Deep learning
Enhancing machine learning performance in cardiac surgery ICU: Hyperparameter optimization with metaheuristic algorithm
PLoS One. 2025 Feb 10;20(2):e0311250. doi: 10.1371/journal.pone.0311250. eCollection 2025.
ABSTRACT
The healthcare industry is generating a massive volume of data, promising a potential goldmine of information that can be extracted through machine learning (ML) techniques. The Intensive Care Unit (ICU) stands out as a focal point within hospitals and provides a rich source of data for informative analyses. This study examines the cardiac surgery ICU, where the vital topic of patient ventilation takes center stage. In other words, ventilator-supported breathing is a fundamental need within the ICU, and the limited availability of ventilators in hospitals has become a significant issue. A crucial consideration for healthcare professionals in the ICU is prioritizing patients who require ventilators immediately. To address this issue, we developed a prediction model using four ML and deep learning (DL) models-LDA, CatBoost, Artificial Neural Networks (ANN), and XGBoost-that are combined in an ensemble model. We utilized Simulated Annealing (SA) and Genetic Algorithm (GA) to tune the hyperparameters of the ML models constructing the ensemble. The results showed that our approach enhanced the sensitivity of the tuned ensemble model to 85.84%, which are better than the results of the ensemble model without hyperparameter tuning and those achieved using AutoML model. This significant improvement in model performance underscores the effectiveness of our hybrid approach in prioritizing the need for ventilators among ICU patients.
PMID:39928609 | DOI:10.1371/journal.pone.0311250
Addressing imbalanced data classification with Cluster-Based Reduced Noise SMOTE
PLoS One. 2025 Feb 10;20(2):e0317396. doi: 10.1371/journal.pone.0317396. eCollection 2025.
ABSTRACT
In recent years, the challenge of imbalanced data has become increasingly prominent in machine learning, affecting the performance of classification algorithms. This study proposes a novel data-level oversampling method called Cluster-Based Reduced Noise SMOTE (CRN-SMOTE) to address this issue. CRN-SMOTE combines SMOTE for oversampling minority classes with a novel cluster-based noise reduction technique. In this cluster-based noise reduction approach, it is crucial that samples from each category form one or two clusters, a feature that conventional noise reduction methods do not achieve. The proposed method is evaluated on four imbalanced datasets (ILPD, QSAR, Blood, and Maternal Health Risk) using five metrics: Cohen's kappa, Matthew's correlation coefficient (MCC), F1-score, precision, and recall. Results demonstrate that CRN-SMOTE consistently outperformed the state-of-the-art Reduced Noise SMOTE (RN-SMOTE), SMOTE-Tomek Link, and SMOTE-ENN methods across all datasets, with particularly notable improvements observed in the QSAR and Maternal Health Risk datasets, indicating its effectiveness in enhancing imbalanced classification performance. Overall, the experimental findings indicate that CRN-SMOTE outperformed RN-SMOTE in 100% of the cases, achieving average improvements of 6.6% in Kappa, 4.01% in MCC, 1.87% in F1-score, 1.7% in precision, and 2.05% in recall, with setting SMOTE's neighbors' number to 5.
PMID:39928607 | DOI:10.1371/journal.pone.0317396
Prediction of Intensive Care Length of Stay for Surviving and Nonsurviving Patients Using Deep Learning
Crit Care Med. 2025 Feb 7. doi: 10.1097/CCM.0000000000006588. Online ahead of print.
ABSTRACT
OBJECTIVES: Length of stay (LOS) models support evaluating ICU care; however, current benchmarking models fail to consider differences in LOS between surviving and nonsurviving patients, which can lead to biased predictions toward the surviving population. We aim to develop a model addressing this as well as documentation bias to improve ICU benchmarking.
DESIGN: The Critical Care Outcomes Prediction Model (CCOPM) LOS uses patient characteristics, vitals, and laboratories during the first 24 hours of ICU admission to predict LOS in the hospital and ICU using a deep learning framework for modeling time to events with competing risk. Data was randomly divided into training, validation, and test (hold out) sets in a 2:1:1 ratio.
SETTING: Electronic ICU Research Institute database from participating tele-critical care programs.
