EJNMMI Phys. 2025 Apr 16;12(1):38. doi: 10.1186/s40658-025-00745-4.

ABSTRACT

BACKGROUND: A reduction of dose and/or acquisition duration of PET examinations is desirable in terms of radiation protection, patient comfort and throughput, but leads to decreased image quality due to poorer image statistics. Recently, different deep-learning based methods have been proposed to improve image quality of low-count PET images. For example, one such approach allows the generation of AI-enhanced PET images (AI-PET) based on ultra-low count PET/CT scans. The performance of this algorithm has so far only been clinically evaluated on patient data featuring limited scan statistics and unknown actual activity concentration. Therefore, this study investigates the performance of this deep-learning algorithm using PET measurements of a phantom resembling different lesion sizes and count statistics (from ultra-low to high) to understand the capabilities and limitations of AI-based post processing for improved image quality in ultra-low count PET imaging.

METHODS: A previously trained pix2pixHD Generative Adversarial Network was evaluated. To this end, a NEMA PET body phantom filled with two sphere-to-background activity concentration ratios (4:1 and 10:1) and two attenuation scenarios to investigate the effects of obese patients was scanned in list mode. Images were reconstructed with 13 different acquisition durations ranging from 5 s up to 900 s. Image noise, recovery coefficients, SUV-differences, image quality measurement metrics such as the Structural Similarity Index Metric, and the contrast-to-noise-ratio were assessed. In addition, the benefits of the deep-learning network over Gaussian smoothing were investigated.

RESULTS: The presented AI-algorithm is very well suitable for denoising ultra-low count PET images and for restoring structural information, but increases image noise in ultra-high count PET scans. The generated AI-PET scans strongly underestimate SUV especially in small lesions with a diameter ≤ 17 mm, while quantitative measures of large lesions ≥ 37 mm in diameter were accurately recovered. In ultra-low count or low contrast images, the AI algorithm might not be able to recognize small lesions ≤ 13 mm in diameter. In comparison to standardized image post-processing using a Gaussian filter, the deep-learning network is better suited to improve image quality, but at the same time degrades SUV accuracy to a greater extent than post-filtering and quantitative SUV accuracy varies for different lesion sizes.

CONCLUSIONS: Phantom-based validation of AI-based algorithms allows for a detailed assessment of the performance, limitations, and generalizability of deep-learning based algorithms for PET image enhancement. Here it was confirmed that the AI-based approach performs very well in denoising ultra-low count PET images and outperforms traditional Gaussian post-filtering. However, there are strong limitations in terms of quantitative accuracy and detectability of small lesions.

PMID:40237913 | DOI:10.1186/s40658-025-00745-4

Discov Oncol. 2025 Apr 16;16(1):537. doi: 10.1007/s12672-025-02329-1.

ABSTRACT

BACKGROUND: Breast cancer (BC) remains a leading cause of cancer-related mortality among women globally, with increasing incidence rates posing significant public health challenges. Recent advancements in artificial intelligence (AI) have revolutionized medical imaging, particularly in enhancing diagnostic accuracy and prognostic capabilities for BC. While multimodal imaging combined with AI has shown remarkable potential, a comprehensive analysis is needed to synthesize current research and identify emerging trends and hotspots in AI-assisted multimodal imaging for BC.

METHODS: This study analyzed literature on AI-assisted multimodal imaging in BC from January 2010 to November 2024 in Web of Science Core Collection (WoSCC). Bibliometric and visualization tools, including VOSviewer, CiteSpace, and the Bibliometrix R package, were employed to assess countries, institutions, authors, journals, and keywords.

RESULTS: A total of 80 publications were included, revealing a steady increase in annual publications and citations, with a notable surge post-2021. China led in productivity and citations, while Germany exhibited the highest citation average. The United States demonstrated the strongest international collaboration. The most productive institution and author are Radboud University Nijmegen and Xi, Xiaoming. Publications were predominantly published in Computerized Medical Imaging and Graphics, with Qian, XJ's 2021 study on BC risk prediction under deep learning frameworks being the most influential. Keyword analysis highlighted themes such as "breast cancer", "classification", and "deep learning".

CONCLUSIONS: AI-assisted multimodal imaging has significantly advanced BC diagnosis and management, with promising future developments. This study offers researchers a comprehensive overview of current frameworks and emerging research directions. Future efforts are expected to focus on improving diagnostic precision and refining therapeutic strategies through optimized imaging techniques and AI algorithms, emphasizing international collaboration to drive innovation and clinical translation.

