Detection of Pneumothorax with Deep Learning Models: Learning From Radiologist Labels vs Natural Language Processing Model Generated Labels

BackgroundCurrently, there is no fully automated tool available for evaluating the degree of cervical spinal stenosis. The aim of this study was to develop and validate the use of artificial intelligence (AI) algorithms for the assessment of cervical spinal stenosis.MethodsIn this retrospective multi-center study, cervical spine magnetic resonance imaging (MRI) scans obtained from July 2020 to June 2023 were included. Studies of patients with spinal instrumentation or studies with suboptimal image quality were excluded. Sagittal T2-weighted images were used. The training data from the Fourth People’s Hospital of Shanghai (Hos. 1) and Shanghai Changzheng Hospital (Hos. 2) were annotated by two musculoskeletal (MSK) radiologists following Kang’s system as the standard reference. First, a convolutional neural network (CNN) was trained to detect the region of interest (ROI), with a second Transformer for classification. The performance of the deep learning (DL) model was assessed on an internal test set from Hos. 2 and an external test set from Shanghai Changhai Hospital (Hos. 3), and compared among six readers. Metrics such as detection precision, interrater agreement, sensitivity (SEN), and specificity (SPE) were calculated.ResultsOverall, 795 patients were analyzed (mean age ± standard deviation, 55±14 years; 346 female), with 589 in the training (75%) and validation (25%) sets, 206 in the internal test set, and 95 in the external test set. Four tasks with different clinical application scenarios were trained, and their accuracy (ACC) ranged from 0.8993 to 0.9532. When using a Kang system score of ≥2 as a threshold for diagnosing central cervical canal stenosis in the internal test set, both the algorithm and six readers achieved similar areas under the receiver operating characteristic curve (AUCs) of 0.936 [95% confidence interval (CI): 0.916–0.955], with a SEN of 90.3% and SPE of 93.8%; the AUC of the DL model was 0.931 (95% CI: 0.917–0.946), with a SEN in the external test set of 100%, and a SPE of 86.3%. Correlation analysis comparing the DL method, the six readers, and MRI reports between the reference standard showed a moderate correlation, with R values ranging from 0.589 to 0.668. The DL model produced approximately the same upgrades (9.2%) and downgrades (5.1%) as the six readers.ConclusionsThe DL model could fully automatically and reliably assess cervical canal stenosis using MRI scans.

Sentiment analysis is a subfield of Natural Language Processing (NLP) which tries to process a text to extract opinions or attitudes towards topics or entities. Recently, the use of deep learning methods for sentiment analysis has received noticeable attention from researchers. Generally, different deep learning methods have shown superb performance in sentiment analysis problem. However, deep learning models are different in nature and have different strengths and limitations. For example, convolutional neural networks are useful for extracting local structures from data, while recurrent models are able to learn order dependence in sequential data. In order to combine the advantages of different deep models, in this paper we have proposed a novel approach for aspect-based sentiment analysis which utilizes deep ensemble learning. In the proposed method, we first build four deep learning models, namely CNN, LSTM, BiLSTM and GRU. Then the outputs of these models are combined using stacking ensemble approach where we have used logistic regression as meta-learner. The results of applying the proposed method on the real datasets show that our method has increased the accuracy of aspect-based prediction by 5% to 20% compared to the basic deep learning methods.

