Internal Dataset Research Articles

Constructing a 22-year internal wave dataset for the northern South China Sea: spatiotemporal analysis using MODIS imagery and deep learning

Abstract. Internal waves (IWs) are an important ocean phenomenon facilitating energy transfer between multiscale ocean processes. Understanding such processes necessitates the collection and analysis of extensive observational data. IWs predominantly occur in marginal seas, with the South China Sea (SCS) being one of the most active regions, characterized by frequent and large-amplitude IW activities. In this study, we present a comprehensive IW dataset for the northern SCS (https://doi.org/10.12157/IOCAS.20240409.001, Zhang and Li, 2024), covering the area from 112.40 to 121.32° E and from 18.32 to 23.19° N, spanning the period from 2000 to 2022 with a 250 m spatial resolution. During the 22 years, a total of 15 830 MODIS images were downloaded for further processing. Out of these, 3085 high-resolution MODIS true-color images were identified to contain IW information and were included in the dataset with precise IW positions extracted using advanced deep learning techniques. IWs in the northern SCS are categorized into four regions based on extracted IW spatial distributions. This classification enables detailed analyses of IW characteristics, including their spatial and temporal distributions across the entire northern SCS and its specific sub-regions. Interestingly, our temporal analysis reveals characteristic “double-peak” patterns aligned with the lunar day, highlighting the strong connection between IWs and tidal cycles. Furthermore, our spatial analysis identifies two IW quiescent zones within the IW clusters influenced by underwater topography, highlighting regional variations in IW characteristics and suggesting underlying mechanisms which merit further investigation. There are also three gap regions between distinct IW clusters, which may indicate different IW sources. The constructed dataset holds significant potential for studying IW–environment interactions, developing monitoring and prediction models, validating numerical simulations, and serving as an educational resource to promote awareness and interest in IW research.

Development of a Dual-Plane MRI-Based Deep Learning Model to Assess the 1-Year Postoperative Outcomes in Lumbar Disc Herniation After Tubular Microdiscectomy.

Tubular microdiscectomy (TMD) is a treatment for lumbar disc herniation (LDH). Although the combination of MRI and deep learning (DL) has shown promise, its application in evaluating postoperative outcomes in TMD has not been fully explored. To evaluate whether integrating preoperative dual-plane MRI-based DL features with clinical features can assess 1-year outcomes in TMD for LDH. Retrospective. The study involved 548 patients who underwent TMD between January 2016 and January 2021. Training set (N = 305, mean age 51.85 ± 13.84 years, 56.4% male). Internal validation set (N = 131, mean age 51.85 ± 13.84 years, 54.2% male). External validation set (N = 112, mean age 51.54 ± 14.43 years, 50.9% male). 3 T MRI with sagittal and transverse T2-weighted sequences (Fast Spin Echo). Ground truth labels were based on improvement rate in 1-year Japanese Orthopaedic Association (JOA) scores. Information on 42 preoperative clinical features was collected. The largest protrusions were identified from T2 MRI by three clinicians and were used to train deep learning models (ResNet50, ResNet101, and ResNet152) to extract DL features. After feature selection, three models were built, namely, clinical, DL, and combined models. Chi-square or Fisher's exact tests was used for group comparisons. Quantitative differences were analyzed using the t-test or Mann-Whitney U test. P-values <0.05 were considered significant. Models were validated on internal and external datasets using metrics such as the area under the curve (AUC). The AUCs of the clinical models achieved 0.806 (internal) and 0.779 (external). ResNet152 performed best in three DL models, with AUCs of 0.858 (internal) and 0.834 (external). The combined model achieved AUCs of 0.889 (internal) and 0.857 (external). A model combining preoperative dual-plane MRI DL features and clinical features can assess 1-year outcomes of TMD for LDH. 4 TECHNICAL EFFICACY: Stage 2.

Artificial intelligence-aided colonoscopic differential diagnosis between Crohn's disease and gastrointestinal tuberculosis.

Differentiating between Crohn's disease (CD) and gastrointestinal tuberculosis (GITB) is challenging. We aimed to evaluate the clinical applicability of an artificial intelligence (AI) model for this purpose. The AI model was developed and assessed using an internal dataset comprising 1,132 colonoscopy images of CD and 1,045 colonoscopy images of GITB at a tertiary referral center. Its stand-alone performance was further evaluated in an external dataset comprising 67 colonoscopy images of 17 CD patients and 63 colonoscopy images of 14 GITB patients from other institutions. Additionally, a crossover trial involving three expert endoscopists and three trainee endoscopists compared AI-assisted and unassisted human interpretations. In the internal dataset, the sensitivity, specificity, and accuracy of the AI model in distinguishing between CD and GITB were 95.3%, 100.0%, and 97.7%, respectively, with an area under the ROC curve of 0.997. In the external dataset, the AI model exhibited a sensitivity, specificity, and accuracy of 77.8%, 85.1%, and 81.5%, respectively, with an area under the ROC curve of 0.877. In the human endoscopist trial, AI assistance increased the pooled accuracy of the six endoscopists from 86.2% to 88.8% (P=0.010). While AI did not significantly enhance diagnostic accuracy for the experts (96.7% with AI vs 95.6% without, P=0.360), it significantly improved accuracy for the trainees (81.0% vs 76.7%, P=0.002). This AI model shows potential in aiding the accurate differential diagnosis between CD and GITB, particularly benefiting less experienced endoscopists.

