Deep Transfer Learning Research Articles

The value of CT radiomics combined with deep transfer learning in predicting the nature of gallbladder polypoid lesions.

Computed tomography (CT) radiomics combined with deep transfer learning was used to identify cholesterol and adenomatous gallbladder polyps that have not been well evaluated before surgery. To investigate the potential of various machine learning models, incorporating radiomics and deep transfer learning, in predicting the nature of cholesterol and adenomatous gallbladder polyps. A retrospective analysis was conducted on clinical and imaging data from 100 patients with cholesterol or adenomatous polyps confirmed by surgery and pathology at our hospital between September 2015 and February 2023. Preoperative contrast-enhanced CT radiomics combined with deep learning features were utilized, and t-tests and least absolute shrinkage and selection operator (LASSO) cross-validation were employed for feature selection. Subsequently, 11 machine learning algorithms were utilized to construct prediction models, and the area under the ROC curve (AUC), accuracy, and F1 measure were used to assess model performance, which was validated in a validation group. The Logistic algorithm demonstrated the most effective prediction in identifying polyp properties based on 10 radiomics combined with deep learning features, achieving the highest AUC (0.85 in the validation group, 95% confidence interval = 0.68-1.0). In addition, the accuracy (0.83 in the validation group) and F1 measure (0.76 in the validation group) also indicated strong performance. The machine learning radiomics combined with deep learning model based on enhanced CT proves valuable in predicting the characteristics of cholesterol and adenomatous gallbladder polyps. This approach provides a more reliable basis for preoperative diagnosis and treatment of these conditions.

Deep learning for enhanced brain Tumor Detection and classification

The purpose of this research is to build an automated, robust, intelligent and hybrid system for the early diagnosis and classifying of brain tumor. To serve this purpose, the authors propose the Auto Contrast Enhancer, Tumor Detector and Classifier to efficiently provide on-demand contrast improvement of poor contrast MRI images for the early diagnosis and classification of brain tumors. The classifier accomplishes its task through a two-phase approach. During the initial phase, ODTWCHE is employed to enhance image contrast, facilitating accurate diagnosis of brain tumors. In the subsequent phase, the classifier leverages the power of deep transfer learning, utilizing the pre-trained Inception V3 model to refine the diagnostic process further. tumor classification. Compared to state-of-the-art models, including AlexNet, VGG-16, DenseNet-201, VGG-19, GoogLeNet, and ResNet-50, the proposed system showcased its outstanding performance by achieving the highest accuracy of 98.89 % on a public dataset that consists of MRI images with varying contrast and brightness levels. The precise detection and classification achieved on this multicolored dataset prove the system's robustness. The authors of the article address the usage of metrics in a variety of contexts, including academia, as well as the possible problems that may result from their improper application. They emphasize how crucial it is to create measurements that align with the system's objectives and to reduce any negative consequences that can skew the data or allow people to manipulate the system's incentives. The authors provide a thorough process for creating metrics that takes into account design considerations, countermeasures for unfavorable effects, and crucial requirements. The paper provides answers for the creation of metrics and gives examples of metrics' failures in many fields. The authors emphasize the significance of understanding how the goal and the data at hand relate to one another, as well as the necessity of compromise and clarity when goals are contradictory or incoherent.A comparative analysis with existing models further confirms that the proposed system consistently outperforms the competition.

Image-based fire detection using an attention mechanism and pruned dense network transfer learning

Fire accidents continually threaten lives and property, posing a great risk to public safety. Therefore, timely, accurate fire smoke and flame detection technology provides a technical foundation for ensuring personnel and property safety. This paper proposes a deep transfer learning model that integrates attention mechanisms and pruning techniques into the DenseNet network (P-DenseNet-A-TL) for detecting fire smoke and flame targets. To reduce the computational complexity of the fire detection model, we simplified the DenseNet network structure. To improve the recognition accuracy of the fire detection model, we incorporated attention mechanisms including a channel attention module (CAM) and spatial attention module (SAM) into the pruned dense network structure. To expedite the model training process, we also introduced a transfer learning model. Furthermore, to enhance the generalization ability and robustness of the model, we created a large-scale fire smoke and flame dataset. The experimental results show that our model (P-DenseNet-A-TL) achieved a test accuracy of 99.06%, F1 score of 99.09%, area under the curve (AUC) of 0.97, and a detection speed of 756 frames per second (FPS). The comparison experimental results and ablation experimental results indicate that our method achieves high detection accuracy and efficiency. Additionally, it possesses strong generalization capability and robustness.

