Search

Objectives: This study aimed to propose a deep learning (DL)–based framework for identifying the composition of thyroid nodules and assessing their malignancy risk. Methods: We conducted a retrospective multicenter study using ultrasound images from four hospitals. Convolutional neural network (CNN) models were constructed to classify ultrasound images of thyroid nodules into solid and non-solid, as well as benign and malignant. A total of 11,201 images of 6784 nodules were used for training, validation, and testing. The area under the receiver-operating characteristic curve (AUC) was employed as the primary evaluation index. Results: The models had AUCs higher than 0.91 in the benign and malignant grading of solid thyroid nodules, with the Inception-ResNet AUC being the highest at 0.94. In the test set, the best algorithm for identifying benign and malignant thyroid nodules had a sensitivity of 0.88, and a specificity of 0.86. In the human vs. DL test set, the best algorithm had a sensitivity of 0.93, and a specificity of 0.86. The Inception-ResNet model performed better than the senior physicians (p < 0.001). The sensitivity and specificity of the optimal model based on the external test set were 0.90 and 0.75, respectively. Conclusions: This research demonstrates that CNNs can assist thyroid nodule diagnosis and reduce the rate of unnecessary fine-needle aspiration (FNA). Clinical relevance statement: High-resolution ultrasound has led to increased detection of thyroid nodules. This results in unnecessary fine-needle aspiration and anxiety for patients whose nodules are benign. Deep learning can solve these problems to some extent. Key Points: • Thyroid solid nodules have a high probability of malignancy. • Our models can improve the differentiation between benign and malignant solid thyroid nodules. • The differential performance of one model was superior to that of senior radiologists. Applying this could reduce the rate of unnecessary fine-needle aspiration of solid thyroid nodules. [ABSTRACT FROM AUTHOR]

Objectives: To determine whether fructose-1,6-bisphosphatase 1 (FBP1) expression is associated with [18F]FDG PET uptake and postsurgical outcomes in patients with mesial temporal lobe epilepsy (mTLE) and to investigate whether the molecular mechanism involving gamma-aminobutyric acid type A receptor (GABAAR), glucose transporter-3 (GLUT-3), and hexokinase-II (HK-II). Methods: Forty-three patients with mTLE underwent [18F]FDG PET/CT. Patients were divided into Ia (Engel class Ia) and non-Ia (Engel class Ib–IV) groups according to more than 1 year of follow-up after surgery. The maximum standard uptake value (SUVmax) and asymmetry index (AI) of hippocampus were measured. The relationship among the SUVmax, AI, prognosis, and FBP1 expression was analyzed. A lithium-pilocarpine acute mTLE rat model was subjected to [18F]FDG micro-PET/CT. Hippocampal SUVmax and FBP1, GABAAR, GLUT-3, and HK-II expression were analyzed. Results: SUVmax was higher in the Ia group than in the non-Ia group (7.31 ± 0.97 vs. 6.56 ± 0.96, p < 0.05) and FBP1 expression was lower in the Ia group (0.24 ± 0.03 vs. 0.27 ± 0.03, p < 0.01). FBP1 expression was negatively associated with SUVmax and AI (p < 0.01). In mTLE rats, the hippocampal FBP1 increased (0.26 ± 0.00 vs. 0.17 ± 0.00, p < 0.0001), and SUVmax, GLUT-3 and GABAAR levels decreased significantly (0.73 ± 0.12 vs. 1.46 ± 0.23, 0.20 ± 0.01 vs. 0.32 ± 0.05, 0.26 ± 0.02 vs. 0.35 ± 0.02, p < 0.05); no significant difference in HK-II levels was observed. In mTLE patients and rats, FBP1 negatively correlated with SUVmax and GLUT-3 and GABAAR levels (p < 0.05). Conclusion: FBP1 expression was inversely associated with SUVmax in mTLE, which might inhibit [18F]FDG uptake by regulating GLUT-3 expression. High FBP1 expression was indicative of low GABAAR expression and poor prognosis. Key Points: • It is of paramount importance to explore the deep pathophysiological mechanisms underlying the pathogenesis of mesial temporal lobe epilepsy and find potential therapeutic targets. • [18F]FDG PET has demonstrated low metabolism in epileptic regions during the interictal period, and hypometabolism may be associated with prognosis, but the pathomechanism of this association remains uncertain. • Our results support the possibility that FBP1 might be simultaneously involved in the regulation of glucose metabolism levels and the excitability of neurons and suggest that targeting FBP1 may be a viable strategy in the diagnosis and treatment of mesial temporal lobe epilepsy. [ABSTRACT FROM AUTHOR]

