Risk Of Malignancy In Nodules Research Articles

Enhancing cancer prediction in challenging screen-detected incident lung nodules using time-series deep learning

Lung cancer screening (LCS) using annual computed tomography (CT) scanning significantly reduces mortality by detecting cancerous lung nodules at an earlier stage. Deep learning algorithms can improve nodule malignancy risk stratification. However, they have typically been used to analyse single time point CT data when detecting malignant nodules on either baseline or incident CT LCS rounds. Deep learning algorithms have the greatest value in two aspects. These approaches have great potential in assessing nodule change across time-series CT scans where subtle changes may be challenging to identify using the human eye alone. Moreover, they could be targeted to detect nodules developing on incident screening rounds, where cancers are generally smaller and more challenging to detect confidently.Here, we show the performance of our Deep learning-based Computer-Aided Diagnosis model integrating Nodule and Lung imaging data with clinical Metadata Longitudinally (DeepCAD-NLM-L) for malignancy prediction. DeepCAD-NLM-L showed improved performance (AUC = 88%) against models utilizing single time-point data alone. DeepCAD-NLM-L also demonstrated comparable and complementary performance to radiologists when interpreting the most challenging nodules typically found in LCS programs. It also demonstrated similar performance to radiologists when assessed on out-of-distribution imaging dataset. The results emphasize the advantages of using time-series and multimodal analyses when interpreting malignancy risk in LCS.

Enhancing a deep learning model for pulmonary nodule malignancy risk estimation in chest CT with uncertainty estimation.

To investigate the effect of uncertainty estimation on the performance of a Deep Learning (DL) algorithm for estimating malignancy risk of pulmonary nodules. In this retrospective study, we integrated an uncertainty estimation method into a previously developed DL algorithm for nodule malignancy risk estimation. Uncertainty thresholds were developed using CT data from the Danish Lung Cancer Screening Trial (DLCST), containing 883 nodules (65 malignant) collected between 2004 and 2010. We used thresholds on the 90th and 95th percentiles of the uncertainty score distribution to categorize nodules into certain and uncertain groups. External validation was performed on clinical CT data from a tertiary academic center containing 374 nodules (207 malignant) collected between 2004 and 2012. DL performance was measured using area under the ROC curve (AUC) for the full set of nodules, for the certain cases and for the uncertain cases. Additionally, nodule characteristics were compared to identify trends for inducing uncertainty. The DL algorithm performed significantly worse in the uncertain group compared to the certain group of DLCST (AUC 0.62 (95% CI: 0.49, 0.76) vs 0.93 (95% CI: 0.88, 0.97); p < .001) and the clinical dataset (AUC 0.62 (95% CI: 0.50, 0.73) vs 0.90 (95% CI: 0.86, 0.94); p < .001). The uncertain group included larger benign nodules as well as more part-solid and non-solid nodules than the certain group. The integrated uncertainty estimation showed excellent performance for identifying uncertain cases in which the DL-based nodule malignancy risk estimation algorithm had significantly worse performance. Deep Learning algorithms often lack the ability to gauge and communicate uncertainty. For safe clinical implementation, uncertainty estimation is of pivotal importance to identify cases where the deep learning algorithm harbors doubt in its prediction. • Deep learning (DL) algorithms often lack uncertainty estimation, which potentially reduce the risk of errors and improve safety during clinical adoption of the DL algorithm. • Uncertainty estimation identifies pulmonary nodules in which the discriminative performance of the DL algorithm is significantly worse. • Uncertainty estimation can further enhance the benefits of the DL algorithm and improve its safety and trustworthiness.

Radiomics and artificial intelligence for risk stratification of pulmonary nodules: Ready for primetime?

Pulmonary nodules are ubiquitously found on computed tomography (CT) imaging either incidentally or via lung cancer screening and require careful diagnostic evaluation and management to both diagnose malignancy when present and avoid unnecessary biopsy of benign lesions. To engage in this complex decision-making, clinicians must first risk stratify pulmonary nodules to determine what the best course of action should be. Recent developments in imaging technology, computer processing power, and artificial intelligence algorithms have yielded radiomics-based computer-aided diagnosis tools that use CT imaging data including features invisible to the naked human eye to predict pulmonary nodule malignancy risk and are designed to be used as a supplement to routine clinical risk assessment. These tools vary widely in their algorithm construction, internal and external validation populations, intended-use populations, and commercial availability. While several clinical validation studies have been published, robust clinical utility and clinical effectiveness data are not yet currently available. However, there is reason for optimism as ongoing and future studies aim to target this knowledge gap, in the hopes of improving the diagnostic process for patients with pulmonary nodules.

