Malignant Nodules Research Articles

An early lung cancer diagnosis model for non-smokers incorporating ct imaging analysis and circulating genetically abnormal cells (CACs)

BackgroundAn increase in the prevalence of lung cancer that is not smoking-related has been noticed in recent years. Unfortunately, these patients are not included in low dose computer tomography (LDCT) screening programs and are not actually considered in early diagnosis. Therefore, improved early diagnosis methods are urgently needed for non-smokers. It is necessary to establish a prediction model for non-smoking individuals at intermediate to high risk of developing lung cancer (LC) and develop a tool to address the significant gap in evaluating pulmonary nodules in non-smokers.MethodsWe retrospectively investigated 1121 patients with pulmonary nodules, who underwent LDCT examinations between September 2019 and March 2023. Five artificial intelligence (AI) algorithms were used to build two kinds of models and identify which one was better at diagnosing non-smoking pulmonary nodules patients. In the first model, we assigned 554 non-smoking individuals to a training cohort and 150 non-smoking patients to an independent validation cohort. The second model included 971 patients for the training set and 150 non-smoking patients for an independent validation set. All LDCT images of participants were obtained for AI analysis. AI of LDCT scans, liquid biopsy, and clinical characteristics were collected for model building.ResultsAmong LC patients, 58,4% were non-smokers. Non-smoking patients had a high incidence of LC (71.4%), and women showed a significant excess risk compared with non-smoking men in terms of LC risk. Furthermore, our results indicated that the model built using random forest (RF) method, which integrates clinical characteristics (age, extra-thoracic cancer history, gender), radiological characteristics of pulmonary nodules (nodule diameter, nodule count, upper lobe location, malignant sign at the nodule edge, subsolid status), the artificial intelligence analysis of LDCT data, and liquid biopsy achieved the best diagnostic performance in the independent external non-smokers validation cohort (sensitivity 92%, specificity 97%, area under the curve [AUC] = 0.99).ConclusionsThese results could significantly improve early non-smoker LC diagnosis and treatment for non-smoker patients with malignant nodules. The established multi-omics model is a noninvasive prediction tool for non-smoking malignant pulmonary nodule diagnosis. Validation revealed that these models exhibited excellent discrimination and calibration capacities, especially the first model built using the RF method, suggesting their clinical utility in the early screening and diagnosis of non-smoking LC.

Natural History and Long Term Follow-Up of Incidental Thyroid Nodules on Computed Tomography Imaging.

incidental thyroid nodules (ITNs) are found in up to 25% of CT scans. Increased use of cross-sectional imaging has contributed to the increased incidence of thyroid cancer worldwide. ITNs pose a management dilemma since nodule malignancy rate is 5-15% but most cancers are indolent and prognosis in differentiated thyroid cancer is excellent. Study aims are to determine prevalence of ITNs ≥1 cm on CT scans, evaluate reporting practices, assess for emergence of clinically evident thyroid cancer during 13-year follow-up and assess interim nodule growth and clinical outcomes in nodules that were further investigated. direct image review of 1499 consecutive CT scans that included the thyroid, performed during January 2009 in a large NHS health board was performed. Clinical data up to January 2022 was analysed in 150 patients with at least one ITN > =1cm. ITN prevalence was 11% with mean patient age 70 years and mean nodule diameter 17.5 mm. 30% of ITNs were mentioned in the CT report. During the follow-up period 11% proceeded to thyroid ultrasound, 5% fine needle aspiration and 2% diagnostic hemithyroidectomy with no thyroid malignancy found. 120 patients (80%) were deceased by the study endpoint, none from thyroid malignancy. No patients presented with clinically evident thyroid malignancy during follow-up. None of150 ITN cases developed clinically evident thyroid malignancy in a 13-year follow-up period with 80% of patients deceased by the study endpoint from non-thyroid causes. This would suggest that ITNs detected on CT do not require further investigation unless malignant appearances or significant clinical concern for thyroid malignancy.

A Machine Learning-Based Radiomics Model for the Differential Diagnosis of Benign and Malignant Thyroid Nodules in F-18 FDG PET/CT: External Validation in the Different Scanner.

