Communications medicine最新文献

Background: HIV-tuberculosis (HIV-TB) coinfection poses a significant public health challenge among children in high-burden African regions. Most previous transcriptomic studies have concentrated on adults and non-African populations, primarily analyzing gene-level differential expression. This approach overlooks multi-isoform complexity and may obscure both inherent and pathogen-induced intragenic heterogeneity. This multi-center case-control study aimed to identify and characterize the transcript-level landscape of HIV-TB coinfection in children from different African regions.

Methods: We analyzed whole-blood RNA sequencing data from 97 children with and without tuberculosis from Uganda (East Africa) and from Botswana and Eswatini (Southern Africa). Reads were quality-controlled, and low-abundance transcripts filtered out. Differential transcript expression was estimated using models that adjusted for batch, age, and sex, with multiple testing controlled by the Benjamini-Hochberg procedure. Pathway enrichment was performed on the set of differentially expressed transcripts.

Results: Our analyses show geographic heterogeneity in immune responses; however, the top three gene pathways - immune system, innate immune system, and neutrophil degranulation are consistently conserved across regions. Although there is limited overlap among upregulated transcripts, four of the six shared differentially expressed transcripts (DETs) are enriched in neutrophil degranulation pathways, indicating a conserved transcriptional signature of HIV-TB coinfection. Additionally, we identify five genes with region-specific, non-overlapping isoforms, a distinction not detectable through gene-level analysis.

Conclusions: These findings demonstrate a conserved whole-blood transcriptomic signature in pediatric HIV-TB coinfection, while also highlighting regional variation at the isoform level. This supports the use of transcript-level analyses to identify biomarkers and enhance understanding of host responses in diverse African settings.

Background: As the population ages, a rising mortality burden is attributed to deaths in older adults, particularly deaths from inflammation-related chronic non-communicable diseases. The synergism between aging and inflammation remains unclear. Here, we conducted a study to examine whether a decrease in leukocyte mitochondrial DNA copy number (mtDNA_CN) with age modifies the association between inflammation and the mortality risk in older adults.

Methods: A total of 3520 adults (mean [SD] age, 67.6 [7.4] years) underwent serial leukocyte mtDNA_CN and serum high-sensitivity C-reactive protein (hs-CRP) measurements and ascertainment of subsequent all-cause and cardiovascular diseases (CVD) deaths. Mortality risks were estimated using Cox proportional hazards models.

Results: Compared to participants with both a sustainedly low serum hs-CRP and change in leukocyte mtDNA_CN at the highest tertile, the adjusted hazard ratios (95% CI) of all-cause and CVD death are 3.20 (2.20-4.66) and 5.77 (2.72-12.21) for those with both increased serum hs-CRP and change in leukocyte mtDNA_CN at the lowest tertile, 1.48 (0.93-2.38) and 1.24 (0.44-3.53) for those with increased serum hs-CRP alone, and 1.29 (0.93-1.81) and 1.44 (0.70-2.97) for those with a change in leukocyte mtDNA_CN at the lowest tertile alone. The relative excess risks due to interaction (95% CI) for all-cause and CVD death are 1.42 (0.19-2.65) and 4.08 (0.21-7.96). Similar results are observed for those with a change in leukocyte mtDNA_CN at the middle tertile and in sensitivity analyses.

Conclusions: We demonstrate super-additive interactions between decreases in leukocyte mtDNA_CN and inflammation on the mortality risk in older adults, indicating underlying synergism.

Background: Major Depression (MDD) is a potentially life-threatening condition that ranks among the diseases with the highest global burden. Despite its prevalence, current diagnostic methods remain largely subjective, and first-line treatments exhibit high rates of non-responders.

Methods: This study investigates the application of deep learning (DL) algorithms to electroencephalogram (EEG) data for the MDD-diagnosis and prediction of treatment outcomes following the administration of selective serotonin reuptake inhibitors (SSRIs), using six large, independent datasets with a total of n = 146 for healthy subjects and n = 203 for patients. DL models were trained on one portion of the datasets and tested on unseen data from different subjects. To interpret the classification features, Gradient-weighted Class Activation Mapping (Grad-CAM) was applied.

Results: The models achieve an average accuracy of 67.5% (best fold 70%) in distinguishing MDD patients from healthy controls and mean 79% accuracy (best fold 85%) in predicting SSRI responders. Key EEG markers for both classification tasks revealed by Grad-CAM include alpha activity in the frontal and parietal regions. Simulation of a clinical decision scenario for SSRI treatment selection indicates a number needed to treat (NNT) of five when using a model with 80% predictive accuracy, corresponding to an increase in treatment response from a 50% baseline to 70% with model-guided selection.

