). a Original melting curves of V600D+, V600E+ and wild-type samples in green, red and blue, respectively. b Melting peaks showing the three characteristic shapes of V600D+ (green), V600E+ (red) and wild-type (blue) samples. c Normalization of V600D+ (green), V600E+ (red) and wild-type (blue) melting curves. d The differences between the three normalized curves, belonging to V600D+ (green), V600E+ (red) and wild-type (blue) samples.

Fig. 3

Test of sample stability over time and reliability of the HRM results. We decided to test the stability of the refrozen DNA after a lapse of time by HRM analysis. Thus, 3 groups of samples, thawed and tested by HRM for the first time in April 2012, 2013 and 2014 were rethawed in May 2014 and reanalyzed by HRM, obtaining for all samples good amplification curves and excellent melting profiles, comparable to the control profiles, also confirming the same results of their first HRM analyses. -d(RFU)/dT = Negative derivative (-d) of relative fluorescence units (RFU) over temperature (dT). a The green and the red curves refer to the negative (wild-type) and to the BRAF V600E+ control, respectively. b-d Melting profiles obtained in May 2014, of BRAF- samples that were analyzed by HRM for the first time in April 2012 (b), April 2013 (c) and April 2014 (d), respectively.