Accurate diagnosis is essential to avoid unnecessary procedures for thyroid incidentalomas (TIs). Advances in radiomics and machine learning applied to medical imaging offer promise for assessing thyroid nodules. This study utilized radiomics analysis on F-18 FDG PET/CT to improve preoperative differential diagnosis of TIs. A total of 152 patient cases were retrospectively analyzed and split into training and validation sets (7:3) using stratification and randomization. The least absolute shrinkage and selection operator (LASSO) algorithm identified nine radiomics features from 960 candidates to construct a radiomics signature predictive of malignancy. Performance of the radiomics score was evaluated using receiver operating characteristic (ROC) analysis and area under the curve (AUC). In the training set, the radiomics score achieved an AUC of 0.794 (95% CI: 0.703-0.885, p < 0.001). Validation was performed on internal and external datasets, yielding AUCs of 0.702 (95% CI: 0.547-0.858, p = 0.011) and 0.668 (95% CI: 0.500-0.838, p = 0.043), respectively. These results demonstrate that the selected nine radiomics features effectively differentiate malignant thyroid nodules. Overall, the radiomics model shows potential as a valuable predictive tool for thyroid cancer in patients with TIs, supporting improved preoperative decision-making.

Shear wave elastography for thyroid nodule evaluation in patients with chronic autoimmune thyroiditis.

Shear wave elastography (SWE) is a valuable tool in discerning the malignancy risk of thyroid nodules. This study investigates whether 2D-SWE can reliably differentiate malignant thyroid nodules in patients with chronic autoimmune thyroiditis (CAT), despite the challenges posed by fibrosis, which can increase tissue stiffness and complicate diagnosis. This retrospective observational study evaluated 130 thyroid nodules (91 benign, 39 malignant) in patients with underlying CAT using conventional ultrasound (B-mode) and 2D-SWE with SuperSonic Mach30 equipment (Supersonic Imagine, Aix-en-Provence, France). Pathology reports served as the reference standard. Out of the 130 nodules, 30% were malignant, and 70% were benign. 2D-SWE proved to be an excellent distinguisher between benign and malignant nodules. Malignant nodules had significantly higher elasticity indices compared to benign nodules (mean elasticity index: 47.2 kPa vs. 18.1 kPa, p < 0.0001; maximum elasticity index: 75 kPa vs. 26.2 kPa, p < 0.0001). The mean elasticity index was the most reliable elastographic parameter (AUC 0.907, sensitivity 87.2% [95% CI: 77.3-94.0%], specificity 84.6% [95% CI: 75.4-91.5%], and NPV 93.9% for a cut-off value of 30.5 kPa). Our results confirm that 2D-SWE can accurately diagnose malignant thyroid nodules in cases with CAT (p < 0.0001), supporting its use as a complementary tool to conventional ultrasound.

Development of a clinical prediction model for benign and malignant pulmonary nodules with a CTR ≥ 50% utilizing artificial intelligence-driven radiomics analysis

ObjectiveIn clinical practice, diagnosing the benignity and malignancy of solid-component-predominant pulmonary nodules is challenging, especially when 3D consolidation-to-tumor ratio (CTR) ≥ 50%, as malignant ones are more invasive. This study aims to develop and validate an AI-driven radiomics prediction model for such nodules to enhance diagnostic accuracy.MethodsData of 2,591 pulmonary nodules from five medical centers (Zhengzhou People’s Hospital, etc.) were collected. Applying exclusion criteria, 370 nodules (78 benign, 292 malignant) with 3D CTR ≥ 50% were selected and randomly split 7:3 into training and validation cohorts. Using R programming, Lasso regression with 10-fold cross-validation filtered features, followed by univariate and multivariate logistic regression to construct the model. Its efficacy was evaluated by ROC, DCA curves and calibration plots.ResultsLasso regression picked 18 non-zero coefficients from 108 features. Three significant factors—patient age, solid component volume and mean CT value—were identified. The logistic regression equation was formulated. In the training set, the ROC AUC was 0.721 (95%CI: 0.642–0.801); in the validation set, AUC was 0.757 (95%CI: 0.632–0.881), showing the model’s stability and predictive ability.ConclusionThe model has moderate accuracy in differentiating benign from malignant 3D CTR ≥ 50% nodules, holding clinical potential. Future efforts could explore more to improve its precision and value.Clinical trial numberNot applicable.