Conclusion: These findings underscore the clinical potential of EEG-based DL models for stratified treatment in MDD, facilitating accurate therapy choices and reducing ineffective treatments. The results of the integration of objective neurophysiological markers into clinical psychiatry are indicating the potential for more personalized treatment allocation.

Background: We aimed to elucidate the effects of rectal douching on gut microbial communities and their associated metabolites in men who have sex with men (MSM).

Methods: A community sample of HIV-uninfected MSM were recruited in Guangzhou, China. Participants were stratified into rectal douching and non-douching groups based on their rectal douching behavior within the past three months. Peripheral blood and fecal samples were collected. Serum markers of microbial translocation were quantified using enzyme-linked immunosorbent assay (ELISA). Gut microbiota composition was assessed via 16S rRNA gene sequencing of fecal samples, and microbial metabolites were profiled using a mass spectrometry-based platform.

Results: This study enrolls a total of 51 HIV-uninfected MSM (20 in the rectal douching group and 31 in the non-douching group). The two groups have comparable age distributions [median age: 25 years, interquartile range (IQR): 23-30 vs. 27 years, IQR: 24-31]. Significant structural shifts in microbial community composition are observed at both the phylum and genus levels in the rectal douching group relative to the non-douching group. The relative abundances of the genera Clostridium, Lachnospira, and Turicibacter are significantly lower in the rectal douching group compared to the non-douching group. Furthermore, rectal douching significantly alters a wide range of microbial metabolites. Notably, rectal douching appears to reduce gut barrier integrity, as evidenced by significantly elevated levels of the microbial translocation marker lipopolysaccharide-binding protein (LBP) in the rectal douching group.

Conclusions: Rectal douching among HIV-uninfected MSM is associated with gut microbiota dysbiosis, significant alterations in microbial metabolic profiles, and reduced gut barrier integrity. These findings underscore the need for increased awareness and health education within this population.

Background: The neuroinflammatory cerebral form of X-linked adrenoleukodystrophy (CALD) is among the most severe neurological diseases affecting children. Early intervention by hematopoietic stem cell transplantation (HSCT) or gene therapy halts CALD, justifying its inclusion in newborn screening programmes. Currently, eligibility for treatment is assessed using MRI-based Loes and neurological scoring; however, grading poorly differentiates advanced stages or atypical lesion patterns. We recently identified blood neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP), indicative of axonal and glial damage, as valid biomarkers reflecting CALD activity.

Methods: Here, we investigated how pre-treatment biomarker levels relate to outcome of HSCT or gene therapy in a retrospective cohort of 14 paediatric CALD patients aged 5-13 years. Patients were monitored longitudinally before and for up to 5.5 years following HSCT or gene therapy. Disease progression was assessed using MRI-based Loes scoring and neurological evaluations. Plasma NfL and GFAP were quantified using single molecule array (Simoa) technology.

Results: Here we show that blood NfL and GFAP transiently increase due to myeloablative conditioning and gradually normalize post-treatment. Pre-treatment NfL ≤ 113 pg/ml, but not GFAP, correlate with CALD stabilization or, for Loes >9, only minor progression. In contrast, NfL >243 pg/ml associates with major progression irrespective of baseline Loes score. In three boys with atypical lesions or advanced disease, NfL outperforms Loes-scoring in predicting outcome.

Conclusions: Blood NfL can complement clinical decision-making, particularly in patients with advanced CALD or atypical lesions. These findings are especially relevant for clinical management in countries that have not yet implemented X-linked adrenoleukodystrophy into newborn screening.

Background: Oral squamous cell carcinoma (OSCC) is associated with many risk factors but not all individuals who are exposed to the risk factors develop OSCC. This aspect warrants the need to understand the genetic susceptibility in presence of risk factors.

Methods: The study protocol has been registered with PROSPERO (CRD42023422519). The present study included the English-language literature published from January 2000 to April 2024. The articles were searched using keywords via Ovid platform through various databases. The odds ratio was considered as standard measure of outcome. Meta-analysis was carried out using random effects model with RevMan 5.4, establishing a 95% confidence interval and a significance threshold of p ≤ 0.05.

Results: Taiwan, China, India, and USA share the major bulk of studies. The subgroup analysis shows polymorphism in P53 gene in presence of smoking (Z = 2.15, p = 0.02), alcohol (Z = 2.38, p = 0.02) and mixed habit (tobacco + alcohol) (Z = 3.28, p ≤ 0.001) and CASP 8 (Z = 5.38, p ≤ 0.0001) in presence of alcohol consumption habit has highly significant risk for development of OSCC; however the studies shows moderate heterogeneity (I² = 40-50%). Certain genes such as HIF (Z = 2.82, p = 0.05), MTNR1 (Z = 12.12, p ≤ 0.0001) and DEC (Z = 10.46, p ≤ 0.0001) shows statistically significant correlation in presence of smoking and chewing habit with very low heterogeneity (I² = 0%). The CYP1 gene shows a highly significant (p ≤ 0.0001) correlation (Z = 3.11) in presence of mixed habit with 0% heterogeneity.

