Ultrasonic examination of thyroid nodule: should we follow the guidelines or adopt a computer-aided detection and diagnostic system for patients in Hong Kong?


Grant Data
Project Title
Ultrasonic examination of thyroid nodule: should we follow the guidelines or adopt a computer-aided detection and diagnostic system for patients in Hong Kong?
Principal Investigator
Dr Wong, Kai Pun   (Principal Investigator (PI))
Duration
12
Start Date
2018-06-30
Completion Date
2019-06-29
Amount
56460
Conference Title
Ultrasonic examination of thyroid nodule: should we follow the guidelines or adopt a computer-aided detection and diagnostic system for patients in Hong Kong?
Presentation Title
Keywords
computer-aided, guideline, indetermine, surgeon-performed, thyroid nodule, ultrasound
Discipline
Imaging,Cancer
HKU Project Code
201711159259
Grant Type
Seed Fund for PI Research – Basic Research
Funding Year
2017
Status
Completed
Objectives
Introduction: Thyroid nodules are very common in general population. The are found in 4-8% general population during palpation. (1, 2) The incidence of thyroid nodules are even higher when ultrasonography of thyroid have been performed. 19-68% of randomly selected people have detected thyroid nodules by high-resolution ultrasound. (3-5). The clinical impact of these thyroid nodule implicates the need of excluding thyroid cancer, which is present in 7-15%, depending on sex, age, and exposure to other risk factors. (6)(7) Incidence of thyroid cancer rises in the last few decades all over the world. The yearly incidence has nearly tripled from 4.9 per 100,000 people in 1975 to 14.3 per people in 2009 in United states. (8) According to the Hong Kong Cancer Registry, there were 782 patients newly diagnosed with thyroid cancer in 2013. The incidence of thyroid cancer was increasing in the past 10 years. The crude rate have been nearly double over 10 years, which increased from 6.6 per 100,000 patients in 2004 to 10.9 per 100,000 patient in 2013.(9) This increase have been related to the increased use of ultrasonography or other imaging in detecting and early treatment in small thyroid cancer. To evaluate thyroid nodule, high-resolution ultrasonography have been widely applied. Ultrasound examination of thyroid are nowadays being performed not only by radiologist but also thyroid clinicians including medical endocrinologist and endocrine surgeons (10)(11). According to recent published guideline by American Thyroid Association 2015, it is recommend that ultrasonography should be done in all patient with suspected thyroid nodule, nodular goitre, or any radiographic abnormality in thyroid gland.(12)Certain sonographic feature, like hypo-echogenicity, micro-calcification, irregular margins, absence of halo, intra-nodular vascularity and tall than wide, were traditionally associated with presence of thyroid cancer.(2) However, reliability of any one feature in diagnosing thyroid malignancy was suboptimal. For example, sensitivity of presence of micro calcification ranged from 26.1% to 59.1%. For hypo-echogenicity, it ranged from 26.5% to 87.1%. Diagnostic sensitivity and accuracy in differentiating a malignant thyroid nodule were diverse and non-consistent according to different studies. Different guidelines from international study groups, like American Thyroid Association (ATA), British Thyroid Association (BTA) and Thyroid Imaging Reporting and Data System (TIRADs) have been launched for clinicians treating thyroid nodules. Thyroid nodules were stratified into different grades according to high risk sonographic features and proposed the risk of malignancy. For guideline published by ATA 2015, thyroid nodules were stratified into ""high suspicion"" with estimate risk of >70-90%, , ""intermediate suspicion"" with estimate risk of 10-20%, ""low suspicion"" with estimate risk of 5-10%, ""very low suspicion"" with estimate risk of <3% and ""benign"" with < 1% risk. For each thyroid nodule, risk of malignancy was estimated.(12) The nodule will be recommended for observation, fine needle aspiration cytology or surgery according to estimated risk. Similarly, guideline proposed by BTA and TIRADs categorised thyroid nodules into different risk groups.(13, 14) For daily clinical practise, application of these guidelines was too complicated and estimated risk of malignancy was variable with reference to different guidelines. Furthermore, these suggestions were based on foreign studies, applicability for local population was still questionable. Chang et al. have introduced the usage of computer-aided detection and diagnosis system (CADDS) to detect suspicious lesion for fine need aspiration cytology.(15) After marking the nodule by ultrasonography, four parameters index, including micro-calcification, hypoechoic lesion, heterogeneity and indistinct margin are calculated. The CADDS will predict the risk of malignancy and provide recommendation on necessarity of FNAC. By using the CADDS, potential abuse of FNAC, learning curve of identification of ultrasound image and inter-observer variability can be avoided. In this study, we prospectively collect ultrasound features and finding of thyroid nodule in Chinese patients undergoing thyroidectomy. We aim to assess the accuracy of each ultrasound feature in predicting thyroid malignancy. We will evaluate which guideline have a higher sensitivity and accuracy in predicting thyroid malignancy. Last but not least, we aim to compare if any guidelines or CADDS is a better tool in predicting malignant thyroid nodule for Chinese patients. 
 Objective: To evaluate accuracy of each ultrasound feature in predicting thyroid malignancy To evaluate which guidelines have a higher ability in stratifying benign, indeterminate and malignant thyroid nodule To compare if CADDS is better than guidelines in stratifying benign from indeterminate and malignant thyroid nodule