The incidence of papillary thyroid cancer (PTC) is increasing around the world according to several epidemiological studies. The increase is mainly due to the detection of papillary carcinomas less than 1-2 cm in size, while the incidence of larger tumors is stable [
From July, 2011, to July, 2013, 173 nodules in 157 patients were included in this retrospective study. The inclusion criteria of the target nodules were as follows: (1) suspicions of malignancy based on the features on gray-scale US; (2) the diameter of the nodules ranged from 0.5 cm to 1.0 cm; (3) solid or almost solid (<25% cystic) nodules; (4) the cystic and calcified portions of the nodule can be excluded out of the region of interest (ROI) when using the ARFI examination; (5) no treatment such as ablation or biopsy was performed on the nodules; (6) enough thyroid tissue surrounding the nodule at the same depth; (7) pathological results were obtained. The flowchart for the selection of thyroid nodules was shown in Figure
The flowchart of selection of the patients with thyroid nodules.
All gray-scale US and color Doppler flow imaging (CDFI) images were obtained by using the same scanner (S2000 US machine, Siemens Medical Solutions, Mountain View, CA, USA) and same linear array transducer (9L4, Siemens Medical Solutions, Mountain View, CA, USA) with a center frequency of 7.5 MHz (range, 5.0–14.0 MHz).
The conventional US and ARFI elastography imaging were performed by one investigator who had more than 5-year experience in thyroid US. On gray-scale US, the following features of the nodules were evaluated: echogenicity, margin, calcification, shape, and halo sign. Echogenicity was classified as hyper-, iso-, hypoechogenicity (in comparison with the normal thyroid tissue) or marked hypoechogenicity (when a nodule showed relatively hypoechogenicity compared with the surrounding strap muscle). Margin was classified as well-defined or ill-defined (microlobulated or irregular margin). Shape was classified as ovoid to round, taller than wide, or irregular. Calcification was classified as microcalcification (less than or equal to 1 mm in diameter; tiny, punctate, hyperechoic foci, either with or without acoustic shadows), macrocalcification (including calcification more than 1 mm in diameter and eggshell calcification), or no calcification. Halo sign was classified as presence or absence of halo sign. The color Doppler flow pattern was classified into three types: type I, absence of blood flow; type II, perinodular and absent or slight intranodular blood flow; type III, marked intranodular and absent or slight perinodular blood flow.
ARFI elastography images were obtained after evaluating the nodules with conventional US using the same probe and by the same operator. ARFI elastography included virtual touch tissue quantification (VTQ) and virtual touch tissue imaging (VTI). The patient was asked to hold the breath when the VTI and VTQ mode was initiated. The probe was placed gently on the body surface with light pressure to the thyroid. VTI and VTQ imaging were performed both on the long axis dimension of the nodule.
VTI of ARFI was then carried out firstly. The VTI image reflects the elasticity of tissue with gray-scale image in the field of view (FOV), in which the dark indicates hard tissue whereas the bright indicates soft tissue. The FOV is adjusted to include the whole lesion (approximately occupying 70% of the whole FOV) and some surrounding thyroid tissue (approximately occupying 20–30% of the whole FOV). The VTI images of the thyroid lesions were thereafter scored according to Xu’s scoring system [
After the VTI imaging, VTQ was performed. VTQ can reflect the elasticity of tissue quantitatively with the shear wave velocity (SWV). The ROI was placed on the solid portion of the nodule. The calcified and liquefaction necrosis portions of the nodule were avoided during measurement. The size of ROI was 6 mm × 5 mm and cannot be altered. The SWV value was displayed on the screen as m/s (range from 0 to 9 m/s). If the stiffness of the tissue is beyond the limits of measurement, whether high or low, the SWV would be displayed as “x.xx m/s.” The measurement was repeated for 7 times. The median of all 7 measurements per lesion was calculated and used for further analysis. The value of “x.xx m/s” was allocated to be 9 m/s after excluding the possible influencing factors such as patient’s respiration or motion and operator’s inappropriate gesture [
Statistical analysis was performed by using a software package (SPSS, version 17.0 for Windows; SPSS, Chicago, Ill). Quantitative data were expressed as mean ± standard deviation. The mean ages and mean diameters of benign and malignant nodules were compared with
In the 173 thyroid nodules, there were 96 (55.5%) malignant nodules and 77 (44.5%) benign nodules. All the 96 malignant nodules were found to be papillary microcarcinomas after surgery. As to the 77 benign nodules, there were nodular goiter in 61 (79.2%), follicular adenoma in 1 (1.3%), and Hashimoto nodule caused by Hashimoto thyroiditis in 15 (19.5%), respectively.
