In the literature: Summer 2026 highlights
Discover the recently published papers from our community!
Bührer, Lea, et al. “Disentangling determinants of one-year modified Rankin scale in patients with incidentally detected solitary intracranial aneurysms.” Computers in Biology and Medicine 210 (2026): 111731.
Unruptured intracranial aneurysms (UIAs) affect 3 %–5 % of the population and are increasingly detected incidentally. Although rupture risk is low, UIAs pose clinical challenges, as rupture can cause severe disability or death, and treatments carry complications.
——————————
Argus, F., et al. “Accounting for approximation errors using surrogate-based parameter estimation of cardiac mechanics digital twins.” Computer Methods and Programs in Biomedicine (2026): 109528.
Parameter estimation for complex physics-based cardiac models is computationally demanding. Surrogate models can be used to speed up model evaluations and improve the feasibility of estimation and uncertainty quantification. However, the use of surrogates introduces additional sources of error that, if neglected, can cause bias or overconfidence in inferences. Here, we present a general approach to account for such model errors when carrying out surrogate-based parameter estimation and uncertainty quantification.
——————————
Saqawa, Mohamed, et al. “Next-generation osteosarcoma models for precision medicine.” Communications Biology 9.1 (2026): 966.
Osteosarcoma (OS) is a highly aggressive bone malignancy that predominantly affects adolescents and young adults. Despite the progress in conventional treatment approaches, such as surgery and chemotherapy, patient outcomes remain poor due to OS metastatic potential, chemoresistance and frequent recurrence. Although recent advancements in novel therapeutic strategies, such as molecular inhibitors, gene-based interventions, alongside immuno- and radiotherapies, have emerged in recent years, OS is still not well understood due to its complexity and heterogeneity. Tissue engineering and predictive preclinical models offer a good toolbox to study OS and produce patient-tailored therapeutic protocols. This review discusses the recent development of tissue engineering-based strategies and OS mimicking preclinical models like 3D in vitro models, organ-on-chip technologies, and predictive computational (in silico) models, ranging from mechanistic (white-box) to data-driven (black-box) and hybrid (grey-box) approaches. Based on these recent developments, researchers can better replicate the native OS tumour microenvironment, which opens the doors to a more representative high-throughput drug screening tools and patient-tailored treatment strategies.
——————————
Mukherjee, Satanik, Raphaelle Lesage, and Liesbet Geris. “A multiscale modeling approach to study the role of mechanics and inflammation in pathophysiology of articular cartilage.” bioRxiv (2025): 2025-12.
Mechanical loading regulates chondrocyte health in articular cartilage. While physiological stimuli maintain homeostasis, supra-physiological stimuli from joint injuries disrupt it, leading to osteoarthritis (OA). OA progression involves complex mechanical and biochemical interactions across multiple length scales, which are challenging to investigate experimentally. In silico models provide an effective framework to explore these mechanisms. This study developed an integrated multiscale modeling framework for articular cartilage. It combined finite element (FE) models at tissue and cellular scales with an intracellular gene/protein regulatory network. The network incorporated key chondrocyte mechanotransduction and inflammatory pathways. It was implemented using a semi-quantitative formalism, capturing the directional and qualitative interplay between mechanical and inflammatory stimuli on chondrocyte biology, rather than quantitatively predicting absolute gene expression levels. A Hill’s function was applied to link cellular forces from the FE model to a mechanical loading input to the regulatory network. Hill’s function constants were calibrated through a genetic algorithm by matching simulated and experimental gene expressions of COL-II and ADAMTS5 in cartilage explants under 20% cyclic compression. As a validation step, model simulations were performed at 10% cyclic compression of cartilage explants. Predicted sGAG loss matched the trend of experimental data. COL-II and ACAN were overestimated and ADAMTS5 was underestimated compared with experimental data. These discrepancies are consistent with the semi-quantitative nature of the model and are attributed to the simplified representation of inflammation-dominated catabolic pathways at low mechanical loads in the current framework. Simulated chondrocyte responses at different locations revealed spatial heterogeneity in chondrocyte activity. Overall, the multiscale modeling workflow developed in this study represents a first step towards a powerful platform for mechanistically deciphering the complex interplay of mechanics and inflammation in articular cartilage across multiple length scales.
