Curr Res Microb Sci. 2025 Mar 27;8:100382. doi: 10.1016/j.crmicr.2025.100382. eCollection 2025.

ABSTRACT

Isoprene, a biogenic volatile compound emitted largely by plants, can form greenhouse gases when it reacts with atmospheric radicals. A significant amount of isoprene is absorbed into soil and can be degraded by soil microorganisms, but our understanding of the microbial biodegradation of isoprene in tropical ecosystems remains limited. This study investigated isoprene degradation by soil microbes indigenous to a tropical evergreen forest, focusing on those associated with coffee and tea plants grown as integrated crops and their genome characteristics in relation to their biodegradation capabilities. Following a 96-hour incubation with 7.2 × 10⁵ parts per billion by volume (ppbv) of isoprene, soil samples exhibited degradation levels ranging from 11.95 % to 36.54 %. From these soils, bacterial isolates belonging to the genera Rhodococcus and Gordonia (Actinomycetota) were recovered. These isolates demonstrated high isoprene biodegradation activity (50.3 %-69.1 % over seven days) and carried the isoA gene associated with isoprene metabolism. According to genome analysis, the organization of genes in the iso cluster was homologous, and the encoded amino acid sequences were highly similar to those of previously known isoprene-degrading members of the same genera. These findings emphasized the contribution of these widespread isoprene-degrading bacterial genera in the biodegradation of isoprene and the role of their isoprene monooxygenases in modulating atmospheric isoprene flux.

PMID:40255246 | PMC:PMC12008541 | DOI:10.1016/j.crmicr.2025.100382

OMICS. 2025 Apr 21. doi: 10.1089/omi.2025.0031. Online ahead of print.

ABSTRACT

Network medicine considers the interconnectedness of human diseases and their underlying molecular substrates. In this context, sarcoidosis and pulmonary hypertension (PH) have long been thought of as distinct diseases, but there is growing evidence of shared molecular mechanisms. This study reports on common differentially expressed genes (DEGs), regulatory elements, and pathways between the two diseases. Publicly available transcriptomic datasets for sarcoidosis (GSE157671) and PH (GSE236251) were retrieved from the Gene Expression Omnibus database. DEGs were identified using GEO2R, followed by pathway enrichment and gene interaction analyses via GeneMANIA and STRING. Importantly, a total of 13 common DEGs were identified between sarcoidosis and PH, with 7 upregulated and 6 downregulated genes. The SMAD2/3 nuclear pathway was a shared enriched pathway, suggesting a role in fibrosis and immune regulation. There were also divergences between sarcoidosis and PH. For example, gene set enrichment analysis indicated significant associations of the IFN-gamma signaling pathway with sarcoidosis and the TNF-alpha signaling with PH. miRNA network analysis identified hsa-miR-34a-5p, hsa-let-7g-5p, and hsa-miR-19a-3p as key shared regulators linked to common genes in both sarcoidosis and PH. Finally, DGIdb analysis revealed potential therapeutic candidates targeting these genes in both diseases. This study contributes to the field of drug design and discovery from a network medicine standpoint. The shared molecular links uncovered between sarcoidosis and PH in this study point to several potential biomarkers and therapeutic targets. Further experimental validation and translational medical research are called for diagnostics and drugs, which can effectively and safely help the clinical management of both diseases.

PMID:40255202 | DOI:10.1089/omi.2025.0031

Anal Chem. 2025 Apr 21. doi: 10.1021/acs.analchem.4c06796. Online ahead of print.

NO ABSTRACT

PMID:40254992 | DOI:10.1021/acs.analchem.4c06796

Brief Bioinform. 2025 Mar 4;26(2):bbaf174. doi: 10.1093/bib/bbaf174.