Automatic speech recognition using advanced deep learning approaches: A survey

Recent advancements in deep learning (DL) have posed a significant challenge for automatic speech recognition (ASR). ASR relies on extensive training datasets, including confidential ones, and demands substantial computational and storage resources. Enabling adaptive systems improves ASR performance in dynamic environments. DL techniques assume training and testing data originate from the same domain, which is not always true. Advanced DL techniques like deep transfer learning (DTL), federated learning (FL), and deep reinforcement learning (DRL) address these issues. DTL allows high-performance models using small yet related datasets, FL enables training on confidential data without dataset possession, and DRL optimizes decision-making in dynamic environments, reducing computation costs.This survey offers a comprehensive review of DTL, FL, and DRL-based ASR frameworks, aiming to provide insights into the latest developments and aid researchers and professionals in understanding the current challenges. Additionally, Transformers, which are advanced DL techniques heavily used in proposed ASR frameworks, are considered in this survey for their ability to capture extensive dependencies in the input ASR sequence. The paper starts by presenting the background of DTL, FL, DRL, and Transformers and then adopts a well-designed taxonomy to outline the state-of-the-art (SOTA) approaches. Subsequently, a critical analysis is conducted to identify the strengths and weaknesses of each framework. Additionally, a comparative study is presented to highlight the existing challenges, paving the way for future research opportunities.

Ultrasound-based deep learning radiomics model for differentiating benign, borderline, and malignant ovarian tumours: a multi-class classification exploratory study.

Accurate preoperative identification of ovarian tumour subtypes is imperative for patients as it enables physicians to custom-tailor precise and individualized management strategies. So, we have developed an ultrasound (US)-based multiclass prediction algorithm for differentiating between benign, borderline, and malignant ovarian tumours. We randomised data from 849 patients with ovarian tumours into training and testing sets in a ratio of 8:2. The regions of interest on the US images were segmented and handcrafted radiomics features were extracted and screened. We applied the one-versus-rest method in multiclass classification. We inputted the best features into machine learning (ML) models and constructed a radiomic signature (Rad_Sig). US images of the maximum trimmed ovarian tumour sections were inputted into a pre-trained convolutional neural network (CNN) model. After internal enhancement and complex algorithms, each sample's predicted probability, known as the deep transfer learning signature (DTL_Sig), was generated. Clinical baseline data were analysed. Statistically significant clinical parameters and US semantic features in the training set were used to construct clinical signatures (Clinic_Sig). The prediction results of Rad_Sig, DTL_Sig, and Clinic_Sig for each sample were fused as new feature sets, to build the combined model, namely, the deep learning radiomic signature (DLR_Sig). We used the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC) to estimate the performance of the multiclass classification model. The training set included 440 benign, 44 borderline, and 196 malignant ovarian tumours. The testing set included 109 benign, 11 borderline, and 49 malignant ovarian tumours. DLR_Sig three-class prediction model had the best overall and class-specific classification performance, with micro- and macro-average AUC of 0.90 and 0.84, respectively, on the testing set. Categories of identification AUC were 0.84, 0.85, and 0.83 for benign, borderline, and malignant ovarian tumours, respectively. In the confusion matrix, the classifier models of Clinic_Sig and Rad_Sig could not recognise borderline ovarian tumours. However, the proportions of borderline and malignant ovarian tumours identified by DLR_Sig were the highest at 54.55% and 63.27%, respectively. The three-class prediction model of US-based DLR_Sig can discriminate between benign, borderline, and malignant ovarian tumours. Therefore, it may guide clinicians in determining the differential management of patients with ovarian tumours.

