OBJECTIVE: To evaluate the efficacy of combining predictive artificial intelligence (AI) and image similarity model to risk stratify thyroid nodules, using retrospective external validation study.
METHODS: Two datasets were used to determine efficacy of the AI application. One was Stanford dataset ultrasound images of 192 nodules between April 2017 to May 2018 and the second was private practice consisting of 118 thyroid nodule images between January 2018 to December 2023. The nodules had definitive diagnosis by cytology or surgical pathology. The AI application was used to predict the diagnosis and American College of Radiology Thyroid Imaging and Data System (ACR TI-RADS) score.
RESULTS: In the Stanford dataset, the AI application predicted malignancies with sensitivity of 1.0 and specificity of 0.55. Positive predictive value (PPV) was 0.18 and negative predictive value (NPV) was 1.0. The Area Under the Curve - Receiver Operating Characteristic (AUC-ROC) was 0.78. ACR TI-RADS based clinical recommendation had a polychoric correlation of 0.67. In the private dataset, the AI application predicted malignancies with sensitivity of 0.91 and specificity of 0.95. PPV was 0.8 and NPV was 0.98. AUC-ROC was 0.93 and accuracy was 0.94. ACR TI-RADS based score had a polychoric correlation of 0.94.
CONCLUSIONS: The AI application showed good performance for sensitivity and NPV between the two datasets and demonstrated potential for 61.5% reduction in the need for fine needle aspiration (FNA) and strong correlation to ACR TI-RADS. However, PPV was variable between the datasets possibly from variability in image selection and prevalence of malignancy. If implemented widely and consistently among various clinical settings, this could lead to decreased patient burden associated with an invasive procedure and possibly to decreased health care spending.

OBJECTIVE: To validate the performance of a recently created risk stratification system (RSS) for thyroid nodules on ultrasound, the Artificial Intelligence Thyroid Imaging Reporting and Data System (AI TI-RADS).
METHODS: 378 thyroid nodules from 320 patients were included in this retrospective evaluation. All nodules had ultrasound images and had undergone fine needle aspiration (FNA). 147 nodules were Bethesda V or VI (suspicious or diagnostic for malignancy), and 231 were Bethesda II (benign). Three radiologists assigned features according to the AI TI-RADS lexicon (same categories and features as the American College of Radiology TI-RADS) to each nodule based on ultrasound images. FNA recommendations using AI TI-RADS and ACR TI-RADS were then compared and sensitivity and specificity for each RSS were calculated.
RESULTS: Across three readers, mean sensitivity of AI TI-RADS was lower than ACR TI-RADS (0.69 vs 0.72, p < 0.02), while mean specificity was higher (0.40 vs 0.37, p < 0.02). Overall total number of points assigned by all three readers decreased slightly when using AI TI-RADS (5,998 for AI TI-RADS vs 6,015 for ACR TI-RADS), including more values of 0 to several features.
CONCLUSIONS: AI TI-RADS performed similarly to ACR TI-RADS while eliminating point assignments for many features, allowing for simplification of future TI-RADS versions.

OBJECTIVE: In clinical practice, thyroid nodules are classified according to TI-RADS by B-mode and color-flow Doppler study. The aim of the study is to evaluate the possible added value of Superb microvascular imaging (SMI) and elastosonography in the stratification of malignancy risk of thyroid nodules.
METHODS: All patients with thyroid nodules who were candidates for needle aspiration were enrolled. Experienced operators performed a standard examination with TI-RADS calculation, followed by SMI and elastosonography on the nodules. The needle aspiration outcome was used as the gold standard. Statistical analysis calculated the ROC curves of the techniques applied individually and serially.
RESULTS: In this prospective study, we analysed 260 nodules, found in 251 patients (mean age 58.6 yo ± 14). 11.2% were TI-RADS 1, 18.9% TI-RADS 2, 41.1% TI-RADS 3, 28.1% TI-RADS 4, and 0.8% TI-RADS 5. The SMI technique showed an AUC of 0.57 (95% CI 0.49; 0.66) while elastosonography had an AUC of 0.58 (95% CI 0.49; 0.67) when used individually. SMI together with elastosonography had AUC of 0.62 (95% CI 0.52; 0.71). TI-RADS had AUC of 0.67 (95% CI 0.59; 0.75). SMI and elastosonography applied together with TI-RADS had AUC of 0.69 (95% CI 0.61; 0.77).
CONCLUSIONS: In the real-world cohort of patients, the SMI technique and elastosonography slightly increase the AUC of TI-RADS. Taken individually, SMI and elastosonography do not have a very strong AUC.

