Texture Analysis of Thyroid Nodules Using Computed Tomography: Is it a Viable Method for Objective Assessment of Thyroid Nodules?

Abstract

Objective: Computed aided detection (CAD) systems can be developed to help radiologists in the accurate interpretation of computed tomography (CT) images. The recently popularised texture analysis method allows for qualitative and quantitative evaluation by analysing the grey-level distribution and relationships within an image. We aimed to compare the ratios of texture analysis data in the differentiation of benignmalignant nodules with the proportions of radiologists in the distinction between benign and malignant nodules and to compare the results. Materials and Methods: Retrospectively, the data of 80 patients who underwent thyroidectomy and had contrast-enhanced neck CT preoperatively were analysed. Two radiologists, experienced in head and neck radiology, blinded to the patients’ data evaluated neck CT images. Manual marking was performed and scanned to take tissue sections from the nodule area in transverse contrast-enhanced CT images, and the size of the nodule in the contralateral normal thyroid parenchyma was almost equal. Results: The computed tomography texture analysis (CTTA) model achieved the highest sensitivity of 81.4%, followed by the first radiologist at 51.2% and the second radiologist at 55.8%. Additionally, the CTTA model achieved the highest accuracy at 61.3%, followed by the first radiologist at 41.3% and second radiologist at 47.5%. On average, the CTTA model performed significantly better than the two radiologists, especially with regard to sensitivity. Conclusion: The CTTA model was superior to both radiologists in differentiating between benign and malignant thyroid nodules. Medical experts can benefit from CTTA-based solutions to extend their understanding of thyroid nodules in their routine practise.

Keywords

• Texture analysis
• thyroid cancer
• computer-aided diagnosis
• thyroid nodules

References

1. 1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T, et al. Cancer statistics, 2008. CA 2008;58(2):71-96. google scholar
2. 2. Hundahl SA, Cady B, Cunningham MP, Mazzaferri E, McKee RF, Rosai J, et al. Initial results from a prospective cohort study of 5583 cases of thyroid carcinoma treated in the United States during 1996: an American college of surgeons commission on cancer patient care evaluation study. Cancer 2000;89(1):202-17. google scholar
3. 3. Gopinath B, Shanthi N. Computer-aided diagnosis system for classifying benign and malignant thyroid nodules in multi-stained FNAB cytological images. ACPSEM 2013;36:219-30. google scholar
4. 4. Kwak JY, Han KH, Yoon JH, Moon HJ, Son EJ, Park SH, et al. Thyroid imaging reporting and data system for US features of nodules: a step in establishing better stratification of cancer risk. Radiology 2011;260(3):892-9. google scholar
5. 5. Yoon JH, Kim E-K, Hong SW, Kwak JY, Kim MJ. Sonographic features of the follicular variant of papillary thyroid carcinoma. J Ultrasound Med 2008;27(10):1431-7. google scholar
6. 6. Orhan A, Yumuşakhuylu A, GÜNDOĞMUŞ CA, Çağatay O. Ultrasonographic and Cytological Diagnostic Difficulties of Follicular-Variant Papillary Thyroid Carcinoma. Tr-ENT. 2021;31(1):1-5. google scholar
7. 7. Moschetta M, Ianora AAS, Testini M, Vacca M, Scardapane A, Angelelli G. Multidetector computed tomography in the preoperative evaluation of retrosternal goiters: a useful procedure for patients for whom magnetic resonance imaging is contraindicated. Thyroid 2010;20(2):181-7. google scholar
8. 8. Ishigaki S, Shimamoto K, Satake H, Sawaki A, Itoh S, Ikeda M, et al. Multi-slice CT of thyroid nodules: comparison with ultrasonography. Radiation Med 2004;22(5):346-53. google scholar

9. 9. Bin Saeedan M, Aljohani IM, Khushaim AO, Bukhari SQ, Elnaas ST. Thyroid computed tomography imaging: pictorial review of variable pathologies. Insights imaging 2016;7(4):601-17. google scholar
10. 10. Lubner MG, Smith AD, Sandrasegaran K, Sahani DV, Pickhardt PJ. CT texture analysis: definitions, applications, biologic correlates, and challenges. Radiographics 2017;37(5):1483-503. google scholar
11. 11. Hegedüs L. The thyroid nodule. NEJM 2004;351(17):1764-71. google scholar
12. 12. Li M, Zheng X, Li J, Yang Y, Lu C, Xu H, et al. Dual-energy computed tomography imaging of thyroid nodule specimens: comparison with pathologic findings. Invest Radiol 2012;47(1):58-64. google scholar
13. 13. Yoon DY, Chang SK, Choi CS, Yun EJ, Seo YL, Nam ES, et al. The prevalence and significance of incidental thyroid nodules identified on computed tomography. JCAT 2008;32(5):810-5. google scholar
14. 14. Lee C, Chalmers B, Treister D, Adhya S, Godwin B, Ji L, et al. Thyroid lesions visualized on CT: sonographic and pathologic correlation. Acad Radiol 2015;22(2):203-9. google scholar
15. 15. Sollini M, Cozzi L, Chiti A, Kirienko M. Texture analysis and machine learning to characterize suspected thyroid nodules and differentiated thyroid cancer: Where do we stand? Eur J Radiol 2018;99:1-8. google scholar
16. 16. Jo JI, Im Kim J, Ryu JK, Lee HN. Incidental Thyroid Nodule on Chest Computed Tomography: Application of Computed Tomography Texture Analysis in Prediction of Ultrasound Classification. JCAT 2022;46(3):480-6. google scholar
17. 17. Liu C, Chen S, Yang Y, Shao D, Peng W, Wang Y, et al. The value of the computer-aided diagnosis system for thyroid lesions based on computed tomography images. QIMS 2019;9(4):642. google scholar
18. 18. Peng W, Liu C, Xia S, Shao D, Chen Y, Liu R, et al. Thyroid nodule recognition in computed tomography using first order statistics. Biomed Eng Online 2017;16:1-14. google scholar
19. 19. Nam SJ, Yoo J, Lee HS, Kim E-K, Moon HJ, Yoon JH, et al. Quantitative evaluation for differentiating malignant and benign thyroid nodules using histogram analysis of grayscale sonograms. J Med Ultrasound 2016;35(4):775-82. google scholar
20. 20. Chang Y, Paul AK, Kim N, Baek JH, Choi YJ, Ha EJ, et al. Computer-aided diagnosis for classifying benign versus malignant thyroid nodules based on ultrasound images: a comparison with radiologist-based assessments. Med Phys 2016;43(1):554-67. google scholar