PATIENTS: Six hundred sixty-nine thousand eight hundred seventy-six ICU admissions pertaining to 628,815 patients from 329 ICUs in 194 U.S. hospitals, from 2017 to 2019.
INTERVENTIONS: None.
MEASUREMENTS AND MAIN RESULTS: Model performance was assessed using the coefficient of determination (R2), concordance index, mean absolute error, and calibration. For individual stays in the test set, the ICU LOS model presented R2 = 0.29 and 0.23 for surviving and nonsurviving populations, respectively, at the individual level and R2 = 0.48 and 0.23 at the ICU level. Conversely, hospital LOS model presented R2 = 0.46 and 0.52 at the individual level and R2 = 0.71 and 0.64 at the ICU level. In the subset of the test set containing predictions from Acute Physiology and Chronic Health Evaluation (APACHE) IVb, R2 of ICU LOS for surviving and nonsurviving populations was, respectively, 0.30 and 0.23 for the CCOPM and 0.16 and zero for APACHE IVb. For hospital LOS, the values were R2 = 0.39 and 0.40 for the CCOPM and 0.27 and zero for APACHE IVb.
CONCLUSIONS: This novel LOS model represents a step forward in achieving more equitable benchmarking across diverse ICU settings with varying risk profiles.
PMID:39928543 | DOI:10.1097/CCM.0000000000006588
Deformation registration based on reconstruction of brain MRI images with pathologies
Med Biol Eng Comput. 2025 Feb 10. doi: 10.1007/s11517-025-03319-9. Online ahead of print.
ABSTRACT
Deformable registration between brain tumor images and brain atlas has been an important tool to facilitate pathological analysis. However, registration of images with tumors is challenging due to absent correspondences induced by the tumor. Furthermore, the tumor growth may displace the tissue, causing larger deformations than what is observed in healthy brains. Therefore, we propose a new reconstruction-driven cascade feature warping (RCFW) network for brain tumor images. We first introduce the symmetric-constrained feature reasoning (SFR) module which reconstructs the missed normal appearance within tumor regions, allowing a dense spatial correspondence between the reconstructed quasi-normal appearance and the atlas. The dilated multi-receptive feature fusion module is further introduced, which collects long-range features from different dimensions to facilitate tumor region reconstruction, especially for large tumor cases. Then, the reconstructed tumor images and atlas are jointly fed into the multi-stage feature warping module (MFW) to progressively predict spatial transformations. The method was performed on the Multimodal Brain Tumor Segmentation (BraTS) 2021 challenge database and compared with six existing methods. Experimental results showed that the proposed method effectively handles the problem of brain tumor image registration, which can maintain the smooth deformation of the tumor region while maximizing the image similarity of normal regions.
PMID:39928283 | DOI:10.1007/s11517-025-03319-9
Smart IoT-based snake trapping device for automated snake capture and identification
Environ Monit Assess. 2025 Feb 10;197(3):258. doi: 10.1007/s10661-025-13722-2.
ABSTRACT
The threat of snakebites to public health, particularly in tropical and subtropical regions, requires effective mitigation strategies to avoid human-snake interactions. With the development of an IoT-based smart snake-trapping device, an innovative non-invasive solution for preventing snakebites is presented, autonomously capturing and identifying snakes. Using artificial intelligence (AI) and Internet of Things (IoT) technologies, the entire system is designed to improve the safety and efficiency of snake capture, both in rural and urban areas. A camera and sensors are installed in the device to detect heat and vibration signatures, mimicking the natural prey of snakes using tungsten wire and vibration motors to attract them into the trap. A real-time classification algorithm based on deep learning determines whether a snake is venomous or non-venomous as soon as the device detects it. This algorithm utilizes a transfer learning approach using a convolutional neural network (CNN) and has been trained using snake images, achieving an accuracy of 91.3%. As a result of this identification process, appropriate actions are taken, such as alerting authorities or releasing non-venomous snakes into the environment in a safe manner. Through the integration of IoT technology, users can receive real-time notifications and data regarding the trap via a smartphone application. The system's connectivity allows for timely intervention in case of venomous species, reducing snakebite risks. Additionally, the system provides information regarding snake movement patterns and species distribution, contributing to the study of broader ecological issues. An automated and efficient method of managing snakes could be implemented in snakebite-prone regions with the smart trapping device.