BACKGROUND: Thyroid nodule fine needle aspiration (FNA) biopsies are associated with a low false-negative rate. There is limited data regarding the predictive value of American College of Radiology Thyroid Imaging Reporting and Data System for false-negative FNA.
METHODS: This single-center retrospective study evaluated 119 patients who underwent thyroidectomy. The association of TR category, along with other clinical variables, with false-negative FNA was evaluated.
RESULTS: The overall false-negative rate of FNA was 10.8% (n = 9). False-negative FNAs were associated with younger age (mean 42 years vs 50.6 years, P = .04), larger nodule size (mean 4.4 cm vs 3.2 cm, P = .03), and a lower TR category (median 3 v 4, P = .01).
CONCLUSIONS: Lower TR category, younger age, and larger nodule size were associated with false-negative FNA of thyroid nodules. These findings should be taken into context when counseling patients with thyroid nodules who have a benign FNA.

OBJECTIVE: The process of evaluating pediatric thyroid nodules at our institution was inconsistent with a high rate of negative biopsies raising concern of appropriate patient selection for biopsy. Our aim was to institute a standardized risk stratification reporting system for thyroid nodules to increase utilization and agreement of TI-RADS reporting at our institution.
METHODS: Radiology report data were collected and analyzed as part of a quality improvement project. A standardized TI-RADS dictation template was created, ultrasound technicians were trained, a multi-disciplinary conference initiated, and education provided for radiologists and clinicians. Control charts were used to track utilization and agreement of scoring of TI-RADS reporting based upon review by a radiologist trained in TI-RADS scoring.
RESULTS: From January 2019 to January 2021, 218 patients with a thyroid nodule had a thyroid ultrasound performed at our institution. TI-RADS was utilized in 0 % (0 of 57) of children in the four months prior to project initiation. Following creation of the template, utilization increased to 65 % (39 of 60) over 5 months. Utilization further increased after the first training conference and was maintained above 90 % for 13 months. Ultrasound reports were in agreement in 46.7 % (28 of 60) of children initially. Agreement in reporting improved to 71.4 % (10 of 14) in the 3 months following the first training and to 78.4 % (58 of 74) over 12 months. Agreement in reporting was maintained at 80 % in the following 6 months.
CONCLUSIONS: A quality improvement initiative can improve utilization and agreement of scoring using the TI-RADS system in pediatrics. This may ultimately reduce unnecessary biopsies and sedation in children.
METHODS: Level III.
METHODS: Quality Improvement.

BACKGROUND: The thyroid cancer incidence has been experimenting an accelerated growth all over the world. The serine/threonine-protein kinase (BRAF) V600E gene detection or the DNA ploidy analysis has been employed in the identification of thyroid cancer type. This study aimed to evaluate the diagnostic value of the BRAF V600E gene integrated with DNA ploidy analysis in thyroid cancer.
METHODS: From August 2022 to May 2023, 400 individuals from the thyroid surgery outpatient department of our hospital were enrolled in this study. The participants were divided into low-risk groups (Ⅰ+Ⅱ+Ⅲ group; n = 200) and high-risk groups (Ⅳ+Ⅴ group; n = 200) based on the Thyroid Imaging Reporting and Data System (TI-RADS). A total of the patients were subjected to the DNA ploidy analysis, the BRAF V600E gene detection, or the combination of both techniques. We evaluated the diagnostic value of the above techniques and considered the postoperative pathology results as gold standard for cancer diagnosis. The negative predictive value (NPV), accuracy, specificity, sensitivity, and positive predictive value (PPV) of TI-RADS, BRAF V600E gene detection, DNA ploidy analysis, and BRAF V600E gene detection joined with DNA ploidy analysis were calculated.
RESULTS: Among 400 subjects, 238 presented thyroid cancer and 162 had benign lesions, according to the postoperative pathology results. The obtained sensitivity, specificity, accuracy, PPV, and NPV values of TI-RADS were 55.88%, 58.64%, 57.00%, 66.50%, 47.50%, respectively; of BRAF V600E gene detection were 81.93%, 69.75%, 77.00%, 79.92%, 72.44%, respectively; of DNA ploidy analysis were 83.19%, 72.84%, 79.00%, 81.82%, 74.68%, respectively; of BRAF V600E gene combined with DNA ploidy analysis were 90.34%, 76.54%, 84.75%, 84.98%, 84.35%, respectively. Compared with TI-RADS, the sensitivity, specificity, accuracy, PPV, and NPV values of DNA ploidy analysis, BRAF V600E gene detection, and the conjunction of these last two methods were increased (p < 0.05). The combination of DNA ploidy analysis and BRAF V600E gene detection had the highest values among them all.
CONCLUSIONS: BRAF V600E gene detection in conjunction with DNA ploidy analysis showed a better diagnostic value than both methods separately or TI-RADS.