Quantitative summary on the human pharmacokinetic properties of cannabidiol to accelerate scientific clinical application of cannabis.

Cannabidiol (CBD) is a non-psychoactive substance that exerts numerous pharmacological benefits, including anti-inflammatory and antioxidant properties. It has received attention as a useful substance for the treatment of intractable pain, seizures, and anxiety, and related clinical trials have continued. However, the CBD pharmacokinetic results between reports are highly variable, making it difficult to clearly identify the pharmacokinetic properties of CBD. The main purpose of this study was to identify CBD clinical pharmacokinetic properties through meta-analysis. In particular, we sought to derive valid, interpretable independent variables and interpret their pharmacokinetic parameter correlations in relation to the large inter-individual and inter-study variability in CBD pharmacokinetics. For this study, CBD-related clinical trial reports were extensively screened and intercomparisons were performed between internal data sets through systematic classification and extraction of pharmacokinetic parameter values. The candidate independent variables associated with interpretation of CBD pharmacokinetic diversity established and explored in this study were as follows: diet, tetrahydrocannabinol (THC) combination, sample matrix type, liver and renal function, exposure route, dosage form, CBD exposure dose, cannabis smoking frequency, multiple exposure. The results of this study showed that CBD pharmacokinetics were influenced (increased plasma exposure by approximately 2-5 times) by diet immediately before or during CBD exposure, and that THC was not expected to have an antagonistic effect on the CBD absorption. The influence of changes in liver function would be significant in CBD pharmacokinetic diversity. Due to decreased liver function, the plasma exposure of CBD increased 2.57-5.15 times compared to healthy adults, and the half-life and clearance showed a 2.58-fold increase and a 5.15-fold decrease, respectively. CBD can be rapidly absorbed into the body (time to reach maximum concentration within 3.18h) by oral, transdermal, and inhalation exposures, and lipid emulsification and nanoformulation of CBD will greatly improve CBD bioavailability (up to approximately 2 times). The pharmacokinetics of CBD generally follow linear kinetic characteristics. The importance of this study is that it suggests key factors that should be considered in terms of pharmacokinetics in further clinical trials and formulations of CBD in the future.

A New Intraocular Lens Power Formula Integrating an Artificial Intelligence-Powered Estimation for Effective Lens Position Based on Chinese Eyes.

To develop and evaluate a new intraocular lens (IOL) formula based on Chinese eyes. A training dataset of 709 eyes undergoing uneventful cataract surgery was used to train the algorithm for effective lens position estimation. The algorithm was then integrated with Gaussian optics to develop the new IOL formula (Jin-AI). From the same center, 177 eyes served as an internal test dataset. An independent dataset of 557 eyes served as an external test dataset. The standard deviation (SD) of prediction errors was compared among the Jin-AI formula, traditional third-generation formulas (SRK/T, Holladay 1, Hoffer Q), and newer generation formulas (Kane, Barrett Universal II [BUII], Hill-radial basis function [RBF] 3.0, and PEARL-DGS). In the internal test dataset, the Jin-AI formula showed the lowest SD (0.25 D), followed by the BUII (0.31 D), Kane (0.33 D), and PEARL-DGS (0.33 D) formulas. In the external test dataset, the Jin-AI, Kane, and PEARL-DGS formulas had the lowest SD (0.38 D), followed by the BUII (0.39 D), Hill-RBF 3.0 (0.39 D), SRK/T (0.45 D), Holladay 1 (0.48 D), and Hoffer Q (0.48 D) formulas. The SD of the Jin-AI formula was significantly lower than the third-generation formulas and comparable to the four newer generation formulas. Predictive accuracy of the Jin-AI formula was similar to the newer generation formulas across all axial length, keratometry, and anterior chamber depth ranges. The new formula has exhibited good performance in predicting postoperative refractions. Its overall predictive accuracy was better than the third-generation formulas and comparable to the newer generation ones. The Jin-AI formula could be a reliable alternative for IOL power calculation in Chinese.

Deformable dose prediction network based on hybrid 2D and 3D convolution for nasopharyngeal carcinoma radiotherapy.

