Articles in 2025

Valero-Muñoz, Saw et al. apply a translational model in which adult cardiac myocytes are exposed to serum from people with HIV before and after initiation of ART. The authors identify distinct molecular signatures, indicating that HIV infection and antiretroviral therapy engage separate biological pathways that may precede diastolic dysfunction and the development of HFpEF.

Sánchez-Valle et al. analyze gene expression data from over 100 diseases to explore differences in comorbidity occurrence between sexes. They find that men and women share distinct molecular mechanisms driving disease co-occurrence and distinct drug associations with comorbidities.

Anneberg et al. investigate associations between data-driven maternal dietary patterns during pregnancy and offspring’s risk of inflammatory bowel disease (IBD). Based on 56,097 mother-child pairs followed over 18 years, they show that a diverse dietary pattern may protect offspring against IBD development compared to a Western dietary pattern.

Kim, D’Anniballe et al. examined olfactory tissue from individuals with Long COVID-related anosmia and conducted a randomised pilot trial of topical steroid therapy. They found no persistent viral presence, identified clonally expanded inflammatory T cells, and confirmed the therapy’s safety, supporting the need for larger trials.

Wang et al at introduce a dynamic, network-based tool to predict the risk of hepatic encephalopathy in patients with cirrhosis. The model integrates expert knowledge with data-driven insights to enable early prediction and risk stratification across diverse patient populations.

Yang et al. conduct a prospective cohort study to investigate the association between mental health and the risk of metabolic dysfunction-associated steatotic liver disease (MASLD). Anxiety, depression, mania, and trauma increase MASLD risk with metabolites related to nitrogen and lipid metabolism partially mediating this association.

Chu et al. develop a monthly-calibrated photoplethysmography pipeline that leverages an implicit HbA1c with PPG signals to non-invasively estimate fasting blood glucose across three diabetic cohorts. Improved accuracy is seen over prior approaches, as assessed by mean absolute relative differences and Parkes Error Grid analyses.

Grimani et al. examine whether prosocial messages can optimise the effect on population behaviour to reduce respiratory infection transmission. Prosocial message interventions, particularly those including health consequences or reducing exposure cues, have a small but positive effect on protective behavioural intentions.

Wang et al. propose an in-situ dual-lighting System (ISDLS), an approach that enables the simultaneous signal amplification and detection of HER2 nucleic acid and protein in a single pathological section. ISDLS ensures excellent consistency with clinical outcomes for IHC ≥ 1+ and FISH+ sections, as well as improved accuracy for IHC ≥ 1+ and FISH- sections.

Hellgren et al. develop AEGIS, a transcriptomics-based tool that quantifies apoptosis-linked gene signatures to assess drug-induced liver injury risk across doses, in vitro models, species, and modalities. AEGIS ranks compounds by liver risk, translates from in vitro to in vivo, and can accelerate safer drug development.

Ajmera et al. use brain imaging and machine learning to dissociate young adults with loudness hyperacusis from matched controls. There are intrinsic functional connectivity differences related to emotion and cognitive processing between the two groups, but fewer differences with respect to auditory processing.

Boué et al. longitudinally assess cognition in patients aged ≥70 years receiving hormone therapy for metastatic prostate cancer. Poorer objective and subjective cognition occurs in patients receiving hormone therapy, particularly when androgen receptor pathway inhibitors (ARPI) are added to androgen deprivation therapy (ADT).

Constantinides et al., perform a recall study from a British South Asian genetic cohort to explore the genetic and phenotypic risk of cholestatic liver disease. 55.6% of participants with rare heterozygous ABCB4/ABCB11 variants or a history of intrahepatic cholestasis of pregnancy showed evidence of liver involvement, highlighting the utility of genetic screening and monitoring.

Snelling et al. describe the methods used to establish a pregnancy register in the QResearch database. The resulting algorithm successfully sorts and consolidates delivery and pregnancy loss (incl. terminations) records from primary and secondary care datasets, creating a valuable resource for future research into pregnant populations in England.

Myint, Lim et al evaluate the cost effectiveness, budget impact, and human resource requirements of different dialysis policies in Brunei Darussalam. Adopting a peritoneal dialysis first policy in place of the current haemodialysis dominant policy could generate cost savings while reducing additional workforce demand.

Georgia-Nektaria et al. investigate the levels of optimism and mental resilience among healthcare professionals in one of the largest hospitals in Greece. They find moderate overall levels of optimism and resilience, with older participants exhibiting greater levels of resilience and paramedics showing greater levels of optimism.

Wehnes et al. examine how the oral microbiome is associated with DNA methylation–based biological age and functional health in the oldest old. They find that biologically younger individuals have lower microbial diversity and distinct bacterial patterns related to short-chain fatty-acids and cognition.

Nderitu et al. develop a deep learning system to predict two-year referable diabetic retinopathy and maculopathy, augmented with synthetic retinal images generated by a conditional cascaded diffusion model. Augmenting training data with synthetic images improves internal test prognostic model performance but did not generalise on external testing.

Yu et al. developed AIPatient, a virtual patient that uses a large language model-based AI in curating a medical knowledge base that holds realistic and personality-aware clinical conversations. The system shows high accuracy in answering questions and medical-term recognition, with readable, reliable, and consistent outputs for medical training.

Fontes et al. report a case study of a person with recurrent glioblastoma for whom patient-derived zebrafish xenografts (zAvatars) were used to test 13 off-label therapies with radiotherapy. zAvatars identified pemetrexed and doxorubicin as effective, guiding treatment that stabilized the patient.