PMID:39928180 | DOI:10.1007/s10661-025-13722-2
Qualitative and Quantitative Transformer-CNN Algorithm Models for the Screening of Exhale Biomarkers of Early Lung Cancer Patients
Anal Chem. 2025 Feb 10. doi: 10.1021/acs.analchem.4c06604. Online ahead of print.
ABSTRACT
Electronic nose (E-nose) has been applied many times for exhale biomarker detection for lung cancer, which is a leading cause of cancer-related mortality worldwide. These noninvasive breath testing techniques can be used for the early diagnosis of lung cancer patients and help improve their five year survival. However, there are still many key challenges to be addressed, including accurately identifying the kind of volatile organic compounds (VOCs) biomarkers in human-exhaled breath and the concentrations of these VOCs, which may vary at different stages of lung cancer. Recent research has mainly focused on E-nose based on a metal oxide semiconductor sensor array with proposed single gas qualitative and quantitative algorithms, but there are few breakthroughs in the detection of multielement gaseous mixtures. This work proposes two hybrid deep-learning models that combine the Transformer and CNN algorithms for the identification of VOC types and the quantification of their concentrations. The classification accuracy of the qualitative model reached 99.35%, precision reached 99.31%, recall was 99.00%, and kappa was 98.93%, which are all higher than those of the comparison algorithms, like AlexNet, MobileNetV3, etc. The quantitative model achieved an average R2 of 0.999 and an average RMSE of only 0.109 on the mixed gases. Otherwise, the parameter count and FLOPs of only 0.7 and 50.28 M, respectively, of the model proposed in this work were much lower than those of the comparison models. The detailed experiments demonstrated the potential of our proposed models for screening patients with early stage lung cancer.
PMID:39928114 | DOI:10.1021/acs.analchem.4c06604
Artificial intelligence-assisted diagnosis of early gastric cancer: present practice and future prospects
Ann Med. 2025 Dec;57(1):2461679. doi: 10.1080/07853890.2025.2461679. Epub 2025 Feb 10.
ABSTRACT
Gastric cancer (GC) occupies the first few places in the world among tumors in terms of incidence and mortality, causing serious harm to human health, and at the same time, its treatment greatly consumes the health care resources of all countries in the world. The diagnosis of GC is usually based on histopathologic examination, and it is very important to be able to detect and identify cancerous lesions at an early stage, but some endoscopists' lack of diagnostic experience and fatigue at work lead to a certain rate of under diagnosis. The rapid and striking development of Artificial intelligence (AI) has helped to enhance the ability to extract abnormal information from endoscopic images to some extent, and more and more researchers are applying AI technology to the diagnosis of GC. This initiative has not only improved the detection rate of early gastric cancer (EGC), but also significantly improved the survival rate of patients after treatment. This article reviews the results of various AI-assisted diagnoses of EGC in recent years, including the identification of EGC, the determination of differentiation type and invasion depth, and the identification of borders. Although AI has a better application prospect in the early diagnosis of ECG, there are still major challenges, and the prospects and limitations of AI application need to be further discussed.
PMID:39928093 | DOI:10.1080/07853890.2025.2461679
Human sleep position classification using a lightweight model and acceleration data
Sleep Breath. 2025 Feb 10;29(1):95. doi: 10.1007/s11325-025-03247-w.
ABSTRACT
PURPOSE: This exploratory study introduces a portable, wearable device using a single accelerometer to monitor twelve sleep positions. Targeted for home use, the device aims to assist patients with mild conditions such as gastroesophageal reflux disease (GERD) by tracking sleep postures, promoting healthier habits, and improving both reflux symptoms and sleep quality without requiring hospital-based monitoring.
METHODS: The study developed AnpoNet, a lightweight deep learning model combining 1D-CNN and LSTM, optimized with BN and Dropout. The 1D-CNN captures short-term movement features, while the LSTM identifies long-term temporal dependencies. Experiments were conducted on data from 15 participants performing twelve sleep positions, with each position recorded for one minute at a sampling frequency of 50 Hz. The model was evaluated using 5-Fold cross-validation and unseen participant data to assess generalization.