Malignancy in Thyroid Nodules with Bethesda III Category on Repeat Fine Needle Aspiration Biopsy.

This study aimed to evaluate the risk of malignancy for nodules repeatedly classified as Bethesda category III on fine needle aspiration biopsy (FNAB). A chart review on a series of 59 patients seen with thyroid nodules who underwent both initial and repeat FNAB at the Diabetes, Thyroid and Endocrine Center of St. Luke's Medical Center, Quezon City was conducted. The Thyroid Registry was utilized to collect each patient's demographic and clinical characteristics, ultrasonographic features of thyroid nodules along with the cytopathologic and histopathologic results. The subclassification of atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS) were retrieved from the cytopathology reports using the institution's electronic Healthcare-Results Management System. A total of 59 adult patients with thyroid nodules were included. Nodules which were initially AUS/FLUS turned out to be malignant on repeat FNAB in 38 patients with a prevalence of 64.41% (95% CI: 50.87-76.45%). There was no significant difference with regards to clinical, ultrasonographic and cytopathologic features of malignancy between benign and malignant nodules. Findings of this study support surgical intervention as a reasonable option after a repeat Bethesda III classification on FNAB. However, the small sample size warrants confirmation in future studies with a representative sample of patients.

Influence of CT dose reduction on AI-driven malignancy estimation of incidental pulmonary nodules

ObjectivesThe purpose of this study was to determine the influence of dose reduction on a commercially available lung cancer prediction convolutional neuronal network (LCP-CNN).MethodsCT scans from a cohort provided by the local lung cancer center (n = 218) with confirmed pulmonary malignancies and their corresponding reduced dose simulations (25% and 5% dose) were subjected to the LCP-CNN. The resulting LCP scores (scale 1–10, increasing malignancy risk) and the proportion of correctly classified nodules were compared. The cohort was divided into a low-, medium-, and high-risk group based on the respective LCP scores; shifts between the groups were studied to evaluate the potential impact on nodule management. Two different malignancy risk score thresholds were analyzed: a higher threshold of ≥ 9 (“rule-in” approach) and a lower threshold of > 4 (“rule-out” approach).ResultsIn total, 169 patients with 196 nodules could be included (mean age ± SD, 64.5 ± 9.2 year; 49% females). Mean LCP scores for original, 25% and 5% dose levels were 8.5 ± 1.7, 8.4 ± 1.7 (p > 0.05 vs. original dose) and 8.2 ± 1.9 (p < 0.05 vs. original dose), respectively. The proportion of correctly classified nodules with the “rule-in” approach decreased with simulated dose reduction from 58.2 to 56.1% (p = 0.34) and to 52.0% for the respective dose levels (p = 0.01). For the “rule-out” approach the respective values were 95.9%, 96.4%, and 94.4% (p = 0.12). When reducing the original dose to 25%/5%, eight/twenty-two nodules shifted to a lower, five/seven nodules to a higher malignancy risk group.ConclusionCT dose reduction may affect the analyzed LCP-CNN regarding the classification of pulmonary malignancies and potentially alter pulmonary nodule management.Clinical relevance statementUtilization of a “rule-out” approach with a lower malignancy risk threshold prevents underestimation of the nodule malignancy risk for the analyzed software, especially in high-risk cohorts.Key Points• LCP-CNN may be affected by CT image parameters such as noise resulting from low-dose CT acquisitions.• CT dose reduction can alter pulmonary nodule management recommendations by affecting the outcome of the LCP-CNN.• Utilization of a lower malignancy risk threshold prevents underestimation of pulmonary malignancies in high-risk cohorts.

THU222 Malignancy Risk In Nodules Classified As Atypia Of Undetermined Significance In A Predominantly Hispanic Free-standing Children’s Hospital