Radiotherapy is recognized as the primary treatment for nasopharyngeal carcinoma (NPC). Rapid and accurate dose prediction is crucial for enhancing the quality and efficiency of radiotherapy planning. However, the current dose prediction model based on 2D architecture cannot effectively learn the spatial information among slices. Although some studies have explored the incorporation of interslice features through 3D architecture, the resolution properties of medical image anisotropy significantly limit the predictive performance. To address the issues, we propose a novel deformable dose prediction network based on hybrid 2D and 3D convolution for NPC radiotherapy. Specifically, the proposed model innovatively incorporates a 2.5D architecture based on hybrid 2D and 3D convolution, and effectively utilizes the directional information within anisotropic resolutions to achieve cross-scale feature extraction. Additionally, deformable convolution is introduced into the model to enhance the receptive field and effectively handle multi-scale spatial transformations. To improve channel correlation and reduce redundant features, we design a Residual Deformable Squeeze-and-Excitation Module. We conduct extensive experiments on an internal dataset, and the results show that the proposed model outperforms other existing methods in most dosimetric criteria. The proposed model has superior dose prediction performance in NPC radiotherapy, and has important clinical significance for assisting physicists to optimize the treatment plan and improve standardization of radiotherapy planning. The source code is available at https://github.com/CDUTJ102/2.5D-Deformable-UNet .

Machine learning and single-cell RNA sequencing reveal relationship between intratumor CD8+ T cells and uveal melanoma metastasis

PurposeUveal melanoma (UM) is adults’ most common primary intraocular malignant tumor. It has been observed that 40% of patients experience distant metastasis during subsequent treatment. While there exist multigene models developed using machine learning methods to assess metastasis and prognosis, the immune microenvironment’s specific mechanisms influencing the tumor microenvironment have not been clarified. Single-cell transcriptome sequencing can accurately identify different types of cells in a tissue for precise analysis. This study aims to develop a model with fewer genes to evaluate metastasis risk in UM patients and provide a theoretical basis for UM immunotherapy.MethodsRNA-seq data and clinical information from 79 μm patients from TCGA were used to construct prognostic models. Mechanisms were probed using two single-cell datasets derived from the GEO database. After screening for metastasis-related genes, enrichment analysis was performed using GO and KEGG. Prognostic genes were screened using log-rank test and one-way Cox regression, and prognostic models were established using LASSO regression analysis and multifactor Cox regression analysis. The TCGA-UVM dataset was used as internal validation and dataset GSE22138 as external validation data. A time-dependent subject work characteristic curve (time-ROC) was established to assess the predictive ability of the model. Subsequently, dimensionality reduction, clustering, pseudo-temporal analysis and cellular communication analysis were performed on GSE138665 and GSE139829 to explore the underlying mechanisms involved. Cellular experiments were also used to validate the relevant findings.ResultsBased on clinical characteristics and RNA-seq transcriptomic data from 79 samples in the TCGA-UVM cohort, 247 metastasis-related genes were identified. Survival models for three genes (SLC25A38, EDNRB, and LURAP1) were then constructed using lasso regression and multifactorial cox regression. Kaplan-Meier survival analysis showed that the high-risk group was associated with poorer overall survival (OS) and metastasis-free survival (MFS) in UM patients. Time-dependent ROC curves demonstrated high predictive performance in 6 m, 18 m, and 30 m prognostic models. Cell scratch assay showed that the 24 h and 48 h migration rates of cells with reduced expression of the three genes were significantly higher than those of the si-NC group. CD8 + T cells may play an important role in tumour metastasis as revealed by immune infiltration analysis. An increase in the percentage of cytotoxic CD8 + T cells in the metastatic high-risk group was found in the exploration of single-cell transcriptome data. The communication intensity of cytotoxic CD8 was significantly enhanced. It was also found that the CD8 + T cells in the two groups were in different states, although the number of CD8 + T cells in the high-risk group increased, they were mostly in the exhausted and undifferentiated state, while in the low-risk group, the CD8 + T cells were mostly in the functional state.ConclusionsWe developed a precise and stable 3-gene model to predict the metastatic risk and prognosis of patients. CD8 + T cells exhaustion in the tumor microenvironment play a crucial role in UM metastasis.

Left atrial enlargement: a crucial indicator for identifying atrial fibrillation in patients with hypertension

