J Prim Care Community Health. 2025 Jan-Dec;16:21501319241311567. doi: 10.1177/21501319241311567.

ABSTRACT

INTRODUCTION/OBJECTIVES: Individually rare, rare diseases are collectively common resulting in frequent health system use. Navigating the health system persists as a challenge. Primary care provides longitudinal contact with the health system and is placed to provide integrated rare-disease-care.

METHODS: This scoping review used Joanna Briggs Institute and PRISMA methods with a Consolidated Framework for Implementation Research based data extraction tool to find how integrated rare-disease-care is delivered, enablers and barriers to the same, in primary care settings in contemporary literature in OECD countries.

RESULTS: The Primary Care Provider (PCP) role varies from routine primary care to shared-rare-disease-care models. In the 26 papers, the most frequently cited PCP roles included involvement in diagnosis (n = 14), care coordination (n = 16), primary and preventative care (n = 18), management of components of rare-disease-care (n = 13), and treatment monitoring (n = 10). Individuals whose PCP was actively involved in their care were reported to have shortened diagnostic delay, improved transitions of care across the lifespan, reduced unplanned utilization of emergency and hospital services, comprehensive psychosocial care, improved quality of life across environments including home, school and work and improved palliative care experiences.

CONCLUSIONS: Sufficient communication from specialists, information, resources, time and reimbursement for complex care are still needed. Future integrated-rare-disease-care models should be developed by, or with, PCPs.

PMID:39772949 | DOI:10.1177/21501319241311567

Eur J Clin Pharmacol. 2025 Jan 7. doi: 10.1007/s00228-024-03798-z. Online ahead of print.

ABSTRACT

OBJECTIVE: Several studies have attempted to identify genetic determinants of clinical response to opioids administered during labor or after cesarean section. However, their results were often contrasting. A systematic review and meta-analysis was conducted to quantitatively assess the association between gene polymorphisms and clinical outcomes of opioid administration in the treatment of labor pain and post-cesarean pain.

METHODS: A comprehensive search was performed up to December 2023 using PubMed, Web of Knowledge, Cochrane Library, and OpenGrey databases. The clinical endpoints of interest were pain score after opioid treatment, total opioid consumption, patient's analgesic satisfaction, and incidence of opioid side effects. Random-effects meta-analyses were conducted when data were available in at least three studies.

RESULTS: Twenty-six studies enrolling 7765 patients were included in the systematic review. Overall, a total of 12 candidate polymorphic genes (OPRM1, COMT, CYP2D6, CYP3A4, ABCB1, ABCC3, UGT2B7, CGRP, OPRK1, OPRD1, KCNJ6, KCNJ9) were considered by the included studies, among which the most investigated variant was OPRM1 rs1799971. Overall pooled results indicated that individuals carrying the G allele of OPRM1 rs1799971 required higher opioid doses for pain management in comparison to rs1799971 AA subjects (standardized mean difference: 0.26; 95% CI: 0.09-0.44; P = 0.003). Such an association was confirmed in the subgroups of patients with labor pain and post-cesarean pain.

CONCLUSION: The present meta-analysis provides strong evidence of an association between OPRM1 rs1799971 and opioid dose requirement for relief of labor pain or post-cesarean pain. However, given the insufficient evidence for other polymorphic gene variants, large studies are still needed to investigate the impact of genetic variability on the efficacy and safety of opioid medications for relief of labor pain and post-cesarean pain (INPLASY Registration No. 202410040).

PMID:39774699 | DOI:10.1007/s00228-024-03798-z

Ther Drug Monit. 2025 Feb 1;47(1):141-151. doi: 10.1097/FTD.0000000000001289. Epub 2024 Dec 10.