Traffic prediction via clustering and deep transfer learning with limited data

AbstractThis paper proposes a method based on the clustering algorithm, deep learning, and transfer learning (TL) for short‐term traffic prediction with limited data. To address the challenges posed by limited data and the complex and diverse traffic patterns observed in traffic networks, we propose a profile model based on few‐shot learning to extract each detector's unique profiles. These profiles are then used to cluster detectors with similar patterns into distinct clusters, facilitating effective learning with limited data. A Convolutional Neural Network ‐ Long Short‐Term Memory (CNN‐LSTM)‐based predictive model is proposed to learn and predict traffic volumes for each detector within a cluster. The proposed method demonstrates resilience to detector failures and has been validated using the Performance Measurement System dataset. In scenarios with less than 2 months of training data and 10% failed detectors, the mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE) for station‐level traffic volume prediction increase from 12.7 vehs/5 min, 20.9 vehs/5 min, and 10.5% to 13.9 vehs/5 min, 24.2 vehs/5 min, and 11.7%, respectively. For lane‐level traffic volume prediction, the average MAE, RMSE, and MAPE increase from 4.7 vehs/5 min, 7.7 vehs/5 min, and 15% to 5.6 vehs/5 min, 9.6 vehs/5 min, and 16.8%. Furthermore, the proposed method extends its applicability to traffic speed and occupancy prediction tasks. TL is integrated to improve speed/occupancy prediction accuracy by leveraging knowledge obtained from traffic volume, considering the correlation between traffic flow, speed, and occupancy. When less than 1 month of traffic speed/occupancy data is available for learning, the proposed method achieves an MAE, RMSE, and MAPE of 0.7 km/h, 1.3 km/h, and 1.3% for station‐level traffic speed prediction and 0.5%, 1.1%, and 11% for station‐level traffic occupancy.

Early Parkinson’s Disease Diagnosis through Hand-Drawn Spiral and Wave Analysis Using Deep Learning Techniques

Parkinson’s disease (PD) is a chronic brain disorder affecting millions worldwide. It occurs when brain cells that produce dopamine, a chemical controlling movement, die or become damaged. This leads to PD, which causes problems with movement, balance, and posture. Early detection is crucial to slow its progression and improve the quality of life for PD patients. This paper proposes a handwriting-based prediction approach combining a cosine annealing scheduler with deep transfer learning. It utilizes the NIATS dataset, which contains handwriting samples from individuals with and without PD, to evaluate six different models: VGG16, VGG19, ResNet18, ResNet50, ResNet101, and Vit. This paper compares the performance of these models based on three metrics: accuracy, precision, and F1 score. The results showed that the VGG19 model, combined with the proposed method, achieved the highest average accuracy of 96.67%.

Classification of Covid-19 using Differential Evolution Chaotic Whale Optimization based Convolutional Neural Network

COVID-19, also known as the Coronavirus disease-2019, is an transferrable disease that spreads rapidly, affecting countless individuals and leading to fatalities in this worldwide pandemic. The precise and swift detection of COVID-19 plays a crucial role in managing the pandemic's dissemination. Additionally, it is necessary to recognize COVID-19 quickly and accurately by investigating chest x-ray images. This paper proposed a Differential Evolution Chaotic Whale Optimization Algorithm (DECWOA) based Convolutional Neural Network (CNN) method for identifying and classifying COVID-19 chest X-ray images. The DECWOA based CNN model improves the accuracy and convergence speed of the algorithm. This method is evaluated {by} Chest X-Ray (CXR) dataset and attains better results in terms of accuracy, precision, sensitivity, specificity, and F1-score values of about 99.89}%, 99.83%, 99.81%, 98.92%, and 99.26% correspondingly. The result shows that the proposed DECWOA based CNN model provides accurate and quick identification and classification of COVID-19 compared to existing techniques like ResNet50, VGG-19, and Multi-Model Fusion of Deep Transfer Learning (MMF-DTL) models.

A Zero-Shot Image Classification Method of Ship Coating Defects Based on IDATLWGAN

In recent years, the defect image classification method based on deep transfer learning has been widely explored and researched, and the task of source and target domains with the same painting defect image class has been solved successfully. However, in real applications, due to the complexity and uncertainty of ship painting conditions, it is very likely that there are unknown classes of painting defects, and the traditional deep learning model cannot identify a few classes, which leads to model overfitting and reduces its generalization ability. In this paper, a zero-shot Image classification method for ship painting defects based on IDATLWGAN is proposed to identify new unknown classes of defects in the target domain. The method is based on a deep convolutional neural network combined with adversarial transfer learning. First, a preprocessed ship painting defect dataset is used as input for the domain-invariant feature extractor. Then, the domain invariant feature extractor takes domain invariant features from the source and target domains. Finally, Defect discriminators and domain alignment discriminators are employed to classify the known categories of unlabeled defects and unknown categories of unlabeled defects in the target domain and to further reduce the distance between the edge distributions of the source and target domains. The experimental results show that the proposed model in this paper extracts a better distribution of invariant features in the source and target domains compared to other existing transfer learning models. It can successfully complete the migration task and accurately recognize the painting defects of known categories and new unknown categories, which is a perfect combination of intelligent algorithms and engineering practice.