OBJECTIVE: The influence of age on the malignant cytology rate of thyroid nodules remains uncertain. The American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) is currently used to guide subsequent investigations of thyroid nodules, regardless of clinical variables. This study aimed to investigate the impact of age on the malignant cytology rates of thyroid nodules and the diagnostic performance of ACR TI-RADS across different age groups.
METHODS: A retrospective, single-center, observational study.
METHODS: Patients aged ≥ 20 years with thyroid nodules, who underwent fine-needle aspiration biopsy between 2012 and 2019 were evaluated. Ultrasound images were used to obtain the TI-RADS data. Malignancy was determined based on suspicious for malignancy (Bethesda V) and malignant (Bethesda VI) cytology results or malignancy in cell block analysis.
RESULTS: A total of 1023 nodules from 921 patients (88.2% female) were analyzed. The median age was 58.5 (interquartile range [IQR], 41.1-66.6) years, and the median nodule size was 2.4 (IQR, 1.7-3.6) cm. Stratification by age revealed a decreasing prevalence of malignant cytology across subgroups of 20-39, 40-59, and ≥60 years (10.7%, 8.5%, and 3.7%, respectively; P = .002). After adjusting for sex, multinodularity, nodule size, and ACR TI-RADS category, we observed that each year of age reduced the OR for malignant cytology by 3.0% (95% CI: 0.7%-5.3%; P = .011). When comparing the subgroups of 20-39 and ≥60 years, the malignant cytology rate decreased by half in TI-RADS 4 (from 21.4% to 10.4%) and two-thirds in TI-RADS 5 (from 64.7% to 22.6%).
CONCLUSIONS: Our study demonstrated that as patient age increased, the rate of malignant cytology in thyroid nodules decreased. Moreover, age significantly influences the malignancy rates of thyroid nodules classified according to the ACR TI-RADS.

OBJECTIVE: Our purposes were: 1) to estimate the prediction performance (PP) of cytology in identifying papillary thyroid carcinoma (PTC) subtypes; 2) to explore how the PTC subtypes distribute among the American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) categories.
METHODS: Nodules were included if both the histology with the PTC subtype report and the cytology report with the possible PTC subtype were available. The PP was calculated by making the proportion of True positives/False positives+false negatives.
RESULTS: 309 cytologically \"suspicious for malignancy\" and \"malignant\" thyroid nodules with PTC histology were evaluated. ACR TI-RADS categorization for classical PTC was significantly different from non-classical PTC (p-value 0.02). For the whole cohort the PP of cytologically classical cases was 0.74, while that of cytologically non classical cases was 0.41. ACR TI-RADS categorization was not significantly different for aggressive vs non-aggressive PTC subtypes (p-value 0.1). When considering only aggressive or non-aggressive PTC subtypes, the PP of cytologically classical cases was respectively 0.86 and 0.87, while that of cytologically non classical cases was respectively 0.27 and 0.22. The PP of cytologically classical cases was 0.73 and 0.79, respectively for macroPTCs and microPTCs, while that of cytologically non classical cases was 0.55 and 0.33, respectively for macroPTCs and microPTCs.
CONCLUSIONS: Cytology examination reliably performed in predicting classical PTC versus non classical PTC subtypes. ACR TI-RADS categorization was significantly different among classical PTC versus non classical PTC subtypes.

