Risk stratification of thyroid nodules with Bethesda category III results on fine-needle aspiration cytology: The additional value of acoustic radiation force impulse elastography

Interpretable machine learning models for predicting skip metastasis in cN0 papillary thyroid cancer based on clinicopathological and elastography radiomics features

Background

Skip lymph node metastasis (SLNM) in papillary thyroid cancer (PTC) involves cancer cells bypassing central nodes to directly metastasize to lateral nodes, often undetected by standard preoperative ultrasonography. Although multiple models exist to identify SLNM, they are inadequate for clinically node-negative (cN0) patients, resulting in underestimated metastatic risks and compromised treatment effectiveness. Our study aims to develop and validate a machine learning (ML) model that combines elastography radiomics with clinicopathological data to predict pre-surgical SLNM risk in cN0 PTC patients with increased risk of lymph node metastasis (LNM), improving their treatment strategies.

Methods

Our study conducted a retrospective analysis of 485 newly diagnosed primary PTC patients, divided into training and external validation cohorts. Patients were categorized into SLNM and non-SLNM groups based on follow-up outcomes and postoperative pathology. We collected preoperative clinicopathological data and extracted, standardized radiomics features from elastography imaging to develop various ML models. These models were internally validated using radiomics and clinicopathological data, with the optimal model's feature importance analyzed through the Shapley Additive Explanations (SHAP) approach and subsequently externally validated.

Results

In our study of 485 patients, 67 (13.8%) exhibited SLNM. The extreme gradient boosting (XGBoost) model, integrating elastography radiomics with clinicopathological data, demonstrated superior performance in both internal and external validations. SHAP analysis identified five key determinants of SLNM: three radiomics features from elastography images, one clinical variable, and one pathological variable.

Conclusion

Our evaluation highlights the XGBoost model, which integrates elastography radiomics and clinicopathological data, as the most effective ML approach for the prediction of SLNM in cN0 PTC patients with increased risk of LNM. This innovative model significantly enhances the accuracy of risk assessments for SLNM, enabling personalized treatments that could reduce postoperative metastases in these patients.

Analysis and Comparison of the Malignant Thyroid Nodules Not Recommended for Biopsy in ACR TIRADS and AI TIRADS With a Large Sample of Surgical Series

OBJECTIVE

To determine if the artificial intelligence-based Thyroid Imaging, Reporting and Data System (AI TIRADS) would perform better than the American College of Radiology (ACR) TIRADS in monitoring malignant thyroid nodules not recommended for biopsy using follow-up thresholds.

METHODS

A total of 3499 thyroid nodules with surgical histopathology and ultrasound features were retrospectively reviewed and categorized using ACR TIRADS and AI TIRADS. The recommendations for biopsy and follow-up divided nodules into three groups 1) fine needle aspiration (FNA), 2) follow-up ultrasound, and 3) no further evaluation.

RESULTS

Of the total 1608 malignant nodules in this study, 974 malignant nodules would not be biopsied in ACR TIRADS compared with 967 in AI TIRADS. While 60.0% (584/974) of these non-biopsied malignancies could be followed-up by ultrasound in ACR TIRADS and 62.8% (607/967) in AI TIRADS. For the malignancies of no further evaluation, 97.4% (380/390) were sized <10 mm in ACR TIRADS and 93.3% (336/360) in AI TIRADS. Compared with ACR TIRADS, AI TIRADS had lower unnecessary FNA rate and missing cancer rate (41.0% vs 47.8% and 22.8% vs 27.5%, P < .05, respectively) while having higher specificity and AUC as well as lower sensitivity (65.0% vs 57.9%, 0.895 vs 0.881, and 96.1% vs 97.8%, all P < .05).

CONCLUSIONS

Using the follow-up thresholds, more than half of the malignancies not being biopsied were monitored by ultrasound in both ACR TIRADS and AI TIRADS, and AI TIRADS had lower missing cancer rate. More than 90% of malignancies recommended for no further evaluation were <10 mm in diameter.