Abstract Disclosure: D. Baboun: None. N. Solano: None. A. Martinez Sanchez: None. C. Brathwaite: None. R. Restrepo: None. F. Alkhoury: None. A. Carrillo-Iregui: None. Background: During the last 3 decades, there has been an increase in pediatric thyroid cancer. Ultrasound-guided fine needle aspiration (US-FNA) of thyroid nodules continues to be the main modality for histopathological evaluation. Thyroid nodules reported as atypia of undetermined significance (AUS) present a challenge to clinicians in deciding on surgical intervention versus observation (1). Objective: The purpose of this study is to determine the rate of malignancy in cases classified as AUS in the predominantly Hispanic population treated at a free-standing children’s hospital. Methods: Upon Institutional Review Board approval, we performed a retrospective chart review from 2012-2022 of 211 patients who underwent US-FNAs of thyroid nodules. We analyzed demographics and histopathological diagnoses. Results: The majority of our population identified as Hispanic (80.10%). There were 167 (79.10%) females and 44 (20.90%) males, and the age range was 4-21 years. Two hundred eleven patients underwent a total of 232 US-FNAs. The Bethesda distribution is as follows: 29 (12.50%) nondiagnostic, 98 (42.24%) benign, 45 (19.40%) AUS, 21 (9.05%) follicular neoplasm, 9 (3.88%) suspicious, and 30 (12.93%) malignant. Of the AUS group, 33 of the 43 patients underwent surgery. Of the surgical specimens, 16 (48.48%) were malignant, and 17 (51.52%) were benign. Of the 10 patients who did not undergo surgery, 2 underwent a repeat US-FNA with benign results, 5 were monitored, and 3 did not return for a follow-up. Conclusions: In our experience, patients with US-FNAs classified as AUS have a rate of malignancy of 48.48%, which is consistent with the published rate of higher incidence in the pediatric population than in the adult population. Thus, our results support the current recommendation of surgical intervention rather than observation in these pediatric patients. Molecular markers are a useful tool for determining management of thyroid nodules in adults; however, further research in the pediatric population is needed to have standardized recommendations. Reference: (1) Rossi, E.D. et al., Cancer Cytopathology. 2017 Aug 125:594-603. Presentation: Thursday, June 15, 2023

Assessment of thyroid nodule malignancy risk in medical radiation workers using ACR TI-RADS

BackgroundLow-dose radiation exposure brings medical radiation professionals at risk for thyroid parenchymal disease and nodules. ObjectiveTo classify thyroid nodules in medical radiation workers according to their risk, and to look into parenchymal disease and thyroid dysfunction. Design, setting and participantsRetrospective observational study in a tertiary health facility with 148 healthcare workers working in radiation-containing areas. Main outcome and measuresParticipants were divided into groups based on their sociodemographic characteristics, the department they worked in, and the number of years they had been in the profession. The levels of TSH, parenchymal heterogeneity, and thyroid nodule presence were compared between the groups. American College of Radiology (ACR) Thyroid Imaging – Reporting and Data System (TI-RADS) classification was used to evaluate thyroid nodules. ResultsThe presence of nodules was significantly more prevalent in patients over the age of 33 (p = 0.02). The majority of individuals (98%) had nodules that below the American College of Radiology (ACR) Thyroid Imaging – Reporting and Data System (TI-RADS) category 4. Age and professional experience both elevated TSH results (p = 0.015, p = 0.033). With age, parenchymal heterogeneity became more prominent (p = 0.04). For the parameters, there were no appreciable differences between the occupation and department. The results applied to both genders equally. ConclusionWith increasing long-term low-dose exposure, medical radiation workers are at risk for nodules and parenchymal disease. However, the majority of the nodules in our study were sonographically in the low risk group, therefore routine follow-up is advised.

Zinc in microscopic calcifications isolated from thyroid fine needle aspiration may serve as a biomarker of thyroid nodule malignancy: A promising proof-of-concept.