Deep transfer learning technique to detect white blood cell classification in regular clinical practice using histopathological images

Deep transfer learning technique to detect white blood cell classification in regular clinical practice using histopathological images

A few-shot based phase-batch multi-layer domain adaptation pattern recognition method

Deep transfer learning has been widely applied in the field of intelligent fault diagnosis. However, existing deep transfer learning-based diagnostic methods struggle to train reliable diagnostic models when there is a lack of data and significant distribution differences between the two domains. To address this issue, a few-shot based phase-batch multi-layer domain adaptation pattern recognition method is proposed. This method simultaneously measures the feature distribution differences of both the fully connected layers and the classification layers, thus better correcting the data domain bias. Additionally, a phase-batch training strategy and pseudo-label learning are employed to improve the convergence speed and stability of the training process. The proposed method is validated on two public datasets, Jiang Nan and Paderborn University, as well as a dataset obtained through independent experiments. It is compared with traditional feature-based transfer learning methods, the results show that the proposed method achieves higher diagnostic accuracy, faster convergence, and greater stability. Furthermore, its superior diagnostic performance in the few-shot scenario is demonstrated through experiments on a self-collected dataset.

Deep transfer learning with fuzzy ensemble approach for the early detection of breast cancer

Breast Cancer is a significant global health challenge, particularly affecting women with higher mortality compared with other cancer types. Timely detection of such cancer types is crucial, and recent research, employing deep learning techniques, shows promise in earlier detection. The research focuses on the early detection of such tumors using mammogram images with deep-learning models. The paper utilized four public databases where a similar amount of 986 mammograms each for three classes (normal, benign, malignant) are taken for evaluation. Herein, three deep CNN models such as VGG-11, Inception v3, and ResNet50 are employed as base classifiers. The research adopts an ensemble method where the proposed approach makes use of the modified Gompertz function for building a fuzzy ranking of the base classification models and their decision scores are integrated in an adaptive manner for constructing the final prediction of results. The classification results of the proposed fuzzy ensemble approach outperform transfer learning models and other ensemble approaches such as weighted average and Sugeno integral techniques. The proposed ResNet50 ensemble network using the modified Gompertz function-based fuzzy ranking approach provides a superior classification accuracy of 98.986%.

Punished for Success? A Natural Experiment of Displaying Clinical Hospital Quality on Review Platforms

The healthcare market struggles with information asymmetry, limiting patients’ ability to make informed hospital choices. Aiming to bridge this gap, review platforms like Yelp have begun displaying hospitals’ clinical quality data alongside consumer reviews. However, our research uncovers that Yelp’s introduction of maternity care clinical quality measures unexpectedly resulted in lower subsequent Yelp ratings for high-quality hospitals with insufficient staffing. Employing precise foot traffic data and transfer deep learning, we discovered that high-quality, yet understaffed, hospitals experienced a surge in patient volume, which strained their resources and diminished patient satisfaction, leading to negative reviews. This finding has significant implications, signaling the unintended consequences of revealing clinical quality measures, including potential financial losses for hospitals because of reduced federal funding. This research not only contributes to our understanding the dynamics of patient satisfaction but also, offers actionable insights for high-quality hospitals to mitigate the negative impacts of unexpected visibility on review platforms. Our research underscores the importance for patients to discern between objective clinical quality measures and self-reported subjective ratings in their decision-making process. This research applies machine learning and transfer deep learning techniques to healthcare analytics, offering a deeper understanding of the interplay between information disclosure, online reviews, patient satisfaction, and hospital management.

Deep transfer learning for fine-grained maize leaf disease classification

Machine learning (ML) can enhance agricultural yields by combating plant diseases and climate change. However, traditional image processing techniques for disease detection have limitations in robustness and generalizability. In this study, we investigate deep transfer learning for fine-grained disease classification in maize plants, which is a challenging task due to the subtle and nuanced disease patterns. We use four tailored deep learning frameworks: VGGNET, Inception V3, ResNet50, and InceptionResNetV2. ResNet50 achieves the highest validation accuracy of 87.51%, precision of 90.33%, and recall of 99.80%, demonstrating the efficacy and superiority of our approach. Our study offers an innovative solution for accurate disease classification in maize plants.