Abstract Background Left atrial enlargement resulting from hypertension is closely linked to the development and persistence of atrial fibrillation (AF). The newly proposed staging recognizes AF as disease continuum, which makes us aware that AF prevention should focus on the Pre-AF stage, and atrial enlargement is one of the important manifestations in this stage. Previous scoring systems, such as CHA2DS2-VASc and C2HEST, along with the recently highlighted left atrial diameter (LAD), have been significant tools for predicting AF occurrence. However, a comprehensive assessment of their utility is currently lacking. Purpose This study aims to explore the role of left atrial size in identifying atrial fibrillation (AF) among hospitalized hypertensives, and to compare its recognition effectiveness with previous scoring systems. Methods We conducted a cross-sectional analysis within hospitalized hypertensives. The discovery, internal and external validation datasets were established. The XGBoost was employed to identify key variables related to AF occurrence, which were ranked based on their importance scores. To gauge the predictive prowess of LAD regarding AF occurrence, we plotted the receiver operating characteristic curve (ROC) and calculated the area under the curve (AUC). This enabled us to pinpoint the LAD cutoff value corresponding to the maximum Youden index, indicative of susceptibility to AF. Subsequently, Youden index determined the optimal cutoff value from the ROC curve. Delong’s test compared the identification abilities of different tools within the same dataset. Logistic regression analysis assessed the correlation between clinical variables and left atrial size. Results The XGBoost revealed 17 variables affecting AF occurrence in hypertensives, with LAD, age, and HF ranking among the top 3 (Figure 2a). LAD had a more pronounced impact on AF occurrence than other variables. LAD-based identification of AF in hypertensives had AUC of 0.827 (95%CI: 0.816-0.837), with an optimal cutoff of 38mm. Subgrouping the study population by LAD≥38mm, ROC curves showed AUCs of 0.757 (95%CI: 0.745-0.768) for the discovery set, 0.750 (95%CI: 0.731-0.768) for internal validation, and 0.759 (95%CI: 0.743-0.776) for external validation. LAD≥38mm had a more higher and stable AUC value for judging the AF occurrence than CHA2DS2-VASc and C2HEST scores (Figure 2b). Hypertensives with LAD≥38mm had a 9.72-fold higher AF risk than those with LAD&amp;lt;38mm. Conclusion LAD is a pivot determinant for AF occurrence. LAD≥38mm is a valuable identifier for high-risk hypertensives prone to AF.Figure 1.Center illustrationFigure 2.XGboost and Delong’s test of ROC

An artificial intelligence-enabled electrocardiogram model outperforms the revised cardiac risk index score for predicting major adverse cardiovascular events in patients undergoing noncardiac surgery

Abstract Background Existing clinical preoperative risk assessment tools such as the Revised Cardiac Risk Index (RCRI) have been used extensively to predict perioperative major adverse cardiovascular events (MACE). Electrocardiogram (ECG) analysis using Deep Learning can identify hidden features that human eyes cannot identify and might help further risk stratify patients undergoing noncardiac surgery. Objective We aimed to develop a Deep Learning model that predicts perioperative MACE in patients undergoing elective noncardiac surgeries. Methods We included patients ≥ 18-year-old who underwent elective non-cardiac surgery between 2000-2021. MACE was validated in duplicate and were defined as a composite of in-hospital myocardial infarction, cardiac arrest or mortality. A convolutional neural network was developed using ECGs within 30 days before surgery. We randomly split subjects into training, internal validation, and testing datasets in a 7:1:2 ratio. Performance was evaluated using the area under the receiver operator characteristics curve (AUC) values on the testing dataset. Models trained included AI-ECG models to predict in-hospital MACE and to predict in-hospital mortality, and AI-ECG+RCRI models to predict in-hospital MACE and to predict in-hospital mortality. All models were validated using 30-day validated outcomes (MACE or all-cause mortality) in a community-based cohort and were compared with the Revised Cardiac Risk Index (RCRI) score using AUC values. Findings: We included 195,214 patients who underwent a total of 241,999 surgeries. 169,214 ECGs were used in the training dataset, 24,234 in the internal validation dataset, and 48,551 in the testing dataset. In the test cohort, the AI-ECG algorithm discriminated risk for in-hospital MACE with an AUC of 0.82 (95% CI 0.79-0.85). The algorithm similarly discriminated 30-day MACE with an AUC value of 0.79 (95% CI 0.77-0.81) surpassing the discrimination of the RCRI score AUC of 0.69 (95% CI 0.66-0.72). All models’ performances are shown in Figure1A. The AI-ECG model to predict in-hospital mortality had an AUC of 0.83 (95% CI 0.80-0.86) and predicted 30-day mortality with an AUC of 0.80 (95% CI 0.77-0.82). Other models’ performances are shown Figure1B. Training the AI-ECG models including RCRI data did not improve the AI-ECG models’ performance. Conclusion AI-ECG models with no additional clinical data can improve perioperative risk prediction in non-cardiac surgery, outperforming a conventional and widely used risk stratification tool such as the RCRI.