To develop a deep learning (DL) model to measure the sagittal Cobb angle of the cervical spine on computed tomography (CT). Two VB-Net-based DL models for cervical vertebra segmentation and key-point detection were developed. Four-points and line-fitting methods were used to calculate the sagittal Cobb angle automatically. The average value of the sagittal Cobb angle was manually measured by two doctors as the reference standard. The percentage of correct key points (PCK), matched samples t test, intraclass correlation coefficient (ICC), Pearson correlation coefficient, mean absolute error (MAE), and Bland‒Altman plots were used to evaluate the performance of the DL model and the robustness and generalization of the model on the external test set. A total of 991 patients were included in the internal data set, and 112 patients were included in the external data set. The PCK of the DL model ranged from 78 to 100% in the test set. The four-points method, line-fitting method, and reference standard measured sagittal Cobb angles were - 1.10 ± 18.29°, 0.30 ± 13.36°, and 0.50 ± 12.83° in the internal test set and 4.55 ± 20.01°, 3.66 ± 18.55°, and 1.83 ± 12.02° in the external test set, respectively. The sagittal Cobb angle calculated by the four-points method and the line-fitting method maintained high consistency with the reference standard (internal test set: ICC = 0.75 and 0.97; r = 0.64 and 0.94; MAE = 5.42° and 3.23°, respectively; external test set: ICC = 0.74 and 0.80, r = 0.66 and 0.974, MAE = 5.25° and 4.68°, respectively). The DL model can accurately measure the sagittal Cobb angle of the cervical spine on CT. The line-fitting method shows a higher consistency with the doctors and a minor average absolute error.

Aims: This study aimed to explore the effect of training set diversity on the performance of deep learning models for predicting early gastric cancer (EGC) under endoscopy. Methods: Images of EGC and non-cancerous lesions under narrow-band imaging (ME-NBI) and magnifying blue laser imaging (ME-BLI) were retrospectively collected. Training set 1 was composed of 150 non-cancerous and 309 EGC ME-NBI images, training set 2 was composed of 1505 non-cancerous and 309 EGC ME-BLI images, and training set 3 was the combination of training set 1 and 2. Test set 1 was composed of 376 non-cancerous and 1052 EGC ME-NBI images, test set 2 consisted of 529 non-cancerous and 71 EGC ME-BLI images, and test set 3 was the combination of test set 1 and test set 2. Three deep learning models, convolutional neural network (CNN) 1, CNN 2 and CNN 3 (CNN 1, CNN 2 and CNN 3 were independently trained using training set 1, training set 2 and training set 3, respectively), were constructed, and their performances on each test set were respectively evaluated. One hundred and thirty-eight ME-NBI videos and 17 ME-BLI videos were further collected to evaluate and compare the performance of each model in real time. Results: On the whole, the performance of CNN 3 was the best. The accuracy (Acc), sensitivity (Sn), specificity (Sp) and area under the curve (AUC) of test set 1 in CNN 3 were 87.89% (1255/1428), 90.96% (342/376), 86.79% (913/1052) and 94.60%, respectively. The Acc, Sn, Sp and AUC of test set 2 in CNN 3 were 95% (570/600), 97.92% (518/529), 73.24% (52/71) and 90.93% respectively. The Acc, Sn, Sp and AUC of test set 3 in CNN 3 were 89.99% (1825/2028), 95.03% (860/905), 85.93% (965/1123) and 94.89%, respectively. The performance of CNN 3 was also the best in videos test set. The Acc, Sn and Sp of videos test set in CNN 3 were 91.03% (142/156), 90.58% (125/138) and 94.44% (17/18), respectively. Conclusions: The deep learning model with the most diverse training data has the best diagnostic effect.

The World Health Organization glioma classification highlights genetic profiles, such as isocitrate dehydrogenase (IDH) mutation. This study developed an uncertainty-aware deep learning model to predict IDH mutations in glioma patients, employing conformal prediction (CP) for uncertainty quantification (UQ). The UCSF dataset split into training (70%) and validation (30%) sets, and the UPENN dataset divided into calibration (30%) and external test (70%) sets. We developed various 3D convolutional neural network models and selected the best based on the validation set area under the precision-recall curve (AUPRC). Also, we trained a logistic regression (LR) ensemble classifier on the training set and selected the best ensemble model based on the validation set AUPRC. Finally, we employed CP, with nonconformity thresholds set at the error rate of 0.01. The conformal model was calibrated on the calibration set and was evaluated on the external test set using the AUPRC and the area under the receiver operating characteristic (AUROC). 3D convolutional models achieved a mean AUROC and AUPRC of 0.9004 and 0.8102 on the validation set, and a mean AUROC and AUPRC of 0.7340 and 0.2139 on the external test set, respectively. Using the LR ensemble model, the model achieved an AUROC and AUPRC of 0.79820 and 0.2583 on the external test set, respectively. CP with a 0.01 nonconformity threshold with 0.9917 coverage reached an AUROC and AUPRC of 0.8592 and 0.5217, respectively. Integrating CP for UQ improves the performance of deep learning models for predicting IDH mutation in glioma patients.