Conclusions: Asian countries show a large cluster of patients with a genetic risk for development of OSCC. Genetic factors such as P53, CASP 8, HIF, DEC1, MTNR1 and CYP1A1 show statistically significant risk for development of OSCC in the presence of risky environmental factors such as tobacco and alcohol.

Background: Skin cancer is one of the most prevalent cancers globally, with early detection critical to ensure reduced mortality risk. To aid early detection, machine learning (ML) skin cancer detection models have been proposed, currently with a focus on dermatoscopic imaging only. However, freetext may provide extra diagnostic information that is not present in images alone.

Methods: We constructed a multimodal dataset comprising 5481 dermatoscopic images from 4538 patients, including patient metadata and clinical notes, with binary labels (benign vs. malignant, 7% malignant). To assess and mitigate bias from leading language, we developed a clinical text preprocessing pipeline combining regular expressions and large language models, enabling multiple levels of filtering. We train multimodal ML models on this dataset to explore the effect of freetext on model performance.

Results: Our results show that incorporating unfiltered text significantly improves classification performance (0.970 AUROC) compared to visual data alone (0.909 AUROC); even with leading language removed, performance gains persist (0.948 AUROC).

Conclusions: This work benchmarks clinical freetext inclusion in skin lesion classification, demonstrating that clinical text contributes predictive value beyond that available in images alone. The model's high performance on unfiltered clinical text highlights the high levels of bias, and possible shortcutting, present in this text which may make it unsuitable for inclusion in some ML models. By systematically filtering clinical notes via our proposed technique, we show that multimodal models retain improved accuracy while reducing bias. These results provide practical guidance for integrating clinical text into real-world skin cancer detection systems and establish a foundation for future multimodal research in dermatology.

Background: Hepatitis C infection (HCV) is a leading cause of liver disease and mortality. Despite curative treatment options, HCV elimination remains elusive. Although the US has national screening and treatment recommendations, HCV remains under-screened and under-diagnosed. We utilized two national surveys to estimate trends in overall HCV care utilization and testing in the US.

Methods: Data from the National Ambulatory Medical Care Survey and the National Hospital Ambulatory Medical Care Survey from 2010 to 2019 were analyzed using inverse probability weighting to generate national estimates of visits and testing. Weighted chi-square and logistic regression analyses adjusted for demographics, payor type, and time assessed the primary outcome of ambulatory care utilization and the secondary outcome of office-based hepatitis C screening.

Results: Between 2010-2019, 23,469,344 HCV ambulatory visits are identified with higher overall adjusted visit rates for men (OR 1.54), people born 1945-1965 (OR 4.00), and insured by Medicare (OR 1.98) with increased utilization by White, privately insured patients in the office since 2016. HCV cases with Medicaid (OR 6.05) or have associated substance use disorder (SUD) (OR 3.30) are more likely to utilize the ER than office care. Screening rates are low in initial (2%) and overall (1%) primary care health visits.

Conclusions: In a nationally representative study, we find increasing rates of HCV visits, largely in White, privately insured patients seen in office. Low screening rates and disproportionate ER utilization among rural, racial/ethnic minorities, Medicaid insured, and patients with SUD highlight the importance of policy and practice guideline updates to improve identification and care linkage for HCV.

Background: The implementation of population-based breast cancer screening programs has been pivotal for early cancer detection, yet sociospatial disparities in participation rates may remain. Understanding and monitoring these variations is essential for improving participation, enabled by modern space-time approaches. This study aimed to (1) assess the existence of spatial clustering of participation in a breast cancer screening program, (2) evaluate temporal shifts in spatial patterns, and (3) assess the relative importance of area-level determinants in predicting participation rates.

Methods: We used the emerging hot spot analysis to mine and visualize space-time participation patterns. We assessed the determinants of screening participation using eXtreme Gradient Boosting combined with SHapley Additive exPlanations values for model interpretation. This approach was applied to a dataset of 482,318 georeferenced invitations sent from 2003 to 2020 by the breast cancer screening program in the canton of Geneva, Switzerland.

Results: Here we show that the overall participation rate of 41.5% falls below the national average of 46%, despite increases across all population segments. Initial analysis shows a clear periurban-urban pattern with lower urban participation. Space-time pattern mining further delineates this pattern into 13 distinct profiles, with rates varying from 27.8% in intensifying cold spots to 49.2% in intensifying hot spots. Modeling reveals higher screening participation in socioeconomically deprived areas and a negative association between accessibility to screening centers and participation rates.

Conclusions: The approach applied in this study enables a more nuanced monitoring of screening participation dynamics. Our findings support targeted interventions in prioritized areas to further reduce cancer screening inequalities.