ABSTRACT

Different polymorphisms in genes encoding metabolizing enzymes and drug transporters have been associated with tacrolimus pharmacokinetics. In particular, studies on CYP3A4 and CYP3A5, and their combined cluster have demonstrated their significance in adjusting tacrolimus dosing to minimize under- and overexposure thereby increasing the proportion of patients who achieve tacrolimus therapeutic target. Many factors influence the pharmacokinetics of tacrolimus, contributing to inter-patient variability affecting individual dosing requirements. On the other hand, the growing use of population pharmacokinetic models in solid organ transplantation, including different tacrolimus formulations, has facilitated the integration of pharmacogenetic data and other variables into algorithms to easier implement the personalized dose adjustment in transplant centers. The future of personalized medicine in transplantation lies in implementing these models in clinical practice, with pharmacogenetics as a key factor to account for the high inter-patient variability in tacrolimus exposure. To date, three clinical trials have validated the clinical application of these approaches. The aim of this review is to provide an overview of the current studies regarding the different population pharmacokinetic including pharmacogenetics and those translated to the clinical practice for individualizing tacrolimus dose adjustment in kidney transplantation.

PMID:39774592 | DOI:10.1097/FTD.0000000000001289

Front Mol Neurosci. 2024 Dec 24;17:1509280. doi: 10.3389/fnmol.2024.1509280. eCollection 2024.

ABSTRACT

The 90 kDa Heat shock protein (Hsp90) is a family of ubiquitously expressed molecular chaperones responsible for the stabilization and maturation of >400 client proteins. Hsp90 exhibits dramatic conformational changes to accomplish this, which are regulated by partner proteins termed co-chaperones. One of these co-chaperones is called the activator or Hsp90 ATPase activity homolog 1 (Aha1) and is the most potent accelerator of Hsp90 ATPase activity. In conditions where Aha1 levels are dysregulated including cystic fibrosis, cancer and neurodegeneration, Hsp90 mediated client maturation is disrupted. Accumulating evidence has demonstrated that many disease states exhibit large hetero-protein complexes with Hsp90 as the center. Many of these include Aha1, where increased Aha1 levels drive disease states forward. One strategy to block these effects is to design small molecule disruptors of the Hsp90/Aha1 complex. Studies have demonstrated that current Hsp90/Aha1 small molecule disruptors are effective in both models for cancer and neurodegeration.

PMID:39776493 | PMC:PMC11703849 | DOI:10.3389/fnmol.2024.1509280

Pharmacoepidemiol Drug Saf. 2025 Jan;34(1):e70076. doi: 10.1002/pds.70076.

ABSTRACT

The German Cystic Fibrosis (CF) Registry (GCFR) is a national General Data Protection Regulation-compliant centralised database sponsored by the German Cystic Fibrosis Association (Mukoviszidose e.V.) and based on informed consent for each participating patient, ethical approval, and data protection votes. The aims of the GCFR are to optimise quality of care for CF at the centres, generate epidemiologic overviews, address research questions related to improved CF care, and inform caregivers, patients (aimed at patient empowerment), and health authorities and industry (aimed at care planning and pharmacovigilance). Established in 1995, the Registry has captured data on > 9600 individuals with a combined total of more than 140 000 annual assessments with an estimated coverage rate of > 90%. Patient data are collected after informed consent and confirmed diagnosis of CF, or a CFTR-related disorder, or a screening-positive inconclusive diagnosis of CF (i.e., CFSPID). The registry collects core, encounter, and annual health data. Data include demographics, anthropometrics, lung function, microbiology, CF-specific complications and chronic medications, hospitalisations, demand-oriented antibiotic therapies, and outcomes (death and transplants). Real world and pharmacovigilance studies have been published and additional research underway; there is a formal process for requesting access to the GCFR.

PMID:39775994 | DOI:10.1002/pds.70076

Nat Commun. 2025 Jan 7;16(1):456. doi: 10.1038/s41467-025-55843-9.