The diagnosis of severe COVID-19 lung infection is important because it carries a higher risk for the patient and requires prompt treatment with oxygen therapy and hospitalization while those with less severe lung infection often stay on observation. Also, severe infections are more likely to have long-standing residual changes in their lungs and may need follow-up imaging. We have developed deep learning neural network models for classifying severe vs. non-severe lung infections in COVID-19 patients on chest radiographs (CXR). A deep learning U-Net model was developed to segment the lungs. Inception-v1 and Inception-v4 models were trained for the classification of severe vs. non-severe COVID-19 infection. Four CXR datasets from multi-country and multi-institutional sources were used to develop and evaluate the models. The combined dataset consisted of 5748 cases and 6193 CXR images with physicians’ severity ratings as reference standard. The area under the receiver operating characteristic curve (AUC) was used to evaluate model performance. We studied the reproducibility of classification performance using the different combinations of training and validation data sets. We also evaluated the generalizability of the trained deep learning models using both independent internal and external test sets. The Inception-v1 based models achieved AUC ranging between 0.81 ± 0.02 and 0.84 ± 0.0, while the Inception-v4 models achieved AUC in the range of 0.85 ± 0.06 and 0.89 ± 0.01, on the independent test sets, respectively. These results demonstrate the promise of using deep learning models in differentiating COVID-19 patients with severe from non-severe lung infection on chest radiographs.

e13628 Background: Despite the wide availability and cost-effectiveness of chest X-rays (CXRs), their potential as a biomarker for immune checkpoint inhibitor (ICI) response in non-small cell lung cancer (NSCLC) remains largely unexplored. We developed and validated a novel deep learning model that predicts ICI response from a single pre-treatment CXR, independent of histopathologic biomarkers including PD-L1 expression and immune phenotype. Methods: A retrospective dataset of 2,905 NSCLC patients at Samsung Medical Center was collected, including pretreatment CXRs and H&E whole slide images (WSI). A deep learning model was trained using 1,939 pre-ICI CXRs to predict ICI outcomes, with time to next treatment (TTNT) within 6 months as a proxy. To evaluate regimen-specific performance, the model was tested on pretreatment CXRs from three patient groups defined by first-line regimen: A (N = 119, ICI monotherapy), B (N = 327, platinum-based chemotherapy), and C (N = 520, ICI-chemotherapy combination). To interpret the model's behavior, we generated segmentation masks for tumor regions and lung fields using a separate model and injected white noise at various signal-to-noise ratios (SNR). We assessed tumor infiltrating lymphocyte (TIL) density and immune phenotype on 113 paired H&E WSIs from test set A using Lunit SCOPE IO, a deep learning H&E analyzer, to correlate CXR predictions with the tumor microenvironment. Results: The model achieved an area under the ROC curve (AUC) of 0.745 in test set A, with predicted good responders showing significantly prolonged TTNT (median: 12.0 vs 2.0 months, hazard ratio [HR] 0.41, p < 0.001) and OS (median: 24.7 vs 6.0 months, HR 0.42, p < 0.001) compared to predicted non-responders. The model's AUC was 0.531 in test set B (predicting TTNT after chemotherapy) and 0.654 in test set C (ICI-chemotherapy combination), demonstrating specificity of the model to ICI. Noise injection in the nodule region and lung field led to a decrease in AUC, with the most significant decrease in extra-tumor regions of the ipsilateral lung. Prediction scores were not correlated with intratumoral TIL density (r = 0.07, p = 0.44) or PD-L1 expression (r = -0.08, p = 0.56). The prediction of the CXR model for ICI monotherapy (test set A) was independent of high PD-L1 expression (PD-L1≥50%, HR 0.43, p = 0.02; PD-L1 < 50%, HR 0.38, p = 0.13) or immune phenotypes determined by TIL enrichment (Inflamed: HR 0.48, p < 0.05; Non-inflamed: HR 0.38, p < 0.001). Conclusions: A deep learning CXR model predicted ICI response in NSCLC regardless of PD-L1 expression and immune phenotype. This work suggests that CXRs contain intrinsic information relevant to ICI response, achieving competitive performance with state-of-the-art CT or histology based approaches despite relying on broader lung features.