Systematic Review and Meta-Analysis to Identify the Immunocytochemical Markers Effective in Delineating Benign from Malignant Thyroid Lesions in FNAC Samples

Conventional cytology-based diagnosis for thyroid cancer is limited with more than 30-45% of nodules categorized as indeterminate, necessitating surgery for confirming or refuting the diagnosis. This systematic review and meta-analysis were aimed at identifying immunocytochemical markers effective in delineating benign from malignant thyroid lesions in fine needle aspiration cytology (FNAC) samples, thereby improving the accuracy of cytology diagnosis. A systematic review of relevant articles (2000-2021) from online databases was carried out and the search protocol registered in PROSPERO database (CRD42021229121). The quality of studies was assessed using QUADAS-2. Review Manager 5.4.1 from Cochrane collaboration and MetaDisc Version 1.4 was used to conduct the meta-analysis. Bias in the studies were visually analyzed using funnel plots, and statistical significance was evaluated by Egger's test. Systematic review identified 64 original articles, while meta-analysis in eligible articles (n = 41) identified a panel of 5 markers, Galectin-3, HBME-1, CK-19, CD-56, and TPO. Assessment of the diagnostic performance revealed that Gal-3 (sensitivity: 0.81; CI: 0.79-0.83; specificity: 0.84; CI: 0.82-0.85) and HBME-1 (sensitivity: 0.83; Cl: 0.81-0.86; specificity: 0.85; CI: 0.83-0.86) showed high accuracy in delineating benign from malignant thyroid nodules. Efficacy analysis in indeterminate nodules showed Gal-3 and HBME-1 have high specificity of 0.86 (CI 0.84-0.89) and 0.82 (CI 0.78-0.86), respectively, and low sensitivity of 0.76 (CI 0.72-0.80) and 0.75 (CI 0.70-0.80), respectively. Diagnostic odds ratio (DOR) of Galectin-3 and HBME-1 were 39.18 (CI 23.38-65.65) and 24.44 (CI 11.16-53.54), respectively. Significant publication bias was observed for the markers Galectin-3 and CK-19 (p < 0.05). The panel of 5 markers identified from this meta-analysis are high-confidence candidates that need to be validated in thyroid cytology to establish their efficacy and enable clinical applicability.

Reducing Unnecessary Biopsy of American College of Radiology Thyroid Imaging Reporting and Data System Category 4 Nodules

OBJECTIVES

To assess the effectiveness of strain elastography (SE), Virtual Touch tissue imaging and quantification (VTIQ; Siemens Medical Solutions, Mountain View, CA) and their combination (SE + VTIQ) in distinguishing benign from malignant American College of Radiology Thyroid Imaging Reporting and Data System category 4 nodules (TR4) to reduce unnecessary biopsy.

METHODS

In this retrospective study, 985 thyroid nodules from 920 patients were initially enrolled and examined with conventional ultrasound, SE, and VTIQ. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of SE, VTIQ, and SE + VTIQ were calculated respectively. An area under the receiver operating characteristic curve analysis was applied to evaluate the diagnostic efficiency of SE, VTIQ, and SE + VTIQ in the differentiation of benign and malignant TR4 nodules.

RESULTS

A total of 864 patients with 922 TR4 nodules were enrolled ultimately, as 63 thyroid nodules from 56 patients were excluded because they did not meet the inclusion criteria of this study. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of SE, VTIQ, and SE + VTIQ were 72.4% versus 86.3% versus 82.1%, 77.7% versus 80.5% versus 83.5%, 70.6% versus 76.1% versus 80.2%, 73.4% versus 76.3% versus 83.5%, and 75.5% versus 79.7% versus 82.8%, respectively. The areas under the receiver operating characteristic curve for the diagnosis of TR4 nodules by SE, VTIQ, and SE + VTIQ were 0.751, 0.817, and 0.844.