ABSTRACT

Encoding spatially resolved transcriptomics (SRT) data serves to identify the biological semantics of RNA expression within the tissue while preserving spatial characteristics. Depending on the analytical scenario, one may focus on different contextual structures of tissues. For instance, anatomical regions reveal consistent patterns by focusing on spatial homogeneity, while elucidating complex tumor micro-environments requires more expression heterogeneity. However, current spatial encoding methods lack consideration of the tissue context. Meanwhile, most developed SRT technologies are still limited in providing exact patterns of intact tissues due to limitations such as low resolution or missed measurements. Here, we propose STForte, a novel pairwise graph autoencoder-based approach with cross-reconstruction and adversarial distribution matching, to model the spatial homogeneity and expression heterogeneity of SRT data. STForte extracts interpretable latent encodings, enabling downstream analysis by accurately portraying various tissue contexts. Moreover, STForte allows spatial imputation using only spatial consistency to restore the biological patterns of unobserved locations or low-quality cells, thereby providing fine-grained views to enhance the SRT analysis. Extensive evaluations of datasets under different scenarios and SRT platforms demonstrate that STForte is a scalable and versatile tool for providing enhanced insights into spatial data analysis.

PMID:40254832 | DOI:10.1093/bib/bbaf174

Cureus. 2025 Mar 20;17(3):e80907. doi: 10.7759/cureus.80907. eCollection 2025 Mar.

ABSTRACT

Most patients who undergo percutaneous coronary intervention (PCI) to address coronary artery disease receive antiplatelets and anticoagulants to lower the risk of postoperative thrombotic events. Tirofiban, a glycoprotein IIb/IIIa inhibitor (GPI), has demonstrated remarkable efficacy in reducing morbidity and mortality rates in PCI postoperative care. However, it is crucial to be vigilant about potential complications associated with tirofiban, particularly thrombocytopenia. Thrombocytopenia is a serious complication that requires close monitoring of the patients' platelet count after initiation of the therapy. Regularly monitoring levels in two- to six-hour increments during the initial 24-48 hours after exposure can detect most cases of acute and potentially life-threatening thrombocytopenia. Prompt discontinuation of GPI and timely implementation of other supportive measures can help prevent further adverse events. We present a case of a 70-year-old male who presented to the Emergency Department with chest pain. Following a thorough evaluation, the patient underwent angiography, during which stent placement was performed. Administration of tirofiban resulted in profound thrombocytopenia, with platelets decreasing to 1 g/L within 24 hours. Tirofiban was promptly withdrawn, and a platelet transfusion was initiated in order to stabilize the patient's platelet level.

PMID:40255833 | PMC:PMC12009166 | DOI:10.7759/cureus.80907

Cureus. 2025 Mar 19;17(3):e80832. doi: 10.7759/cureus.80832. eCollection 2025 Mar.

ABSTRACT

Orlistat is an FDA-approved medication for obesity management that functions as a pancreatic lipase inhibitor. This medication is accessible without a prescription in numerous developed nations. As its utilization rises, the likelihood of experiencing adverse events also increases. A thorough understanding of these events is crucial for making informed decisions and ensuring effective management. We describe the case of a 23-year-old female who presented with acute pancreatitis after she started orlistat. We reviewed the association between orlistat use and acute pancreatitis, analyzing clinical cases and potential risk factors. By examining available medical literature and case studies, our study aims to provide insights into the correlation between orlistat therapy and the manifestation of acute pancreatitis.

PMID:40255727 | PMC:PMC12007682 | DOI:10.7759/cureus.80832

Can J Gastroenterol Hepatol. 2025 Apr 10;2025:2702089. doi: 10.1155/cjgh/2702089. eCollection 2025.

ABSTRACT

CKD-495 is a newly developed drug extracted from Cinnamomum cassia Presl. This phase II study assessed the clinical benefits of CKD-495 in the treatment of acute and chronic gastritis. This study randomly assigned 250 patients with endoscopically-proven gastric mucosal erosion to five groups. The groups received either 75 mg or 150 mg of CKD-495, 100 mg of rebamipide, 60 mg of Artemisiae argyi folium 95% ethanol ext. (20 ⟶ 1) (Stillen; Dong-A ST Co., Ltd., Seoul, Korea), or placebo for 2 weeks, respectively. The primary endpoint was the erosion improvement rate, and the secondary endpoints were erosion cure rates, improvement rates of gastrointestinal symptoms, edema, redness, and hemorrhage. Drug-related adverse events were evaluated. The endoscopic erosion improvement rate was significantly higher in the 75 mg CKD-495 group than in the other groups in both the full analysis set (73% vs. 41%, 45%, 52%, 48% for the 75 mg CKD-495, 150 mg CKD-495, placebo, 60 mg Stillen, and 100 mg rebamipide groups, respectively) and the per-protocol set (PPS) (75% vs. 37%, 45%, 51%, 50%). The cure rate of gastric erosion was significantly higher in the 75 mg CKD-495 group than in the other groups. The improvement rates of hemorrhage erosion were significantly higher in the 150-mg CKD-495 group. No significant differences were observed in the safety profiles. No serious adverse events or drug reactions were observed. These results demonstrate that 75 mg of CKD-495 has excellent efficacy for the treatment of endoscopic and symptomatic improvements for acute and chronic gastritis. Trial Registration: ClinicalTrials.gov identifier: NCT03437785.