PMID:40237900 | DOI:10.1007/s12672-025-02329-1

Radiol Artif Intell. 2025 Apr 16:e240097. doi: 10.1148/ryai.240097. Online ahead of print.

ABSTRACT

"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To present and evaluate Dafne (deep anatomic federated network), a freely available decentralized, collaborative deep learning system for the semantic segmentation of radiologic images through federated incremental learning. Materials and Methods Dafne is free software with a client-server architecture. The client side is an advanced user interface that applies the deep learning models stored on the server to the user's data and allows the user to check and refine the prediction. Incremental learning is then performed at the client's side and sent back to the server, where it is integrated into the root model. Dafne was evaluated locally, by assessing the performance gain across model generations on 38 MRI datasets of the lower legs, and through the analysis of real-world usage statistics (n = 639 use-cases). Results Dafne demonstrated a statistically improvement in the accuracy of semantic segmentation over time (average increase of the Dice Similarity Coefficient by 0.007 points/generation on the local validation set, P < .001). Qualitatively, the models showed enhanced performance on various radiologic image types, including those not present in the initial training sets, indicating good model generalizability. Conclusion Dafne showed improvement in segmentation quality over time, demonstrating potential for learning and generalization. ©RSNA, 2025.

PMID:40237599 | DOI:10.1148/ryai.240097

Radiol Artif Intell. 2025 May;7(3):e250125. doi: 10.1148/ryai.250125.

NO ABSTRACT

PMID:40237597 | DOI:10.1148/ryai.250125

J Agric Food Chem. 2025 Apr 16. doi: 10.1021/acs.jafc.4c11492. Online ahead of print.

ABSTRACT

Quality inspection of fruits and vegetables linked to food safety monitoring and quality control. Traditional chemical analysis and physical measurement techniques are reliable, they are also time-consuming, costly, and susceptible to environmental and sample changes. Hyperspectral imaging technology combined with deep learning methods can effectively overcome these problems. Compared with human evaluation, automated inspection improves inspection efficiency, reduces subjective error, and promotes the intelligent and precise fruit and vegetable quality inspection. This paper reviews reports on the application of hyperspectral imaging technology combined to deep learning methods in various aspects of fruits and vegetables quality assessment. In addition, the latest applications of these technologies in the fields of fruit and vegetable safety, internal quality, and external quality inspection are reviewed, and the challenges and future development directions of hyperspectral imaging technology combined with deep learning in this field are prospected. Hyperspectral imaging combined with deep learning has shown significant advantages in fruit and vegetable quality inspection, especially in improving inspection accuracy and efficiency. Future research should focus on reducing costs, optimizing equipment, personalizing feature extraction, and model generalizability. In addition, the development of lightweight models and the balance of accuracy, the enhancement of the database and the importance of quantitative research should also be brought to attention. These efforts will promote the wide application of hyperspectral imaging technology in fruit and vegetable inspection, improve its practicability in the actual production environment, and bring important progress for food safety and quality management.

PMID:40237548 | DOI:10.1021/acs.jafc.4c11492

mSystems. 2025 Apr 16:e0145524. doi: 10.1128/msystems.01455-24. Online ahead of print.

ABSTRACT

Deep learning is revolutionizing biomedical research by facilitating the integration of multi-omics data sets while bridging classical bioinformatics with existing knowledge. Building on this powerful potential, Zhang et al. proposed a semi-supervised learning framework called Autoencoder-Based Subtypes Detector for Cancer (ASD-cancer) to improve the multi-omics data analysis (H. Zhang, X. Xiong, M. Cheng, et al., 2024, mSystems 9:e01395-24, https://doi.org/10.1128/msystems.01395-24). By utilizing autoencoders pre-trained on The Cancer Genome Atlas data, the ASD-cancer framework outperforms the baseline model. This approach also makes the framework scalable, enabling it to process new data sets through transfer learning without retraining. This commentary explores the methodological innovations and scalability of ASD-cancer while suggesting future directions, such as the incorporation of additional data layers and the development of adaptive AI models through continuous learning. Notably, integrating large language models into ASD-cancer could enhance its interpretability, providing more profound insights into oncological research and increasing its influence in cancer subtyping and further analysis.