ABSTRACT

Self-replicating RNA (srRNA) technology, in comparison to mRNA vaccines, has shown dose-sparing by approximately 10-fold and more durable immune responses. However, no improvements are observed in the adverse events profile. Here, we develop an srRNA vaccine platform with optimized non-coding regions and demonstrate immunogenicity and safety in preclinical and clinical development. Optimized srRNA vaccines generate protective immunity (according to the WHO defined thresholds) at doses up to 1,000,000-fold lower than mRNA in female mouse models of influenza and rabies. Clinically, safety and immunogenicity of RBI-4000, an srRNA vector encoding the rabies glycoprotein, was evaluated in a Phase I study (NCT06048770). RBI-4000 was able to elicit de novo protective immunity in the majority of healthy participants when administered at a dose of 0.1, 1, or 10 microgram (71%, 94%, 100%, respectively) in a prime-boost schedule. Similarly, we observe immunity above the WHO benchmark of protection following a single administration in most participants at both 1 and 10 microgram doses. There are no serious adverse events reported across all cohorts. These data establish the high therapeutic index of optimized srRNA vectors, demonstrating feasibility of both low dose and single dose approaches for vaccine applications.

PMID:39774967 | DOI:10.1038/s41467-025-55843-9

Genes Immun. 2025 Jan 7. doi: 10.1038/s41435-024-00318-y. Online ahead of print.

ABSTRACT

Immunoglobulin GM (γ marker) and KM (κ marker) allotypes-encoded by immunoglobulin heavy chain G (IGHG) and immunoglobulin κ constant (IGKC) genes-have been shown to be associated with immune responsiveness to a variety of self and nonself antigens. The aim of the present investigation was to determine whether allelic variation at the GM and KM loci was associated with antibody responsiveness to poly-N-acetyl-D-glucosamine (PNAG), a broadly-conserved surface polysaccharide expressed by many microbial pathogens. In addition, we wished to determine whether Fcγ receptor 2 A (FCGR2A) genotypes, which have been shown to be risk factors for some pathogens, also influenced antibody responses to PNAG. DNA from 257 patients with various pulmonary diseases (PD) was genotyped for several GM, KM, and FCGR2A alleles, and plasma were characterized for anti-PNAG IgG antibodies. The levels of IgG4 antibodies to PNAG were associated with FCGR2A genotypes (p = 0.01). Also, KM and FCGR2A alleles epistatically contributed to anti-PNAG IgG3 antibody responses: subjects with KM 1/1 or KM 1/3 and homozygous for the R allele of FCGR2A had the highest levels of anti-PNAG IgG3 antibodies compared to all other genotype combinations. If confirmed by larger studies, these results are potentially relevant to immunotherapy against many PNAG-expressing infectious pathogens.

PMID:39774260 | DOI:10.1038/s41435-024-00318-y

Sci Rep. 2025 Jan 7;15(1):1143. doi: 10.1038/s41598-024-84003-0.

ABSTRACT

Continuous glucose monitors (CGM) provide valuable insights about glycemic control that aid in diabetes management. However, interpreting metrics and charts and synthesizing them into linguistic summaries is often non-trivial for patients and providers. The advent of large language models (LLMs) has enabled real-time text generation and summarization of medical data. The objective of this study was to assess the strengths and limitations of using an LLM to analyze raw CGM data and produce summaries of 14 days of data for patients with type 1 diabetes. We first evaluated the ability of GPT-4 to compute quantitative metrics specific to diabetes found in an Ambulatory Glucose Profile (AGP). Then, using two independent clinician graders, we evaluated the accuracy, completeness, safety, and suitability of qualitative descriptions produced by GPT-4 across five different CGM analysis tasks. GPT-4 performed 9 out of the 10 quantitative metrics tasks with perfect accuracy across all 10 cases. The clinician-evaluated CGM analysis tasks had good performance across measures of accuracy [lowest task mean score 8/10, highest task mean score 10/10], completeness [lowest task mean score 7.5/10, highest task mean score 10/10], and safety [lowest task mean score 9.5/10, highest task mean score 10/10]. Our work serves as a preliminary study on how generative language models can be integrated into diabetes care through data summarization and, more broadly, the potential to leverage LLMs for streamlined medical time series analysis.

