Diagnosis And Management Of Different Kinds Of Thyroid Nodules: A Comparative Study With 5 International Guidelines

Part One Associations between suspicious ultrasonography features and malignancy risk of thyroid nodules:categorization based on composition and echogenicityObjective:This study aimed to explore the associations between suspicious ultrasonography(US)features and malignancy risk of thyroid nodules among different categorizations based on composition and echogenicity.Methods:From June 2016 to December 2019,a total of 1762 thyroid nodules from1636 consecutive patients who had undergone fine-needle aspiration biopsy(FNAB)in Wuhan Tongji hospital were enrolled in this study.Univariable and multivariable logistic regression analyses were used to determine independent US features for malignancy risk.The?²or Fisher's exact tests were used to compare the frequency and malignancy risk for each US feature,and for any of 0?4 features in all and subgroups according to composition and echogenicity.Logistic regression models for interaction were designed to determine the differences in associations of US features and malignancy risk.Results:The suspicious US features of microcalcification,taller-than-wide,irregular margin,isthmus,upper or middle nodule location and under the capsule were all significant in solid hypoechoic nodules(all P<0.01),but only microcalcifications,taller-than-wide,irregular margin and under the capsule were significant in partially cystic or isohyperechoic nodules(all P_multi<0.05).Compared with the partially cystic or isohyperechoic nodules,the frequency of microcalcification,taller-than-wide,irregular margin,and under the capsule in solid hypoechoic nodules and their malignancy risks were significantly higher(all P_multi<0.01).Compared with the solid hypoechoic nodules,the partially cystic or isohyperechoic nodules had the higher frequencies with any 0?1 of suspicious US feature,but lower frequencies with any 2 or more of suspicious US features(all P<0.001).Solid hypoechoic nodules with any 0?3 of suspicious US features had a significantly higher malignancy risk than the partially cystic or isohyperechoic nodules,respectively,and when those had all 4 suspicious US features,malignancy risks were both 100%.The associations of irregular margin and microcalcification with malignancy risk were significantly influenced by categorization of thyroid nodules based on composition and echogenicity(all Pfor interaction<0.05).Conclusions:The frequencies of suspicious US features in thyroid nodules and their malignancy risks are significantly different in subgroups according to composition and echogenicity.The differences in associations of suspicious US features with malignancy risk of thyroid nodules deserve special clinical concerns.Part Two Comparison of diagnosis and management of guidelines for thyroid nodules: categorization based on composition and echogenicityObjective: This study compared the diagnostic performance of guidelines from international societies in ultrasonography(US)-based malignancy risk stratifications and fine-needle aspiration biopsy(FNAB)criteria for thyroid nodules,including the Kwak-Thyroid Imaging Reporting and Data System(Kwak-TIRADS),the American Thyroid Association(ATA)guidelines,the American Association of Clinical Endocrinologists/American College of Endocrinology/Associazione Medici Endocrinologi(AACE/ACE/AME)guidelines,the American College of Radiology-Thyroid Imaging Reporting and Data System(ACR-TIRADS),and the Chinese-Thyroid Imaging Reporting and Data System(C-TIRADS).Methods: A total of 1762 thyroid nodules from 1636 consecutive patients,who had undergone FNAB in Wuhan Tongji hospital from June 2016 to December 2019,with final diagnoses were enrolled in this retrospective study.Nodules were categorized according to the US-based risk stratifications defined by guidelines.The diagnostic performances for US detection of malignancy were calculated using the receiver operating characteristic curve(ROC).The performance in FNAB recommendation,as well as unnecessary FNAB rates were compared among the nodules ? 