Journal List > Int J Thyroidol > v.10(1) > 1082700

TOOLS
Yoon, Lee, Lee, Kim, Moon, and Kwak: Ultrasonographic Evaluation of Diffuse Thyroid Disease: a Study Comparing Grayscale US and Texture Analysis of Real-Time Elastography (RTE) and Grayscale US
Original Article

Int J Thyroidol 2017;10(1):14-23.

Published online: 5 May 2017

DOI: https://doi.org/10.11106/ijt.2017.10.1.14

Ultrasonographic Evaluation of Diffuse Thyroid Disease: a Study Comparing Grayscale US and Texture Analysis of Real-Time Elastography (RTE) and Grayscale US

Jung Hyun Yoon1, Eunjung Lee2, Hye Sun Lee3, Eun-Kyung Kim1, Hee Jung Moon1, Jin Young Kwak1

1Department of Radiology, Severance Hospital, Research Institute of Radiological Science, Yonsei University College of Medicine, Seoul, Korea

2Department of Computational Science and Engineering, Yonsei University, Seoul, Korea

3Department of Research Affairs, Biostatistics Collaboration Unit, Yonsei University College of Medicine, Seoul, Korea

Correspondence: Jin Young Kwak, MD, Department of Radiology, Severance Hospital, Research Institute of Radiological Science, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 03722, Korea Tel: 82-2-2228-7400, Fax: 82-2-393-3035, E-mail: docjin@yuhs.ac

Received 20 October 20169 February 2017       Accepted 9 February 2017

Copyright © 2017, the Korean Thyroid Association

This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creative-commons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

초록

Background and Objectives

To evaluate and compare the diagnostic performances of grayscale ultrasound (US) and quantitative parameters obtained from texture analysis of grayscale US and elastography images in evaluating patients with diffuse thyroid disease (DTD).

Materials and Methods

From September to December 2012, 113 patients (mean age, 43.4±10.7 years) who had undergone preoperative staging US and elastography were included in this study. Assessment of the thyroid parenchyma for the diagnosis of DTD was made if US features suggestive of DTD were present. Nine histogram parameters were obtained from the grayscale US and elastography images, from which ‘ grayscale index’ and ‘ elastography index’ were calculated. Diagnostic performances of grayscale US, texture analysis using grayscale US and elastography were calculated and compared.

Results

Of the 113 patients, 85 (75.2%) patients were negative for DTD and 28 (24.8%) were positive for DTD on pathology. The presence of US features suggestive of DTD showed significantly higher rates of DTD on pathology, 60.7% to 8.2% (p＜0.001). Specificity, accuracy, and positive predictive value was highest in US features, 91.8%, 84.1%, and 87.6%, respectively (all ps＜0.05). Grayscale index showed higher sensitivity and negative predictive value (NPV) than US features. All diagnostic performances were higher for grayscale index than the elastography index. Area under the curve of US features was the highest, 0.762, but without significant differences to grayscale index or mean of elastography (all ps＞0.05).

Conclusion

Diagnostic performances were the highest for grayscale US features in diagnosis of DTD. Grayscale index may be used as a complementary tool to US features for improving sensitivity and NPV.

Keywords: Thyroid, Ultrasound, Diffuse thyroid disease, Elastography, Texture analysis