To develop a deep learning model to automatically measure the curvature-related sagittal parameters on cervical spinal X-ray images. This retrospective study collected a total of 700 lateral cervical spine X-ray images from three hospitals, consisting of 500 training sets, 100 internal test sets, and 100 external test sets. 6 measured parameters and 34 landmarks were measured and labeled by two doctors and averaged as the gold standard. A Convolutional neural network (CNN) model was built by training on 500 images and testing on 200 images. Statistical analysis is used to evaluate labeling differences and model performance. The percentages of the difference in distance between landmarks within 4mm were 96.90% (Dr. A vs. Dr. B), 98.47% (Dr. A vs. model), and 97.31% (Dr. B vs. model); within 3mm were 94.88% (Dr. A vs. Dr. B), 96.43% (Dr. A vs. model), and 94.16% (Dr. B vs. model). The mean difference of the algorithmic model in labeling landmarks was 1.17 ± 1.14mm. The mean absolute error (MAE) of the algorithmic model for the Borden method, Cervical curvature index (CCI), Vertebral centroid measurement cervical lordosis (CCL), C0-C7 Cobb, C1-C7 Cobb, C2-C7 Cobb in the test sets are 1.67mm, 2.01%, 3.22°, 2.37°, 2.49°, 2.81°, respectively; symmetric mean absolute percentage error (SMAPE) was 20.06%, 21.68%, 20.02%, 6.68%, 5.28%, 20.46%, respectively. Also, the algorithmic model of the six cervical sagittal parameters is in good agreement with the gold standard (intraclass correlation efficiency was 0.983; p < 0.001). Our deep learning algorithmic model had high accuracy in recognizing the landmarks of the cervical spine and automatically measuring cervical spine-related parameters, which can help radiologists improve their diagnostic efficiency.

To develop and validate a deep learning model for the detection of aberrant anterior tibial artery (AATA) on axial T2-weighted knee MRI, given the surgical relevance of unrecognized AATA and the lack of automated detection tools. This retrospective study included 70,260 MRI images from 2315 examinations (1441 without AATA and 874 with AATA) collected after institutional review board approval. Musculoskeletal radiologists performed image-level annotations. Data were split at the patient level into training, validation, and internal test sets; an independent dataset from another institution served as an external test set. A convolutional neural network was implemented in Python and PyTorch. Model performance was assessed at the patient level. At the slice level, the model achieved an F1-score of 0.838 on the internal test set. Patient-level classification using the validation-derived threshold (0.17) yielded F1-scores of 0.966 on the validation set, 0.979 on the internal test set, and 0.786 on the external test set. The area under the receiver operating characteristic curve for the external cohort was 0.97, indicating strong generalization despite a decrease in precision due to false positives. To our knowledge, this is the first study to apply artificial intelligence for automated detection of AATA on knee MRI. The proposed deep learning model performs this task with high sensitivity. Despite reduced precision in the external cohort, it demonstrates strong potential for enhancing preoperative risk assessment and surgical planning. Broader multicenter validation is warranted before clinical deployment.