Thyroid nodules (TNs) are common neck ultrasonography (US) findings, yet only 5-10% of these nodules harbor thyroid cancer (TC). When US characteristics are consistent with an intermediate or high suspicion for TN malignancy, fine needle aspiration for cytology (FNAC) is indicated. The main limitation of FNAC is that cytological results can be indeterminate in up to 30% of cases, necessitating reevaluation through repeated FNAC, expensive molecular testing, or diagnostic thyroid lobe resection. As such, there is a need for further refinement of current diagnostic algorithms for TNs without subjecting patients to additional invasive procedures. As calcifications detected during thyroid US are considered a high-risk feature for malignancy, we used the material remaining following routine thyroid FNAC to isolate microscopic calcifications (MCs). We then characterized the elemental composition, morphology, and crystal phases of these MCs, ultimately revealing differences between the MCs from benign and malignant TNs. Specifically, thyroid MCs were identified as calcium phosphate crystals containing varying levels of magnesium, sodium, iron, and zinc. MCs obtained from malignant TNs, mainly papillary thyroid carcinoma, were composed of sub-micrometer spherical particles, whereas MCs from benign TNs consisted of faceted particles. While samples from most patients with a final diagnosis of malignant TNs (50% of them with indeterminate cytology) harbored zinc-containing MCs, zinc was largely absent in MCs from benign TNs (23% with indeterminate or non-diagnostic cytology). Together, these data suggest that the presence of zinc in MCs isolated from samples collected during routine FNAC may potentially offer value as a biomarker of TN malignancy. STATEMENT OF SIGNIFICANCE: As up to 40% of patients assessed for thyroid malignancy do not receive a definite diagnosis following thyroid nodule (TN) fine needle aspiration (FNA), there is a pressing need to improve the accuracy of current diagnostic algorithms. Chemical analyses of microscopic calcifications (MCs) may serve as a diagnostic target. We developed a straightforward protocol to chemically characterize MCs from excess material collected from TNs during routine FNA and found that these MCs differed between benign and malignant TNs. Specifically, zinc in TN-derived MCs may indicate a higher nodule malignancy risk, thus increasing the diagnostic accuracy of the FNA procedure, reducing the need for recurrent biopsies and diagnostic surgical procedures, and decreasing the costs, uncertainty, and stress faced by affected patients.

ACR TI-RADS Score combined with cytopathology classification improves the risk stratification of indeterminate thyroid nodules.

The aim of this retrospective study was the elaboration of a new diagnostic model that integrate cytological reports (2017 Bethesda System for Reporting Thyroid Cytopathology) with ultrasonographic features (based on ACR TI-RADS score) to achieve a more accurate definition of indeterminate thyroid nodule malignancy risk. Ninety patients submitted to thyroidectomy were divided in three classes: low malignancy risk (AUS/FLUS with TI-RADS 2/TI-RADS 3 and FN/SFN with TI-RADS 2), intermediate malignancy risk (AUS/FLUS with TI-RADS 4/TI-RADS 5 and FN/SFN with TI-RADS 3/TI-RADS 4), and high malignancy risk (FN/SFN with TI-RADS 5). The surgical approach should be recommended in high-risk patients (81.82% of malignancies), carefully evaluated in intermediate risk (25.42%), whereas a conservative approach can be adopted in low-risk patients (0.00%). The integration of these two multiparametric systems in a Cyto-US score has proven to be a feasible and reliable aid to achieve a more accurate definition of malignancy risk.

Increased workload without clinical benefit: Results following implementation of the ACR-TIRADS system for thyroid nodules.

The ACR-TIRADS system for stratifying thyroid nodule malignancy risk has been widely promoted and implemented. We audited its introduction at a large public hospital in Auckland, New Zealand. Audit of outcomes following thyroid nodule fine needle aspiration (FNA) before/after ACR-TIRADS. Individuals undergoing thyroid FNA 2017-2019. From medical records, we obtained details from the pre-FNA ultrasound (nodule size, TIRADS points/levels, radiologist recommendation for FNA), Bethesda (B) cytology classification, histology and post-FNA follow-up. Four hundred and twenty-two individuals had 564 FNAs, 163 had surgery and 54 (13%) had cancer in the primary nodule. 37/54 (69%) cancers were papillary thyroid carcinoma (median size 25 mm, 87% ≥10 mm, 61% ≥20 mm). Following ACR-TIRADS introduction, FNA recommendations increased greater than twofold, FNAs performed by 71%-83%, and the monthly rate of FNAs and operations by 60% and 40%, respectively. However, the proportion of cancers/FNA remained similar (9.9% post-TIRADS vs. 8.7% pre-TIRADS). The proportions of FNA results remained stable for B2-B4 categories, but doubled (11% vs. 5%) for B5-B6: 15 FNAs were needed to identify an additional B5/B6 lesion. TIRADS-5 nodules had a higher proportion of B5/B6 (20%) and a lower proportion of B2 (30%) than TIRADS-3 (2%, 57%, respectively) and TIRADS-4 (9%, 56%) nodules. About 5 additional cancers/year were diagnosed, but they were more often small (49% vs. 8% <2 cm, 17% vs. 0% <1 cm). ACR-TIRADS introduction increased workload (FNAs and operations), without increasing the proportion of cancers/FNA. It led to a few more cancers being diagnosed, but many were small and of uncertain clinical significance.