REFERENCES

1). Moon WK, Choi JW, Cho N, Park SH, Chang JM, Jang M, et al. Computer-aided analysis of ultrasound elasticity images for classification of benign and malignant breast masses. AJR Am J Roentgenol. 2010; 195(6):1460–5.
crossref
2). Kim I, Kim EK, Yoon JH, Han KH, Son EJ, Moon HJ, et al. Diagnostic role of conventional ultrasonography and shear-wave elastography in asymptomatic patients with diffuse thyroid disease: initial experience with 57 patients. Yonsei Med J. 2014; 55(1):247–53.
crossref
3). Marcocci C, Vitti P, Cetani F, Catalano F, Concetti R, Pinchera A. Thyroid ultrasonography helps to identify patients with diffuse lymphocytic thyroiditis who are prone to develop hypothyroidism. J Clin Endocrinol Metab. 1991; 72(1):209–13.
crossref
4). Pedersen OM, Aardal NP, Larssen TB, Varhaug JE, Myking O, Vik-Mo H. The value of ultrasonography in predicting autoimmune thyroid disease. Thyroid. 2000; 10(3):251–9.
crossref
5). Moon HJ, Sung JM, Kim EK, Yoon JH, Youk JH, Kwak JY. Diagnostic performance of grayscale US and elastography in solid thyroid nodules. Radiology. 2012; 262(3):1002–13.
crossref
6). Park SH, Kim SJ, Kim EK, Kim MJ, Son EJ, Kwak JY. Interobserver agreement in assessing the sonographic and elasto- graphic features of malignant thyroid nodules. AJR Am J Roentgenol. 2009; 193(5):W416–23.
7). Rago T, Santini F, Scutari M, Pinchera A, Vitti P. Elasto-graphy: new developments in ultrasound for predicting malignancy in thyroid nodules. J Clin Endocrinol Metab. 2007; 92(8):2917–22.
crossref
8). Castellano G, Bonilha L, Li LM, Cendes F. Texture analysis of medical images. Clin Radiol. 2004; 59(12):1061–9.
crossref
9). Friedrich-Rust M, Ong MF, Herrmann E, Dries V, Samaras P, Zeuzem S, et al. Real-time elastography for noninvasive assessment of liver fibrosis in chronic viral hepatitis. AJR Am J Roentgenol. 2007; 188(3):758–64.
crossref
10). Gao S, Peng Y, Guo H, Liu W, Gao T, Xu Y, et al. Texture analysis and classification of ultrasound liver images. Biomed Mater Eng. 2014; 24(1):1209–16.
crossref
11). Gomez W, Pereira WC, Infantosi AF. Analysis of cooccurrence texture statistics as a function of gray-level quantization for classifying breast ultrasound. IEEE Trans Med Imaging. 2012; 31(10):1889–99.
crossref
12). Itoh A, Ueno E, Tohno E, Kamma H, Takahashi H, Shiina T, et al. Breast disease: clinical application of US elastography for diagnosis. Radiology. 2006; 239(2):341–50.
crossref
13). Xie P, Xiao Y, Liu F. Real-time ultrasound elastography in the diagnosis and differential diagnosis of subacute thyroiditis. J Clin Ultrasound. 2011; 39(8):435–40.
crossref
14). Yoon JH, Yoo J, Kim EK, Moon HJ, Lee HS, Seo JY, et al. Real-time elastography in the evaluation of diffuse thyroid disease: a study based on elastography histogram parameters. Ultrasound Med Biol. 2014; 40(9):2012–9.
crossref
15). Mazziotti G, Sorvillo F, Iorio S, Carbone A, Romeo A, Piscopo M, et al. Grey-scale analysis allows a quantitative evaluation of thyroid echogenicity in the patients with Hashimoto's thyroiditis. Clin Endocrinol (Oxf). 2003; 59(2):223–9.