PMID:39774031 | DOI:10.1038/s41598-024-84003-0

Cell Commun Signal. 2025 Jan 7;23(1):12. doi: 10.1186/s12964-024-02024-8.

ABSTRACT

BACKGROUND: Neutrophils are the most abundant leukocytes in human blood, and their recruitment is essential for innate immunity and inflammatory responses. The initial and critical step of neutrophil recruitment is their adhesion to vascular endothelium, which depends on G protein-coupled receptor (GPCR) triggered integrin inside-out signaling that induces β2 integrin activation and clustering on neutrophils. Kindlin-3 and talin-1 are essential regulators for the inside-out signaling induced β2 integrin activation. However, their contribution in the inside-out signaling induced β2 integrin clustering is unclear because conventional assays on integrin clustering are usually performed on adhered cells, where integrin-ligand binding concomitantly induces integrin outside-in signaling.

METHODS: We used flow cytometry and quantitative super-resolution stochastic optical reconstruction microscopy (STORM) to quantify β2 integrin activation and clustering, respectively, in kindlin-3 and talin-1 knockout leukocytes. We also tested whether wildtype or Pleckstrin homology (PH) domain deleted kindlin-3 can rescue the kindlin-3 knockout phenotypes.

RESULTS: GPCR-triggered inside-out signaling alone can induce β2 integrin clustering. As expected, both kindlin-3 and talin-1 knockout decreases integrin activation. Interestingly, only kindlin-3 but not talin-1 contributes to integrin clustering in the scenario of inside-out-signaling, wherein a critical role of the PH domain of kindlin-3 was highlighted.

CONCLUSIONS: Since talin was known to facilitate integrin clustering in outside-in-signaling-involved cells, our finding provides a paradigm shift by suggesting that the molecular mechanisms of integrin clustering upon inside-out signaling and outside-in signaling are different. Our data also contradict the conventional assumption that integrin activation and clustering are tightly inter-connected by showing separated regulation of the two during inside-out signaling. Our study provides a new mechanism that shows kindlin-3 regulates β2 integrin clustering and suggests that integrin clustering should be assessed independently, aside from integrin activation, when studying leukocyte adhesion in inflammatory diseases.

PMID:39773732 | DOI:10.1186/s12964-024-02024-8

Orphanet J Rare Dis. 2025 Jan 7;20(1):6. doi: 10.1186/s13023-024-03522-1.

ABSTRACT

BACKGROUND: Patients with cystic fibrosis (CF) are rare in China and differ significantly from the Caucasian populations in terms of clinical and genetic characteristics. However, the progression and mortality of Chinese patients with CF have not been well described.

RESULTS: This study included all 67 patients from the Peking Union Medical College Hospital CF cohort, with a median followed up time of 5.2 years. Compared to patients diagnosed with CF in childhood, adult-diagnosed patients exhibit a lower proportion of pancreatic exocrine insufficiency (25.0% vs. 77.8%, P = 0.001) and a higher body mass index (19.6 vs. 17.7 kg/m2, P = 0.045). According to the mixed-effects model, for patients ≤ 30 years of age at diagnosis, FEV1% predicted decreased 1.17% per year. The generalized linear regression model showed that higher baseline FEV1% predicted and occurrence of pulmonary exacerbations were associated with the progression of patients with CF. The survival rates at 5 years and 10 years after the diagnosis were 96.7% and 80.6%, respectively. The log-rank test showed baseline FEV1% predicted < 50%, and high CF-ABLE and 3-year prognostic scores were associated with mortality in patients with CF in China.

CONCLUSIONS: We reported the progression and mortality of patients with CF in China, which was a rare and relatively unknown population in the past. Baseline FEV1% predicted is associated with progression and mortality. Pulmonary exacerbations can accelerate the decline in lung function. The CF-ABLE and 3-year prognostic scores are applicable for predicting poor prognosis in patients with CF in China.