PMID:40237527 | DOI:10.1128/msystems.01455-24

Clin Nucl Med. 2025 Apr 16. doi: 10.1097/RLU.0000000000005898. Online ahead of print.

ABSTRACT

PURPOSE: Convolutional neural networks (CNNs) have been studied for detecting bone metastases on bone scans; however, the application of ConvNeXt and transformer models has not yet been explored. This study aims to evaluate the performance of various deep learning models, including the ConvNeXt and transformer models, in diagnosing metastatic lesions from bone scans.

MATERIALS AND METHODS: We retrospectively analyzed bone scans from patients with cancer obtained at 2 institutions: the training and validation sets (n=4626) were from Hospital 1 and the test set (n=1428) was from Hospital 2. The deep learning models evaluated included ResNet18, the Data-Efficient Image Transformer (DeiT), the Vision Transformer (ViT Large 16), the Swin Transformer (Swin Base), and ConvNeXt Large. Gradient-weighted class activation mapping (Grad-CAM) was used for visualization.

RESULTS: Both the validation set and the test set demonstrated that the ConvNeXt large model (0.969 and 0.885, respectively) exhibited the best performance, followed by the Swin Base model (0.965 and 0.840, respectively), both of which significantly outperformed ResNet (0.892 and 0.725, respectively). Subgroup analyses revealed that all the models demonstrated greater diagnostic accuracy for patients with polymetastasis compared with those with oligometastasis. Grad-CAM visualization revealed that the ConvNeXt Large model focused more on identifying local lesions, whereas the Swin Base model focused on global areas such as the axial skeleton and pelvis.

CONCLUSIONS: Compared with traditional CNN and transformer models, the ConvNeXt model demonstrated superior diagnostic performance in detecting bone metastases from bone scans, especially in cases of polymetastasis, suggesting its potential in medical image analysis.

PMID:40237349 | DOI:10.1097/RLU.0000000000005898

ArXiv [Preprint]. 2025 Mar 31:arXiv:2504.00232v1.

ABSTRACT

Pancreatic ductal adenocarcinoma (PDAC) is a highly aggressive cancer, with most cases diagnosed at stage IV and a five-year overall survival rate below 5%. Early detection and prognosis modeling are crucial for improving patient outcomes and guiding early intervention strategies. In this study, we developed and evaluated a deep learning fusion model that integrates radiology reports and CT imaging to predict PDAC risk. The model achieved a concordance index (C-index) of 0.6750 (95% CI: 0.6429, 0.7121) and 0.6435 (95% CI: 0.6055, 0.6789) on the internal and external dataset, respectively, for 5-year survival risk estimation. Kaplan-Meier analysis demonstrated significant separation (p<0.0001) between the low and high risk groups predicted by the fusion model. These findings highlight the potential of deep learning-based survival models in leveraging clinical and imaging data for pancreatic cancer.

PMID:40236838 | PMC:PMC11998856

Front Physiol. 2025 Apr 1;16:1584299. doi: 10.3389/fphys.2025.1584299. eCollection 2025.

ABSTRACT

OBJECTIVE: This study aims to develop a multimodal deep learning-based stress detection method (MMFD-SD) using intermittently collected physiological signals from wearable devices, including accelerometer data, electrodermal activity (EDA), heart rate (HR), and skin temperature. Given the unique demands and high-intensity work environment of the nursing profession, stress measurement in nurses serves as a representative case, reflecting stress levels in other high-pressure occupations.

METHODS: We propose a multimodal deep learning framework that integrates time-domain and frequency-domain features for stress detection. To enhance model robustness and generalization, data augmentation techniques such as sliding window and jittering are applied. Feature extraction includes statistical features derived from raw time-domain signals and frequency-domain features obtained via Fast Fourier Transform (FFT). A customized deep learning architecture employs convolutional neural networks (CNNs) to process time-domain and frequency-domain features separately, followed by fully connected layers for final classification. To address class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) is utilized. The model is trained and evaluated on a multimodal physiological signal dataset with stress level labels.

RESULTS: Experimental results demonstrate that the MMFD-SD method achieves outstanding performance in stress detection, with an accuracy of 91.00% and an F1-score of 0.91. Compared to traditional machine learning classifiers such as logistic regression, random forest, and XGBoost, the proposed method significantly improves both accuracy and robustness. Ablation studies reveal that the integration of time-domain and frequency-domain features plays a crucial role in enhancing model performance. Additionally, sensitivity analysis confirms the model's stability and adaptability across different hyperparameter settings.