PMID:40255536 | PMC:PMC12006708 | DOI:10.1155/cjgh/2702089

Drug Metab Dispos. 2025 Mar 12;53(5):100063. doi: 10.1016/j.dmd.2025.100063. Online ahead of print.

ABSTRACT

Fluoroquinolone tissue distribution and cellular accumulation are hindered by efflux transporters, including ATP-binding cassette subfamily B member 1 (ABCB1), ATP-binding cassette subfamily G member 2 (ABCG2), and ATP-binding cassette subfamily C member 4 (ABCC4). Genetic polymorphisms (single-nucleotide polymorphisms) can impact transporter activity, leading to interindividual variability in the systemic and cellular pharmacokinetics of their substrates. This study assesses the impact of these transporters on moxifloxacin and ciprofloxacin (CIP) cellular accumulation in vitro, and the effect of common single-nucleotide polymorphisms in ABCB1 [c.1199G>A (rs2229109); common haplotype c.1236C>T (rs1128503), c.2677G>T/A (rs2032582), and c.3435C>T (rs1045642)] and ABCG2 [c.421C>A (rs2231142)]. Recombinant human embryonic kidney (HEK) cell lines overexpressing wild-type or variant transporters were generated via stable plasmid transfection. The impact of transporter overexpression on fluoroquinolone cell disposition was assessed through accumulation experiments in the presence of specific inhibitors to establish the link between transporter expression and differential accumulation. Results indicated that ABCB1 overexpression reduced moxifloxacin cellular concentration by 30% but inconsistently with that of CIP and that zosuquidar or elacridar reversed these effects. ABCG2 had no impact. ABCC4 markedly reduced CIP accumulation by 25%, even at the basal level, an effect reversed by MK517. Contrarily to the wild-type and the c.1199A carriers, ABCB1 CGT and TTT variants did not reduce antibiotic accumulation. In conclusion, moxifloxacin and CIP are substrates of the wild-type and 1199G>A ABCB1, while CGT and TTT haplotypes had a marginal impact on fluoroquinolone transport by ABCB1. CIP is a preferential ABCC4 substrate. Because of the large body distribution of these transporters, our findings may help rationalize their role and the impact of their polymorphisms in fluoroquinolone disposition in tissues and cells. SIGNIFICANCE STATEMENT: This study demonstrates that moxifloxacin and ciprofloxacin are substrates of ABCB1, with ciprofloxacin also transported by ABCC4. Specific ABCB1 polymorphisms (CGT and TTT haplotypes) reduce the ABCB1 transport capacity toward fluoroquinolones. These findings highlight the importance of considering ABCB1 and ABCC4 inducers or inhibitors, which may affect fluoroquinolone disposition in tissues and cells, as well as ABCB1 polymorphisms that could explain interindividual variability in pharmacokinetic profiles.

PMID:40253817 | DOI:10.1016/j.dmd.2025.100063

Clin Pharmacokinet. 2025 Apr 21. doi: 10.1007/s40262-025-01505-4. Online ahead of print.