Customer sentiment analytics has emerged as a critical tool for retail businesses aiming to understand consumer perceptions, improve customer experience, and enhance decision-making. This paper explores the application of Natural Language Processing (NLP) and deep learning techniques in analyzing unstructured textual data such as customer reviews, social media posts, and feedback to extract actionable insights for the retail industry. Traditional sentiment analysis methods often rely on keyword-based approaches, which lack the contextual understanding necessary to interpret nuanced emotions and implicit meanings. In contrast, NLP combined with advanced deep learning models, such as Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and Transformers (e.g., BERT), enables more accurate sentiment classification by capturing semantic relationships, context, and sentiment polarity. The study presents a comprehensive framework that leverages data preprocessing, tokenization, embedding layers (e.g., Word2Vec, GloVe), and model training pipelines to analyze large volumes of customer-generated content. Experimental results from real-world retail datasets demonstrate the superiority of deep learning models over traditional machine learning techniques in sentiment prediction accuracy, especially in multilingual and noisy data scenarios. Moreover, the paper discusses the integration of sentiment scores with customer segmentation and product performance metrics to drive personalized marketing, inventory adjustments, and service enhancements. Challenges such as data labeling, domain adaptation, and interpretability of deep models are also addressed, along with strategies to overcome them using transfer learning, attention mechanisms, and explainable AI tools. The research concludes that customer sentiment analytics using NLP and deep learning is not merely a technological advancement but a strategic enabler for data-driven retail transformation. By systematically capturing and interpreting customer voices, retail businesses can proactively respond to market demands, enhance loyalty, and gain a competitive edge in an increasingly digital consumer landscape.

BackgroundDeep learning has the potential to augment the use of chest radiography in clinical radiology, but challenges include poor generalizability, spectrum bias, and difficulty comparing across studies.PurposeTo develop and evaluate deep learning models for chest radiograph interpretation by using radiologist-adjudicated reference standards.Materials and MethodsDeep learning models were developed to detect four findings (pneumothorax, opacity, nodule or mass, and fracture) on frontal chest radiographs. This retrospective study used two data sets. Data set 1 (DS1) consisted of 759 611 images from a multicity hospital network and ChestX-ray14 is a publicly available data set with 112 120 images. Natural language processing and expert review of a subset of images provided labels for 657 954 training images. Test sets consisted of 1818 and 1962 images from DS1 and ChestX-ray14, respectively. Reference standards were defined by radiologist-adjudicated image review. Performance was evaluated by area under the receiver operating characteristic curve analysis, sensitivity, specificity, and positive predictive value. Four radiologists reviewed test set images for performance comparison. Inverse probability weighting was applied to DS1 to account for positive radiograph enrichment and estimate population-level performance.ResultsIn DS1, population-adjusted areas under the receiver operating characteristic curve for pneumothorax, nodule or mass, airspace opacity, and fracture were, respectively, 0.95 (95% confidence interval [CI]: 0.91, 0.99), 0.72 (95% CI: 0.66, 0.77), 0.91 (95% CI: 0.88, 0.93), and 0.86 (95% CI: 0.79, 0.92). With ChestX-ray14, areas under the receiver operating characteristic curve were 0.94 (95% CI: 0.93, 0.96), 0.91 (95% CI: 0.89, 0.93), 0.94 (95% CI: 0.93, 0.95), and 0.81 (95% CI: 0.75, 0.86), respectively.ConclusionExpert-level models for detecting clinically relevant chest radiograph findings were developed for this study by using adjudicated reference standards and with population-level performance estimation. Radiologist-adjudicated labels for 2412 ChestX-ray14 validation set images and 1962 test set images are provided.© RSNA, 2019Online supplemental material is available for this article.See also the editorial by Chang in this issue.