Malignant Risk Assessment of Cystic-Solid Thyroid Nodules Based on Multimodal Ultrasound Features: A Systematic Review and Meta-Analysis

Background: The malignant risk of cystic-solid thyroid nodules may be underestimated in the ultrasound assessment. Objective: This systematic review and meta-analysis aimed to evaluate the value of multimodal ultrasound characteristics in the malignant risk assessment of cystic-solid thyroid nodules. Methods: We conducted a comprehensive search of PubMed, Web of Science, and Cochrane Library databases for studies depicting the ultrasound characteristics of cystic-solid thyroid nodules published prior to October 2023. The Review Manager 5.4 software was utilized to evaluate the ultrasound features suggestive of malignancy and to determine their sensitivity and specificity. Additionally, the Sata 12.0 software was utilized to construct summary receiver operating characteristic curves (SROC), estimate the area under the curve (AUC), and evaluate any potential publication bias. Results: This review included 16 studies comprising 5,655 cystic-solid thyroid nodules. Nine ultrasound features were identified as risk factors for tumor malignancy. Among the ultrasound features, microcalcification in the solid portion, heterogeneous hypoenhancement on Contrast-Enhanced Ultrasound (CEUS), and sharp angles in the solid portion exhibited higher malignant predictive value in cystic-solid thyroid nodules, with AUC values of 0.91, 0.84, and 0.81, respectively. Conclusion: Our findings indicate that microcalcification and sharp angles in the solid part of the nodule, along with heterogeneous hypoenhancement on contrast-enhanced ultrasound (CEUS), can better predict malignant cystic-solid thyroid nodules. The systematic review and meta-analysis was registered prospectively in the International Prospective Register of Systematic Reviews (No. CRD42024602893).

UMAM-NET: ultrasound thyroid nodule malignancy grading network based on multi-subnet attention mechanism

UMAM-NET: ultrasound thyroid nodule malignancy grading network based on multi-subnet attention mechanism

The value of non-punctate echogenic foci in the ultrasonic diagnosis of thyroid nodules.

To evaluate the diagnostic value of different subtypes of non-punctate echogenic foci in thyroid malignancy. Retrospective research of 342 thyroid nodules with calcification was performed. The echogenic foci were divided into punctate echogenic foci (type I) and non-punctate echogenic foci (type II), and type II were further divided into four subtypes: macrocalcification (type IIa), continuous peripheral calcification (type IIb), discontinuous peripheral calcification (type IIc) and isolated calcification (type IId). Postoperative histopathological results were used as the gold standard to evaluate the correlation between non-punctate echogenic foci subtypes and thyroid malignancy. The malignant risk of nodules with echogenic foci was type I (82.1%) > type IIa (66.2%) > type IIc (52.9%) > type IId (16.7%) > type IIb (13.9%), P < 0.001. Type I and type IIa echogenic foci were independent risk factors for thyroid cancer (OR = 16.593, 7.785). Solid, hypoechogenicity/marked hypoechogenicity and a single lesion in a unilateral thyroid lobe were independently associated with malignant thyroid nodules with macrocalcification(OR = 6.825, 40.042, 5.201). Irregular margins and uneven calcification thickness were independent factors for malignant thyroid nodules with peripheral calcification (OR = 5.676, 2.750). Type IIa echogenic foci could independently predict thyroid malignancy. The diagnostic value of non-punctate echogenic foci depended on the differentiated combination of ultrasound characteristics. Type IIa nodules with solid composition, irregular margins, and a single lesion in a unilateral thyroid lobe implied a higher risk of malignancy; peripheral calcified nodules with irregular margins and uneven calcification thickness suggested an increased risk of malignancy.

Discrimination of Benign and Malignant Thyroid Nodules through Comparative Analyses of Human Saliva Samples via Metabolomics and Deep-Learning-Guided Label-free SERS.

Thyroid nodules are a very common entity. The overall prevalence in the populace is estimated to be around 65-68%, among which a small portion (less than 5%) is malignant (cancerous). Therefore, it is important to discriminate benign thyroid nodules from malignant thyroid nodules. In this study, an equal number of participants with benign and malignant thyroid nodules (N = 10/group) were recruited. Saliva samples were collected from each participant, and SERS spectra were acquired, followed by validation using a metabolomics approach. An additional equal number of patients (N = 40/group) were recruited to construct diagnostic models. The performance of various machine learning (ML) algorithms was assessed using multiple evaluation metrics. Finally, the reliability of the optimal model was tested using blind test data (N = 10/group for benign and malignant thyroid nodules). The results showed a consistent trend between the SERS metabolic profile and the metabolites identified through MS analysis. The Multi-ResNet algorithm was optimal, achieving a 95% accuracy in sample discrimination. Additionally, blind test data sets yielded an overall accuracy of 83%. In summary, the deep-learning-guided SERS technique holds great potential in the accurate discrimination of benign and malignant thyroid nodules via human saliva samples, which facilitates the noninvasive diagnosis of malignant thyroid nodules in clinical settings.