“Deep Learning Based Classification of Single-Hand South Indian Sign Language Gestures”

Gesture recognition is a branch of computer science and language technology dedicated to utilizing mathematical algorithms for the analysis of human gestures. Within the realm of non-verbal communication, the pivotal role of human arm movements and gestures remains a focal point. This research introduces advanced multi-stream deep transfer learning models tailored for identifying signs from South Indian languages, specifically Kannada, Tamil, and Telugu. The primary aim is to offer support to individuals encountering speech disorders or disabilities. The key deep transfer learning models utilized include Inception-V3, VGG-16, and ResNet-50, which have been modified and improved to attain heightened classification efficacy. The dataset comprises 35,000 images capturing single-hand gestures. In the realm of models, Inception-V3 exhibits the utmost test accuracy, standing at 91.45%, alongside a validation accuracy of 93.45% when tasked with the classification of single-hand gesture images across thirty-five distinct categories. The importance of this study lies in its prospective utility for creating an automated system that can support and improve the functional capabilities of individuals facing speech disorders or disabilities.

Deep Learning Approaches for Age-based Gesture Classification in South Indian Sign Language

This study focuses on recognizing and categorizing South Indian Sign Language gestures based on different age groups through transfer learning models. Sign language serves as a natural and expressive communication method for individuals with hearing impairments. This study intends to develop deep transfer learning models, namely Inception-V3, VGG-16, and ResNet-50, to accurately identify and classify double-handed gestures in South Indian languages, like Kannada, Tamil, and Telugu. A dataset comprising 30,000 images of double-handed gestures, with 10,000 images for each considered age group (1-7, 8-25, and 25 and above), is utilized to enhance and modify the models for improved classification performance. Amongst the tested models, Inception-V3 achieves the best performance with a test precision of 95.20% and validation accuracy of 92.45%, demonstrating its effectiveness in accurately categorizing images of double-handed gestures into ten different classes.

Multimodal deep transfer learning to predict retinal vein occlusion macular edema recurrence after anti-VEGF therapy

PurposeTo develop a multimodal deep transfer learning (DTL) fusion model using optical coherence tomography angiography (OCTA) images to predict the recurrence of retinal vein occlusion (RVO) and macular edema (ME) after three consecutive anti-VEGF therapies. MethodsThis retrospective cross-sectional study consisted of 2800 B-scan OCTA macular images collected from 140 patients with RVO-ME. The central macular thickness (CMT) > 250 μm was used as a criterion for recurrence in the three-month follow-up after three injections of anti-VEGF therapy. The qualified OCTA image preprocessing and the lesion area segmentation were performed by senior ophthalmologists. We developed and validated the clinical, DTL, and multimodal fusion models based on clinical and extracted OCTA imaging features. The performance of the models and experts predictions were evaluated using several performance metrics, including the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity. ResultsThe DTL models exhibited higher prediction efficacy than the clinical models and experts’ predictions. Among the DTL models, the Vgg19 performed better than that of the other models, with an AUC of 0.968 (95 % CI, 0.943–0.994), accuracy of 0.913, sensitivity of 0.922, and specificity of 0.902 in the validation cohort. Moreover, the fusion Vgg19 model showed the highest prediction efficacy among all the models, with an AUC of 0.972 (95 % CI, 0.946–0.997), accuracy of 0.935, sensitivity of 0.935, and specificity of 0.934 in the validation cohort. ConclusionsMultimodal fusion DTL models showed robust performance in predicting RVO-ME recurrence and may be applied to assist clinicians in determining patients’ follow-up time after anti-VEGF therapy.