Objective: We used radiomics feature-based machine learning classifiers of apparent diffusion coefficient (ADC) maps to differentiate small round cell malignant tumors (SRCMTs) and non-SRCMTs of the nasal and paranasal sinuses.Materials: A total of 267 features were extracted from each region of interest (ROI). Datasets were randomized into two sets, a training set (∼70%) and a test set (∼30%). We performed dimensional reductions using the Pearson correlation coefficient and feature selection analyses (analysis of variance [ANOVA], relief, recursive feature elimination [RFE]) and classifications using 10 machine learning classifiers. Results were evaluated with a leave-one-out cross-validation analysis.Results: We compared the AUC for all the pipelines in the validation dataset using FeAture Explorer (FAE) software. The pipeline using RFE feature selection and Gaussian process classifier yielded the highest AUCs with ten features. When the "one-standard error" rule was used, FAE produced a simpler model with eight features, including Perc.01%, Perc.10%, Perc.90%, Perc.99%, S(1,0) SumAverg, S(5,5) AngScMom, S(5,5) Correlat, and WavEnLH_s-2. The AUCs of the training, validation, and test datasets achieved 0.995, 0.902, and 0.710, respectively. For ANOVA, the pipeline with the auto-encoder classifier yielded the highest AUC using only one feature, Perc.10% (training/validation/test datasets: 0.886/0.895/0.809, respectively). For the relief, the AUCs of the training, validation, and test datasets that used the LRLasso classifier using five features (Perc.01%, Perc.10%, S(4,4) Correlat, S(5,0) SumAverg, S(5,0) Contrast) were 0.892, 0.886, and 0.787, respectively. Compared with the RFE and relief, the results of all algorithms of ANOVA feature selection were more stable with the AUC values higher than 0.800.Conclusions: We demonstrated the feasibility of combining artificial intelligence with the radiomics from ADC values in the differential diagnosis of SRCMTs and non-SRCMTs and the potential of this non-invasive approach for clinical applications.Key Points: • The parameter with the best diagnostic performance in differentiating SRCMTs from non-SRCMTs was the Perc.10% ADC value. • Results of all the algorithms of ANOVA feature selection were more stable and the AUCs were higher than 0.800, as compared with RFE and relief. • The pipeline using RFE feature selection and Gaussian process classifier yielded the highest AUC. [ABSTRACT FROM AUTHOR]

Objectives: Stratification of microsatellite instability (MSI) status in patients with colorectal cancer (CRC) improves clinical decision-making for cancer treatment. The present study aimed to develop a radiomics nomogram to predict the pre-treatment MSI status in patients with CRC. Methods: A total of 762 patients with CRC confirmed by surgical pathology and MSI status determined with polymerase chain reaction (PCR) method were retrospectively recruited between January 2013 and May 2019. Radiomics features were extracted from routine pre-treatment abdominal pelvic computed tomography (CT) scans acquired as part of the patients' clinical care. A radiomics nomogram was constructed using multivariate logistic regression. The performance of the nomogram was evaluated using discrimination, calibration, and decision curves. Results: The radiomics nomogram incorporating radiomics signatures, tumor location, patient age, high-density lipoprotein expression, and platelet counts showed good discrimination between patients with non-MSI-H and MSI-H, with an area under the curve (AUC) of 0.74 [95% CI, 0.68–0.80] in the training cohort and 0.77 [95% CI, 0.68–0.85] in the validation cohort. Favorable clinical application was observed using decision curve analysis. The addition of pathological characteristics to the nomogram failed to show incremental prognostic value. Conclusions: We developed a radiomics nomogram incorporating radiomics signatures and clinical indicators, which could potentially be used to facilitate the individualized prediction of MSI status in patients with CRC. Key Points: • There is an unmet need to non-invasively determine MSI status prior to treatment. However, the traditional radiological evaluation of CT is limited for evaluating MSI status. • Our non-invasive CT imaging-based radiomics method could efficiently distinguish patients with high MSI disease from those with low MSI disease. • Our radiomics approach demonstrated promising diagnostic efficiency for MSI status, similar to the commonly used IHC method. [ABSTRACT FROM AUTHOR]

Objectives: To evaluate retrospectively the feasibility and safety of simultaneous multiple microcoil localizations of multiple pulmonary nodules prior to video-assisted thoracoscopic surgery (VATS). Methods: This retrospective cohort study enrolled 288 consecutive patients, who underwent computed tomography (CT)–guided microcoil localization and subsequent VATS at our academic hospital between July 2017 and June 2018. Of these patients, 36 with 79 pulmonary nodules undergoing simultaneous multiple microcoil localizations in the ipsilateral lung were designated the multiple localization group; the remaining 252 with 252 pulmonary nodules undergoing single microcoil localization were designated the single localization group. The main outcomes were the technical success and complication rates of the localization procedures. The Student t test and Mann-Whitney U test were used for continuous variables. The chi-squared test and logistic regression analysis were used to assess dichotomous variables. Results: The localization technical success rates of the multiple and single localization groups were 96.2% (76/79) and 98.0% (247/252), respectively (p = 0.326). The rate of any complication (pneumothorax or pulmonary hemorrhage) was significantly higher in the multiple localization than in the single localization group (55.6% vs 21.8%, respectively; p < 0.001). The incidence of pneumothorax was significantly higher in the multiple localization than in the single localization group (p < 0.001). The difference between the incidence of pulmonary hemorrhage in the 2 groups was not significant (p = 0.385). Conclusions: Although preoperative CT-guided simultaneous microcoil localizations of multiple pulmonary nodules produced a significantly higher incidence of pneumothorax, the localizations were clinically feasible and safe. Key Points: • Simultaneous preoperative CT-guided microcoil localizations of multiple pulmonary nodules are clinically feasible and safe. • Simultaneous microcoil localizations of multiple pulmonary nodules produced a significantly higher incidence of pneumothorax. [ABSTRACT FROM AUTHOR]