PMID:39773272 | DOI:10.1186/s13023-024-03522-1

Am J Respir Cell Mol Biol. 2025 Jan 7. doi: 10.1165/rcmb.2024-0130OC. Online ahead of print.

ABSTRACT

In asthma, tissue factor (TF) levels are elevated in the lung. In our previous studies using mechanically compressed human bronchial epithelial (HBE) cells, which are a well-defined in vitro model of bronchoconstriction during asthma exacerbations, we detected TF within extracellular vesicles (EVs) released from compressed HBE cells. Here, to better characterize the potential role of this mechanism in asthma, we tested the extent to which the transcriptional regulation of epithelial cell-derived TF varied between donors with and without asthma. Using RNA in situ hybridization, we detected epithelial expression of F3, the TF protein-encoding gene, in human airways. Next, to determine the role of TGF-β receptor (TGF-βR) in the regulation of TF, we exposed well-differentiated HBE cells to mechanical compression in the presence or absence of a pharmacological inhibitor of TGF-β receptor. Furthermore, to identify the protein cargo of EVs released from HBE cells, we used Tandem Mass Tag mass spectrometry. Our findings revealed significantly higher F3 expression in the airways of patients with asthma compared to healthy controls. However, we observed no differences in F3 expression or TF release between asthmatic and non-asthmatic HBE cells, both at baseline and after compression. Mechanistically, compression-induced F3 expression in HBE cells depended on TGF-βR. Our proteomic analysis identified 22 differentially released proteins in EVs, with higher levels in compressed cells compared to controls. Gene ontology analysis indicates these proteins are involved in diverse biological processes, highlighting a potential role for epithelial cell-derived EVs during asthma exacerbations.

PMID:39773168 | DOI:10.1165/rcmb.2024-0130OC

Am J Physiol Lung Cell Mol Physiol. 2025 Jan 8. doi: 10.1152/ajplung.00289.2024. Online ahead of print.

ABSTRACT

Lung infection is one of the leading causes of morbidity and mortality worldwide. Even with appropriate antibiotic and antiviral treatment, mortality in hospitalized patients often exceeds 10%, highlighting the need for the development of new therapeutic strategies. Of late, cystic fibrosis transmembrane conductance regulator (CFTR) is - in addition to its well-established roles in the lung airway and extrapulmonary organs - increasingly recognized as a key regulator of alveolar homeostasis and defense. In the alveolar epithelium, CFTR mediates alveolar fluid secretion and liquid homeostasis; in the microvascular endothelium, CFTR maintains vascular barrier function. CFTR also contributes to alveolar immunity. Yet, in lung infection, diverse molecular mechanisms reduce CFTR abundance and otherwise impair its function, promoting alveolar inflammation, edema, and cell death. Preservation or restoration of CFTR function by CFTR modulator drugs thus presents a promising avenue to combat lung infection in a pathogen-independent manner.

PMID:39772994 | DOI:10.1152/ajplung.00289.2024

J Asthma. 2025 Jan 8:1-11. doi: 10.1080/02770903.2024.2448317. Online ahead of print.