CONCLUSION: The proposed MMFD-SD model provides an accurate and robust stress detection approach by integrating time-domain and frequency-domain features. Designed for occupational environments with intermittent data collection, it effectively addresses real-world stress monitoring challenges. Future research can explore the fusion of additional modalities, real-time stress detection, and improvements in model generalization to enhance its practical applicability.

PMID:40236827 | PMC:PMC11997569 | DOI:10.3389/fphys.2025.1584299

MethodsX. 2025 Mar 25;14:103278. doi: 10.1016/j.mex.2025.103278. eCollection 2025 Jun.

ABSTRACT

Accurate and early diagnosis of Autism Spectrum Disorder (ASD) in toddlers is crucial for effective intervention. Traditional models have shown limited success, while deep learning approaches achieve higher accuracies. Our study proposes a hybrid model combining VGG16, a pre-trained deep CNN, with an RF classifier to leverage high-level image feature extraction using the ACD image dataset on Kaggle alongside robust decision-making on the ACD Questionnaire dataset. The proposed model achieves an accuracy of 88.34 %, outperforming both standalone deep learning models like VGG16, EfficientNetB0, and AlexNet-based models as well as conventional ML models. This improvement demonstrates the effectiveness of combining feature-rich deep learning outputs with RF's ensemble-based classification. Our findings suggest that this hybrid approach is highly suitable for ASD classification tasks, enhancing the reliability of predictions in clinical settings. This research not only establishes the model as an option for ASD diagnosis but also underscores the potential of hybrid architectures that fuse deep learning with machine learning. Future research will focus on integrating multi-modal data (e.g., genetic and socio-demographic) and further testing on diverse datasets to improve generalizability. The study contributes to the growing body of evidence supporting advanced ML techniques in healthcare diagnostics, especially in neurodevelopmental disorders like ASD.

PMID:40236798 | PMC:PMC11999578 | DOI:10.1016/j.mex.2025.103278

MethodsX. 2025 Feb 17;14:103232. doi: 10.1016/j.mex.2025.103232. eCollection 2025 Jun.

ABSTRACT

Diabetic retinopathy (DR) is a serious complication of diabetes that can result in vision loss if untreated, often progressing silently without warning symptoms. Elevated blood glucose levels damage the retina's microvasculature, initiating the condition. Early detection through retinal fundus imaging, supported by timely analysis and treatment, is critical for managing DR effectively. However, manually inspecting these images is a labour-intensive and time-consuming process, making computer-aided diagnosis (CAD) systems invaluable in supporting ophthalmologists. This research introduces the Fundus Imaging Diabetic Retinopathy Classification using Deep Learning and Fennec Fox Optimization (FIDRC-DLFFO) model, which automates the identification and classification of DR. The model integrates several advanced techniques to enhance performance and accuracy.1.The proposed FIDRC-DLFFO model automates DR detection and classification by combining median filtering for noise reduction, Inception-ResNet-v2 for feature extraction, and a gated recurrent unit (GRU) for classification.2.Fennec Fox Optimization (FFO) fine-tunes the GRU hyperparameters, boosting classification accuracy, with its effectiveness demonstrated on benchmark datasets.3.The results provide insights into the model's effectiveness and potential for real-world application.

PMID:40236797 | PMC:PMC11999632 | DOI:10.1016/j.mex.2025.103232

MethodsX. 2025 Mar 29;14:103291. doi: 10.1016/j.mex.2025.103291. eCollection 2025 Jun.