ABSTRACT

Pulmonary complications are the leading cause of morbidity and mortality in pediatric patients with cystic fibrosis. Altered pharmacokinetic parameters in this population, as well as high inter- and intra-individual variability, complicate the optimization of anti-infective treatments. In this review, we aim to summarize and describe all anti-infective population pharmacokinetic (popPK) models applied to pediatric populations with cystic fibrosis. Our objectives were to identify the most-reported structural models and retained covariates and to compare the dosing regimens used in clinical routine with those recommended in literature and guidelines. A literature search was done through the PubMed database from inception to August 2024. Studies were retained only if they complied with the inclusion and exclusion criteria. The review included 21 popPK models covering the pharmacokinetic profiles of eight different molecules. Among these, five are recommended antibiotics for treating pulmonary infections in patients with cystic fibrosis. All models incorporated body composition and/or renal function measures as covariates in their pharmacokinetic parameter equations. Standard dosing regimens in the studies were consistent with guidelines and literature recommendations. This is the first review summarizing and describing all anti-infective popPK models in pediatric patients with cystic fibrosis. Improved estimation of pharmacokinetic parameters and a clearer understanding of variability sources will enhance the optimization of antibiotic treatment in clinical practice. Finally, the impact of new targeted therapies on the management of this population will have to be closely monitored in the years ahead.

PMID:40254714 | DOI:10.1007/s40262-025-01505-4

J Cyst Fibros. 2025 Apr 19:S1569-1993(25)00772-6. doi: 10.1016/j.jcf.2025.04.006. Online ahead of print.

ABSTRACT

BACKGROUND: Cystic fibrosis related diabetes (CFRD) is the most common comorbidity of cystic fibrosis (CF) still, its pathogenesis is poorly understood. Recent studies have suggested that although pancreatic insufficiency is an important explanation for CFRD development, inherent pancreatic islet cell dysfunction may play a role. This study aimed to systematically compile current data regarding the impact of pancreatic islet cell dysfunction on the development of CFRD.

METHODS: A systematic search was conducted in PubMed and Embase. The resulting articles were screened for relevant experimental design and outcomes. Articles underwent data extraction and quality assessment before compilation and analysis of the results.

RESULTS: A total of 268 articles were initially screened and 19 studies conducted between 2006-2022 were finally included in this review. Half of the studies in human tissue and most of the studies in animal tissue could detect CFTR in the islets. Similarly, half of the publications in human islets and most studies in animal islets detect decreased insulin secretion with inhibition/mutation of CFTR.

CONCLUSIONS: The literature on the role of islet CFTR is contradictory. However, a pattern emerges where CFTR loss-of-function mutations have the potential to negatively affect islet cell function in a way that, together with previously described exocrine damage occurring in CF, could play a part in the development of CFRD.

PMID:40254519 | DOI:10.1016/j.jcf.2025.04.006

Ann Surg Oncol. 2025 Apr 20. doi: 10.1245/s10434-025-17290-0. Online ahead of print.

ABSTRACT

BACKGROUND: Deep learning can assess the individual survival prognosis in sizeable datasets with intricate underlying processes. However, studies exploring the performance of deep learning survival in papillary thyroid cancer (PTC) are lacking. This study aimed to construct a deep learning model based on clinical risk factors for survival prediction in patients with PTC.

METHODS: A Cox proportional hazards deep neural network (DeepSurv) was developed and validated by using consecutive patients with PTC from 17 US Surveillance, Epidemiology, and End Results Program (SEER) cancer registries (2000-2020). The performance of the DeepSurv model was further validated on two external test datasets from the University of Texas MD Anderson Cancer Center (MDACC) and The Cancer Genome Atlas (TCGA). Using the survival risk scores at 10 years predicted by the DeepSurv model, we classified patients with PTC into low-risk and high-risk groups and explored their overall survival (OS).

RESULTS: The concordance index of the DeepSurv model for predicting OS was 0.798 in the SEER test dataset, 0.893 in the MDACC dataset, and 0.848 in the TCGA dataset. The DeepSurv model was capable of classifying patients with PTC into low-risk and high-risk groups according to the survival risk scores at 10 years. Patients in the high-risk group had significantly worse OS than patients in the low-risk group in all three test datasets (all P < 0.001).

CONCLUSION: The DeepSurv model was capable of classifying patients with PTC into low-risk and high-risk groups, which may provide important prognostic information for personalized treatment in patients with PTC.

PMID:40254654 | DOI:10.1245/s10434-025-17290-0

Sci Rep. 2025 Apr 21;15(1):13657. doi: 10.1038/s41598-025-98111-y.