ABSTRACT

ObjectivesPatient perception of treatment effectiveness is key to optimizing adherence. This is potentially impacted by color, yet no such studies have been conducted in asthma. This study assessed the influence of pink vs. white pressurized metered-dose inhaler (pMDI) actuators on asthma symptoms perception.MethodsIn this double-blind, randomized, multicenter, crossover study, adults with moderate-to-severe asthma received extrafine formulation beclomethasone dipropionate/formoterol furoate (BDP/FF) pMDI for a two-week run-in. During two, two-week treatment periods they received BDP/FF pMDI, white in one, pink in the other, using an 'authorized deception' approach (patients were told the inhalers contained the same active ingredients, but device characteristics differed). Endpoints included patient-reported asthma symptoms and psychopharmacological aspects (prior to use: did patients expect an improvement; after use: did they think there had been an improvement; both on 100-point visual analog scales [VAS]).ResultsOf 74 patients analyzed, 72 completed the study. There were no statistically significant differences between inhalers for asthma symptoms, with minimal changes from baseline. Patients were numerically more likely to expect symptoms improvement with the white than pink inhaler (mean VAS 64.5 vs. 60.8). Perceived improvements were lower than expected with both, numerically favoring the pink inhaler (mean VAS 41.1 vs. 44.6); 46.6% believed a change had been made, 51.9% of whom believed this impacted symptoms.ConclusionsChanging inhaler color had no impact on asthma symptoms, but did have a numerical impact on patients' expectations of subsequent treatment effect. This emphasizes the importance of communication between patients and healthcare practitioners when changing inhalers.

PMID:39772981 | DOI:10.1080/02770903.2024.2448317

Quant Plant Biol. 2024 Dec 23;5:e12. doi: 10.1017/qpb.2024.11. eCollection 2024.

ABSTRACT

Hormonal mechanisms associated with cell elongation play a vital role in the development and growth of plants. Here, we report Nextflow-root (nf-root), a novel best-practice pipeline for deep-learning-based analysis of fluorescence microscopy images of plant root tissue from A. thaliana. This bioinformatics pipeline performs automatic identification of developmental zones in root tissue images. This also includes apoplastic pH measurements, which is useful for modeling hormone signaling and cell physiological responses. We show that this nf-core standard-based pipeline successfully automates tissue zone segmentation and is both high-throughput and highly reproducible. In short, a deep-learning module deploys deterministically trained convolutional neural network models and augments the segmentation predictions with measures of prediction uncertainty and model interpretability, while aiming to facilitate result interpretation and verification by experienced plant biologists. We observed a high statistical similarity between the manually generated results and the output of the nf-root.

PMID:39777028 | PMC:PMC11706687 | DOI:10.1017/qpb.2024.11

Front Psychol. 2024 Dec 24;15:1509392. doi: 10.3389/fpsyg.2024.1509392. eCollection 2024.

ABSTRACT

INTRODUCTION: While the fact that visual stimuli synthesized by Artificial Neural Networks (ANN) may evoke emotional reactions is documented, the precise mechanisms that connect the strength and type of such reactions with the ways of how ANNs are used to synthesize visual stimuli are yet to be discovered. Understanding these mechanisms allows for designing methods that synthesize images attenuating or enhancing selected emotional states, which may provide unobtrusive and widely-applicable treatment of mental dysfunctions and disorders.

METHODS: The Convolutional Neural Network (CNN), a type of ANN used in computer vision tasks which models the ways humans solve visual tasks, was applied to synthesize ("dream" or "hallucinate") images with no semantic content to maximize activations of neurons in precisely-selected layers in the CNN. The evoked emotions of 150 human subjects observing these images were self-reported on a two-dimensional scale (arousal and valence) utilizing self-assessment manikin (SAM) figures. Correlations between arousal and valence values and image visual properties (e.g., color, brightness, clutter feature congestion, and clutter sub-band entropy) as well as the position of the CNN's layers stimulated to obtain a given image were calculated.

RESULTS: Synthesized images that maximized activations of some of the CNN layers led to significantly higher or lower arousal and valence levels compared to average subject's reactions. Multiple linear regression analysis found that a small set of selected image global visual features (hue, feature congestion, and sub-band entropy) are significant predictors of the measured arousal, however no statistically significant dependencies were found between image global visual features and the measured valence.

CONCLUSION: This study demonstrates that the specific method of synthesizing images by maximizing small and precisely-selected parts of the CNN used in this work may lead to synthesis of visual stimuli that enhance or attenuate emotional reactions. This method paves the way for developing tools that stimulate, in a non-invasive way, to support wellbeing (manage stress, enhance mood) and to assist patients with certain mental conditions by complementing traditional methods of therapeutic interventions.