Basic characteristics, gray-scale US features, and CDFI pattern of benign and malignant nodules were shown in the Table
The basic characteristics and ultrasound features for the 157 patients with 173 thyroid nodules.
Characteristics | Benign | Malignant |
|
---|---|---|---|
Patients ( |
|||
Sex (Male/Female) | 20/51 | 15/71 | 0.108 |
Age (yrs) | 54 ± 8 (32–71) | 49 ± 11 (22–78) | 0.001 |
Single nodule/multiple nodules | 11/60 | 20/66 | 0.224 |
Nodules ( |
|
|
|
Diameter (mm) | 8.2 ± 1.4 (5–10) | 7.7 ± 1.5 (5–10) | 0.031 |
Location | 0.507 | ||
Left lobe | 31 | 39 | |
Right lobe | 44 | 51 | |
Isthmus | 2 | 6 | |
Echogenicity | 0.001* | ||
Markedly hypoechoic | 20 | 46 | |
Hyperechoic | 1 | 0 | |
Isoechoic | 16 | 4 | |
Hypoechoic | 36 | 45 | |
Mixed | 4 | 1 | |
Calcifications | 0.001* | ||
None | 39 | 34 | |
Microcalification | 22 | 53 | |
Macrocalcification | 16 | 9 | |
Shape | <0.001* | ||
Ovoid to round | 55 | 39 | |
Taller than wide | 13 | 52 | |
Irregular | 9 | 5 | |
Margin | 0.176 | ||
Well-defined | 48 | 50 | |
Ill-defined | 29 | 46 | |
Halo sign | 0.023 | ||
Present | 7 | 1 | |
Absent | 70 | 95 | |
Vascularity | <0.001* | ||
Type I | 54 | 79 | |
Type II | 14 | 0 | |
Type III | 9 | 17 |
Caption:
Significant differences were found in age of patients, diameter of nodules, echogenicity of nodules, calcification of nodules, shape of nodules, halo sign (all
VTI scores of thyroid nodules with different pathology types were presented in Table
Pathology types and virtual tissue imaging scores of the thyroid nodules.
VTI |
Benign nodules | Malignant nodules | |||
---|---|---|---|---|---|
Nodule |
Follicular |
Hashimoto |
Total* |
Papillary | |
1 ( |
1 | 0 | 0 | 1 | 1 |
2 ( |
23 | 0 | 2 | 25 | 19 |
3 ( |
25 | 0 | 10 | 35 | 29 |
4 ( |
10 | 0 | 3 | 13 | 30 |
5 ( |
2 | 0 | 0 | 2 | 15 |
6 ( |
0 | 1 | 0 | 1 | 2 |
Caption:
The distribution of VTI score (score 1 to score 6) of benign nodules was significant different from that of the malignant ones. Most of benign nodules were classified to be score 1 to score 3 on VTI score and most of malignant nodules were classified to be score 4 to score 6 (
The mean SWV value of benign and malignant thyroid nodules was 2.57 ± 0.79 m/s (range: 0.90–4.92 m/s) and 3.88 ± 2.24 m/s (range: 1.49–9.00 m/s), respectively, and significant difference was found between them (
ROC curve was used to assess the diagnostic performance of gray-scale US feature, CDFI pattern, VTI score and SWV value in the differential diagnosis between benign and malignant thyroid nodules (Table
Predictive value of conventional US features and ARFI in 173 thyroid lesions.