ABSTRACT

The integration of IoT and deep learning has revolutionized real-time monitoring systems, particularly in public health applications such as face mask detection. With increasing public reliance on these technologies, robust and efficient frameworks are critical for ensuring compliance with health measures. Existing models, on the other hand, often have problems, such as being slow to compute, not being able to work well in a wide range of environments, and not being able to adapt well to IoT devices with limited resources. These shortcomings highlight the need for an optimized and scalable solution. To address these issues, this study utilizes three datasets: the Kaggle Face Mask Dataset, the Public Places Dataset, and the Public Videos Dataset, encompassing varied environmental conditions and use cases. The proposed framework integrates ResNet50 and MobileNetV2 architectures, optimized using the Adaptive Flame-Sailfish Optimization (AFSO) algorithm. This hybrid approach enhances detection accuracy and computational efficiency, making it suitable for real-time deployment. The novelty of the paper lies in combining AFSO with a hybrid deep learning architecture for parameter optimization and improved scalability. Performance metrics such as accuracy, sensitivity, precision, and F1-score were used to evaluate the model. The proposed framework achieved an accuracy of 97.8 % on the Kaggle dataset, significantly outperforming baseline models and demonstrating its robustness and efficiency for IoT-enabled face mask detection systems.•The article introduces a novel hybrid framework that combines ResNet50 and MobileNetV2 architectures optimized with Adaptive Flame-Sailfish Optimization (AFSO).•The system demonstrates superior performance, achieving 97.8 % accuracy on the Kaggle dataset, with improved efficiency for IoT-based real-time applications.•Validates the framework's robustness and scalability across diverse datasets, addressing computational and environmental challenges.

PMID:40236795 | PMC:PMC11999645 | DOI:10.1016/j.mex.2025.103291

MethodsX. 2025 Mar 31;14:103285. doi: 10.1016/j.mex.2025.103285. eCollection 2025 Jun.

ABSTRACT

The worldwide prevalence of glaucoma makes it a major reason for blindness thus proper early diagnosis remains essential for preventing major vision deterioration. Current glaucoma screening methods that need expert handling prove to be time-intensive and complicated before yielding appropriate diagnosis and treatment. Our system addresses these difficulties through an automated glaucoma screening platform which combines advanced segmentation methods with classification approaches. A hybrid segmentation method combines Grey Wolf Optimization Algorithm with U-Shaped Networks to obtain precise extraction of the optic disc regions in retinal fundus images. Through GWOA the network achieves optimal segmentation by adopting wolf-inspired behaviors such as circular and jumping movements to identify diverse image textures. The glaucoma classification depends on CapsNet as a deep learning model that provides exceptional image detection to ensure precise diagnosis. The combination of our method delivers 96.01 % segmentation together with classification precision which outstrips traditional approaches while indicating strong capabilities for discovering glaucoma at early stages. This automated diagnosis system elevates clinical accuracy levels through an automated screening method that solves manual process limitations. The detection framework produces better accuracy to improve clinical results in a strong effort to minimize glaucoma-induced blindness worldwide and display its capabilities in real clinical environments.•Hybrid GWOA-UNet++ for precise optic disc segmentation.•CapsNet-based classification for robust glaucoma detection.•Achieved 96.01 % accuracy, surpassing existing methods.

PMID:40236793 | PMC:PMC11999292 | DOI:10.1016/j.mex.2025.103285

Sens Actuators Rep. 2024 Dec;8:100212. doi: 10.1016/j.snr.2024.100212. Epub 2024 Jun 19.

ABSTRACT

For the 29.8 million people in the world living with HIV/AIDS and receiving antiretroviral therapy, it is crucial to monitor their HIV viral loads. To this end, rapid and portable diagnostic tools that can quantify HIV RNA are critically needed. We report herein a rapid and quantitative digital CRISPR-assisted HIV RNA detection assay that has been implemented within a portable smartphone-based device as a potential solution. Specifically, we first developed a fluorescence-based reverse transcription recombinase polymerase amplification (RT-RPA)-CRISPR assay that can efficiently detect HIV RNA at 42 °C. We then implemented this assay within a commercial stamp-sized digital chip, where RNA molecules were quantified as strongly fluorescent digital reaction wells. The isothermal reaction condition and the strong fluorescence in the digital chip simplified the design of thermal and optical modules, allowing us to engineer a palm-size device measuring 70 × 115 × 80 mm and weighing less than 0.6 kg. We also capitalized the smartphone by developing a custom app to control the device, perform the digital assay, and capture fluorescence images throughout the assay using the smartphone's camera. Moreover, we trained and verified a deep learning-based algorithm for analyzing fluorescence images and identifying positive digital reaction wells with high accuracy. Using our smartphone-enabled digital CRISPR device, we successfully detected as low as 75 copies of HIV RNA in just 15 min, showing its potential toward monitoring of HIV viral loads and aiding the global effort to combat the HIV/AIDS epidemic.