Quantitative CT analysis of lung parenchyma to improve malignancy risk estimation in incidental pulmonary nodules

ObjectivesTo assess the value of quantitative computed tomography (QCT) of the whole lung and nodule-bearing lobe regarding pulmonary nodule malignancy risk estimation.MethodsA total of 251 subjects (median [IQR] age, 65 (57–73) years; 37% females) with pulmonary nodules on non-enhanced thin-section CT were retrospectively included. Twenty percent of the nodules were malignant, the remainder benign either histologically or at least 1-year follow-up. CT scans were subjected to in-house software, computing parameters such as mean lung density (MLD) or peripheral emphysema index (pEI). QCT variable selection was performed using logistic regression; selected variables were integrated into the Mayo Clinic and the parsimonious Brock Model.ResultsWhole-lung analysis revealed differences between benign vs. malignant nodule groups in several parameters, e.g. the MLD (−766 vs. −790 HU) or the pEI (40.1 vs. 44.7 %). The proposed QCT model had an area-under-the-curve (AUC) of 0.69 (95%-CI, 0.62−0.76) based on all available data. After integrating MLD and pEI into the Mayo Clinic and Brock Model, the AUC of both clinical models improved (AUC, 0.91 to 0.93 and 0.88 to 0.91, respectively). The lobe-specific analysis revealed that the nodule-bearing lobes had less emphysema than the rest of the lung regarding benign (EI, 0.5 vs. 0.7 %; p < 0.001) and malignant nodules (EI, 1.2 vs. 1.7 %; p = 0.001).ConclusionsNodules in subjects with higher whole-lung metrics of emphysema and less fibrosis are more likely to be malignant; hereby the nodule-bearing lobes have less emphysema. QCT variables could improve the risk assessment of incidental pulmonary nodules.Key Points• Nodules in subjects with higher whole-lung metrics of emphysema and less fibrosis are more likely to be malignant.• The nodule-bearing lobes have less emphysema compared to the rest of the lung.• QCT variables could improve the risk assessment of incidental pulmonary nodules.

Prognostic Indicators of EIF1AX-Mutated Thyroid Tumor Malignancy and Cancer Aggressiveness.

The risk of malignancy (ROM) of EIF1AX-mutated thyroid nodules has been theorized to be contingent on the position of the mutation within the gene and the presence of co-existing mutations. However, due to EIF1AX's low mutation frequency, sample sizes currently reported in the literature are too diminutive to appraise the clinical utility of molecular diagnostic testing. The objective of this study was to elucidate prognostic indicators of EIF1AX-mutated thyroid tumors and cancer aggressiveness by examining a large cohort of cytologically indeterminate thyroid nodules (CITNs) that underwent molecular testing and subsequent surgical resection. This is a multicenter study involving 764 subtotal and total thyroidectomy patients that underwent preoperative molecular testing at two quaternary care hospitals. A five-year retrospective review was performed on the 42 charts of patients that opted for surgery following a positive EIF1AX mutation on ThyroseqV3 results from January 2018 to May 2022. Patient demographics, cytopathology results, molecular testing results, and postoperative histopathology were reviewed. Of the 42 surgically resected nodules that harbored an EIF1AX mutation, 16 (38.1%) were benign, six (14.3%) were non-invasive follicular thyroid neoplasms with papillary-like nuclear features (NIFTPs) or well-differentiated thyroid neoplasms of uncertain malignant potential (WDT-UMPs), and 20 (47.6%) were malignant. An isolated EIF1AX mutation conferred a ROM of 47.6%, whereas the ROM for nodules with at least one additional molecular alteration was 72.7%. The ROM increased to 100% for nodules with at least one additional molecular alteration and the A113_splice site mutation. Six malignant nodules were aggressive, with five having variegated components of poorly differentiated thyroid carcinoma (PDTC). EIF1AX-mutated thyroid nodules are more susceptible to malignancy in the presence of the A113_splice site mutation and when co-mutated with RAS and/or TP53. This deleterious amalgam is associated with aggressive disease and renders these nodules PDTC. A preoperative molecular test finding of an EIF1AX mutation can be a useful tool for thyroid specialists to optimize clinical management.