CONCLUSIONS

In spite of a slight decrease in the sensitivity, the application of combining SE and VTIQ techniques can improve the specificity and accuracy of TR4 nodule diagnosis, which can help reduce unnecessary biopsies of TR4 nodules.

The values of shear wave elastography in avoiding repeat fine-needle aspiration for thyroid nodules with nondiagnostic and undetermined cytology

OBJECTIVE

To evaluate the value of shear wave elastography (SWE) in avoiding repeat fine-needle aspiration of thyroid nodules with nondiagnostic and undetermined cytology.

METHODS

A total of 232 thyroid nodules with nondiagnostic (n = 132) and undetermined (n = 100) cytology underwent ultrasound (US) and SWE, followed by repeat ultrasound (US)-guided fine-needle aspiration cytology (FNAC). The final diagnosis was based on cytological or pathological findings. The US and SWE characteristics of benign and malignant nodules were compared using the χ² -test. The receiver operating characteristic (ROC) curves of the thyroid imaging reporting and data system (TI-RADS) categories from the US and the E_Mean and E_SD from the SWE were graphed, and the areas under the curves (AUCs) were compared using a Z test.

RESULTS

There were significant differences between the benign and malignant nodules in terms of the echogenicity, shape, margin, calcification and TI-RADS categories (all P < 0.05). The differences were significant between the malignant and benign nodules for E_Mean [(34.57 ± 14.81) kPa vs. (19.18 ± 7.09) kPa] and E_SD [(13.68 ± 13.01) kPa vs. (3.97 ± 2.58) kPa] (both P < 0.001). Though the difference in the AUCs of E_Mean (0.864) and E_SD (0.876) was not significant (P = 0.745), they both had higher diagnostic performances in comparison with TI-RADS categories (0.762) (all P < 0.05). Moreover, E_SD attained a sensitivity of 100% with a relatively higher specificity of 49.75% when its cut-off value was 3.3 kPa.

CONCLUSIONS

Shear wave elastography is a promising imaging method for reducing repeat FNAC for benign thyroid nodules with nondiagnostic and undetermined cytology when using E_SD as an index.

Elastography for the evaluation of thyroid nodules in pediatric patients

Objective

To evaluate the usefulness of elastography (using manual compression) as an additional diagnostic tool for children and adolescents with thyroid nodules.

Materials and Methods

This was a prospective study conducted between September 2012 and August 2013 at a hospital in Brazil. We performed elastography, ultrasound, and fine-needle aspiration biopsy in 32 patients between 6 and 18 years of age who had, in total, 38 thyroid nodules.

Results

The elastography findings correlated with the histopathological diagnosis in 78.5% of cases. In three patients, an unnecessary thyroidectomy could have been avoided if the elastography results had been prioritized. Only one malignant thyroid nodule was found to show high elasticity.

Conclusion

Our results suggest that high elasticity of a nodule on elastography is associated with a low risk of thyroid cancer. If further confirmed in other studies, elastography may prove useful as a complementary test for screening thyroid nodules in children.

Risk stratification of the thyroid nodule with Bethesda indeterminate cytology, category III, IV, V on the one surgeon-performed US-guided fine-needle aspiration with 27-gauge needle, verified by histopathology of thyroidectomy: the additional value of one surgeon-performed elastography

BACKGROUND

The aim is to assess the value of strain elastography (SE) in differentiating likelihood of malignancy for the thyroid nodules, possessing the Bethesda Category III, IV, and V indeterminate cytology.

METHODS

The data was obtained by ultrasonography (US)-guided fine-needle aspiration (US-g-FNA) via 27-gauge needle, with the verification of indicated thyroidectomies in a retrospective analysis, from April 2010 to April 2014, by enrolling the documents of 262 consecutive patients, with 327 thyroid nodules, subjected to one-surgeon performed neck US, SE, and US-g-FNA with 27-G needle to rule out the malignancy.