Deep transfer learning with gravitational search algorithm for enhanced plant disease classification

Plant diseases annually cause damage and loss of much of the crop, if not its complete destruction, and this constitutes a significant challenge for farm owners, governments, and consumers alike. Therefore, identifying and classifying diseases at an early stage is very important in order to sustain local and global food security. In this research, we designed a new method to identify plant diseases by combining transfer learning and Gravitational Search Algorithm (GSA). Two state-of-the-art pretrained models have been adopted for extracting features in this study, which are MobileNetV2 and ResNe50V2. Multilayer feature extraction is applied in this study to ensure representations of plant leaves from different levels of abstraction for precise classification. These features are then concatenated and passed to GSA for optimizing them. Finally, optimized features are passed to Multinomial Logistic Regression (MLR) for final classification. This integration is essential for categorizing 18 different types of infected and healthy leaf samples. The performance of our approach is strengthened by a comparative analysis that incorporates features optimized by the Genetic Algorithm (GA). Additionally, the MLR algorithm is contrasted with K-Nearest Neighbors (KNN). The empirical findings indicate that our model, which has been refined using GSA, achieves very high levels of precision. Specifically, the average precision for MLR is 99.2%, while for KNN it is 98.6%. The resulting results significantly exceed those achieved with GA-optimized features, thereby highlighting the superiority of our suggested strategy. One important result of our study is that we were able to decrease the number of features by more than 50%. This reduction greatly reduces the processing requirements without sacrificing the quality of the diagnosis. This work presents a robust and efficient approach to the early detection of plant diseases. The work demonstrates the utilization of sophisticated computational methods in agriculture, enabling the development of novel data-driven strategies for plant health management, therefore enhancing worldwide food security.

CT-based intratumoral and peritumoral deep transfer learning features prediction of lymph node metastasis in non-small cell lung cancer.

The main metastatic route for lung cancer is lymph node metastasis, and studies have shown that non-small cell lung cancer (NSCLC) has a high risk of lymph node infiltration. This study aimed to compare the performance of handcrafted radiomics (HR) features and deep transfer learning (DTL) features in Computed Tomography (CT) of intratumoral and peritumoral regions in predicting the metastatic status of NSCLC lymph nodes in different machine learning classifier models. We retrospectively collected data of 199 patients with pathologically confirmed NSCLC. All patients were divided into training (n = 159) and validation (n = 40) cohorts, respectively. The best HR and DTL features in the intratumoral and peritumoral regions were extracted and selected, respectively. Support Vector Machine (SVM), k-Nearest Neighbors (KNN), Light Gradient Boosting Machine (Light GBM), Multilayer Perceptron (MLP), and Logistic Regression (LR) models were constructed, and the performance of the models was evaluated. Among the five models in the training and validation cohorts, the LR classifier model performed best in terms of HR and DTL features. The AUCs of the training cohort were 0.841 (95% CI: 0.776-0.907) and 0.955 (95% CI: 0.926-0.983), and the AUCs of the validation cohort were 0.812 (95% CI: 0.677-0.948) and 0.893 (95% CI: 0.795-0.991), respectively. The DTL signature was superior to the handcrafted radiomics signature. Compared with the radiomics signature, the DTL signature constructed based on intratumoral and peritumoral areas in CT can better predict NSCLC lymph node metastasis.

An advanced nomogram model using deep learning radiomics and clinical data for predicting occult lymph node metastasis in lung adenocarcinoma

PurposeTo evaluate the effectiveness of deep learning radiomics nomogram in distinguishing the occult lymph node metastasis (OLNM) status in clinical stage IA lung adenocarcinoma. MethodsA cohort of 473 cases of lung adenocarcinomas from two hospitals was included, with 404 cases allocated to the training cohort and 69 cases to the testing cohort. Clinical characteristics and semantic features were collected, and radiomics features were extracted from the computed tomography (CT) images. Additionally, deep transfer learning (DTL) features were generated using RseNet50. Predictive models were developed using the logistic regression (LR) machine learning algorithm. Moreover, gene analysis was conducted on RNA sequencing data from 14 patients to explore the underlying biological basis of deep learning radiomics scores. ResultThe training and testing cohorts achieved AUC values of 0.826 and 0.775 for the clinical model, 0.865 and 0.801 for the radiomics model, 0.927 and 0.885 for the DTL-radiomics model, and 0.928 and 0.898 for the nomogram model. The nomogram model demonstrated superiority over the clinical model. The decision curve analysis (DCA) revealed a net benefit in predicting OLNM for all models. The investigation into the biological basis of deep learning radiomics scores identified an association between high scores and pathways related to tumor proliferation and immune cell infiltration in the microenvironment. ConclusionsThe nomogram model, incorporating clinical-semantic features, radiomics, and DTL features, exhibited promising performance in predicting OLNM. It has the potential to provide valuable information for non-invasive lymph node staging and individualized therapeutic approaches.