1.0 cm using the FNAB criteria of guidelines.Mc Nemar tests were used to compare the diagnosis and management of guidelines for all nodules and nodules categorized by composition and echogenicity.Results: Based on data from relatively high-risk thyroid nodules in Chinese patients,the Kwak-TIRADS and C-TIRADS presented the largest AUC(0.838,all P < 0.001).The Kwak-TIRADS exhibited higher diagnostic sensitivity(86.3%,P < 0.001),and the C-TIRADS exhibited higher diagnostic specificity(88.7%,P < 0.001).When nodules were categorized,C-TIRADS presented the largest AUC(0.820,all P < 0.001)in solid hypoechoic nodules,Kwak-TIRADS and ACR-TIRADS presented the largest AUC in partially cystic or isohyperechoic nodules(0.863 and 0.859,respectively,all P < 0.001),and C-TIRADS presented the largest AUC(0.813,all P < 0.001)in nodules smaller than 1cm.The ACR-TIRADS and AACE/ACE/AME guidelines indicated the highest accuracy of diagnostic performance and lowest unnecessary FNAB rate in the FNAB recommendation.Conclusions: In this population,compared with another three guidelines,the C-TIRADS and Kwak-TIRADS show a better diagnostic performance in US-based risk stratifications in general.The ACR-TIRADS and AACE/ACE/AME guidelines show a better performance of the FNAB criteria.The C-TIRADS also performs well in US diagnosis,FNAB management and convenient usage.Guidelines show the different diagnostic efficacy in different nodules categorized based on composition and echogenicity.Clinicians require a clear understanding of the strengths and weaknesses associated with the different guidelines related to different nodules that are available,as well as their potential impact on the diagnosis of this less aggressive cancer.Part Three Combined diagnosis of guidelines and BRAFV600 E mutation in fine-needle aspiration biopsy indeterminate thyroid nodulesObjective: This study aimed to compare the diagnostic efficacy of combined models of guidelines and BRAFV600 E mutation in fine-needle aspiration biopsy(FNAB)indeterminate(Bethesda III and IV)thyroid nodules.Methods: A total of 1216 consecutive thyroid nodules with BRAFV600 E tests were enrolled in this retrospective study from June 2016 to December 2019 in Wuhan Tongji hospital.The combined diagnosis models of guidelines and BRAFV600 E mutation were constructed by Logistic regression analyses.The diagnostic cutoffs and efficacy for detection of malignancy were calculated using the receiver operating characteristic curve(ROC).Mc Nemar tests were used to compare the diagnosis efficacy.Results: In FNAB indeterminate(Bethesda III and IV)thyroid nodules with BRAFV600 E mutation,the optimal cutoffs for suspicious malignancy were “4a” in Kwak-TIRADS,“low suspicion” in ATA guidelines,“low risk” in AACE/ACE/AME guidelines,“mildly suspicious” in ACR-TIRADS,and “4A” in C-TIRADS.In FNAB indeterminate(Bethesda III and IV)nodules without BRAFV600 E mutation,the optimal cutoffs of the five guidelines were “5”,“high suspicion”,“high risk”,“highly suspicious”,and “4C”,respectively.The sensitivity,negative predictive value(NPV),and area under receiver operating characteristic curve(AUC)of combined diagnosis were higher than those of diagnosis by guidelines or BRAFV600 E test alone(all P < 0.05).Compared with another three guidelines,the Kwak-TIRADS and C-TIRADS combined with BRAFV600 E test presented the largest AUC(0.765 and 0.766,respectively,all P < 0.001).The Kwak-TIRADS exhibited higher diagnostic sensitivity than the C-TIRADS(79.4% vs.67.4%,P < 0.001)when combined with BRAFV600 E test,and the C-TIRADS exhibited higher diagnostic specificity and positive predictive value(PPV)than Kwak-TIRADS when combined with BRAFV600 E test(78.3% vs.55.2%,68.3% vs.55.2%,all P < 0.001).The difference of NPVs was not significant(77.6% vs.79.4%,P > 0.05).Conclusions: The combination of guidelines and BRAFV600 E mutation test distinctly improved diagnosis efficacy of FNAB indeterminate(Bethesda III and IV)thyroid nodules.The Kwak-TIRADS and C-TIRADS combined with BRAFV600 E test show a better diagnostic performance.