Gastric cancer (GC) is one of the leading causes of cancer-related deaths globally, with a 5-year survival rate of less than 40%. While immune checkpoint inhibitors have provided promising therapeutic options for advanced GC, only a small proportion of patients benefit. In this study, we developed a deep learning model using whole-slide images to predict prognoses and sensitivity to immune checkpoint inhibitors in GC patients by predicting a novel marker. Formalin-fixed, paraffin-embedded whole-slide images from 292 patients in the Cancer Genome Atlas-Stomach Adenocarcinoma cohort were analyzed. Tumor regions were identified using a ResNet50-based tumor detection model and validated in HiESD dataset. Tiles classified as malignant were extracted for subsequent analysis. Risk score prediction models were developed using convolutional neural networks, clustering-constrained attention multiple-instance learning (CLAM), and dual-stream multiple-instance learning (DSMIL). Attention heatmap visualization was used to interpret tumor microenvironment (TME) features and a multi-model classification framework utilizing a support vector machine (SVM) was developed to assess the impact of clinical variables in model performances. The results were evaluated using the area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, specificity, and F1 score. The tumor detection model achieved an AUROC of 0.99 on the training set, 0.92 on the test set, and 0.87 on external test set. Among risk score prediction models, DSMIL demonstrated the highest performance, with an AUROC of 0.73 and accuracy of 0.73 on the training set and AUROC of 0.70 and accuracy of 0.68 on the internal test set. High-risk patients exhibited worse survival outcomes and lower immunotherapy response rates compared with low-risk patients. Feature attribution analysis using attention heatmaps confirmed that the model prioritized TME components, specifically regions with dense lymphocytic infiltration. The muti-modal analysis showed that the image features alone were superior to the model combining image features with clinicopathological data. Deep learning models leveraging whole-slide images show potential in predicting prognoses and immunotherapy responses in GC. By integrating tumor-specific and tumor microenvironmental features, this approach offers a scalable, objective tool for personalized treatment planning, improving precision oncology strategies.

This study aims to develop a deep learning (DL) model based on conventional ultrasound images for differentiating pancreatic serous neoplasms (SCN) from mucinous cystic neoplasms (MCN). We seek to determine if such a model can surpass the diagnostic performance of routine sonographic evaluation and provide a valuable, cost-effective decision-support tool. Data from 459 patients with histopathologically confirmed SCN and MCN from center 1 (n = 345) and center 2 (n = 114) were retrospectively collected. Five DL models (EfficientNet-B3, Resnet-101, Densenet-121, Se-resnext-50, Inception-v3) were constructed for differential diagnosis. Simultaneously compare the diagnostic performance of DL model with that of junior radiologists (JR) and senior radiologists (SR), and determine whether DL can assist radiologists in making better diagnoses. The EfficientNet-B3 model demonstrated the optimal predictive performance, with an area under the receiver operating curve (AUC) of 0.904 (95% CI:0.814–0.973) in the internal validation set and 0.866 (95% CI:0.783–0.936) in the external test set. In the external test set, the EfficientNet-B3 model demonstrated significantly superior diagnostic performance compared to both JR in both F1-score and specificity (all P < 0.0125), while showing fully comparable performance to both SR in these precise metrics (all P > 0.0125). With the EfficientNet-B3 model assistance, both JR showed statistically significant improvements in F1-score and specificity (all P < 0.0125), whereas both SR exhibited no statistically significant changes in either performance metric (all P > 0.0125). The EfficientNet-B3 model demonstrated comparable diagnostic performance to SR while significantly enhancing the diagnostic accuracy of JR. A deep learning-based ultrasound model was developed to distinguish pancreatic serous cystic neoplasms from mucinous cystic neoplasms, outperforming junior radiologists and matching the diagnostic accuracy of senior radiologists. With model assistance, junior radiologists demonstrated significantly enhanced diagnostic performance.