RESULTS

122 of 327 cases were Bethesda Category III, IV, and V with histopathologically benign, 110 (90.2%); PTC, 7 (5.7%); FTC, 4 (3.3%); HCC, 1 (0.8%). Tsukuba Elasticity Score (TES) 1, 2, 3, 4, and 5 were detected as 38 (31.1%), 8 (6.6%), 59 (48.4%), 4 (3.3%), and 13 (10.7%), respectively for the cases with the indeterminate cytology. No significant difference was detected between TES 4 and 5 and malign histopathology by McNemar test (p = .727) with a good level of concordance, the kappa coefficient, 0.737.

CONCLUSION

SE may be a useful tool in differentiating malign from benign thyroid nodules by selecting surgery adaptation even for Bethesda indeterminate cytology on FNAC.

Effect of anatomo-topographic and sonographic Polarity of the thyroid nodules on the thyroid malignancy by evaluating its impression on the relationship between the Bethesda System, TBSRTC, Strain Elastography Score and the thyroid histopathology

Background: The goal is to evaluate the association between the topo-sonographic polarity of the thyroid nodules and the thyroid malignancy by analyzing its effect on The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC), score of Strain Elastography (SE) for thyroid as Tsukuba Elasticity Score (TES), and histopathologic assessment. Material and Methods: A preliminary single-center retrospective study was carried out by including the documents of 641 consecutive eligible patients, possessing 770 thyroid nodules which undergone neck ultrasonography (US), Doppler US, SE, and US-guided-fine needle aspiration (FNA) during April 2011 to April 2017. The stiffness had been measured by TES of SE. The ability of the prediction of the malignancy by the polarity of 770 thyroid nodules considering the association between; i) TBSRTC and histopathology and ii) TES and histopathology had been evaluated. Results: Of the 770 thyroid nodules evaluated, 408 (53.0%) were located at the superior pole (Pol 1) while 362 (47.0%) were at the inferior pole (Pol 0) with 0.9046 AUC and 0.8171 AUC for the association between TBSRTC and histopathology and 0.9280 AUC and 0.7888 AUC for the association between TES and histopathology, respectively. However, those difference were not significant for Pol 1, topographically. Conclusion: The topograghic and sonographic polarity of the thyroid nodules may not be useful for estimating the thyroid malignancy by using the association between TBSRTC and histopathology with TES and histopathology. However, the association with Pol 1, the superior thyroid pole, was stronger though the difference was not significant.

Association Between Preoperative US, Elastography Features and Prognostic Factors of Papillary Thyroid Cancer With BRAFV600E Mutation

Purpose: To investigate the value of US and elastography for predicting prognostic factors of papillary thyroid cancer (PTC) in the positive BRAF^V600E Mutation group. Materials and Methods: A total of 116 BRAF^V600E Mutation patients with PTCs were enrolled in this prospective study, who were preoperatively evaluated by US, US elasticity imaging (EI), and Virtual Touch tissue imaging (VTI) and Virtual Touch tissue quantification (VTQ) of acoustic radiation force impulse (ARFI) imaging. Multivariate logistic regression analysis was performed to assess 23 independent variables for predicting prognostic factors. Diagnostic performance was evaluated with receiver operating characteristic (ROC) curve analysis. Results: Forty-two (36.2%) of 116 PTC patients with BRAF^V600E Mutation had central lymph node metastasis (LNM). Nine (7.8%) and fifty-six (48.3%) had lateral LNM and extra-thyroidal extension (ETE), respectively. In multivariate logistic regression analyses, rich internal flow [odds ratio [OR]: 6.66] was the best predictor for central LNM, followed by male sex (OR: 4.22), halo sign absence (OR: 2.78) (all P < 0.05). VTQ ratio (OR: 1.57) was the only predictor for lateral LNM (P = 0.02). Rich internal flow (OR: 6.33) was the strongest predictor for ETE, followed by male sex (OR: 3.29), halo sign absence (OR: 2.90), and VTQ ratio (OR: 1.63) (all P < 0.05). Conclusion: VTQ ratio on ARFI imaging, rich internal flow and halo sign absence on US are the predicting prognostic factors in PTC patients with BRAF^V600E Mutation. The specificities were significantly increased by combining ARFI imaging and US features, which has a potential to avoid unnecessary therapeutic neck dissection in the high-risk PTC patients.