PMID:39776961 | PMC:PMC11703666 | DOI:10.3389/fpsyg.2024.1509392

Patterns (N Y). 2024 Dec 9;5(12):101116. doi: 10.1016/j.patter.2024.101116. eCollection 2024 Dec 13.

ABSTRACT

To achieve adequate trust in patient-critical medical tasks, artificial intelligence must be able to recognize instances where they cannot operate confidently. Ensemble methods are deployed to estimate uncertainty, but models in an ensemble often share the same vulnerabilities to adversarial attacks. We propose an ensemble approach based on feature decorrelation and Fourier partitioning for teaching networks diverse features, reducing the chance of perturbation-based fooling. We test our approach against white-box attacks in single- and multi-channel electrocardiogram classification and adapt adversarial training and DVERGE into an ensemble framework for comparison. Our results indicate that the combination of decorrelation and Fourier partitioning maintains performance on unperturbed data while demonstrating superior uncertainty estimation on projected gradient descent and smooth adversarial attacks of various magnitudes. Furthermore, our approach does not require expensive optimization with adversarial samples during training. These methods can be applied to other tasks for more robust models.

PMID:39776851 | PMC:PMC11701855 | DOI:10.1016/j.patter.2024.101116

Infect Drug Resist. 2025 Jan 3;18:31-42. doi: 10.2147/IDR.S482584. eCollection 2025.

ABSTRACT

BACKGROUND: Early differentiation between spinal tuberculosis (STB) and acute osteoporotic vertebral compression fracture (OVCF) is crucial for determining the appropriate clinical management and treatment pathway, thereby significantly impacting patient outcomes.

OBJECTIVE: To evaluate the efficacy of deep learning (DL) models using reconstructed sagittal CT images in the differentiation of early STB from acute OVCF, with the aim of enhancing diagnostic precision, reducing reliance on MRI and biopsies, and minimizing the risks of misdiagnosis.

METHODS: Data were collected from 373 patients, with 302 patients recruited from a university-affiliated hospital serving as the training and internal validation sets, and an additional 71 patients from another university-affiliated hospital serving as the external validation set. MVITV2, Efficient-Net-B5, ResNet101, and ResNet50 were used as the backbone networks for DL model development, training, and validation. Model evaluation was based on accuracy, precision, sensitivity, F1 score, and area under the curve (AUC). The performance of the DL models was compared with the diagnostic accuracy of two spine surgeons who performed a blinded review.

RESULTS: The MVITV2 model outperformed other architectures in the internal validation set, achieving accuracy of 98.98%, precision of 100%, sensitivity of 97.97%, F1 score of 98.98%, and AUC of 0.997. The performance of the DL models notably exceeded that of the spine surgeons, who achieved accuracy rates of 77.38% and 93.56%. The external validation confirmed the models' robustness and generalizability.

CONCLUSION: The DL models significantly improved the differentiation between STB and OVCF, surpassing experienced spine surgeons in diagnostic accuracy. These models offer a promising alternative to traditional imaging and invasive procedures, potentially promoting early and accurate diagnosis, reducing healthcare costs, and improving patient outcomes. The findings underscore the potential of artificial intelligence for revolutionizing spinal disease diagnostics, and have substantial clinical implications.

PMID:39776757 | PMC:PMC11706012 | DOI:10.2147/IDR.S482584

Clin Med Insights Oncol. 2025 Jan 6;19:11795549241311408. doi: 10.1177/11795549241311408. eCollection 2025.