RESULTS: AnpoNet achieved a classification accuracy of 94.67% ± 0.80% and an F1-score of 92.94% ± 1.35%, outperforming baseline models. Accuracy was computed as the mean of accuracies obtained for three participants in the test set, averaged over five independent random seeds. This evaluation approach ensures robustness by accounting for variability in both individual participant performance and model initialization, underscoring its potential for real-world, home-based applications.
CONCLUSION: This study provides a foundation for a portable system enabling continuous, non-invasive sleep posture monitoring at home. By addressing the needs of GERD patients, the device holds promise for improving sleep quality and supporting positional therapy. Future research will focus on larger cohorts, extended monitoring durations, and user-friendly interfaces for broader adoption.
PMID:39928075 | DOI:10.1007/s11325-025-03247-w
Novel pre-spatial data fusion deep learning approach for multimodal volumetric outcome prediction models in radiotherapy
Med Phys. 2025 Feb 10. doi: 10.1002/mp.17672. Online ahead of print.
ABSTRACT
BACKGROUND: Given the recent increased emphasis on multimodal neural networks to solve complex modeling tasks, the problem of outcome prediction for a course of treatment can be framed as fundamentally multimodal in nature. A patient's response to treatment will vary based on their specific anatomy and the proposed treatment plan-these factors are spatial and closely related. However, additional factors may also have importance, such as non-spatial descriptive clinical characteristics, which can be structured as tabular data. It is critical to provide models with as comprehensive of a patient representation as possible, but inputs with differing data structures are incompatible in raw form; traditional models that consider these inputs require feature engineering prior to modeling. In neural networks, feature engineering can be organically integrated into the model itself, under one governing optimization, rather than performed prescriptively beforehand. However, the native incompatibility of different data structures must be addressed. Methods to reconcile structural incompatibilities in multimodal model inputs are called data fusion. We present a novel joint early pre-spatial (JEPS) fusion technique and demonstrate that differences in fusion approach can produce significant model performance differences even when the data is identical.
PURPOSE: To present a novel pre-spatial fusion technique for volumetric neural networks and demonstrate its impact on model performance for pretreatment prediction of overall survival (OS).
METHODS: From a retrospective cohort of 531 head and neck patients treated at our clinic, we prepared an OS dataset of 222 data-complete cases at a 2-year post-treatment time threshold. Each patient's data included CT imaging, dose array, approved structure set, and a tabular summary of the patient's demographics and survey data. To establish single-modality baselines, we fit both a Cox Proportional Hazards model (CPH) and a dense neural network on only the tabular data, then we trained a 3D convolutional neural network (CNN) on only the volume data. Then, we trained five competing architectures for fusion of both modalities: two early fusion models, a late fusion model, a traditional joint fusion model, and the novel JEPS, where clinical data is merged into training upstream of most convolution operations. We used standardized 10-fold cross validation to directly compare the performance of all models on identical train/test splits of patients, using area under the receiver-operator curve (AUC) as the primary performance metric. We used a two-tailed Student t-test to assess the statistical significance (p-value threshold 0.05) of any observed performance differences.
RESULTS: The JEPS design scored the highest, achieving a mean AUC of 0.779 ± 0.080. The late fusion model and clinical-only CPH model scored second and third highest with 0.746 ± 0.066 and 0.720 ± 0.091 mean AUC, respectively. The performance differences between these three models were not statistically significant. All other comparison models scored significantly worse than the top performing JEPS model.
CONCLUSION: For our OS evaluation, our JEPS fusion architecture achieves better integration of inputs and significantly improves predictive performance over most common multimodal approaches. The JEPS fusion technique is easily applied to any volumetric CNN.
PMID:39928034 | DOI:10.1002/mp.17672
Deep Learning for Antimicrobial Peptides: Computational Models and Databases
J Chem Inf Model. 2025 Feb 10. doi: 10.1021/acs.jcim.5c00006. Online ahead of print.
ABSTRACT
Antimicrobial peptides are a promising strategy to combat antimicrobial resistance. However, the experimental discovery of antimicrobial peptides is both time-consuming and laborious. In recent years, the development of computational technologies (especially deep learning) has provided new opportunities for antimicrobial peptide prediction. Various computational models have been proposed to predict antimicrobial peptide. In this review, we focus on deep learning models for antimicrobial peptide prediction. We first collected and summarized available data resources for antimicrobial peptides. Subsequently, we summarized existing deep learning models for antimicrobial peptides and discussed their limitations and challenges. This study aims to help computational biologists design better deep learning models for antimicrobial peptide prediction.
