In the literature: December 2025 highlights
Discover the recently published papers from our community!
Emmert-Streib, Frank, et al. “The role of digital twins in P4 medicine: A paradigm for modern healthcare.” npj Digital Medicine 8.1 (2025): 735.
P4 medicine (Predictive, Preventive, Personalized, and Participatory) offers a comprehensive approach to personalized healthcare, emphasizing both the transition from disease to wellness and the importance of preventive care. In this perspective, we propose a novel extension to P4 medicine. Specifically, we argue that combining P4 medicine with digital twins (DTs) introduces capabilities that elevate it far beyond its current scope. While P4 medicine provides a conceptual framework grounded in medical principles, digital twins offer a complementary framework for methodological realization. Together, these concepts form a synergistic pairing, interfacing through individual patients and their corresponding data. Furthermore, we emphasize that digital twins represent a new paradigm—not merely a method—characterized by four critical features: explainability, intervenability, learnability, and diversability. We believe that this integration unlocks capabilities that go beyond the reach of traditional bioinformatics and systems biology approaches.
——————————
Dominguez-Gomez, Paula, et al. “Can in silico models predict drug-induced cardiac risk in vulnerable populations?.” Toxicology Reports (2025): 102179.
These findings demonstrate that computational models capture variability in drug response across pathologies and sexes, offering a predictive framework for preclinical safety evaluations and supporting safer, more personalized drug development strategies.
——————————
Zheng, Yumin, et al. “A deep generative model for deciphering cellular dynamics and in silico drug discovery in complex diseases.” Nature Biomedical Engineering (2025): 1-26.
Human diseases are characterized by intricate cellular dynamics. Single-cell transcriptomics provides critical insights, yet a persistent gap remains in computational tools for detailed disease progression analysis and targeted in silico drug interventions. Here we introduce UNAGI, a deep generative neural network tailored to analyse time-series single-cell transcriptomic data. This tool captures the complex cellular dynamics underlying disease progression, enhancing drug perturbation modelling and screening. When applied to a dataset from patients with idiopathic pulmonary fibrosis, UNAGI learns disease-informed cell embeddings that sharpen our understanding of disease progression, leading to the identification of potential therapeutic drug candidates. Validation using proteomics reveals the accuracy of UNAGI’s cellular dynamics analysis, and the use of the fibrotic cocktail-treated human precision-cut lung slices confirms UNAGI’s predictions that nifedipine, an antihypertensive drug, may have anti-fibrotic effects on human tissues. UNAGI’s versatility extends to other diseases, including COVID, demonstrating adaptability and confirming its broader applicability in decoding complex cellular dynamics beyond idiopathic pulmonary fibrosis, amplifying its use in the quest for therapeutic solutions across diverse pathological landscapes.
——————————
Lissoni, Vittorio, et al. “Uncertainty quantification for patient-specific domain in virtual aortic procedures: application to thoracic endovascular aortic repair.” Biomechanics and Modeling in Mechanobiology 25.1 (2026): 6.
Simulating medical procedures requires accounting for inherent uncertainty in many numerical model parameters, such as material properties. Evaluating the impact of these uncertainties is crucial for identifying parameters needing precise definition and correctly interpreting simulation results. This study explores how uncertainties in modelling the aorta affect finite element outcomes of a thoracic endovascular aortic repair (TEVAR) procedure. Based on literature data, aortic wall thickness and mechanical properties were identified as the most uncertain. The aorta was modelled using shell elements with homogeneous thickness and assumed to behave as a linear elastic isotropic material. A design of experiments approach was used for uncertainty quantification and sensitivity analysis: wall thickness and Young’s modulus were varied over 11 levels in a full factorial design, resulting in 121 simulations. Uncertainty was quantified using statistical metrics such as mean, standard deviation, coefficient of variation, and 95% confidence intervals. Results indicate wall thickness significantly affects aortic wall stress (σaorta), with minimal influence on stent stress (σstent) and device opening area (OA). Conversely, Young’s modulus has limited impact on σaorta but affects σstent and OA to a greater extent. The highest uncertainty was observed in σaorta (~ 25% coefficient of variation), while σstent and OA showed lower variability (2.6% and 6.9%, respectively). These findings suggest that, in this model, accurate wall thickness definition is more critical than precise Young’s modulus for reducing uncertainty in wall stress predictions. Therefore, literature-based averages for Young’s modulus may be sufficient for simulating this procedure.
——————————
Savelli, Giacomo, et al. “In silico prediction of hip fractures: improved fall modeling and expanded validation across cohorts with diverse risk profiles.” Journal of the Mechanical Behavior of Biomedical Materials (2025): 107182.
Osteoporosis constitutes a significant global health concern, however the development of novel treatments is challenging due to the limited cost-effectiveness and ethical concerns inherent to placebo-controlled clinical trials. Computational approaches are emerging as alternatives for the development and assessment of biomedical interventions. The aim of this study was to evaluate the ability of an In Silico trial technology (BoneStrength) to predict hip fracture incidence by implementing a novel approach designed to reproduce the phenomenology of falls as reported in clinical data, and by testing its accuracy in three virtual cohorts characterised by different risk profiles. Three cohorts of 1270, 1249 and 1262 virtual patients (Finite Element models of proximal femur) were generated based on a statistical anatomy atlas. Fall events were modelled using a negative binomial distribution, which replicated the over-dispersed nature of falls among the elderly population. A multiscale stochastic model was employed to estimate the impact force for each fall event, and subject-specific FE models were used to determine fall-specific femur strength. Patients were classified as fractured if the impact force exceeded femur strength. Fracture incidence over a two- or three-years follow-up was predicted with a Markov chain approach. The model predicted 12 ± 4, 16 ± 3 and 37 ± 7 fractures for the three cohorts, in alignment with clinical data (8, 14 and 41 fractures reported respectively). In conclusion, BoneStrength could reproduce fall phenomenology and fracture incidence in diverse populations. These results highlight its potential for future applications in the development of hip fracture prevention strategies.
——————————
Baroni, Sofia, et al. “In Silico Model for Aseptic Loosening Prediction in Cementless Hip Stems: A Design of Experiments.” Journal of Biomechanical Engineering 147.12 (2025): 121008.
Hip arthroplasty is a common orthopedic surgery. Cementless hip prostheses are currently more common, especially in young and active patients. The number of procedures and revisions is expected to increase with life expectancy. Aseptic loosening (AL) is the main cause of failure. Osteoinductive coatings improve long-term implant stability by enhancing osseointegration. This study aimed to develop a computational framework for predicting AL, considering both the biomechanical factors involved in the osseointegration process and the biological response to osteoinductive materials. A finite element model of a human femur implanted with a cementless hip stem was coupled with a Finite State Machine to simulate osseointegration and tissue fibrotization. The osteoinductive coating was modeled by adjusting the maximum gap at the bone-implant interface that can be bridged by newly formed bone, as well as the bone growth rate. To explore population variability, a total of 27 cases were simulated, including three different stem sizes, initial stem fit, and loading conditions. For each case, both uncoated and hydroxyapatite (HA)-coated stems were evaluated. Overall, this modeling framework was able to predict improved osseointegration with the osteoinductive coating at two years of follow-up. In two cases, the HA coating prevented AL, which occurred for the uncoated stem. Osseointegration patterns were consistent with previously reported data. The availability of this pipeline enables the simulation of large virtual cohorts and, therefore, the development of an In Silico technology for stem design to estimate the risk of AL associated with different designs and/or specific coatings.
——————————