crossref
16). Acharya UR, Vinitha Sree S, Mookiah MR, Yantri R, Molinari F, Zieleznik W, et al. Diagnosis of Hashimoto's thyroiditis in ultrasound using tissue characterization and pixel classification. Proc Inst Mech Eng H. 2013; 227(7):788–98.
crossref
17). Schiemann U, Avenhaus W, Konturek JW, Gellner R, Hengst K, Gross M. Relationship of clinical features and laboratory parameters to thyroid echogenicity measured by standardized grey scale ultrasonography in patients with Hashimoto's thyroiditis. Med Sci Monit. 2003; 9(4):MT13–7.
18). Loy M, Cianchetti ME, Cardia F, Melis A, Boi F, Mariotti S. Correlation of computerized grayscale sonographic findings with thyroid function and thyroid autoimmune activity in patients with Hashimoto's thyroiditis. J Clin Ultrasound. 2004; 32(3):136–40.
crossref
19). Kim SY, Kim EK, Moon HJ, Yoon JH, Kwak JY. Application of texture analysis in the differential diagnosis of benign and malignant thyroid nodules: comparison with grayscale ultrasound and elastography. AJR Am J Roentgenol. 2015; 205(3):W343–51.
crossref
20). Kim EY, Kim WG, Kim WB, Kim TY, Kim JM, Ryu JS, et al. Coexistence of chronic lymphocytic thyroiditis is associated with lower recurrence rates in patients with papillary thyroid carcinoma. Clin Endocrinol (Oxf). 2009; 71(4):581–6.
crossref
21). Wang J, Guo L, Shi X, Pan W, Bai Y, Ai H. Real-time elastography with a novel quantitative technology for assessment of liver fibrosis in chronic hepatitis B. Eur J Radiol. 2012; 81(1):e31–6.
crossref
22). Willms A, Bieler D, Wieler H, Willms D, Kaiser KP, Schwab R. Correlation between sonography and antibody activity in patients with Hashimoto thyroiditis. J Ultrasound Med. 2013; 32(11):1979–86.
crossref
23). Emblem KE, Nedregaard B, Nome T, Due-Tonnessen P, Hald JK, Scheie D, et al. Glioma grading by using histogram analysis of blood volume heterogeneity from MR-derived cerebral blood volume maps. Radiology. 2008; 247(3):808–17.
crossref
24). Ganeshan B, Abaleke S, Young RC, Chatwin CR, Miles KA. Texture analysis of non-small cell lung cancer on unenhanced computed tomography: initial evidence for a relationship with tumour glucose metabolism and stage. Cancer Imaging. 2010; 10:137–43.
crossref
25). Ganeshan B, Skogen K, Pressney I, Coutroubis D, Miles K. Tumour heterogeneity in oesophageal cancer assessed by CT texture analysis: preliminary evidence of an association with tumour metabolism, stage, and survival. Clin Radiol. 2012; 67(2):157–64.
crossref
26). Sivaramakrishna R, Powell KA, Lieber ML, Chilcote WA, Shekhar R. Texture analysis of lesions in breast ultrasound images. Comput Med Imaging Graph. 2002; 26(5):303–7.
crossref
27). Sandrin L, Fourquet B, Hasquenoph JM, Yon S, Fournier C, Mal F, et al. Transient elastography: a new noninvasive method for assessment of hepatic fibrosis. Ultrasound Med Biol. 2003; 29(12):1705–13.
crossref