Background: Risk stratification systems for thyroid nodules are limited by low specificity. The fine-needle aspiration (FNA) biopsy size thresholds and stratification criteria are based on evidence from the literature and expert consensus. Our aims were to investigate the optimal FNA biopsy size thresholds in the American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) and artificial intelligence (AI) TI-RADS and to revise the stratification criteria in AI TI-RADS. Methods: A total of 2596 thyroid nodules (in 2511 patients) on ultrasound examination with definite pathological diagnoses were retrospectively identified from January 2017 to September 2021 in 6 participating Chinese hospitals. The modified criteria for ACR TI-RADS were as follows: (1) no FNA for TR3; (2) FNA threshold for TR4 increased to 2.5 cm. The modified criteria for AI TI-RADS were as follows: (1) 6-point nodules upgraded to TR5; (2) no FNA for TR3; (3) FNA threshold for TR4 increased to 2.5 cm. The diagnostic performance and the unnecessary FNA rate (UFR) of modified versions were compared with the original ACR TI-RADS. Results: Compared with the original ACR TI-RADS, the modified ACR (mACR) TI-RADS yielded higher specificity (73% vs. 46%), accuracy (74% vs. 51%), area under the receiver operating characteristic curve (AUC; 0.80 vs. 0.70), and lower UFR (25% vs. 48%; all p < 0.001), although the sensitivity was slightly decreased (87% vs. 93%, p = 0.057). Compared with the original ACR TI-RADS, the modified AI (mAI) TI-RADS yielded higher specificity (73% vs. 46%), accuracy (75% vs. 51%), AUC (0.81 vs. 0.70), and lower UFR (24% vs. 48%; all p < 0.001), although the sensitivity tended to be slightly decreased (89% vs. 93%, p = 0.13). There was no significant difference between the mACR TI-RADS and mAI TI-RADS in the diagnostic performance and UFR (all p > 0.05). Conclusions: The revised FNA thresholds and the stratification criteria of the mACR TI-RADS and mAI TI-RADS may be associated with improvements in specificity and accuracy, without significantly sacrificing sensitivity for malignancy detection.

UNASSIGNED: The American College of Radiology Thyroid Image Reporting and Data System (ACR TI-RADS) was developed to predict malignancy risk in thyroid nodules using ultrasound features. TI-RADS was derived from a database of patients already selected for fine-needle aspiration (FNA), raising uncertainty about applicability to unselected patients.
UNASSIGNED: We aimed to assess the effect of ACR TI-RADS reporting in unselected patients presenting for thyroid ultrasound in a real-world setting.
UNASSIGNED: Records for all patients presenting for thyroid ultrasonography in Canterbury, New Zealand, were reviewed across two 18-month periods, prior to and after implementation of TI-RADS reporting. Patient outcomes were compared between the 2 periods. Malignancy rates were calculated for nodules 10 mm or larger with a definitive FNA or histology result.
UNASSIGNED: A total of 1210 nodules were identified in 582 patients prior to implementation of TI-RADS; 1253 nodules were identified in 625 patients after implementation of TI-RADS. TI-RADS category was associated with malignancy rate (0% in TR1 and TR2, 3% in TR3, 5% in TR4, 12% in TR5; P = .02); however, 63% of nodules were graded TR3 or TR4, for which malignancy rate did not meaningfully differ from baseline risk. After implementation of TI-RADS there was a small reduction in the proportion of patients proceeding to FNA (49% vs 60%; P < .01) or surgery (14% vs 18%; P < .05), with no difference in cancer diagnoses (3% vs 4%, not significant).
UNASSIGNED: TI-RADS category is associated with malignancy rate and may alter clinical decision-making in a minority of patients; however, it is nondiscriminatory in the majority of nodules. In this study of unselected patients, nodules classified as TR5 and thus considered \"highly suspicious\" for cancer had only a modest risk of malignancy.