BN |
CA |
Sensitivity |
Specificity |
PPV |
NPV |
Accuracy |
Az | |
---|---|---|---|---|---|---|---|---|
US features | ||||||||
Hypoechoic |
94.8 |
27.3 |
61.9 |
80.8 |
64.7 |
0.610* | ||
Yes | 56 | 91 | ||||||
No | 21 | 5 | ||||||
Spot microcalcification | 55.2 |
71.4 |
70.7 |
56.1 |
62.4 |
0.633* | ||
Yes | 22 | 53 | ||||||
No | 55 | 43 | ||||||
Shape (taller than wide) | 54.2 |
83.1 |
80 |
59.3 |
67.1 |
0.686 | ||
Yes | 13 | 52 | ||||||
No | 64 | 44 | ||||||
Halo sign | 99.0 |
9.1 |
57.6 |
87.5 |
59.0 |
0.540* | ||
Yes | 7 | 1 | ||||||
No | 70 | 95 | ||||||
Type III vascularity | 17.7 |
88.3 |
65.4 |
46.2 |
49.1 |
0.455* | ||
Yes | 9 | 17 | ||||||
No | 68 | 79 | ||||||
ARFI features | ||||||||
VTI |
61.4 |
88.3 |
86.8 |
64.8 |
73.4 |
0.746 | ||
Yes | 9 | 59 | ||||||
No | 68 | 37 | ||||||
VTQ |
56.2 |
79.2 |
77.1 |
59.2 |
66.5 |
0.702 | ||
Yes | 16 | 54 | ||||||
No | 61 | 42 |
Caption: BN, benign; CA, carcinoma; US, ultrasound; PPV, positive predictive value; NPV, negative predictive value.
After multivariate logistic regression analysis, gender, hypoechogenicity, shape (taller than wide), VTI score ≥ 4, and SWV > 3.10 m/s were found to be independent risk factors in predicting PTMC. After excluding the factors that had no predictive value on multivariate analysis, ORs of gender, hypoechogenicity, shape (taller than wide), VTI score ≥ 4, and SWV > 3.10 m/s were 3.591 (95% CI 1.116–11.558), 4.838 (95% CI 1.288–18.169), 5.478 (95% CI 2.161–13.887), 15.133 (95% CI 5.546–41.296), and 5.891 (95% CI 2.417–14.362), respectively (Table
Multivariate logistic regression analysis for predicting malignant nodules.
Characteristic | OR | 95% CI |
|
---|---|---|---|
Gender | 3.591 | 1.116–11.558 | 0.032 |
Hypoechogenicity | 4.838 | 1.288–18.169 | 0.020 |
Shape (taller than wide) | 5.478 | 2.161–13.887 | 0.000 |
VTI score ≥ 4 | 15.133 | 5.546–41.296 | 0.000 |
VTQ SWV > 3.10 m/s | 5.891 | 2.417–14.362 | 0.000 |
Caption: OR, odd ratio; CI, confidence interval
Gender, hypoechogenicity, shape (taller than wide), VTI elastography score ≥ 4, and SWV > 3.10 m/s were found to be independent risk factors in predicting PTMC.