Fig. 1.
Example of texture analysis of elastography and grayscale US. One longitudinal elastographic image was selected for texture analysis. Images are automatically displayed in split-screen mode to show both grayscale US and corresponding color-scale elastographic images. A region of interest (ROI) was previously set on the elastography by the radiologist who performed US (left, upper row, box). The same ROI was set on the grayscale US (right, upper row, box) transferring from elastographic ROI. From these ROIs, histogram and cooccurrence matrix parameters are automatically calculated with an in-house built software. Histogram analysis (bottom row) show the distribution of the number pixels (y-axis) according to the pixel intensity value (x-axis) within the ROIs.
ijt-10-14f1.tif
Table 1.
Comparison of US features, grayscale US and RTE parameters according to the presence of DTD
Negative for DTD (n=85) Positive for DTD (n=28) p
Gray scale US     ＜0.001
Negative on US 78 (91.8%) 11 (39.3%)  
Positive on US 7 (8.2%) 17 (60.7%)  
Texture analysis of gray scale US
Mean 79.6±13.7 (46.4, 112.9) 77.1±14.4 (52.3, 115.6) 0.408
SD 14.3±2.4 (9.7, 23.5) 15.0±1.9 (11.5, 19.5) 0.127
Skewness 0.2±0.3 (−0.8, 1.0) 0.3±0.3 (−0.5, 1.0) 0.157
Kurtosis 4.0±1.4 (2.3, 11.6) 4.1±1.2 (2.6, 7.2) 0.909
Contrast 0.1±0.01 (0.1, 0.2) 0.1±0.02 (0.1, 0.2) 0.209
Correlation 0.717±0.059 (0.589, 0.838) 0.739±0.039 (0.660, 0.799) 0.026
Uniformity 0.506±0.106 (0.302, 0.805) 0.466±0.069 (0.354, 0.618) 0.024
Homogeneity 0.94±0.01 (0.91, 0.97) 0.94±0.01 (0.92, 0.95) 0.174
Entropy 1.037±0.190 (0.507, 1.562) 1.116±0.121 (0.854, 1.291) 0.013
Texture analysis of RTE
Mean 172.5±10.5 (151.8, 198.6) 165.2±14.9 (130.1, 195.7) 0.021
SD 46.9±5.8 (33.1, 65.8) 47.6±6.6 (33.5, 60.4) 0.578
Skewness −1.2±0.6 (−2.7, −0.4) −1.2±0.5 (−2.3, −0.4) 0.775
Kurtosis 7.9±2.7 (3.7, 17.7) 7.3±2.8 (2.3, 14.7) 0.400
Contrast 0.7±0.2 (0.4, 1.2) 0.7±0.2 (0.5, 1.2) 0.237
Correlation 0.770±0.052 (0.636, 0.872) 0.788±0.060 (0.7, 0.9) 0.111
Uniformity 0.263±0.052 (0.185, 0.398) 0.259±0.056 (0.145, 0.383) 0.692
Homogeneity 0.900.02 (0.86, 0.93) 0.90±0.01 (0.88, 0.93) 0.879
Entropy 1.839±0.193 (1.395, 2.248) 1.891±0.204 (1.448, 2.347) 0.230