Quantitative texture analysis using machine learning for predicting interpretable pulmonary perfusion from non-contrast computed tomography in pulmonary embolism patients

BackgroundPulmonary embolism (PE) is life-threatening and requires timely and accurate diagnosis, yet current imaging methods, like computed tomography pulmonary angiography, present limitations, particularly for patients with contraindications to iodinated contrast agents. We aimed to develop a quantitative texture analysis pipeline using machine learning (ML) based on non-contrast thoracic computed tomography (CT) scans to discover intensity and textural features correlated with regional lung perfusion (Q) physiology and pathology and synthesize voxel-wise Q surrogates to assist in PE diagnosis.MethodsWe retrospectively collected 99mTc-labeled macroaggregated albumin Q-SPECT/CT scans from patients suspected of PE, including an internal dataset of 76 patients (64 for training, 12 for testing) and an external testing dataset of 49 patients. Quantitative CT features were extracted from segmented lung subregions and underwent a two-stage feature selection pipeline. The prior-knowledge-driven preselection stage screened for robust and non-redundant perfusion-correlated features, while the data-driven selection stage further filtered features by fitting ML models for classification. The final classification model, trained with the highest-performing PE-associated feature combination, was evaluated in the testing cohorts based on the Area Under the Curve (AUC) for subregion-level predictability. The voxel-wise Q surrogate was then synthesized using the final selected feature maps (FMs) and model score maps (MSMs) to investigate spatial distributions. The Spearman correlation coefficient (SCC) and Dice similarity coefficient (DSC) were used to assess the spatial consistency between FMs or MSMs and Q-SPECT scans.ResultsThe optimal model performance achieved an AUC of 0.863 during internal testing and 0.828 on the external testing cohort. The model identified a combination containing 14 intensity and textural features that were non-redundant, robust, and capable of distinguishing between high- and low-functional lung regions. Spatial consistency assessment in the internal testing cohort showed moderate-to-high agreement between MSMs and reference Q-SPECT scans, with median SCC of 0.66, median DSCs of 0.86 and 0.64 for high- and low-functional regions, respectively.ConclusionsThis study validated the feasibility of using quantitative texture analysis and a data-driven ML pipeline to generate voxel-wise lung perfusion surrogates, providing a radiation-free, widely accessible alternative to functional lung imaging in managing pulmonary vascular diseases.Clinical trial numberNot applicable.

A novel endoscopic artificial intelligence system to assist in the diagnosis of autoimmune gastritis: a multicenter study.

Autoimmune gastritis (AIG), distinct from Helicobacter pylori-associated atrophic gastritis (HpAG), is underdiagnosed due to limited awareness. This multicenter study aims to develop a novel endoscopic artificial intelligence (AI) system assisting in AIG diagnosis. Patients diagnosed with AIG, as well as HpAG and non-atrophic gastritis (NAG), were retrospectively enrolled from six centers. Endoscopic images with relevant demographic and medical data, were collected for the development of AI-assisted system, SEER-SCOPE AI, based on multi-site feature fusion model. The diagnostic performance of SEER-SCOPE AI was evaluated in the internal and external datasets. Endoscopists' performance with and without AI support was tested and compared using Mann-Whitney U test. Heatmap analysis was performed to interpret SEER-SCOPE AI. 1 070 patients (294 AIG, 386 HpAG, 390 NAG) with 18 828 endoscopy images were collected. SEER-SCOPE AI achieved strong performance for identifying AIG, with 96.9% sensitivity, 92.2% specificity and an AUROC of 0.990 internally, and 90.3% sensitivity, 93.1% specificity and an AUROC of 0.973 externally. The performance of SEER-SCOPE AI (sensitivity 91.3%) was comparable to experts (87.3%) and significantly outperformed non-experts (70.0%). With AI support, the overall performance of endoscopists was improved (sensitivity: 90.3% [95% CI 86.0%-93.2%] vs. 78.7% [95% CI 73.6%-83.2%], p=0.008). Heatmap analysis revealed consistent focus of SEER-SCOPE AI on regions corresponding to atrophic areas. SEER-SCOPE AI demonstrated expert-level performance in identifying AIG, and enhanced the diagnostic ability of endoscopists. Its application holds promise as a potent endoscopy-assisted tool for guiding biopsy sampling and early detection of AIG.

Automated quantification of cerebral microbleeds in susceptibility-weighted MRI: association with vascular risk factors, white matter hyperintensity burden, and cognitive function.