This study aims to develop an attention-based deep learning model for distinguishing benign from malignant breast lesions on CESM. Preoperative CESM images of 1239 patients, which were definitely diagnosed on pathology in a multicentre cohort, were divided into training and validation sets, internal and external test sets. The regions of interest of the breast lesions were outlined manually by a senior radiologist. We adopted three conventional convolutional neural networks (CNNs), namely, DenseNet 121, Xception, and ResNet 50, as the backbone architectures and incorporated the convolutional block attention module (CBAM) into them for classification. The performance of the models was analysed in terms of the receiver operating characteristic (ROC) curve, accuracy, the positive predictive value (PPV), the negative predictive value (NPV), the F1 score, the precision recall curve (PRC), and heat maps. The final models were compared with the diagnostic performance of conventional CNNs, radiomics models, and two radiologists with specialised breast imaging experience. The best-performing deep learning model, that is, the CBAM-based Xception, achieved an area under the ROC curve (AUC) of 0.970, a sensitivity of 0.848, a specificity of 1.000, and an accuracy of 0.891 on the external test set, which was higher than those of other CNNs, radiomics models, and radiologists. The PRC and the heat maps also indicated the favourable predictive performance of the attention-based CNN model. The diagnostic performance of two radiologists improved with deep learning assistance. Using an attention-based deep learning model based on CESM images can help to distinguishing benign from malignant breast lesions, and the diagnostic performance of radiologists improved with deep learning assistance.

Category: Midfoot/Forefoot; Hindfoot Introduction/Purpose: Evaluation of plain radiographs is of fundamental importance in the diagnosis and treatment planning of foot and ankle disorders. These assessments are essential, but can be difficult for the following reasons: First, many angles must be measured using various reference lines. Second, there may be inter-observer differences in the measurements. Consistent and standardized measurement methods are required to increase the efficiency and accuracy of radiographic readings. Several recent studies have reported good results using deep learning (DL) models for medical image analysis. It is expected that efficiency and accuracy can be increased through radiographic readings using DL. This study aims to develop a DL model that recognizes the bones of the foot and ankle in plain radiographs and automatically measures angles frequently used for diagnosis. Methods: The DL model was developed using two architectures: segmenting objects by locations (SOLO) for bone segmentation and Global Universal U-Net (GU2Net) for reference line detection. The angles measured were hallux valgus angle (HVA) and 1st- 2nd intermetatarsal angle (IMA) on anterior-posterior (AP) X-ray; calcaneal pitch angle (CPA) and lateral Meary's angle (LMA) on lateral X-ray. Only data obtained from external test set were used for statistical analysis to evaluate the model's performance. The test set used a total of 558 images collected from January 2022 to September 2022. Angles measured by the DL model were compared with angles measured manually by human researchers. The intraclass correlation coefficient (ICC) was used to evaluate the accuracy of angle measurements. The error between the angle measured by the DL model and the angle measured by human researchers was evaluated using root mean square error (RMSE) and mean absolute difference (MAE). Results: As shown in the attached table, the ICCs of HVA were 0.957 (right) and 0.933 (left). The ICCs of IMA were 0.881 (right) and 0.902 (left). The ICCs of LMA were 0.895 (right) and 0.937 (left). The ICCs of CPA were 0.918 (right) and 0.868 (left). All ICCs were greater than 0.8, indicating excellent correlations between angles measured with the DL model and angles measured manually (ground truth). For RMSE and MAE, the left IMA was the smallest (1.50 and 1.16, respectively), and the right LMA was the largest (4.50 and 2.47, respectively). Conclusion: The DL model in our study was able to perform various angle measurements in both AP and lateral X-rays with considerable accuracy. Through additional model training and development, it is expected that it will be useful for analyzing various foot and ankle disorders such as hallux valgus, flatfoot, pes cavus and other alignment deformities.