DTD: diffuse thyroid disease, RTE: realtime elastography, SD: standard deviation, US: ultrasound Minimum and maximum values are in parentheses.

Table 2.
Diagnostic performances of US features, grayscale US and RTE parameters in the diagnosis of DTD
Cutoff Sensitivity p Specificity p Accuracy p PPV p NPV p A z (95% CI) p∗
Gray scale US - 60.7 - 91.8 - 84.1 - 70.8 - 87.6 - 0.762 -
(17/28) (78/85) (95/113) (17/24) (78/89) (0.666, 0.860)
Texture analysis of gray scale US 0.005† ＜ ＜0.001† 0.002† ＜0.001† 0.316† 0.108†
Mean ≤86.9 85.7 0.705∗ ＜ 29.4 ＜0.001∗ 43.4 0.002∗ 28.6 0.241∗ 86.2 0.444∗ 0.558 0.264∗
(24/28) (25/85) (49/113) (24/84) (25/29) (0.433, 0.683)
SD ＞13.9 78.6 0.067∗ 49.4 0.053∗ 56.6 0.368∗ 33.9 0.881∗ 87.5 0.187∗ 0.632 0.403∗
(22/28) (42/85) (64/113) (22/65) (42/48) (0.517, 0.747)
Skewness ＞0.3 57.1 0.005∗ 64.7 - 62.8 - 34.8 - 82.1 0.079∗ 0.585 0.363∗
(16/28) (55/85) (71/113) (16/46) (55/67) (0.459, 0.711)
Kurtosis ＞3.8 53.6 0.002∗ 57.6 0.270∗ 56.6 0.260∗ 29.4 0.336∗ 79.0 0.032∗ 0.530 0.133∗
(15/28) (49/85) (63/113) (15/51) (49/62) (0.402, 0.658)
Contrast ＞1 71.4 0.014∗ 48.2 0.023∗ 54.0 0.178∗ 31.3 0.576∗ 83.7 0.048∗ 0.582 0.084∗
(20/28) (41/85) (61/113) (20/64) (41/49) (0.469, 0.694)
Correlation ＞0.7 85.7 0.654∗ 45.9 0.010∗ 55.8 0.274∗ 34.3 0.930∗ 90.7 0.783∗ 0.624 0.485∗
(24/28) (39/85) (63/113) (24/70) (39/43) (0.516, 0.731)
Uniformity ≤0.5 85.7 0.309∗ 41.2 0.001∗ 52.2 0.105∗ 32.4 0.681∗ 89.8 0.321∗ 0.604 0.080∗
(24/28) (35/85) (59/113) (24/74) (35/39) (0.493, 0.716)
Homogeneity ≤0.9 71.4 0.014∗ 49.4 0.037∗ 54.9 0.230∗ 31.8 0.635∗ 84.0 0.054∗ 0.586 0.105∗
(20/28) (42/85) (62/113) (20/63) (42/50) (0.473, 0.670)
Entropy ＞1.0 89.3＞ ＞0.999∗ 40.0＜ ＜0.001∗ 52.2 0.105∗ 32.9 0.734∗ 91.9 0.581∗ 0.640 0.180∗
(25/28) (34/85) (59/113) (25/76) (34/37) (0.529, 0.751)
Grayscale index ＞−0.3 89.3 - 41.2 0.001∗ 53.1 0.134∗ 33.3 0.794∗ 92.1 - 0.661 -
(25/28) (35/85) (60/113) (25/75) (35/38) (0.554, 0.769)
Texture analysis of RTE 0.002† 0.244† 0.057† 0.074† 0.584† 0.093†
Mean ≤167.7 60.7 0.002∗ ＜ 65.9 ＜0.001∗ 64.6 0.039∗ 37.0 0.108∗ 83.6 0.242∗ 0.645 -
(17/28) (56/85) (73/113) (17/46) (56/67) (0.525, 0.766)
SD ＞51.7 32.1＜ ＜0.001∗ 82.4 0.314∗ 70.0 0.158∗ 37.5 0.154∗ 78.7 0.034∗ 0.540 0.251∗
(9/28) (70/85) (79/113) (9/24) (70/89) (0.409, 0.671)
Skewness ≤−0.9 78.6 0.142∗ 34.1＜ ＜0.001∗ 45.1＜ ＜0.001∗ 28.2 0.020∗ 82.9 0.326∗ 0.490 0.077∗
(22/28) (29/85) (51/113) (22/78) (29/35) (0.368, 0.612)
Kurtosis ≤5.6 28.6＜ ＜0.001∗ 87.1＞ ＞0.999∗ 72.6 0.315∗ 42.1 0.283∗ 78.7 0.035∗ 0.546 0.242∗
(8/28) (74/85) (82/113) (8/19) (74/94) (0.415, 0.678)
Contrast ≤0.6 42.9＜ ＜0.001∗ 80.0 0.152∗ 70.8 0.218∗ 41.4 0.239∗ 81.0 0.147∗ 0.612 0.648∗
(12/28) (68/85) (80/113) (12/29) (68/84) (0.487, 0.737)
Correlation ＞0.8 42.9＜ ＜0.001∗ 87.1 - 76.1 - 52.2 - 82.2 0.116∗ 0.577 0.438∗
(12/28) (74/85) (86/113) (12/23) (74/90) (0.441, 0.713)
Uniformity ＞0.2 75.0 0.079∗ 37.7 ＜0.001∗ 46.9 ＜0.001∗ 28.4 0.020∗ 82.1 0.276∗ 0.510 0.141∗
(21/28) (32/85) (53/113) (21/74) (32/39) (0.385, 0.635)
Homogeneity ≤0.9 92.9 - 24.7 ＜0.001∗ 41.6 ＜0.001∗ 28.9 0.009∗ 91.3 - 0.523 0.130∗
(26/28) (21/85) (47/113) (26/90) (21/23) (0.406, 0.640)
Entropy ＞1.8 82.1 0.067∗ 32.9 ＜0.001∗ 45.1 ＜0.001∗ 28.8 0.006∗ 84.9 0.206∗ 0.548 0.211∗
(23/28) (28/85) (51/113) (23/80) (28/33) (0.423, 0.665)
Elastography index ＞−1.1 85.7 0.142∗ 28.2 ＜0.001∗ 42.5 ＜0.001∗ 28.2 0.006∗ 85.7 0.241∗ 0.515 0.138∗
(24/28) (24/85) (48/113) (24/85) (24/28) (0.394, 0.633)

A z: area under the receiver operating characteristics curve, CI: confidence interval, DTD: diffuse thyroid disease, RTE: realtime elastography, SD: standard deviation, US: ultrasound Raw data are in parentheses

∗ p values compared to the highest value within each modality.

† p values compared to US features, using the parameter showing highest value.

TOOLS
Similar articles