Web-based thyroid imaging reporting and data system: Malignancy risk of atypia of undetermined significance or follicular lesion of undetermined significance thyroid nodules calculated by a combination of ultrasonography features and biopsy results

BACKGROUND

The purpose of this study was to construct a web-based predictive model using ultrasound characteristics and subcategorized biopsy results for thyroid nodules of atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS) to stratify the risk of malignancy.

METHODS

Data included 672 thyroid nodules from 656 patients from a historical cohort. We analyzed ultrasound images of thyroid nodules and biopsy results according to nuclear atypia and architectural atypia. Multivariate logistic regression analysis was performed to predict whether nodules were diagnosed as malignant or benign.

RESULTS

The ultrasound features, including spiculated margin, marked hypoechogenicity, calcifications, biopsy results, and cytologic atypia, showed significant differences between groups. A 13-point risk scoring system was developed, and the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the development and validation sets were 0.837 and 0.830, respectively (http://www.gap.kr/thyroidnodule_b3.php).

CONCLUSION

We devised a web-based predictive model using the combined information of ultrasound characteristics and biopsy results for AUS/FLUS thyroid nodules to stratify the malignant risk.

A Proposal to Stratify the Intermediate-Risk Thyroid Nodules According to the AACE/ACE/AME Guidelines with Ultrasound Features

To propose a risk stratification system for intermediate-risk thyroid nodules (TNs) according to American Association of Clinical Endocrinologists, American College of Endocrinology and Associazione Medici Endocrinologi Medical (AACE/ACE/AME) Guideline with ultrasound (US) features. 1000 patients with 1000 nodules (902 benign nodules and 98 malignant nodules) were included. All the nodules were confirmed with either fine needle aspiration (FNA) cytology and follow-up or histology results after surgery. Univariate analysis and binary multivariate logic regression analysis were applied to analyze the possible risk US features associated with malignancy. Receiver operating characteristic curves (ROC) were drew and compared. Univariate analysis and binary multivariate logistic regression analysis showed that indeterminate hyper-echoic spot (OR = 4.544), slightly ill-defined margin (OR = 2.559), slight hyper-echogenicity (OR = 1.992) and no macro-calcification (OR = 1.921) were risk factors for the intermediate-risk thyroid nodules (TNs). A predicting model was established based on the 4 risk factors. The risk rates of malignancy were 5.7% (26/455) in Stage I, 11.0% (49/445) in Stage II, 23.1% (21/91) in Stage III, 33.3% (3/9) in Stage IV. In conclusion, for the intermediate-risk TNs, special attention should be paid to the TNs with indeterminate hyper-echoic spot, slightly ill-defined margin, slight hyper-echogenicity, or no macro-calcification. The probability of malignancy increased with the number of risk factors increasing.

Detection of Malignancy Among Suspicious Thyroid Nodules <1 cm on Ultrasound with Various Thyroid Image Reporting and Data Systems

BACKGROUND

In patients undergoing active surveillance of papillary thyroid microcarcinoma, definitive therapy-usually preceded by a definitive diagnostic procedure-is not recommended until evidence of disease progression is obtained, as stated in the American Thyroid Association guidelines. This is because the deferring of definitive diagnosis and therapy until disease progression has no impact on the disease-specific survival. This study evaluated the malignancy rate and probability of thyroid nodules, which was further stratified based on the size cutoff value of 1 cm, with suspicious findings on ultrasonography (US), by using various malignant stratification systems.