Malignancy risk stratification of thyroid nodules according to echotexture and degree of hypoechogenicity: a retrospective multicenter validation study

Various risk stratification systems show discrepancies in the ultrasound lexicon of nodule echotexture and hypoechogenicity. This study aimed to determine the malignancy risk of thyroid nodules according to their echotexture and degree of hypoechogenicity. From June to September 2015, we retrospectively evaluated 5601 thyroid nodules with final diagnoses from 26 institutions. Nodules were stratified according to the echotexture (homogeneous vs. heterogeneous) and degree of hypoechogenicity (mild, moderate, or marked). We calculated the malignancy risk according to composition and suspicious features. Heterogeneous hypoechoic nodules showed a significantly higher malignancy risk than heterogeneous isoechoic nodules (P ≤ 0.017), except in partially cystic nodules. Malignancy risks were not significantly different between homogeneous versus heterogeneous nodules in both hypoechoic (P ≥ 0.086) and iso- hyperechoic nodules (P ≥ 0.05). Heterogeneous iso-hyperechoic nodules without suspicious features showed a low malignancy risk. The malignancy risks of markedly and moderately hypoechoic nodules were not significantly different in all subgroups (P ≥ 0.48). Marked or moderately hypoechoic nodules showed a significantly higher risk than mild hypoechoic (P ≤ 0.016) nodules. The predominant echogenicity effectively stratifies the malignancy risk of nodules with heterogeneous echotexture. The degree of hypoechogenicity could be stratified as mild versus moderate to marked hypoechogenicity.

Solitary focal lung lesions: modern views on the diagnosis and medical tactics of their management. Literature review

A review of the literature is devoted to highlighting of the main points relating to the modern actual multidisciplinary problem — single or solitary lung nodules with incidental detection. The main terms that are used in the study of such lesions or the more common term — nodules, as well as their definition are given. Known pathology causes for the development of solitary pulmonary nodules are also presented. The morphological characteristics of such nodules are described by X-ray or tomogram. The main problem if such nodules are detected is to establish their nature and, above all, their malignant or benign nature. Since the further medical tactic of their management are exactly different. Now the main attention is paid to malignant tumors, since they require the fastest most reliable identification and appropriate medical actions.It has been established that accurate determination of nodules etiology by the main indicators of their radiological images is impossible. Although certain dependences of the malignancy risk of nodules on their size, contours and features of optical density have been identified. And only a morphological study allows us to establish the true nature of the nodules. It is believed that over diagnostics of malignancy of pulmonary nodules leads to excessive use of expensive invasive diagnostic procedures, and to some extent harms the patient in cases of benign nodules. The review ends with a list of possible options for medical management when solitary lung nodules of unknown origin are detected. Their disadvantages and advantages are briefly presented.

Novel Human Artificial Intelligence Hybrid Framework Pinpoints Thyroid Nodule Malignancy and Identifies Overlooked Second-Order Ultrasonographic Features.

Simple SummaryDeep learning-based computer-aided diagnosis has gained momentum in the radiology field thanks to the technological advances of convolutional neural networks (CNN). However, how to utilize the black-box predictions of these CNN models to the clinical routine still relies on radiologists’ personal judgements. In addition, existing CNN models only improve radiologists’ diagnosis when they outperform the radiologists, thereby limiting their added values for possible efficiency enhancement and improving mostly the diagnostic performances of junior radiologists.We present a Human Artificial Intelligence Hybrid (HAIbrid) integrating framework that reweights Thyroid Imaging Reporting and Data System (TIRADS) features and the malignancy score predicted by a convolutional neural network (CNN) for nodule malignancy stratification and diagnosis. We defined extra ultrasonographical features from color Doppler images to explore malignancy-relevant features. We proposed Gated Attentional Factorization Machine (GAFM) to identify second-order interacting features trained via a 10 fold distribution-balanced stratified cross-validation scheme on ultrasound images of 3002 nodules all finally characterized by postoperative pathology (1270 malignant ones), retrospectively collected from 131 hospitals. Our GAFM-HAIbrid model demonstrated significant improvements in Area Under the Curve (AUC) value (p-value < 10−5), reaching about 0.92 over the standalone CNN (~0.87) and senior radiologists (~0.86), and identified a second-order vascularity localization and morphological pattern which was overlooked if only first-order features were considered. We validated the advantages of the integration framework on an already-trained commercial CNN system and our findings using an extra set of ultrasound images of 500 nodules. Our HAIbrid framework allows natural integration to clinical workflow for thyroid nodule malignancy risk stratification and diagnosis, and the proposed GAFM-HAIbrid model may help identify novel diagnosis-relevant second-order features beyond ultrasonography.