——————————
Valerio, Thomas, and Enrico Dall’Ara. “Sensitivity analysis of musculoskeletal parameters and motion variability on the mouse tibial loading.” Journal of Biomechanics (2026): 113377.
Accurate prediction of bone adaptation in mouse hindlimbs requires reliable estimates of forces applied to the bones during locomotion. Musculoskeletal models are commonly used for this purpose, but their predictions are sensitive to variations in input parameters. Identifying which parameters influence the bone loading most is essential to improve predictions of mechano-regulated bone adaptation based on personalized finite element models. In this study a global sensitivity analysis of a mouse musculoskeletal models was conducted using the Morris screening technique to assess the importance of 173 input parameters (musculotendon properties, joint angles, and ground reaction force) on the tibial forces estimated during the stance phase. The absolute mean and standard deviation of elementary effects were used to rank parameters according to their overall influence and potential nonlinear or interaction effects. The sensitivity analysis was done for two cases: with and without the force-length-velocity (F-L-V) relationship. Excluding the F-L-V relationship led to realistic results while including F-L-V relationship led to high reserve actuators and unrealistic simulations. When the F-L-V relationship was excluded, a subset of 19 parameters was found to have a significant effect on the magnitude of the loads applied to the tibia. These parameters were related to maximal isometric force of 14 muscles of the mouse hindlimb, hip and knee joint flexion, vertical and anteroposterior ground reaction force. These results suggest that personalizing these parameters is critical for accurate determination of the physiological loads that act on the tibia and should therefore be considered when modelling bone adaptation over time.
——————————
Arminio, Mariachiara, et al. “Patient-specific Biomechanical Investigation of Percutaneous Pulmonary Valves: Towards the Integration of Routinely Acquired Clinical Data and Fluid-structure Interaction Simulations.” IEEE Journal of Biomedical and Health Informatics (2026).
Percutaneous pulmonary valves (PPVs) offer a minimally invasive option to treat pulmonary valve dysfunction in patients with congenital heart disease. Fluid structure interaction (FSI) simulations can provide a comprehensive assessment of heart valves biomechanics by capturing the coupled hemodynamics and leaflet mechanics. However, their application to PPVs remains limited. As the range of commercially available devices expands, robust characterization of PPV biomechanics may support the evaluation of short and long-term treatment performance. In this context, this study presents a patient-specific FSI framework for PPV simulation based on routinely acquired clinical data. Pulmonary artery and bifurcation geometries from four patients with implanted PPV were reconstructed from computed tomography. A PPV model was created and positioned within each patient-specific geometry. Boundary conditions were prescribed using patient-specific flow-rate waveforms and flow splits. Two-way, strongly coupled FSI simulations were performed to quantify transvalvular pressure drop, PPV-related flow patterns, and leaflets mechanics. Simulations reproduced peak transvalvular pressure drop in good agreement with available clinical measurements. Hemodynamics varied markedly across patients, with inter patient differences up to 393% in vorticity stretching and 337% in helicity intensity. Leaflet mechanics provided insights into the distribution of wall shear stress and first principal stress on PPV leaflets. Overall, the FSI approach captured PPV biomechanics across diverse patient-specific scenarios matching clinical pressure measurements and providing additional information on PPV biomechanical behavior. These results support the potential use of patient-specific FSI modelling to investigate PPV behavior under different patient-specific operating conditions.
——————————
Rouhizadeh, Hossein, et al. “The detectability paradox: bilingual medical report generation with open-weight models and the limits of human oversight.” Journal of the American Medical Informatics Association 33.7 (2026): 1303-1313.
The automation of medical report generation using large language models (LLMs) could significantly reduce physicians’ documentation burden while enhancing healthcare efficiency. However, the misuse of generative artificial intelligence in medical reporting can lead to important safety risks for patients. We addressed 2 questions: (1) What is the quality of medical reports generated by LLMs in English and French? and (2) Can we distinguish between human-written and LLM-generated medical reports?
——————————
Catalano, Alessandra, et al. “Microvascular health status as a novel predictor of acute side effects after radiotherapy: a quantitative analysis across three cancer sites.” Radiotherapy and Oncology (2026): 111693.
To quantitatively assess how microvascular conditions estimated via a non-invasive method can predict acute side effects after radiotherapy in patients with breast, prostate, and head and neck cancers.
——————————