PMID:40236689 | PMC:PMC11997716 | DOI:10.1016/j.snr.2024.100212

JAMIA Open. 2025 Apr 15;8(2):ooaf028. doi: 10.1093/jamiaopen/ooaf028. eCollection 2025 Apr.

ABSTRACT

[This corrects the article DOI: 10.1093/jamiaopen/ooaf007.].

PMID:40236682 | PMC:PMC11999061 | DOI:10.1093/jamiaopen/ooaf028

Digit Biomark. 2025 Feb 28;9(1):67-74. doi: 10.1159/000544913. eCollection 2025 Jan-Dec.

ABSTRACT

The physical basis of speech production in humans requires the coordination of multiple anatomical systems, where inhalation and exhalation of air through lungs is at the core of the phenomenon. Vocalization happens during exhalation, while inhalation typically happens between speech pauses. We use deep learning models to predict respiratory signals during speech-breathing, from which the respiration rate is estimated. Bilingual data from a large clinical study (N = 1,005) are used to develop and evaluate a multivariate time series transformer model with speech encoder embeddings as input. The best model shows the predicted respiration rate from speech within ±3 BPM for 82% of test subjects. A noise-aware algorithm was also tested in a simulated hospital environment with varying noise levels to evaluate the impact on performance. This work proposes and validates speech as a virtual sensor for respiration rate, which can be an efficient and cost-effective enabler for remote patient monitoring and telehealth solutions.

PMID:40236620 | PMC:PMC11999658 | DOI:10.1159/000544913

Head Neck Tumor Segm MR Guid Appl (2024). 2025;15273:123-135. doi: 10.1007/978-3-031-83274-1_9. Epub 2025 Mar 3.

ABSTRACT

Magnetic Resonance Imaging (MRI) plays a crucial role in MRI-guided adaptive radiotherapy for head and neck cancer (HNC) due to its superior soft-tissue contrast. However, accurately segmenting the gross tumor volume (GTV), which includes both the primary tumor (GTVp) and lymph nodes (GTVn), remains challenging. Recently, two deep learning segmentation innovations have shown great promise: UMamba, which effectively captures long-range dependencies, and the nnU-Net Residual Encoder (ResEnc), which enhances feature extraction through multistage residual blocks. In this study, we integrate these strengths into a novel approach, termed 'UMambaAdj'. Our proposed method was evaluated on the HNTS-MRG 2024 challenge test set using pre-RT T2-weighted MRI images, achieving an aggregated Dice Similarity Coefficient ( D S C a g g ) of 0.751 for GTVp and 0.842 for GTVn, with a mean D S C a g g of 0.796. This approach demonstrates potential for more precise tumor delineation in MRI-guided adaptive radiotherapy, ultimately improving treatment outcomes for HNC patients. Team: DCPT-Stine's group.

PMID:40236615 | PMC:PMC11997962 | DOI:10.1007/978-3-031-83274-1_9

Eur J Med Res. 2025 Apr 15;30(1):288. doi: 10.1186/s40001-025-02501-x.

ABSTRACT

BACKGROUND: The diagnosis of idiopathic pulmonary fibrosis (IPF) is complex, which requires lung biopsy, if necessary, and multidisciplinary discussions with specialists. Clinical diagnosis of the two ailments is particularly challenging due to the impact of interobserver variability. Several studies have endeavored to utilize image-based machine learning to diagnose IPF and its subtype of usual interstitial pneumonia (UIP). However, the diagnostic accuracy of this approach lacks evidence-based support.

OBJECTIVE: We conducted a systematic review and meta-analysis to explore the diagnostic efficiency of image-based machine learning (ML) for IPF.

DATA SOURCES AND METHODS: We comprehensively searched PubMed, Cochrane, Embase, and Web of Science databases up to August 24, 2024. During the meta-analysis, we carried out subgroup analyses by imaging source (computed radiography/computed tomography) and modeling type (deep learning/other) to evaluate its diagnostic performance for IPF.

RESULTS: The meta-analysis findings indicated that in the diagnosis of IPF, the C-index, sensitivity, and specificity of ML were 0.93 (95% CI 0.89-0.97), 0.79 (95% CI 0.73-0.83), and 0.84 (95% CI 0.79-0.88), respectively. The sensitivity of radiologists/clinicians in diagnosing IPF was 0.69 (95% CI 0.56-0.79), with a specificity of 0.93 (95% CI 0.74-0.98). For UIP diagnosis, the C-index of ML was 0.91 (95% CI 0.87-0.94), with a sensitivity of 0.92 (95% CI 0.80-0.97) and a specificity of 0.92 (95%CI 0.82-0.97). In contrast, the sensitivity of radiologists/clinicians in diagnosing UIP was 0.69 (95% CI 0.50-0.84), with a specificity of 0.90 (95% CI 0.82-0.94).