METHODS

The data were retrospectively collected between January 2003 and June 2003 from nine university hospitals that had previously participated in the Korean Society of Thyroid Radiology multicenter study on the ultrasonographic differentiation between benign and malignant thyroid nodules. In total, 829 thyroid nodules from 711 patients (620 women, 91 men; M_age = 48.7 years; range 6-98 years; 351 malignant and 478 benign nodules) were included. The probability for malignancy of thyroid nodules was calculated, which was further stratified by size, by using four different types of malignant risk-stratification systems. The factors that could differentiate benign from malignant nodules were assessed using the chi-square test.

RESULTS

In the suspicious thyroid nodules <1 cm on US, the malignancy probability ranged from 77.4% to 82.8%; the lowest rate was found in the Korean Society of Thyroid Radiology multicenter study, whereas the highest rate was noted in the Web-based system. Thus, the probability of benign nodules among suspicious thyroid nodules <1 cm on US was 17.2-22.6%.

CONCLUSION

A biopsy should be considered before active surveillance to exclude benign nodules with suspicious US features, and could thus prevent unnecessary active surveillance and patient anxiety.

Malignancy risk stratification of thyroid nodules: comparisons of four ultrasound Thyroid Imaging Reporting and Data Systems in surgically resected nodules

To compare the efficiency of four different ultrasound (US) Thyroid Imaging Reporting and Data Systems (TI-RADS) in malignancy risk stratification in surgically resected thyroid nodules (TNs). The study included 547 benign TNs and 464 malignant TNs. US images of the TNs were retrospectively reviewed and categorized according to the TI-RADSs published by Horvath E et al. (TI-RADS H), Park et al. (TI-RADS P), Kwak et al. (TI-RADS K) and Russ et al. (TI-RADS R). The diagnostic performances for the four TI-RADSs were then compared. At multivariate analysis, among the suspicious US features, marked hypoechogenicity was the most significant independent predictor for malignancy (OR: 15.344, 95% CI: 5.313-44.313) (P < 0.05). Higher sensitivity was seen in TI-RADS H, TI-RADS K, TI-RADS R comparing with TI-RADS P (P < 0.05 for all), whereas the specificity, accuracy and area under the ROC curve (Az) of TI-RADS P were the highest (all P < 0.05). Higher specificity, accuracy and Az were seen in TI-RADS K compared with TI-RADS R (P = 0.003). With its higher sensitivity, TI-RADS K, a simple predictive model, is practical and convenient for the management of TNs in clinical practice. The study indicates that there is a good concordance between TI-RADS categories and histopathology.

Cytologically indeterminate thyroid nodules: increased diagnostic performance with combination of US TI-RADS and a new scoring system

To investigate the diagnostic performance of combination of ultrasound (US) thyroid imaging reporting and data system (TI-RADS) and a new US scoring system for diagnosing thyroid nodules (TNs) with indeterminate results (Bethesda categories III, IV and V) on fine-needle aspiration (FNA) cytology. 453 patients with 453 cytologically indeterminate TNs were included in this study. Multivariate analyses were performed to construct the scoring system. The diagnostic performances of TI-RADS and the combined method were evaluated and compared. Multivariate analyses revealed that marked hypoechogenicity, taller than wide shape and absence of halo sign were independent predictors for malignancy in cytologically indeterminate TNs. Scoring system was thereafter defined as follows: risk score (RS) = 3.2 x (if marked hypoechogenicity) + 2.8 x (if taller than wide shape) + 1.3 x (if absence of halo sign). Compared with TI-RADS alone, the areas under the receiver operating characteristic curves (AUC), specificity, accuracy and positive predictive value (PPV) of the combined method increased significantly with 0.731 versus 0.569, 48.5% versus 14.1%, 76.2% versus 62.3%, and 70.9% versus 59.9%, respectively (all P < 0.05). The combination of TI-RADS and new US scoring system showed superior diagnostic performances in predicting malignant TNs with indeterminate FNA cytology results in comparison with TI-RADS alone.