To train and validate a deep learning (DL)-based segmentation model for cerebral microbleeds (CMB) on susceptibility-weighted MRI; and to find associations between CMB, cognitive impairment, and vascular risk factors. Participants in this single-institution retrospective study underwent brain MRI to evaluate cognitive impairment between January-September 2023. For training the DL model, the nnU-Net framework was used without modifications. The DL model's performance was evaluated on independent internal and external validation datasets. Linear regression analysis was used to find associations between log-transformed CMB numbers, cognitive function (mini-mental status examination [MMSE]), white matter hyperintensity (WMH) burden, and clinical vascular risk factors (age, sex, hypertension, diabetes, lipid profiles, and body mass index). Training of the DL model (n = 287) resulted in a robust segmentation performance with an average dice score of 0.73 (95% CI, 0.67-0.79) in an internal validation set, (n = 67) and modest performance in an external validation set (dice score = 0.46, 95% CI, 0.33-0.59, n = 68). In a temporally independent clinical dataset (n = 448), older age, hypertension, and WMH burden were significantly associated with CMB numbers in all distributions (total, lobar, deep, and cerebellar; all P <.01). MMSE was significantly associated with hyperlipidemia (β = 1.88, 95% CI, 0.96-2.81, P <.001), WMH burden (β = -0.17 per 1% WMH burden, 95% CI, -0.27-0.08, P <.001), and total CMB number (β = -0.01 per 1 CMB, 95% CI, -0.02-0.001, P = .04) after adjusting for age and sex. The DL model showed a robust segmentation performance for CMB. In all distributions, CMB had significant positive associations with WMH burden. Increased WMH burden and CMB numbers were associated with decreased cognitive function. CMB = cerebral microbleed; DL = deep learning, DSC = dice similarity coefficient; MMSE = mini-mental status examination; SVD = small vessel disease; SWI = susceptibility-weighted image; WMH = white matter hyperintensity.

Fully automated measurement of noise, signal-to-noise ratio, and contrast-to-noise ratio on chest CT images: feasibility and efficiency.

Rapid and accurate measurement of computed tomography (CT) image noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) is a clinical challenge. To explore the feasibility of intelligent measurement of chest CT image noise, SNR, and CNR. A total of 300 chest CT scans were included in the study, which was divided into research dataset, internal test dataset, and external test dataset. Based on the research dataset, automatically segment and measure the average CT values and standard deviation (SD) of CT values for background air and lung field under different thresholds to obtain noise, SNR, and CNR results. Using the results of manual measurements as the reference standard, we determine the optimal threshold with the highest consistency. Using internal and external test datasets, validate the consistency of automated measurements of noise, SNR, and CNR at the optimal CT threshold with reference standards. With background air set at -900 HU and lung field at -800 HU as thresholds, the automated measurements of noise, SNR, and CNR demonstrate the highest consistency with the reference standards. At the optimal threshold, the noise, SNR, and CNR measured automatically on both the internal (intraclass correlation coefficient [ICC] = 0.85-0.96) and external (ICC = 0.75-0.85) test datasets exhibit high consistency with their respective reference standards. The method we explored can intelligently measure the noise, SNR, and CNR of chest CT images, exhibits high consistency with radiologists, and offers a novel tool for image quality evaluation and analysis.

Addressing the Generalizability of AI in Radiology Using a Novel Data Augmentation Framework with Synthetic Patient Image Data: Proof-of-Concept and External Validation for Classification Tasks in Multiple Sclerosis.

Artificial intelligence (AI) models often face performance drops after deployment to external datasets. This study evaluated the potential of a novel data augmentation framework based on generative adversarial networks (GANs) that creates synthetic patient image data for model training to improve model generalizability. Model development and external testing were performed for a given classification task, namely the detection of new fluid-attenuated inversion recovery lesions at MRI during longitudinal follow-up of patients with multiple sclerosis (MS). An internal dataset of 669 patients with MS (n = 3083 examinations) was used to develop an attention-based network, trained both with and without the inclusion of the GAN-based synthetic data augmentation framework. External testing was performed on 134 patients with MS from a different institution, with MR images acquired using different scanners and protocols than images used during training. Models trained using synthetic data augmentation showed a significant performance improvement when applied on external data (area under the receiver operating characteristic curve [AUC], 83.6% without synthetic data vs 93.3% with synthetic data augmentation; P = .03), achieving comparable results to the internal test set (AUC, 95.0%; P = .53), whereas models without synthetic data augmentation demonstrated a performance drop upon external testing (AUC, 93.8% on internal dataset vs 83.6% on external data; P = .03). Data augmentation with synthetic patient data substantially improved performance of AI models on unseen MRI data and may be extended to other clinical conditions or tasks to mitigate domain shift, limit class imbalance, and enhance the robustness of AI applications in medical imaging. Keywords: Brain, Brain Stem, Multiple Sclerosis, Synthetic Data Augmentation, Generative Adversarial Network Supplemental material is available for this article. © RSNA, 2024.