RNA analysis of patients with benign and malignant pulmonary nodules.

Pulmonary nodules are the main manifestations of early lung cancer. Non-small cell lung cancer is the most common histological type of lung cancer, and the main histological classification of non-small cell lung cancer is lung adenocarcinoma. The present study aimed to analyze the differentially expressed genes between patients with benign and malignant pulmonary nodules, and to identify potential therapeutic targets for lung adenocarcinoma. Sequencing data for benign and malignant pulmonary nodule samples and samples with no nodules were obtained from the National Center for Biotechnology Information Gene Expression Omnibus GSE135304 dataset. Differential gene analysis showed that S100 calcium binding protein P (S100P), ribonuclease A family member 2 (RNASE2), cytochrome c oxidase subunit 7C and mast cell expressed membrane protein 1 (C19orf59) were significantly upregulated among the blood samples collected from patients with malignant pulmonary nodules. Results from Kaplan-Meier plotter datasets showed that S100P, RNASE2 and C19orf59 were associated with the prognosis of lung cancer. RNASE2 expression was positively associated with nodule size and negatively associated with lung cancer prognosis. Moreover, RNASE2 was highly expressed in lung adenocarcinoma tissues compared with that in normal tissues. CCK-8 and Transwell assays indicated that overexpressed RNASE2 promoted the proliferation, migration and invasion of lung adenocarcinoma cells. In lung adenocarcinoma cells, RNASE2 was identified as a downstream target of microRNA (miR)-185-5p and was regulated by it. Inhibited cell proliferation, migration and invasion were observed following overexpression of miR-185-5p. Overexpression of RNASE2 reversed the inhibitory effect of miR-185-5p overexpression. In conclusion, in blood samples from patients with malignant pulmonary nodules and lung adenocarcinoma tissues, RNASE2 was found to be upregulated. High RNASE2 expression was associated with poor overall survival. miR-185-5p inhibited the proliferation, migration and invasion of lung adenocarcinoma cells by downregulating RNASE2 expression. These findings have implications for guiding therapeutic strategies.

Application of Ultrasound Radiomics in Differentiating Benign from Malignant Breast Nodules in Women with Post-Silicone Breast Augmentation.

To evaluate the diagnostic value of ultrasound radiomics in distinguishing between benign and malignant breast nodules in women who have undergone silicone breast augmentation. A retrospective study was conducted of 99 breast nodules detected by ultrasound in 93 women who had undergone silicone breast augmentation. The ultrasound data were collected between 1 January 2006 and 1 September 2023. The nodules were allocated into a training set (n = 69) and a validation set (n = 30). Regions of interest (ROIs) were manually delineated using 3D Slicer software, and radiomic features were extracted and selected using Python programming. Eight machine learning algorithms were applied to build predictive models, and their performance was assessed using sensitivity, specificity, area under the ROC curve (AUC), accuracy, Brier score, and log loss. Model performance was further evaluated using ROC curves and calibration curves, while clinical utility was assessed via decision curve analysis (DCA). The random forest model exhibited superior performance in differentiating benign from malignant nodules in the validation set, achieving sensitivity of 0.765, specificity of 0.838, and an AUC of 0.787 (95% CI: 0.561-0.960). The model's accuracy, Brier score, and log loss were 0.796, 0.197, and 0.599, respectively. DCA suggested potential clinical utility of the model. Ultrasound radiomics demonstrates promising diagnostic accuracy in differentiating benign from malignant breast nodules in women with silicone breast prostheses. This approach has the potential to serve as an additional diagnostic tool for patients following silicone breast augmentation.

Differentiating Pulmonary Nodule Malignancy Using Exhaled Volatile Organic Compounds: A Prospective Observational Study.

