ARFI in DIFFERENTIATION of MALIGNANT from BENIGN THYROID NODULES with a cut-off value of Elastic V=2.545m/s

-----------------------

Acoustic Radiation Force Impulse (ARFI) Imaging of Thyroid Nodules 

Vo Mai Khanh, Nguyen Thien Hung, Phan Thanh Hai 

MEDIC MEDICAL CENTER, HCMC, Vietnam

Purpose: The aim of the present study was to evaluate the feasibility of ARFI-measurements in combining of VTI in the thyroid nodule.

Methods and materials: All patients underwent conventional ultrasound, ARFI-imaging and cytological assessment. ARFI-imaging (VTI and VTQ technology) were performed with 9L4 probe, using Siemens (ACUSON S2000) B-mode-ARFI combination transducer. 

Results and Discussions: 130 nodules were available for analysis. 103 nodules were benign on cytology, 20 nodules were malignant (papillary carcinomas), and 7 follicular lesions. The average velocity of ARFI-imaging in benign and malignant thyroid nodules was of 2.4 m/s,  and of 3.2 m/s, respectively. A sensitivity of 79.4% and specificity  of 53.7% of ARFI-imaging could be achieved using a cut-off of 2.19 m/s (area under ROC curve is 0.731, p under  0.0001).
Conclusions: ARFI can be performed in thyroid nodule with reliable results. ARFI might be the reference criteria for differentiation of benign and malignant thyroid nodules. 

PURPOSES:

Acoustic radiation force impulse (ARFI) imaging is a novel ultrasound-based elastography method enabling quantitative measurement and qualitative assessment of tissue stiffness. In some recent studies, the feasibility of ARFI for evaluating the thyroid gland was shown. Most of those studies used VTQ (Virtual Touch Tissue Quantification) of ARFI-Imaging to measure the shear wave speed of tissue. However, the VTI (Virtual Touch Tissue Imaging) was still available, but it was subjective and dependent on the experience of the sonographer.

The aim of the present study was to evaluate the VTQ of normal thyroid tissues, benign and malignant thyroid nodules. Besides, VTI (Virtual Touch Tissue Imaging) of ARFI-Imaging was assessed in these objects as well.

MATERIALS AND METHODS:

A cross-sectional study was done from August 2011 to October 2012 at Medic Medical Center.  One hundred and thirty nodules  underwent conventional ultrasound, including Color Doppler ultrasound using a 7.5MHz linear transducer. Next, nodule stiffness were measured and assessed by VTQ and VTI of ARFI-Imaging (Acuson Siemens S2000).  The Region-of-interest (ROI) placed at the center of nodules and in the healthy thyroid gland away from thyroid nodules. In addition, five measurements were performed with the ROI for each nodule.  Exclusion criteria were “X.XX m/s” measurements. With VTI assessment, each nodule was assessed by 2 separate examiners. Afterward, FNAC (Fine needle aspiration cytology) under ultrasound guide was used as reference method for the diagnosis of benign and malignant thyroid nodules.

Statistical analysis was performed using Medcalc for Windows.

RESULTS:

Patient characteristics:

Age: 45 (range 16 – 69)

Size of nodule: 14mm (range 5 – 47mm)

Table 1: ARFI velocity characteristics:

ARFI velocity (m/s)	Normal thyroid	Benign nodule	Malignant nodule
Mean ± standard deviation	1.51 ± 0.07	2.15 ± 0.09	3.21 ± 0.46
Median	1.41	2.11	2.59
Minimum	0.84	0.80	0.90
Maximum	3.00	4.04	9.22

Figure 1: Receiver-operating characteristic (ROC) curve for VTQ values for diagnosis of benign and malignant thyroid nodules (AUROC 0.731, p under 0.0001).

Figure 2: Receiver-operating characteristic (ROC) curve for the difference of VTQ between normal thyroid tissue and thyroid nodule for diagnosis of benign and malignant thyroid nodules (AUROC 0.72, p under 0.0001)

Table 2:  Frequency table and Chi-square test for independence of VTI and the differentiation of benign and malignant nodules:

Codes X	VTI
Codes Y	FNAC

	Codes X
Codes Y	1	2	3	4
0	30	10	89	70	199 (76.5%)
1	51	1	4	5	61 (23.5%)
	81 (31.0%)	11 (4.2%)	93 (35.7%)	75 (29.1%)	260

Chi-square	102.553
DF	3

 Chi-square statistic = 102.553, predetermined alpha level of significance = 0.001,  degrees of freedom (DF=3), there is a relationship between VTI and the differentiation of benign and malignant thyroid nodules.

DISCUSSION:

-          The mean VTQ value of  benign thyroid nodules was higher than the one of normal thyroid tissue and lower than malignant thyroid nodules. It was reasonable because most of malignant nodules had harder stiffness than benign nodules.

-          However, the VTQ values were fluctuant and overlapped among these groups.  A malignant nodule was not completely solid, there were some necrosis regions inside it. Similar to benign nodule, some calcifications could make it become harder. Besides, the region of interest (ROI) of Acuson Siemens S2000 was rather big (D=6x5mm) and unadjustable. With some nodules were smaller than 6mm in size, the ROI could involve normal thyroid tissue in measuring VTQ. Anyhow, because AUROC was 0.731, VTQ of ARFI-Imaging could be considered as a helpful method in differentiating of benign and malignant thyroid nodules.

-           Some thyroid nodules combined with diffuse thyroid diseases (Basedow-Graves’ disease, chronic autoimmune thyroiditis,…) might influent to the result of VTQ. Recent study by Sporea I. et al, the stiffness of normal thyroid tissue was lower than in Graves’ disease and chronic autoimmune thyroiditis. In this study, we got VTQ values of thyroid tissue of the other lobe and measured the difference of VTQ values between thyroid nodule and thyroid tissue of the other lobe. The AUROC was 0.72, it meant the difference of VTQ between them would be considered to be “fairly good” at separating benign and malignant nodules. The harder nodule and the larger difference of VTQ between nodule and thyroid tissue of the other lobe were, the larger probability of malignant nodule was.

-          VTI was a qualitative variable therefore it depended on examiners. We had 2 examiners working separately and blinded with FNAC results. Each nodule had 2 evaluations in the classification (including 4  groups: dark, iso, bright and mixed color). The darker VTI was, the more malignant thyroid nodule was. The Chi-square test result showed a dependance between VTI and the differentiation of benign and malignant nodules, alpha level of significance = 0.001.

-          But this study existed some disadvantages:

+ This is a prospective study and FNAC was used as a reference method in differentiating benign and malignant nodules. As we all know, FNAC was not a gold standard of thyroid nodule’s diagnosis, histology was. But in our conditions of an out-patient clinic, it was impossible to have all postoperative results. Multicenters studies are awaited.

+ Our inclusion criteria was any size thyroid nodules detected by ultrasound. This might affect the VTQ result of Acuson Medison S2000 because of the big ROI. We suggested chosing nodules with > 10mm in size or waiting for another improvement of smaller ROI are.

+ The diffuse thyroid diseases combined thyroid nodules were popular, but in this study, they were not enough to do a statistical analysis. Larger studies are awaited.