PMID:39927895 | DOI:10.1021/acs.jcim.5c00006
Artificial Intelligence Applications in Cardiac CT Imaging for Ischemic Disease Assessment
Echocardiography. 2025 Feb;42(2):e70098. doi: 10.1111/echo.70098.
ABSTRACT
Artificial intelligence (AI) has transformed medical imaging by detecting insights and patterns often imperceptible to the human eye, enhancing diagnostic accuracy and efficiency. In cardiovascular imaging, numerous AI models have been developed for cardiac computed tomography (CCT), a primary tool for assessing coronary artery disease (CAD). CCT provides comprehensive, non-invasive assessment, including plaque burden, stenosis severity, and functional assessments such as CT-derived fractional flow reserve (FFRct). Its prognostic value in predicting major adverse cardiovascular events (MACE) has increased the demand for CCT, consequently adding to radiologists' workloads. This review aims to examine AI's role in CCT for ischemic heart disease, highlighting its potential to streamline workflows and improve the efficiency of cardiac care through machine learning and deep learning applications.
PMID:39927866 | DOI:10.1111/echo.70098
Introducing TEC-LncMir for prediction of lncRNA-miRNA interactions through deep learning of RNA sequences
Brief Bioinform. 2024 Nov 22;26(1):bbaf046. doi: 10.1093/bib/bbaf046.
ABSTRACT
The interactions between long noncoding RNA (lncRNA) and microRNA (miRNA) play critical roles in life processes, highlighting the necessity to enhance the performance of state-of-the-art models. Here, we introduced TEC-LncMir, a novel approach for predicting lncRNA-miRNA interaction using Transformer Encoder and convolutional neural networks (CNNs). TEC-LncMir treats lncRNA and miRNA sequences as natural languages, encodes them using the Transformer Encoder, and combines representations of a pair of microRNA and lncRNA into a contact tensor (a three-dimensional array). Afterward, TEC-LncMir treats the contact tensor as a multi-channel image, utilizes a four-layer CNN to extract the contact tensor's features, and then uses these features to predict the interaction between the pair of lncRNA and miRNA. We applied a series of comparative experiments to demonstrate that TEC-LncMir significantly improves lncRNA-miRNA interaction prediction, compared with existing state-of-the-art models. We also trained TEC-LncMir utilizing a large training dataset, and as expected, TEC-LncMir achieves unprecedented performance. Moreover, we integrated miRanda into TEC-LncMir to show the secondary structures of high-confidence interactions. Finally, we utilized TEC-LncMir to identify microRNAs interacting with lncRNA NEAT1, where NEAT1 performs as a competitive endogenous RNA of the microRNAs' targets (mRNAs) in brain cells. We also demonstrated the regulatory mechanism of NEAT1 in Alzheimer's disease via transcriptome analysis and sequence alignment analysis. Overall, our results demonstrate the effectivity of TEC-LncMir, suggest a potential regulation of miRNAs by NEAT1 in Alzheimer's disease, and take a significant step forward in lncRNA-miRNA interaction prediction.
PMID:39927859 | DOI:10.1093/bib/bbaf046
Advancements in Nanobody Epitope Prediction: A Comparative Study of AlphaFold2Multimer vs AlphaFold3
J Chem Inf Model. 2025 Feb 10. doi: 10.1021/acs.jcim.4c01877. Online ahead of print.