Advances in imaging technology have enhanced the detection of pulmonary nodules. However, determining malignancy often requires invasive procedures or repeated radiation exposure, underscoring the need for safer, noninvasive diagnostic alternatives. Analyzing exhaled volatile organic compounds (VOCs) shows promise, yet its effectiveness in assessing the malignancy of pulmonary nodules remains underexplored. Employing a prospective study design from June 2023 to January 2024 at the Affiliated Hospital of Yangzhou University, we assessed the malignancy of pulmonary nodules using the Mayo Clinic model and collected exhaled breath samples alongside lifestyle and health examination data. We applied five machine learning (ML) algorithms to develop predictive models which were evaluated using area under the curve (AUC), sensitivity, specificity, and other relevant metrics. A total of 267 participants were enrolled, including 210 with low-risk and 57 with moderate-risk pulmonary nodules. Univariate analysis identified 11 exhaled VOCs associated with nodule malignancy, alongside two lifestyle factors (smoke index and sites of tobacco smoke inhalation) and one clinical metric (nodule diameter) as independent predictors for moderate-risk nodules. The logistic regression model integrating lifestyle and health data achieved an AUC of 0.91 (95% CI: 0.8611-0.9658), while the random forest model incorporating exhaled VOCs achieved an AUC of 0.99 (95% CI: 0.974-1.00). Calibration curves indicated strong concordance between predicted and observed risks. Decision curve analysis confirmed the net benefit of these models over traditional methods. A nomogram was developed to aid clinicians in assessing nodule malignancy based on VOCs, lifestyle, and health data. The integration of ML algorithms with exhaled biomarkers and clinical data provides a robust framework for noninvasive assessment of pulmonary nodules. These models offer a safer alternative to traditional methods and may enhance early detection and management of pulmonary nodules. Further validation through larger, multicenter studies is necessary to establish their generalizability. Number ChiCTR2400081283.

Dual biomarkers CT-based deep learning model incorporating intrathoracic fat for discriminating benign and malignant pulmonary nodules in multi-center cohorts.

Recent studies in the field of lung cancer have emphasized the important role of body composition, particularly fatty tissue, as a prognostic factor. However, there is still a lack of practice in combining fatty tissue to discriminate benign and malignant pulmonary nodules. This study proposes a deep learning (DL) approach to explore the potential predictive value of dual imaging markers, including intrathoracic fat (ITF), in patients with pulmonary nodules. We enrolled 1321 patients with pulmonary nodules from three centers. Image feature extraction was performed on computed tomography (CT) images of pulmonary nodules and ITF by DL, multimodal information was used to discriminate benign and malignant in patients with pulmonary nodules. Here, the areas under the receiver operating characteristic curve (AUC) of the model for ITF combined with pulmonary nodules were 0.910(95% confidence interval [CI]: 0.870-0.950, P=0.016), 0.922(95% CI: 0.883-0.960, P=0.037) and 0.899(95% CI: 0.849-0.949, P=0.033) in the internal test cohort, external test cohort1 and external test cohort2, respectively, which were significantly better than the model for pulmonary nodules. Intrathoracic fat index (ITFI) emerged as an independent influencing factor for benign and malignant in patients with pulmonary nodules, correlating with a 9.4% decrease in the risk of malignancy for each additional unit. This study demonstrates the potential auxiliary predictive value of ITF as a noninvasive imaging biomarker in assessing pulmonary nodules.

MSTD: A Multi-scale Transformer-based Method to Diagnose Benign and Malignant Lung Nodules

MSTD: A Multi-scale Transformer-based Method to Diagnose Benign and Malignant Lung Nodules

Addendum to: The effectiveness of deep learning model in differentiating benign and malignant pulmonary nodules on spiral CT.

Addendum to: The effectiveness of deep learning model in differentiating benign and malignant pulmonary nodules on spiral CT.

Insights in biomarkers complexity and routine clinical practice for the diagnosis of thyroid nodules and cancer.