Objective: To compare the classification based on contrast-enhanced ultrasound (CEUS) Liver Imaging Reporting and Data System (LI-RADS) with that of contrast-enhanced CT and MRI (CECT/MRI) LI-RADS for liver nodules in patients at high risk of hepatocellular carcinoma.Methods: Two hundred thirty-nine patients with 273 nodules were enrolled in this retrospective study. Each nodule was categorized according to the CEUS LI-RADS version 2017 and CECT/MRI LI-RADS version 2017. The diagnostic performance of CEUS and CECT/MRI was compared. The reference standard was histopathology diagnosis. Inter-modality agreement was assessed with Cohen's kappa.Results: The inter-modality agreement for CEUS LI-RADS and CECT/MRI LI-RADS was fair with a kappa value of 0.319 (p < 0.001). The positive predictive values (PPVs) of hepatocellular carcinoma (HCC) in LR-5, LR-4, and LR-3 were 98.3%, 60.0%, and 25.0% in CEUS, and 95.9%, 65.7%, and 48.1% in CECT/MRI, respectively. The sensitivities and specificities of LR-5 for diagnosing HCC were 75.6% and 93.8% in CEUS, and 83.6% and 83.3% in CECT/MRI, respectively. The positive predictive values of non-HCC malignancy in CEUS LR-M and CECT/MRI LR-M were 33.9% and 93.3%, respectively. The sensitivity, specificity, and accuracy for diagnosing non-HCC malignancy were 90.9%, 84.5%, and 85.0% in CEUS LR-M and 63.6%, 99.6%, and 96.7% in CECT/MRI LR-M, respectively.Conclusions: The inter-modality agreement of the LI-RADS category between CEUS and CECT/MRI is fair. The positive predictive values of HCCs in LR-5 of the CEUS and CECT/MRI LI-RADS are comparable. CECT/MRI LR-M has better diagnostic performance for non-HCC malignancy than CEUS LR-M.Key Points: • The inter-modality agreement for the final LI-RADS category between CEUS and CECT/MRI is fair. • The LR-5 of CEUS and CECT/MRI LI-RADS corresponds to comparable positive predictive values (PPVs) of HCC. For LR-3 and LR-4 nodules categorized by CECT/MRI, CEUS examination should be performed, at least if they can be detected on plain ultrasound. • CECT/MRI LR-M has better diagnostic performance for non-HCC malignancy than CEUS LR-M. For LR-M nodules categorized by CEUS, re-evaluation by CECT/MRI is necessary. [ABSTRACT FROM AUTHOR]

Objectives: To compare the efficacy and safety of localization of small pulmonary nodules with microcoil and hookwire prior to surgical resection.Methods: A total of 112 patients who underwent preoperative computed tomography (CT)-guided localization of small pulmonary nodules were enrolled in this single-center retrospective non-randomized cohort study between June 2016 and June 2017. Seventy-nine patients who underwent percutaneous localization with microcoils formed the microcoil group; the remaining 33 patients underwent percutaneous localization with hookwires (hookwire group). The primary outcomes were the success and complication rates of the procedure. Student's t test was used for continuous variables, whereas chi-square analysis and logistic regression were used for dichotomous variables.Results: Video-assisted thoracoscopic surgery (VATS) was successfully performed in all cases, without conversion to thoracotomy. The localization success rate was 94.9% (75/79) in the microcoil group and 93.9% (31/33) in the hookwire group (p = 0.836). Hookwire group (p = 0.000) and nodule location of the lower lobe (p = 0.012) were associated with an increased incidence of pneumothorax. Hookwire group (p = 0.027) and decreased nodule diameter (p = 0.024) were associated with an increased incidence of moderate to severe chest pain, as well as an increased incidence of overall complications.Conclusions: Although the deployment of the microcoil was more complex and required more time than hookwire placement, microcoil localization was associated with fewer complications.Key Points: • CT-guided percutaneous localization using a microcoil and that using a hookwire are equally effective for localizing small pulmonary nodules prior to resection with video-assisted thoracoscopic surgery. • Lung nodule localization using a microcoil was associated with fewer complications than localization using a hookwire. [ABSTRACT FROM AUTHOR]