CONCLUSIONS: Image-based machine learning techniques demonstrate robust data processing and recognition capabilities, providing strong support for accurate diagnosis of idiopathic pulmonary fibrosis and usual interstitial pneumonia. Future multicenter large-scale studies are warranted to develop more intelligent evaluation tools to further enhance clinical diagnostic efficiency. Trial registration This study protocol was registered with PROSPERO (CRD42022383162).

PMID:40235000 | DOI:10.1186/s40001-025-02501-x

Eur J Med Res. 2025 Apr 15;30(1):293. doi: 10.1186/s40001-025-02553-z.

ABSTRACT

BACKGROUND: There is no standard practice for intensive care admission after non-small cell lung cancer surgery. In this study, we aimed to determine the need for intensive care admission after non-small cell lung cancer surgery with deep learning models.

METHODS: The data of 953 patients who were operated for non-small cell lung cancer between January 2001 and 2023 was analyzed. Clinical, laboratory, respiratory, tumor's radiological and surgical features were included as input data in the study. The outcome data was intensive care unit admission. Deep learning was performed with the Fully Connected Neural Network algorithm and k-fold cross validation method.

RESULTS: The training accuracy value was 92.0%, the training F1 1 score of the algorithm was 86.7%, the training F1 0 value was 94.2%, and the training F1 average score was 90.5%. The test sensitivity value of the algorithm was 67.7%, the test positive predictive value was 84.0%, and the test accuracy value was 85.3%. Test F1 1 score was 75.0%, test F1 0 score was 89.5%, and test F1 average score was 82.3%. The AUC in the ROC curve created for the success analysis of the algorithm's test data was 0.83.

CONCLUSIONS: Using our method deep learning models predicted the need for intensive care unit admission with high success and confidence values. The use of artificial intelligence algorithms for the necessity of intensive care hospitalization will ensure that postoperative processes are carried out safely using objective decision mechanisms.

PMID:40234958 | DOI:10.1186/s40001-025-02553-z

BMC Med Inform Decis Mak. 2025 Apr 15;25(1):165. doi: 10.1186/s12911-025-02981-1.

ABSTRACT

BACKGROUND: Mortality prediction is critical in clinical care, particularly in intensive care units (ICUs), where early identification of high-risk patients can inform treatment decisions. While deep learning (DL) models have demonstrated significant potential in this task, most suffer from limited generalizability, which hinders their widespread clinical application. Additionally, the class imbalance in electronic health records (EHRs) complicates model training. This study aims to develop a causally-informed prediction model that incorporates underlying causal relationships to mitigate class imbalance, enabling more stable mortality predictions.

METHODS: This study introduces the CRISP model (Causal Relationship Informed Superior Prediction), which leverages native counterfactuals to augment the minority class and constructs patient representations by incorporating causal structures to enhance mortality prediction. Patient data were obtained from the public MIMIC-III and MIMIC-IV databases, as well as an additional dataset from the West China Hospital of Sichuan University (WCHSU).

RESULTS: A total of 69,190 ICU cases were included, with 30,844 cases from MIMIC-III, 27,362 cases from MIMIC-IV, and 10,984 cases from WCHSU. The CRISP model demonstrated stable performance in mortality prediction across the 3 datasets, achieving AUROC (0.9042-0.9480) and AUPRC (0.4771-0.7611). CRISP's data augmentation module showed predictive performance comparable to commonly used interpolation-based oversampling techniques.

CONCLUSION: CRISP achieves better generalizability across different patient groups, compared to various baseline algorithms, thereby enhancing the practical application of DL in clinical decision support.

TRIAL REGISTRATION: Trial registration information for the WCHSU data is available on the Chinese Clinical Trial Registry website ( http://www.chictr.org.cn ), with the registration number ChiCTR1900025160. The recruitment period for the data was from August 5, 2019, to August 31, 2021.

PMID:40234903 | DOI:10.1186/s12911-025-02981-1