The differential diagnosis between benign and malignant thyroid nodules continues to be a major challenge in clinical practice. The rising incidence of thyroid neoplasm and the low incidence of aggressive thyroid carcinoma, urges the exploration of strategies to improve the diagnostic accuracy in a pre-surgical phase, particularly for indeterminate nodules, and to prevent unnecessary surgeries. Only in 2022, the 5th WHO Classification of Endocrine and Neuroendocrine Tumors, and in 2023, the 3rd Bethesda System for Reporting Thyroid Cytopathology and the European Thyroid Association included biomarkers in their guidelines. In this review, we discuss the integration of biomarkers within the routine clinical practice for diagnosis of thyroid nodules and cancer. The literature search for this review was performed through Pub Med, Science Direct, and Google Scholar. We selected 156 publications with significant contributions to this topic, with the majority (86, or 55.1%) published between January 2019 and March 2024, including some publications from our group during those periods. The inclusion criteria were based on articles published in recognized scientific journals with high contributions to the proposed topic. We excluded articles not emphasizing molecular biomarkers in refine the pre-surgical diagnosis of thyroid nodules. We explored genetic biomarkers, considering the division of thyroid neoplasm into BRAF-like tumor and RAS-like tumor. The specificity of BRAF mutation in the diagnosis of papillary thyroid carcinoma (PTC) is nearly 100% but its sensitivity is below 35%. RAS mutations are found in a broad spectrum of thyroid neoplasm, from benign to malignant follicular-patterned tumors, but do not increase the ability to distinguish benign from malignant lesions. The overexpression of miRNAs is correlated with tumor aggressiveness, high tumor node metastasis (TMN) stage, and recurrence, representing a real signature of thyroid cancer, particularly PTC. In addition, associations between the expression levels of selected miRNAs and the presence of specific genetic mutations have been related with aggressiveness and worse prognosis. The knowledge of genetic and molecular biomarkers has achieved a high level of complexity, and the difficulties related to its applicability determine that their implementation in clinical practice is not yet a reality. More studies with larger series are needed to optimize their use in routine practice. Additionally, the improvement of new techniques, such as liquid biopsy and/or artificial intelligence, may be the future for a better understanding of molecular biomarkers in thyroid nodular disease.

Exploration of optimal thresholds for predicting the invasive nature of stage T1 lung adenocarcinoma using artificial intelligence-based 3D solid component volume segmentation.

The accuracy of intraoperative rapid frozen pathology is suboptimal, and the assessment of invasiveness in malignant pulmonary nodules significantly influences surgical resection strategies. Predicting the invasiveness of lung adenocarcinoma based on preoperative imaging is a clinical challenge, and there are no established standards for the optimal threshold value using the threshold segmentation method to predict the invasiveness of stage T1 lung adenocarcinoma. This study aimed to explore the efficacy of three-dimensional solid component volume (3D SCV) [calculated by artificial intelligence (AI) threshold segmentation method] in predicting the aggressiveness of T1 lung adenocarcinoma and to determine its optimal threshold and cut-off point. A retrospective case-control analysis was conducted on 1,179 confirmed T1 lung adenocarcinoma nodules from two centers. AI-based threshold segmentation was used to calculate seven sets of solid component volume data (V-550, V-450, V-350, V-250, V-150, V-50, V0) at seven computed tomography (CT) threshold values (-550, -450, -350, -250, -150, -50, 0 HU). The receiver operating characteristic (ROC) curve was employed to explore the optimal threshold value for predicting the invasiveness of T1 lung adenocarcinoma based on solid component volume. Subgroup analysis was performed according to the diameter of the pulmonary nodule, divided into T1a (≤10 mm), T1b (10 mm < T1b ≤20 mm), and T1c (20 mm < T1b ≤30 mm), to investigate the optimal threshold for each subgroup. The solid component volume was a stable predictive factor for the invasiveness of T1 lung adenocarcinoma. The optimal threshold value for predicting the invasiveness of T1 lung adenocarcinoma using AI-based 3D SCV segmentation was -350 HU, with an area under the curve (AUC) and 95% confidence interval (CI) of 0.930 (0.915-0.945), a cut-off value of 106.5 mm3, a sensitivity of 88.462%, and a specificity of 83.853%. For the T1a subgroup, the optimal threshold was -450 HU, with an AUC of 0.926, a cut-off value of 204.5 mm3, a sensitivity of 85.269%, and a specificity of 87.013%. For the T1b subgroup, the optimal threshold was -250 HU, with an AUC of 0.922, a cut-off value of 106 mm3, a sensitivity of 89.726%, and a specificity of 82.266%. For the T1c subgroup, the optimal threshold was -350 HU, with an AUC of 0.961, a cut-off value of 472 mm3, a sensitivity of 95.652%, and a specificity of 83.529%. The -350 HU threshold is reasonable for predicting the invasiveness of T1 lung adenocarcinoma based on solid component volume using the threshold segmentation method. However, there are certain differences in the threshold values depending on the diameter of the pulmonary nodule.