ABSTRACT

Accurate traffic sign recognition (TSR) is critical for enhancing the safety and reliability of autonomous driving systems. This study proposes an optimized YOLOv5-based framework to address challenges such as small-scale detection, environmental variability, and real-time processing constraints. Three key innovations are introduced: (1) k-means++ clustering for anchor box optimization, achieving a 77.55% average IoU (vs. 75.95% for traditional k-means) to enhance small-target detection; (2) comprehensive comparative analysis of YOLOv5 variants (s/m/x), revealing precision-speed trade-offs (99.3-99.5% mAP@0.5 vs. 32-45 ms inference time) for deployment flexibility; and (3) systematic hyperparameter tuning to maximize robustness across diverse scenarios. Leveraging the CCTSDB dataset (13,830 annotated images), experiments demonstrate the framework's superiority: it attains 98.1% mean average precision (mAP), 98.6% recall, and 99.3% precision, outperforming Faster-RCNN and SSD by 5-8% in mAP while maintaining 45 FPS throughput. The YOLOv5s variant achieves optimal balance with 99.3% mAP@0.5 and 32 ms per-image inference, validated through rigorous statistical analysis (Tukey HSD). Robust performance in challenging conditions (e.g., small sample, backlit sample, foggy scenes) is evidenced by detection confidence exceeding 0.90. These results highlight the framework's applicability in latency-sensitive intelligent transportation systems.

PMID:40254650 | DOI:10.1038/s41598-025-98111-y

J Biophotonics. 2025 Apr 20:e70050. doi: 10.1002/jbio.70050. Online ahead of print.

ABSTRACT

This study introduces a 200 kHz swept-source optical coherence tomography system-based fast optical coherence tomography angiography (OCTA) protocol for intraoral imaging by integrating an efficient Intraoral Micro-Angiography Super-Resolution Transformer (IMAST) model. This protocol reduces acquisition time to ~0.3 s by reducing the spatial sampling resolution, thereby minimizing motion artifacts while maintaining a field of view and image quality. The IMAST model utilizes a transformer-based architecture combined with convolutional operations to reconstruct high-resolution intraoral OCTA images from reduced-resolution scans. Experimental results from various intraoral sites and conditions show the model's robustness and high performance in enhancing image quality compared to existing deep-learning methods. Besides, IMAST shows advantages in model complexity, inference time, and computational cost, underscoring its suitability for clinical environments. These findings support the potential of our approach for noninvasive oral disease diagnosis, reducing patient discomfort and facilitating early detection of malignancies, thus serving as a valuable tool for oral assessment.

PMID:40254547 | DOI:10.1002/jbio.70050

Prog Retin Eye Res. 2025 Apr 18:101358. doi: 10.1016/j.preteyeres.2025.101358. Online ahead of print.

ABSTRACT

The Bruch's membrane (BM) is an acellular, extracellular matrix that lies between the choroid and retinal pigment epithelium (RPE). The BM plays a critical role in retinal health, performing various functions including biomolecule diffusion and RPE support. The BM is also involved in many retinal diseases, and insights into BM dysfunction allow for further understanding of the pathophysiology of various chorioretinal pathologies. Thus, characterization of the BM serves as an important area of research to further understand its involvement in retinal disease. In this article, we provide a review of various advancements in characterizing and visualizing the BM. We provide an overview of the BM in retinal health, as well as changes observed in aging and disease. We then describe current state-of-the-art imaging modalities and advances to further visualize the BM including various types of optical coherence tomography imaging, near-infrared reflectance (NIR), and autofluorescence imaging and tissue matrix-assisted laser desorption/ionization imaging mass spectrometry (MALDI-IMS). Following advances in imaging of the BM, we describe animal, cellular, and synthetic models that have been developed to further visualize the BM. Following this section, we provide an overview of deep learning in retinal imaging and describe advances in computational and artificial intelligence (AI) techniques to provide automated segmentation of the BM and BM opening. We conclude this section considering the clinical implications of these segmentation techniques. Ultimately, the diverse advances aimed to further characterize the BM may allow for deeper insights into the involvement of this critical structure in retinal health and disease.