ABSTRACT

Despite the expanding therapeutic options available to cancer patients, therapeutic resistance, disease recurrence, and metastasis persist as hallmark challenges in the treatment of cancer. The rise to prominence of generative artificial intelligence (GenAI) in many realms of human activities is compelling the consideration of its capabilities as a potential lever to advance the development of effective cancer treatments. This article presents a hypothetical case study on the application of generative pre-trained transformers (GPTs) to the treatment of metastatic prostate cancer (mPC). The case explores the design of GPT-supported adaptive intermittent therapy for mPC. Testosterone and prostate-specific antigen (PSA) are assumed to be repeatedly monitored while treatment may involve a combination of androgen deprivation therapy (ADT), androgen receptor-signalling inhibitors (ARSI), chemotherapy, and radiotherapy. The analysis covers various questions relevant to the configuration, training, and inferencing of GPTs for the case of mPC treatment with a particular attention to risk mitigation regarding the hallucination problem and its implications to clinical integration of GenAI technologies. The case study provides elements of an actionable pathway to the realization of GenAI-assisted adaptive treatment of metastatic prostate cancer. As such, the study is expected to help facilitate the design of clinical trials of GenAI-supported cancer treatments.

PMID:39776668 | PMC:PMC11701910 | DOI:10.1177/11795549241311408

J Med Imaging (Bellingham). 2025 Jan;12(1):014501. doi: 10.1117/1.JMI.12.1.014501. Epub 2025 Jan 6.

ABSTRACT

PURPOSE: The prevalence of type 2 diabetes mellitus (T2DM) has been steadily increasing over the years. We aim to predict the occurrence of T2DM using mammography images within 5 years using two different methods and compare their performance.

APPROACH: We examined 312 samples, including 110 positive cases (developed T2DM after 5 years) and 202 negative cases (did not develop T2DM) using two different methods. In the first method, a radiomics-based approach, we utilized radiomics features and machine learning (ML) algorithms. The entire breast region was chosen as the region of interest for extracting radiomics features. Then, a binary breast image was created from which we extracted 668 features and analyzed them using various ML algorithms. In the second method, a complex convolutional neural network (CNN) with a modified ResNet architecture and various kernel sizes was applied to raw mammography images for the prediction task. A nested, stratified five-fold cross-validation was done for both parts A and B to compute accuracy, sensitivity, specificity, and area under the receiver operating curve (AUROC). Hyperparameter tuning was also done to enhance the model's performance and reliability.

RESULTS: The radiomics approach's light gradient boosting model gave 68.9% accuracy, 30.7% sensitivity, 89.5% specificity, and 0.63 AUROC. The CNN method achieved an AUROC of 0.58 over 20 epochs.

CONCLUSION: Radiomics outperformed CNN by 0.05 in terms of AUROC. This may be due to the more straightforward interpretability and clinical relevance of predefined radiomics features compared with the complex, abstract features learned by CNNs.

PMID:39776665 | PMC:PMC11702674 | DOI:10.1117/1.JMI.12.1.014501

Biol Imaging. 2024 Nov 15;4:e13. doi: 10.1017/S2633903X24000096. eCollection 2024.

ABSTRACT

We propose a neural network architecture and a training procedure to estimate blurring operators and deblur images from a single degraded image. Our key assumption is that the forward operators can be parameterized by a low-dimensional vector. The models we consider include a description of the point spread function with Zernike polynomials in the pupil plane or product-convolution expansions, which incorporate space-varying operators. Numerical experiments show that the proposed method can accurately and robustly recover the blur parameters even for large noise levels. For a convolution model, the average signal-to-noise ratio of the recovered point spread function ranges from 13 dB in the noiseless regime to 8 dB in the high-noise regime. In comparison, the tested alternatives yield negative values. This operator estimate can then be used as an input for an unrolled neural network to deblur the image. Quantitative experiments on synthetic data demonstrate that this method outperforms other commonly used methods both perceptually and in terms of SSIM. The algorithm can process a 512 512 image under a second on a consumer graphics card and does not require any human interaction once the operator parameterization has been set up.1.

PMID:39776610 | PMC:PMC11704139 | DOI:10.1017/S2633903X24000096