Investigating the possible association between thyroid nodule location and the malignancy risk of the nodules in FNA samples

IntroductionIn this study we aimed to investigate the association between thyroid nodule location and the malignancy risk of the nodules in FNA samples. MethodsIn this cross sectional study, 400 patients with thyroid nodule size >1.5 cm, thyroid nodule size >1 cm with micro-calcification, and thyroid nodule size <1 cm with sonographic or clinical signs of malignancy who referred to a training hospital, were involved. Morphologic characteristics of nodules, including nodule size, laterality and polarity were recorded. In patients with multinodular goiter, the anatomical location was recorded and analyzed. ResultsIn our study, 345 female patients (86.3 %) and 55 male patients (13.8 %) with thyroid nodule were involved. The majority of thyroid nodules were identified in the right lobe (n = 182, 45.5 %) and left lobe (n = 177, 44.3 %), respectively. A total of 323 nodules were benign and the remaining (n = 77) were malignant. The frequency of malignancy in the upper pole of the thyroid (31.6 %) was higher compared to that in the middle pole (10.7 %), lower pole (19.1 %), and isthmus region (26.8 %), leading to a significant difference (p.value = 0.001). The mean size of thyroid nodules was 15.68 mm in benign nodules and 20.08 mm in malignant nodules. ConclusionOur results showed that thyroid nodules found in the upper lobe carried the greatest risk for malignancy. Accordingly, it seems that the location of thyroid nodules can be an independent risk factor in determining the malignancy risk of thyroid nodules and can be used to enhance clinical decision making in thyroid FNA samples.

Benign cytology does not rule out malignancy in thyroid nodules larger than 4cm.

For thyroid nodules ≥4cm, the accuracy of fine-needle aspiration biopsy (FNAB) is controversial and the approach is unclear. We aimed to compare FNAB and operation of thyroid nodules and to determine the accuracy of FNAB. All total thyroidectomies performed between January 2015 and December 2021 were evaluated. In the study, 301 patients were included. Euthyroid patients with preoperative thyroid ultrasound, FNAB results and operation results were recorded retrospectively. The nodule size was <4cm in 79.1% of the patients, and ≥4cm in 20.9%. In patients with nodule size ≥4cm, 50.8% of FNAB results were reported as benign, and 43.7% of these patients were found to be malignant at the end of the operation. In nodules <4cm, 36.8% of the patients whose FNAB results were found to be benign were malignant. False-negativity rate was found to be quite high in ≥4cm nodules. For thyroid nodules, diagnostic lobectomy may be necessary because the false-negative rate of FNAB was high, especially in nodules ≥4cm. In addition, intermediate results, such as AUS/FN, have a higher risk of malignancy in nodules of ≥4cm compared to nodules of <4cm.

A Practical CEUS Thyroid Reporting System for Thyroid Nodules.

Background The role of contrast-enhanced US (CEUS) in reducing unnecessary biopsies of thyroid nodules has received little attention. Purpose To construct and externally validate a thyroid imaging reporting and data system (TI-RADS) based on nonenhanced US and CEUS to stratify the malignancy risk of thyroid nodules. Materials and Methods This retrospective study evaluated 756 patients with 801 thyroid nodules who underwent nonenhanced US, CEUS, and fine-needle aspiration and received a final diagnosis from January 2018 to December 2019. Qualitative US features of the thyroid nodules were analyzed with univariable and multivariable logistic regression to construct a CEUS TI-RADS. The CEUS TI-RADS was validated with use of internal cross-validation and external validation. Results A total of 801 thyroid nodules in 590 female (mean age, 44 years ± 13) and 166 male (mean age, 47 years ± 13 [SD]) patients were included. Independent predictive US features included nodule composition at CEUS, echogenicity, nodule shape, nodule margin, echogenic foci, extrathyroidal extension, enhancement direction, peak intensity, and ring enhancement. The CEUS TI-RADS showed a higher area under the receiver operating characteristic curve of 0.93 (95% CI: 0.92, 0.95; P < .001 in comparison with all other systems), a biopsy yield of malignancy of 66% (157 of 239 nodules), and an unnecessary biopsy rate of 34% (82 of 239 nodules). In the external validation, the area under the receiver operating characteristic curve, biopsy yield of malignancy, and unnecessary biopsy rate of CEUS TI-RADS were 0.89 (95% CI: 0.84, 0.92), 61% (65 of 106 nodules), and 39% (41 of 106 nodules) for the first external validation set and 0.90 (95% CI: 0.85, 0.94), 57% (56 of 99 nodules), and 43% (43 of 99 nodules) for the second external validation set. Conclusion A contrast-enhanced US (CEUS) thyroid imaging reporting and data system was created with thyroid nodule malignancy risk stratification according to the simplified regression coefficients of nonenhanced US and qualitative features of CEUS. Clinical trials registration no. ChiCTR2000028712 Published under a CC BY 4.0 license. Online supplemental material is available for this article.