PMID:40254245 | DOI:10.1016/j.preteyeres.2025.101358

SLAS Technol. 2025 Apr 18:100287. doi: 10.1016/j.slast.2025.100287. Online ahead of print.

ABSTRACT

BACKGROUND: Processing scanned documents in electronic health records (EHR) was one of the problem in hospital network information management systems (HNIMS). To overcome this difficulty, the complex interactions among natural language processing (NLP), optical character recognition (OCR) and image preprocessing was used.

OBJECTIVE: The goal is to investigate the possibilities of improving diagnostic efficiency and accuracy in healthcare settings by using NLP technologies into HNIMS. These individuals received diagnoses for a wide range of sleep problems. The data collected were converted into scanned PDF images which were then preprocessed by using gray scaling and OCR. Bag of Words (BoW) is used to extract the featured data.

METHOD: Reports are divided among 70% training and 30% test sets for NLP model evaluation. By employing a hidden Bayesian technique on the development set, we suggest a novel hidden Bayesian integrated dense Bi-LSTM (HB-DBi-LSTM) strategy for optimizing bag-of-words models. A 6:1 ratio is further separated for training and validation sets in deep learning-based sequence models because of their high computing requirements. After 100 epochs of Adam optimization, the dense Bi-LSTM model is trained.

RESULT: The models are evaluated assessed at the segment level for AHI and SaO2 for ROC and AUROC on test sets. In the finding assessment phase, the detection capacity of the suggested model is evaluated using many criteria, such as F1-score (0.9637), accuracy (0.9321), recall (0.9421) and precision (0.9532). To evaluate information extraction, a document-level examination is also carried out.

CONCLUSION: To improve diagnostic speed and accuracy, especially when handling scanned documents in EHR, it emphasizes the critical need for strong natural language processing (NLP) systems inside HNIMS.

PMID:40254184 | DOI:10.1016/j.slast.2025.100287

Exp Eye Res. 2025 Apr 18:110394. doi: 10.1016/j.exer.2025.110394. Online ahead of print.

ABSTRACT

In certain ocular conditions, such as in eyes with keratoconus or after corneal laser surgery, Higher Order Aberrations (HOAs) may be dramatically elevated. Accurately recording interpretable wavefronts in such highly aberrated eyes using Shack-Hartmann sensor is a challenging task. While there are studies that have applied deep neural networks to Shack-Hartmann wavefront reconstructions, they have been limited to low resolution and small dynamic range cases. In this study, we introduce a multi-task learning scheme for High-Resolution and High Dynamic Range Shack-Hartmann wavefront reconstruction using a modified attention-UNet (HR-HDR-SHUNet), which outputs a wavefront map along with Zernike coefficients simultaneously. The HR-HDR-SHUNet was evaluated on three large datasets with different levels of HOAs (regularly, highly, and extremely aberrated), with successful reconstruction of all aberrated wavefronts, at the same time achieving significantly higher accuracy than both traditional methods and other deep learning networks; it is also computationally more efficient than the latter.

PMID:40254120 | DOI:10.1016/j.exer.2025.110394

J Environ Manage. 2025 Apr 19;382:125441. doi: 10.1016/j.jenvman.2025.125441. Online ahead of print.

ABSTRACT

The use of data driven deep learning models to predict and monitor Harmful Algal Blooms (HABs) has evolved over the years due to increasing technologies, availability of high frequency data, and statistical prowess. Despite the prowess of these data driven models, they are limited by inherent model structure and uncertainty in the generating process. To overcome the limitations of data driven models, in this research, we introduced the concept of data assimilation (DA) to account for model errors and incorporate new observations into the data driven deep learning HABs prediction model. Data assimilation is a computational method that enhances the precision of predictions in dynamic systems by combining real-time observations with model forecasts. In this study, we developed 100 Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU) to make one-day ahead prediction of chlorophyll-a, an indicator of HABs, using high-frequency pH, temperature, specific conductivity, turbidity, dissolved oxygen, saturated dissolved oxygen, and oxidation-reduction potential (ORP) data. We used an Ensemble Kalman Filter (EnKF) approach to assimilate chlorophyll-a observations of greater confidence into the HABs prediction model. We explored different assimilation frequencies to observe the appropriate timesteps required for introducing new information into the modeling system. The results showed improved chlorophyll-a prediction, as forecasted by the system when DA is applied. We found that increasing assimilation frequency tends to provide improved chlorophyll-a prediction, with daily assimilation having RMSE of 0.03 μg/l for GRU and 0.02 μg/l for LSTM, while monthly assimilation resulted in RMSE of 3.63 μg/l for GRU and 3.59 μg/l for LSTM. The study revealed the potential application of DA strategy to enhance the accuracy and reliability of deep learning models for HABs monitoring. In the presence of new chlorophyll-a observations, findings from this research inform on the appropriate frequency to which such information can be incorporated into a HABs prediction model framework. This process ensures that the model provides timely and accurate predictions to support effective HABs management and decision-making efforts.