Low-Rank Adaptation of Pre-trained Large Vision Models for Improved Lung Nodule Malignancy Classification

Low-Rank Adaptation of Pre-trained Large Vision Models for Improved Lung Nodule Malignancy Classification

Circulating large extracellular vesicles as diagnostic biomarkers of indeterminate thyroid nodules: multi-platform omics analysis.

While most thyroid nodules are benign, 7-15% are malignant. Patients with indeterminate thyroid nodules (specifically Bethesda IV/Thy3f) often undergo diagnostic hemithyroidectomy to reach a diagnosis on final histology. The aim of this study was to assess the feasibility of circulating large extracellular vesicles as diagnostic biomarkers in patients presenting with Thy3f thyroid nodules. This was a two-gate diagnostic accuracy study; patients with Thy3f thyroid nodules were age, sex and body mass index matched to healthy individuals. Final histology confirmed benign and malignant diagnoses. Plasma large extracellular vesicle counts were quantified using flow cytometry. Large extracellular vesicle microRNA and protein profiles were identified using next generation sequencing and mass spectrometry, respectively. A total of 42 patients with Thy3f nodules (22 with cancer, 20 with non-cancer diagnosis) and 16 healthy controls were included. Total large extracellular vesicle concentrations and the concentrations of extracellular vesicles expressing epithelial cell adhesion molecule and the cancer markers atypical chemokine receptor type 7, extracellular matrix metalloproteinase inducer and syndecan-4 were significantly higher in patients with Thy3f nodules (cancer and non-cancer) compared with healthy individuals. In patients with cancerous versus non-cancer Thy3f nodules, one microRNA was upregulated: mir-195-3p (P < 0.001). Five were downregulated: mir-3176 (P < 0.001), mir-205-5p (P < 0.001), novel-hsa-mir-208-3p (P < 0.001), mir-3529-3p (P = 0.01) and let-7i-3p (P = 0.02). Furthermore, three large extracellular vesicle proteins (kallikrein-related peptidase11 (KLK11) (P = 0.001), alpha-1-acid glycoprotein 2 (A1AG2) (P <0.001) and small integral membrane protein 1 (SMIM1) (P = 0.04)) were significantly upregulated, while 20 large extracellular vesicle proteins were significantly downregulated (most downregulated: chemokine (C-X-C motif) ligand 7 (CXCL7), tubulin beta chain 1 (TBB1), binding immunoglobulin protein (BIP) and actinin alpha 1 (ACTN1) (P < 0.001)) in cancerous compared with non-cancer Thy3f nodules. Circulating large extracellular vesicle miRNA and protein profiles have a high diagnostic value to discriminate between benign and malignant nodules for patients with Thy3f cytology. Further validation for clinical performance will be needed.

Attention-based image segmentation and classification model for the preoperative risk stratification of thyroid nodules.

Despite widespread use of standardized classification systems, risk stratification of thyroid nodules is nuanced and often requires diagnostic surgery. Genomic sequencing is available for this dilemma however, costs and access restricts global applicability. Artificial intelligence (AI) has the potential to overcome this issue nevertheless, the need for black-box interpretability is pertinent. We aimed to create an ultrasonographic segmentation and classification model that offers explainability and risk accountability. Four hundred and fourteen ultrasonography images were collected from 105 patients undergoing thyroidectomy, divided into training and testing groups. Classification ground truth used is exclusively surgical histopathology. Relevant nodules were manually annotated by a dedicated study radiologist and surgeon. Three AI architectures with and without block attention modules were trained to identify the relevant nodule and the best performing was selected for the subsequent task in classifying identified nodules into benign or malignant. Gradient-Weighted Class Activation Map is used to provide saliency mapping for visual interpretability. Superior performance was recorded by the block attention model which stratified thyroid nodules into benign versus malignant with an accuracy of 93% versus 90%, F-score 90% versus 89%, sensitivity 93% versus 91% and specificity 92% versus 91% on a training dataset versus a testing dataset respectively. Visual interpretability maps demonstrate salient areas for a benign nodule diagnosis overlaps spongiform areas and malignant diagnosis salient areas overlap solid components of a partially cystic-solid nodule and microcalcifications within nodules. These findings are consistent with established diagnostic criteria for benign and malignant nodules. We developed an image segmentation and classification model for the risk stratification of thyroid nodules benchmarking surgical histopathology as ground truth and providing visual interpretability.