Artificial Intelligence Tool for Assessment of Indeterminate Pulmonary Nodules Detected with CT.

Background Limited data are available regarding whether computer-aided diagnosis (CAD) improves assessment of malignancy risk in indeterminate pulmonary nodules (IPNs). Purpose To evaluate the effect of an artificial intelligence-based CAD tool on clinician IPN diagnostic performance and agreement for both malignancy risk categories and management recommendations. Materials and Methods This was a retrospective multireader multicase study performed in June and July 2020 on chest CT studies of IPNs. Readers used only CT imaging data and provided an estimate of malignancy risk and a management recommendation for each case without and with CAD. The effect of CAD on average reader diagnostic performance was assessed using the Obuchowski-Rockette and Dorfman-Berbaum-Metz method to calculate estimates of area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Multirater Fleiss κ statistics were used to measure interobserver agreement for malignancy risk and management recommendations. Results A total of 300 chest CT scans of IPNs with maximal diameters of 5-30 mm (50.0% malignant) were reviewed by 12 readers (six radiologists, six pulmonologists) (patient median age, 65 years; IQR, 59-71 years; 164 [55%] men). Readers' average AUC improved from 0.82 to 0.89 with CAD (P < .001). At malignancy risk thresholds of 5% and 65%, use of CAD improved average sensitivity from 94.1% to 97.9% (P = .01) and from 52.6% to 63.1% (P < .001), respectively. Average reader specificity improved from 37.4% to 42.3% (P = .03) and from 87.3% to 89.9% (P = .05), respectively. Reader interobserver agreement improved with CAD for both the less than 5% (Fleiss κ, 0.50 vs 0.71; P < .001) and more than 65% (Fleiss κ, 0.54 vs 0.71; P < .001) malignancy risk categories. Overall reader interobserver agreement for management recommendation categories (no action, CT surveillance, diagnostic procedure) also improved with CAD (Fleiss κ, 0.44 vs 0.52; P = .001). Conclusion Use of computer-aided diagnosis improved estimation of indeterminate pulmonary nodule malignancy risk on chest CT scans and improved interobserver agreement for both risk stratification and management recommendations. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Yanagawa in this issue.

Accuracy of the ‘CUT’ Score for Assessing Malignancy in Bethesda 3 and 4 Thyroid Nodules in North American population: a retrospective study.

Background The CUT score is a thyroid nodule malignancy risk assessment scoring system intended to guide surgeons in treating Bethesda 3 and 4 thyroid nodules. It is based on clinical (C) and ultrasonographic (U) features and a five-tiered (T) representing cytology. Purpose Our study aimed to assess the utility of the CUT score in predicting thyroid malignancy in the North American population. The main reason for creating this score is to reduce unnecessary surgeries on these challenging thyroid nodules. Materials and Methods A retrospective record review study applied the CUT score to 219 Bethesda 3 and 4 thyroid nodules. A total of 203 Bethesda 3 and 16 Bethesda 4 nodules from patients treated between January 2015 and December 2019 at a single institution were assessed. A receiver operating characteristic (ROC) curve analysis was performed to evaluate the CUT diagnostic test. Binary logistic regression analysis was performed. Iteration of analysis was performed after stratification according to body mass index to assess CUT score accuracy in obese and non-obese patients. Results Of 219 nodules analyzed, 148 were characterized as benign and 71 as malignant. Prevalence rates of malignancy were 29.6% (n = 60) and 68.8% (n = 11) in Bethesda 3 and 4 nodules, respectively. The mean CU (clinical, ultrasonography) score was 5.35 ± 1.38 in benign nodules versus 4.96 ± 1.5 in malignant nodules (p = 0.08). The area under the curve (AUC =0.433) for the association of CUT scores with nodule malignancy was not significant (p = 0.13). The CUT score was insignificant as a diagnostic test for nodule malignancy in obese (AUC =0.45; p = 0.72) and non-obese patients (AUC =0.39; p = 0.08). Conclusion The CUT score did not correlate with preoperative malignancy risk estimates in Bethesda 3 thyroid nodules and, therefore, may have limited utility as a predictor of malignancy in these thyroid nodules.