PMID:40254001 | DOI:10.1016/j.jenvman.2025.125441

Comput Biol Med. 2025 Apr 19;191:110141. doi: 10.1016/j.compbiomed.2025.110141. Online ahead of print.

ABSTRACT

In many medical and pharmaceutical processes, continuous hygiene monitoring is crucial, often involving the manual detection of microorganisms in agar dishes by qualified personnel. Although deep learning methods hold promise for automating this task, they frequently encounter a shortage of sufficient training data, a prevalent challenge in colony detection. To overcome this limitation, we propose a novel pipeline that combines generative data augmentation with few-shot detection. Our approach aims to significantly enhance detection performance, even with (very) limited training data. A main component of our method is a diffusion-based generator model that inpaints synthetic bacterial colonies onto real agar plate backgrounds. This data augmentation technique enhances the diversity of training data, allowing for effective model training with only 25 real images. Our method outperforms common training-techniques, demonstrating a +0.45 mAP improvement compared to training from scratch, and a +0.15 mAP advantage over the current SOTA in synthetic data augmentation. Additionally, we integrate a decoupled feature classification strategy, where class-agnostic detection is followed by lightweight classification via a feed-forward network, making it possible to detect and classify colonies with minimal examples. This approach achieves an AP50 score of 0.7 in a few-shot scenario on the AGAR dataset. Our method also demonstrates robustness to various image corruptions, such as noise and blur, proving its applicability in real-world scenarios. By reducing the need for large labeled datasets, our pipeline offers a scalable, efficient solution for colony detection in hygiene monitoring and biomedical research, with potential for broader applications in fields where rapid detection of new colony types is required.

PMID:40253923 | DOI:10.1016/j.compbiomed.2025.110141

Comput Med Imaging Graph. 2025 Apr 14;123:102552. doi: 10.1016/j.compmedimag.2025.102552. Online ahead of print.

ABSTRACT

Radiographic imaging is a non-invasive technique of considerable importance for evaluating tumor treatment response. However, redundancy in CT data and the lack of labeled data make it challenging to accurately assess the response of locally advanced rectal cancer (LARC) patients to neoadjuvant chemoradiotherapy (nCRT) using current imaging indicators. In this study, we propose a novel learning framework to automatically predict the response of LARC patients to nCRT. Specifically, we develop a deep learning network called the Expand Intensive Attention Network (EIA-Net), which enhances the network's feature extraction capability through cascaded 3D convolutions and coordinate attention. Instance-oriented collaborative self-supervised learning (IOC-SSL) is proposed to leverage unlabeled data for training, reducing the reliance on labeled data. In a dataset consisting of 1,575 volumes, the proposed method achieves an AUC score of 0.8562. The dataset includes two distinct parts: the self-supervised dataset containing 1,394 volumes and the supervised dataset comprising 195 volumes. Analysis of the lifetime predictions reveals that patients with pathological complete response (pCR) predicted by EIA-Net exhibit better overall survival (OS) compared to non-pCR patients with LARC. The retrospective study demonstrates that imaging-based pCR prediction for patients with low rectal cancer can assist clinicians in making informed decisions regarding the need for Miles operation, thereby improving the likelihood of anal preservation, with an AUC of 0.8222. These results underscore the potential of our method to enhance clinical decision-making, offering a promising tool for personalized treatment and improved patient outcomes in LARC management.

PMID:40253816 | DOI:10.1016/j.compmedimag.2025.102552