ABSTRACT
Nanobodies have emerged as a versatile class of biologics with promising therapeutic applications, driving the need for robust tools to predict their epitopes, a critical step for in silico affinity maturation and epitope-targeted design. While molecular docking has long been employed for epitope identification, it requires substantial expertise. With the advent of AI driven tools, epitope identification has become more accessible to a broader community increasing the risk of models' misinterpretation. In this study, we critically evaluate the nanobody epitope prediction performance of two leading models: AlphaFold3 and AlphaFold2-Multimer (v.2.3.2), highlighting their strengths and limitations. Our analysis revealed that the overall success rate remains below 50% for both tools, with AlphaFold3 achieving a modest overall improvement. Interestingly, a significant improvement in AlphaFold3's performance was observed within a specific nanobody class. To address this discrepancy, we explored factors influencing epitope identification, demonstrating that accuracy heavily depends on CDR3 characteristics, such as its 3D spatial conformation and length, which drive binding interactions with the antigen. Additionally, we assessed the robustness of AlphaFold3's confidence metrics, highlighting their potential for broader applications. Finally, we evaluated different strategies aimed at improving the prediction success rate. This study can be extended to assess the accuracy of emerging deep learning models adopting an approach similar to that of AlphaFold3.
PMID:39927847 | DOI:10.1021/acs.jcim.4c01877
NavBLIP: a visual-language model for enhancing unmanned aerial vehicles navigation and object detection
Front Neurorobot. 2025 Jan 24;18:1513354. doi: 10.3389/fnbot.2024.1513354. eCollection 2024.
ABSTRACT
INTRODUCTION: In recent years, Unmanned Aerial Vehicles (UAVs) have increasingly been deployed in various applications such as autonomous navigation, surveillance, and object detection. Traditional methods for UAV navigation and object detection have often relied on either handcrafted features or unimodal deep learning approaches. While these methods have seen some success, they frequently encounter limitations in dynamic environments, where robustness and computational efficiency become critical for real-time performance. Additionally, these methods often fail to effectively integrate multimodal inputs, which restricts their adaptability and generalization capabilities when facing complex and diverse scenarios.
METHODS: To address these challenges, we introduce NavBLIP, a novel visual-language model specifically designed to enhance UAV navigation and object detection by utilizing multimodal data. NavBLIP incorporates transfer learning techniques along with a Nuisance-Invariant Multimodal Feature Extraction (NIMFE) module. The NIMFE module plays a key role in disentangling relevant features from intricate visual and environmental inputs, allowing UAVs to swiftly adapt to new environments and improve object detection accuracy. Furthermore, NavBLIP employs a multimodal control strategy that dynamically selects context-specific features to optimize real-time performance, ensuring efficiency in high-stakes operations.
RESULTS AND DISCUSSION: Extensive experiments on benchmark datasets such as RefCOCO, CC12M, and Openlmages reveal that NavBLIP outperforms existing state-of-the-art models in terms of accuracy, recall, and computational efficiency. Additionally, our ablation study emphasizes the significance of the NIMFE and transfer learning components in boosting the model's performance, underscoring NavBLIP's potential for real-time UAV applications where adaptability and computational efficiency are paramount.
PMID:39927288 | PMC:PMC11802496 | DOI:10.3389/fnbot.2024.1513354
Diagnosis and detection of bone fracture in radiographic images using deep learning approaches
Front Med (Lausanne). 2025 Jan 24;11:1506686. doi: 10.3389/fmed.2024.1506686. eCollection 2024.
ABSTRACT
INTRODUCTION: Bones are a fundamental component of human anatomy, enabling movement and support. Bone fractures are prevalent in the human body, and their accurate diagnosis is crucial in medical practice. In response to this challenge, researchers have turned to deep-learning (DL) algorithms. Recent advancements in sophisticated DL methodologies have helped overcome existing issues in fracture detection.
METHODS: Nevertheless, it is essential to develop an automated approach for identifying fractures using the multi-region X-ray dataset from Kaggle, which contains a comprehensive collection of 10,580 radiographic images. This study advocates for the use of DL techniques, including VGG16, ResNet152V2, and DenseNet201, for the detection and diagnosis of bone fractures.
RESULTS: The experimental findings demonstrate that the proposed approach accurately identifies and classifies various types of fractures. Our system, incorporating DenseNet201 and VGG16, achieved an accuracy rate of 97% during the validation phase. By addressing these challenges, we can further improve DL models for fracture detection. This article tackles the limitations of existing methods for fracture detection and diagnosis and proposes a system that improves accuracy.
CONCLUSION: The findings lay the foundation for future improvements to radiographic systems used in bone fracture diagnosis.