Comprehensive Approach to TB Diagnosis Powered by Machine Learning in Endemic Countries

Abstract Introduction/Objective Tuberculosis (TB), a leading cause of mortality worldwide, is primarily diagnosed through methods that face numerous challenges, including lengthy result times and diagnostic inaccuracy, particularly in the context of drug-resistant strains and latent infections. This study proposes a novel approach that integrates multiple diagnostic modalities with machine learning to enhance TB diagnosis accuracy. Methods/Case Report The Machine learning model was tailored for the classification of radiological images, specifically CT scans, using original DICOM images to preserve valuable voxel information. The study involved a curated dataset of CT scans from Pakistan, TB-positive (smear+/Culture+ n=72; smear+/culture- n= 3; smear-/Culture+ n=16; smear-/Culture- n=14; clinical and radiological diagnosis only n= 12) and healthy (n=103) scans from COVID-CTSet were included in the training to assess the specificity of the model. Data preprocessing included noise reduction, histogram matching, and data augmentation techniques to ensure a robust training dataset. The methodology also incorporated using k-fold cross validation and a custom learning rate scheduling strategy to optimize the model’s training process. To further test its robustness, the model was evaluated by swapping random patches and analyzing 60 mm patches from the entire CT scan for local feature categorization. Results (if a Case Study enter NA) The model has demonstrated outstanding performance, with sensitivity and specificity exceeding 95% on our internal dataset (n=117) and aligning well with clinical, smear, culture, and anti-TB antibody findings. It also showed promising results on the external NIH TB Portal’s dataset (n= 1,460). The model is robust to variations in CT scanner image quality and noise profiles, maintaining strong generalization across datasets. Even when analyzing random 60mm patches, effectively classifying all local features Conclusion Our study enhances tuberculosis diagnostic methods in endemic regions by blending machine learning with conventional diagnostic techniques. This approach improves TB detection, particularly the early and accurate identification in areas impacted by drug-resistant strains and asymptomatic latent infections. By integrating radiological and clinico-pathological data, we aim to develop a comprehensive diagnostic score. Moving forward, we plan to further integrate this method into clinical decision support systems and adapt the approach to other infectious diseases.

Improving Predictive Efficacy for Drug Resistance in Novel HIV-1 Protease Inhibitors through Transfer Learning Mechanisms.

The human immunodeficiency virus presents a significant global health challenge due to its rapid mutation and the development of resistance mechanisms against antiretroviral drugs. Recent studies demonstrate the impressive performance of machine learning (ML) and deep learning (DL) models in predicting the drug resistance profile of specific FDA-approved inhibitors. However, generalizing ML and DL models to learn not only from isolates but also from inhibitor representations remains challenging for HIV-1 infection. We propose a novel drug-isolate-fold change (DIF) model framework that aims to predict drug resistance score directly from the protein sequence and inhibitor representation. Various ML and DL models, inhibitor representations, and protein representations were analyzed through realistic validation mechanisms. To enhance the molecular learning capacity of DIF models, we employ a transfer learning approach by pretraining a graph neural network (GNN) model for activity prediction on a data set of 4855 HIV-1 protease inhibitors (PIs). By performing various realistic validation strategies on internal and external genotype-phenotype data sets, we statistically show that the learned representations of inhibitors improve the predictive ability of DIF-based ML and DL models. We achieved an accuracy of 0.802, AUROC of 0.874, and r of 0.727 for the unseen external PIs. By comparing the DIF-based models with a null model consisting of isolate-fold change (IF) architecture, it is observed that the DIF models significantly benefit from molecular representations. Combined results from various testing strategies and statistical tests confirm the effectiveness of DIF models in testing novel PIs for drug resistance in the presence of an isolate.

A deep learning model to enhance the classification of primary bone tumors based on incomplete multimodal images in X-ray, CT, and MRI

BackgroundAccurately classifying primary bone tumors is crucial for guiding therapeutic decisions. The National Comprehensive Cancer Network guidelines recommend multimodal images to provide different perspectives for the comprehensive evaluation of primary bone tumors. However, in clinical practice, most patients’ medical multimodal images are often incomplete. This study aimed to build a deep learning model using patients’ incomplete multimodal images from X-ray, CT, and MRI alongside clinical characteristics to classify primary bone tumors as benign, intermediate, or malignant.MethodsIn this retrospective study, a total of 1305 patients with histopathologically confirmed primary bone tumors (internal dataset, n = 1043; external dataset, n = 262) were included from two centers between January 2010 and December 2022. We proposed a Primary Bone Tumor Classification Transformer Network (PBTC-TransNet) fusion model to classify primary bone tumors. Areas under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity were calculated to evaluate the model’s classification performance.ResultsThe PBTC-TransNet fusion model achieved satisfactory micro-average AUCs of 0.847 (95% CI: 0.832, 0.862) and 0.782 (95% CI: 0.749, 0.817) on the internal and external test sets. For the classification of benign, intermediate, and malignant primary bone tumors, the model respectively achieved AUCs of 0.827/0.727, 0.740/0.662, and 0.815/0.745 on the internal/external test sets. Furthermore, across all patient subgroups stratified by the distribution of imaging modalities, the PBTC-TransNet fusion model gained micro-average AUCs ranging from 0.700 to 0.909 and 0.640 to 0.847 on the internal and external test sets, respectively. The model showed the highest micro-average AUC of 0.909, accuracy of 84.3%, micro-average sensitivity of 84.3%, and micro-average specificity of 92.1% in those with only X-rays on the internal test set. On the external test set, the PBTC-TransNet fusion model gained the highest micro-average AUC of 0.847 for patients with X-ray + CT.ConclusionsWe successfully developed and externally validated the transformer-based PBTC-Transnet fusion model for the effective classification of primary bone tumors. This model, rooted in incomplete multimodal images and clinical characteristics, effectively mirrors real-life clinical scenarios, thus enhancing its strong clinical practicability.