The treatment for PTMC, which is a tumor measuring less than 1 cm, is still a subject of controversy. PTMC had been considered as a “silent” cancer because most cases followed an indolent course with an excellent prognosis compared to the thyroid carcinoma > 1 cm [
The development of high-resolution US has resulted in a significant increase in the detection of nonpalpable small thyroid nodules, but the diagnostic value of US in nodules ≤1 cm was relatively lower than that in lesions >1 cm [
ARFI is a new technique of elasticity imaging that has been introduced into clinical practice in recent years. ARFI includes VTQ and VTI. VTQ allows quantification measurement of tissue elasticity in a selected region in the lesion whereas VTI estimates the elasticity of the whole lesion. Previous studies have shown that ARFI elastography seems to be a new tool for the differential diagnosis between benign and malignant thyroid nodules. VTQ can reflect the change of elasticity quantitatively in malignant nodules. The mean SWV value of malignant thyroid nodules was significantly higher than that of normal tissue or benign nodules [
Several US features, including presence of calcification, hypoechogenicity, irregular margin, absence of a halo, predominantly solid composition, and intranodule vascularity, have been found to be associated with an increased risk of thyroid cancer (including cancers >1 cm and ≤1 cm) [
Images in a 54-year-old man with nodular goiter. (a) Conventional US shows marked hypoechogenicity, irregular shape, marked shadow, and ill-defined margin. (b) The color flow Doppler shows absent blood flow. (c) Score 2 is assigned at VTI. (d) The SWV of the nodule is “2.37” m/s. The conventional US features indicate probably malignant lesion; however, the ARFI characteristics show it probably benign. (e) Histology of the lesion confirms the diagnosis of nodular goiter. Hematoxylin and eosin stain, ×100.
Images in a 39-year-old woman with papillary thyroid microcarcinoma. (a) Conventional US shows marked hypoechogenicity, a taller-than-wide shape, and well-defined margin. (b) The color flow Doppler shows perinodular blood flow. (c) Score 5 is assigned at VTI. (d) The SWV of the nodule is “4.18” m/s. The conventional US features indicate suspicious diagnosis; however, the ARFI characteristics help make a malignant diagnosis. (e) Histology of the lesion confirms the diagnosis of papillary thyroid carcinoma. Hematoxylin and eosin stain, ×400.
To find the independent risk factors in predicting malignant nodules, multivariate logistic regression analysis was done and gender, hypoechogenicity, shape (taller than wide), VTI score ≥ 4, and SWV > 3.10 m/s were found to be independent risk factors in predicting PTMC. The size of nodules, age of patients, single or multiple nodules, absence of halo sign, calcification, vascularity type were found to be of no value in predicting PTMC. The possible reason for why spot microcalcification was not a risk factor for PTMC may be that some PTMCs were in the early stage of carcinoma; thus, spot microcalcifications had not yet formed.
Moon et al. [
Images in a 62-year-old woman with papillary thyroid microcarcinoma. (a) Conventional US shows hypoechogenicity, a taller-than-wide shape, and well-defined margin. (b) The color flow Doppler shows absent blood flow. (c) Score 2 is assigned at VTI. (d) The SWV of the nodule is “3.12” m/s. The conventional US features and ARFI characteristics both indicate suspicious diagnosis. (e) Histology of the lesion confirms the diagnosis of papillary thyroid microcarcinoma. Hematoxylin and eosin stain, ×100.
The present study had some limitations. Firstly, cervical lymph node metastasis is related to higher incidence of recurrence; it is considered to be an aggressive clinical feature of PTMC [
In summary, VTQ can provide elasticity information of PTMC quantitatively and VTI can directly reflect the overall elastic property of thyroid nodule. VTI score 4 or greater and SWV value more than 3.10 m/s are highly suggestive of malignancy. Gender, hypoechogenicity, shape (taller than wide), VTI score ≥ 4, and SWV > 3.10 m/s are independent risk factors in predicting PTMC. ARFI elastography seems to be a valuable tool for the differential diagnosis between benign and malignant thyroid nodules ≤1 cm.
The authors declare that there is no conflict of interests regarding the publication of this paper.
Yi-Feng Zhang and Chang Liu contributed equally to this paper. Our paper has been reviewed and approved by all authors and all authors have made significant contribution to the content of this paper.
This work was supported in part by Grant SHDC12014229 from Shanghai Hospital Development Center and Grant 2012045 of Shanghai Talent Development Project from Shanghai Human Resource and Social Security Bureau.