PMID:39927268 | PMC:PMC11803505 | DOI:10.3389/fmed.2024.1506686
New rectum dose surface mapping methodology to identify rectal subregions associated with toxicities following prostate cancer radiotherapy
Phys Imaging Radiat Oncol. 2025 Jan 20;33:100701. doi: 10.1016/j.phro.2025.100701. eCollection 2025 Jan.
ABSTRACT
BACKGROUND AND PURPOSE: Growing evidence suggests that spatial dose variations across the rectal surface influence toxicity risk after radiotherapy. Existing methodologies employ a fixed, arbitrary physical extent for rectal dose mapping, limiting their analysis. We developed a method to standardise rectum contours, unfold them into 2D cylindrical surface maps, and identify subregions where higher doses increase rectal toxicities.
MATERIALS AND METHODS: Data of 1,048 patients with prostate cancer from the REQUITE study were used. Deep learning based automatic segmentations were generated to ensure consistency. Rectum length was standardised using linear transformations superior and inferior to the prostate. The automatic contours were validated against the manual contours through contour variation assessment with cylindrical mapping. Voxel-based analysis of the dose surface maps for the manual and automatic contours against individual rectal toxicities was performed using Student's t permutation test and Cox Proportional Hazards Model (CPHM). Significance was defined by permutation testing.
RESULTS: Our method enabled the analysis of 1,048 patients using automatic segmentation. Student's t-test showed significance (p < 0.05) in the lower posterior for clinical-reported proctitis and patient-reported bowel urgency. Univariable CPHM identified a 3 % increased risk per Gy for clinician-reported proctitis and a 2 % increased risk per Gy for patient-reported bowel urgency in the lower posterior. No other endpoints were significant.
CONCLUSION: We developed a methodology that unfolds the rectum to a 2D surface map. The lower posterior was significant for clinician-reported proctitis and patient-reported bowel urgency, suggesting that reducing the dose in the region could decrease toxicity risk.
PMID:39927213 | PMC:PMC11803856 | DOI:10.1016/j.phro.2025.100701
Inverse design of nanophotonic devices enabled by optimization algorithms and deep learning: recent achievements and future prospects
Nanophotonics. 2025 Jan 27;14(2):121-151. doi: 10.1515/nanoph-2024-0536. eCollection 2025 Feb.
ABSTRACT
Nanophotonics, which explores significant light-matter interactions at the nanoscale, has facilitated significant advancements across numerous research fields. A key objective in this area is the design of ultra-compact, high-performance nanophotonic devices to pave the way for next-generation photonics. While conventional brute-force, intuition-based forward design methods have produced successful nanophotonic solutions over the past several decades, recent developments in optimization methods and artificial intelligence offer new potential to expand these capabilities. In this review, we delve into the latest progress in the inverse design of nanophotonic devices, where AI and optimization methods are leveraged to automate and enhance the design process. We discuss representative methods commonly employed in nanophotonic design, including various meta-heuristic algorithms such as trajectory-based, evolutionary, and swarm-based approaches, in addition to adjoint-based optimization. Furthermore, we explore state-of-the-art deep learning techniques, involving discriminative models, generative models, and reinforcement learning. We also introduce and categorize several notable inverse-designed nanophotonic devices and their respective design methodologies. Additionally, we summarize the open-source inverse design tools and commercial foundries. Finally, we provide our perspectives on the current challenges of inverse design, while offering insights into future directions that could further advance this rapidly evolving field.
PMID:39927200 | PMC:PMC11806510 | DOI:10.1515/nanoph-2024-0536
A comparative analysis of the binary and multiclass classified chest X-ray images of pneumonia and COVID-19 with ML and DL models
Open Med (Wars). 2025 Feb 4;20(1):20241110. doi: 10.1515/med-2024-1110. eCollection 2025.
ABSTRACT
BACKGROUND: The highly infectious coronavirus disease 2019 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2, the seventh coronavirus. It is the longest pandemic in recorded history worldwide. Many countries are still reporting COVID-19 cases even in the fifth year of its emergence.
OBJECTIVE: The performance of various machine learning (ML) and deep learning (DL) models was studied for image-based classification of the lungs infected with COVID-19, pneumonia (viral and bacterial), and normal cases from the chest X-rays (CXRs).