Development and Validation of a Computed Tomography-Based Model for Noninvasive Prediction of the T Stage in Gastric Cancer: Multicenter Retrospective Study.

As part of the TNM (tumor-node-metastasis) staging system, T staging based on tumor depth is crucial for developing treatment plans. Previous studies have constructed a deep learning model based on computed tomographic (CT) radiomic signatures to predict the numberoflymphnodemetastases and survival in patients with resected gastric cancer (GC). However, few studies have reported the combination of deep learning and radiomics in predicting T staging in GC. This study aimed to develop a CT-based model for automatic prediction of the T stage of GC via radiomics and deep learning. A total of 771 GC patients from 3 centers were retrospectively enrolled and divided into training, validation, and testing cohorts. Patients with GC were classified into mild (stage T1 and T2), moderate (stage T3), and severe (stage T4) groups. Three predictive models based on the labeled CT images were constructed using the radiomics features (radiomics model), deep features (deep learning model), and a combination of both (hybrid model). The overall classification accuracy of the radiomics model was 64.3% in the internal testing data set. The deep learning model and hybrid model showed better performance than the radiomics model, with overall classification accuracies of 75.7% (P=.04) and 81.4% (P=.001), respectively. On the subtasks of binary classification of tumor severity, the areas under the curve of the radiomics, deep learning, and hybrid models were 0.875, 0.866, and 0.886 in the internal testing data set and 0.820, 0.818, and 0.972 in the external testing data set, respectively, for differentiating mild (stage T1~T2) from nonmild (stage T3~T4) patients, and were 0.815, 0.892, and 0.894 in the internal testing data set and 0.685, 0.808, and 0.897 in the external testing data set, respectively, for differentiating nonsevere (stage T1~T3) from severe (stage T4) patients. The hybrid model integrating radiomics features and deep features showed favorable performance in diagnosing the pathological stage of GC.

Deep learning model using planar whole-body bone scintigraphy for diagnosis of skull base invasion in patients with nasopharyngeal carcinoma

PurposeThis study assesses the reliability of deep learning models based on planar whole-body bone scintigraphy for diagnosing Skull base invasion (SBI) in nasopharyngeal carcinoma (NPC) patients.MethodsIn this multicenter study, a deep learning model was developed using data from one center with a 7:3 allocation to training and internal test sets, to diagnose SBI in patients newly diagnosed with NPC using planar whole-body bone scintigraphy. Patients were diagnosed based on a composite reference standard incorporating radiologic and follow-up data. Ten different convolutional neural network (CNN) models were applied to both whole-image and partial-image input modes to determine the optimal model for each analysis. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration, decision curve analysis (DCA), and compared with expert assessments by two nuclear medicine physicians.ResultsThe best-performing model using partial-body input achieved AUCs of 0.80 (95% CI: 0.73, 0.86) in the internal test set, 0.84 (95% CI: 0.77, 0.91) in the external cohort, and 0.78 (95% CI: 0.73, 0.83) in the treatment test cohort. Calibration curves and DCA confirmed the models’ excellent discrimination, calibration, and potential clinical utility across internal and external datasets. The AUCs of both nuclear medicine physicians were lower than those of the best-performing deep learning model in external test set (AUC: 0.75 vs. 0.77 vs. 0.84).ConclusionDeep learning models utilizing partial-body input from planar whole-body bone scintigraphy demonstrate high discriminatory power for diagnosing SBI in NPC patients, surpassing experienced nuclear medicine physicians.

The role of artificial intelligence in optimizing management of atrial fibrillation in acute ischemic stroke.

Atrial fibrillation (AF) is a severe condition associated with high morbidity and mortality, including an increased risk of stroke and poor outcomes poststroke. Our understanding of the prognosis in AF remains poor. Machine learning (ML) has been applied to the diagnosis, management, and prognosis of AF in the context of stroke but remains suboptimal for clinical use. This article endeavors to provide a comprehensive overview of current ML applications to AF patients at risk of stroke, as well as poststroke patients without AF. Strategies to develop effective ML involve the validation of a variety of ML algorithms across internal and external datasets as well as exploring their predictive powers in hypothetical and realistic settings. Recent literature of this rapidly evolving field has displayed much promise. However, further testing and innovation of medical artificial intelligence are required before its imminent introduction to ensure complete patient trust within the community. Prioritizing this research is imperative for advancing the optimization of ongoing care for AF patients, as well as the management of stroke patients with AF.