METHODS: The K-nearest neighbour and logistics regression as the two ML models, and Visual Geometry Group-19, Vision transformer, and ConvMixer as the three DL models were included in the investigation to compare the brevity of the detection and classification of the cases.
RESULTS: Among the investigated models, ConvMixer returned the best result in terms of accuracy, recall, precision, F1-score and area under the curve for both binary as well as multiclass classification. The pre-trained ConvMixer model outperformed the other four models in classifying. As per the performance observations, there was 97.1% accuracy for normal and COVID-19 + pneumonia-infected lungs, 98% accuracy for normal and COVID-19 infected lungs, 82% accuracy for normal + bacterial + viral infected lungs, and 98% accuracy for normal + pneumonia infected lungs. The DL models performed better than the ML models for binary and multiclass classification. The performance of these studied models was tried on other CXR image databases.
CONCLUSION: The suggested network effectively detected COVID-19 and different types of pneumonia by using CXR imagery. This could help medical sciences for timely and accurate diagnoses of the cases through bioimaging technology and the use of high-end bioinformatics tools.
PMID:39927166 | PMC:PMC11806240 | DOI:10.1515/med-2024-1110
Artificial Intelligence - Blessing or Curse in Dentistry? - A Systematic Review
J Pharm Bioallied Sci. 2024 Dec;16(Suppl 4):S3080-S3082. doi: 10.4103/jpbs.jpbs_1106_24. Epub 2024 Dec 10.
ABSTRACT
This systematic review examines the diverse applications of AI in all areas of dentistry. The search was conducted using the terms "Artificial Intelligence," "Dentistry," "Machine learning," "Deep learning," and "Diagnostic System." Out of 607 publications analyzed from 2010 to 2024, only 13 were selected for inclusion based on their relevance and publication year. AI in dentistry offers both advantages and challenges. It enhances diagnosis, therapy, and patient outcomes through complex algorithms and massive datasets. However, issues such as data privacy, dental professional job displacement, and the necessity for thorough validation and regulation to ensure safety and efficacy remain significant concerns.
PMID:39926925 | PMC:PMC11804999 | DOI:10.4103/jpbs.jpbs_1106_24
Deep learning-assisted diagnosis of acute mesenteric ischemia based on CT angiography images
Front Med (Lausanne). 2025 Jan 24;12:1510357. doi: 10.3389/fmed.2025.1510357. eCollection 2025.
ABSTRACT
PURPOSE: Acute Mesenteric Ischemia (AMI) is a critical condition marked by restricted blood flow to the intestine, which can lead to tissue necrosis and fatal outcomes. We aimed to develop a deep learning (DL) model based on CT angiography (CTA) imaging and clinical data to diagnose AMI.
METHODS: A retrospective study was conducted on 228 patients suspected of AMI, divided into training and test sets. Clinical data (medical history and laboratory indicators) was included in a multivariate logistic regression analysis to identify the independent factors associated with AMI and establish a clinical factors model. The arterial and venous CTA images were utilized to construct DL model. A Fusion Model was constructed by integrating clinical factors into the DL model. The performance of the models was assessed using receiver operating characteristic (ROC) curves and decision curve analysis (DCA).
RESULTS: Albumin and International Normalized Ratio (INR) were associated with AMI by univariate and multivariate logistic regression (P < 0.05). In the test set, the area under ROC curve (AUC) of the clinical factor model was 0.60 (sensitivity 0.47, specificity 0.86). The AUC of the DL model based on CTA images reached 0.90, which was significantly higher than the AUC values of the clinical factor model, as confirmed by the DeLong test (P < 0.05). The Fusion Model also showed exceptional performance in terms of AUC, accuracy, sensitivity, specificity, and precision, with values of 0.96, 0.94, 0.94, 0.95, and 0.98, respectively. DCA indicated that the Fusion Model provided a greater net benefit than those of models based solely on imaging and clinical information across the majority of the reasonable threshold probabilities.
CONCLUSION: The incorporation of CTA images and clinical information into the model markedly enhances the diagnostic accuracy and efficiency of AMI. This approach provides a reliable tool for the early diagnosis of AMI and the subsequent implementation of appropriate clinical intervention.
PMID:39926426 | PMC:PMC11802816 | DOI:10.3389/fmed.2025.1510357