In Dialogue with: Vit Nováček
In this edition of In Dialogue with, we speak to Vit Nováček, Scientific Researcher, University of West Bohemia, Pilsen, Czechia.
Vit has over 25 years of experience in experimental and computational biomechanics from both academia and industry. His work spans a broad spectrum of projects dealing with medical devices and their interaction with the human body. At the core of his expertise lies a commitment to advancing the concept of the Virtual Human Twin, leveraging cutting-edge modelling and simulation to transform personalised healthcare and improve patient outcomes.
For readers who may be unfamiliar with the concept, what is a Virtual Human Twin, and how does it differ from traditional medical modelling approaches?
A Virtual Human Twin (VHT) is a personalised digital model of a person's body that uses their own health and medical data to simulate how their body functions and how it may respond to different conditions or treatments. According to the European Commission's EDITH initiative, it is a dynamic digital representation of an individual's health that can be updated over time.
Unlike traditional medical models, which often rely on general population data and a "one-size-fits-all" approach, a Virtual Human Twin is tailored to the individual. It can help predict health outcomes, support clinical decision-making, and enable more personalised care by forecasting how a person's condition may change or respond to treatment.
Can you share examples of how Virtual Human Twins are already being used, or could soon be used, to improve diagnosis, treatment planning, or patient monitoring?
Several VHT technologies are currently in clinical use or research:
HeartFlow is a non-invasive diagnostic tool that creates a personalized 3D model of the heart to evaluate coronary artery blockages.
Bologna Biomechanical CT (BBCT) estimates individual hip fracture risk in osteoporotic patients by simulating various fall scenarios. The project reached TRL 7.
Persyst ESI is a neuroimaging solution that automatically combines scalp EEG data with a patient's MRI to perform Electrical Source Imaging (ESI). It pinpoints the origin of brain activity linked to seizures, helping clinicians accurately localize the epileptogenic zone—critical for surgical planning in epilepsy patients.
InSiliCare i an AI-powered twin providing personalized glycemic control for intensive care unit patients to prevent hypoglycemia.
PrediSurge is a French MedTech company based in Saint-Étienne that develops patient-specific digital twins for cardiovascular interventions, particularly aortic aneurysm repair and endovascular surgery. The company was cofounded by Stéphane Avril, a member of the PELVITRACT consortium. We are happy to have Stéphane and his team aboard.
How can modelling and simulation help clinicians move from a "one-size-fits-all" approach to truly personalised medicine?
In general, VHTs move beyond population averages by integrating multi-modal data (genetics, imaging, lifestyle, and medical history) to create a comprehensive view of an individual. Clinicians can then virtually test various treatment options on the computer model to identify the most effective intervention for that specific patient before implementing it in real life.
What are the biggest challenges in accurately modelling the biomechanics of childbirth, and how has the field evolved in recent years?
The primary challenge is the lack of in vivo biomechanical data during labour. Current computational models also rely on heavy simplifications and assumptions regarding the entire process of child delivery. Recently, the field has evolved from generic understanding toward patient-specific modelling, especially thanks to advanced medical imaging techniques including dynamic MRI or Shear Wave Elastography.
What are the current limitations in translating childbirth simulations from research environments into routine clinical practice?
Major hurdles include high computational time, as many simulations are currently too slow for immediate clinical decision-making. Additionally, there is a lack of rigorous clinical validation against real-world data. It is the ambition of the PELVITRACK project to make a difference in these aspects. We united three major European teams working on child delivery simulations for decades and top clinicians across the EU.
How might computational models support clinicians in choosing the safest delivery strategy for individual patients?
Models can provide personalized risk assessments for specific injuries, such as perineal tears, allowing clinicians to identify high-risk individuals. They enable clinicians to virtually test "what-if" scenarios for different delivery strategies, emergency procedures birthing positions or instrumental deliveries.
What role could patient-specific simulations play in improving outcomes for women during pregnancy and childbirth?
Patient-specific simulations can reduce the incidence of birth trauma, thereby preventing chronic postpartum conditions like incontinence, prolapse, and sexual dysfunction. By tailoring care to an individual’s unique anatomy, these models move childbirth care from a "one-size-fits-all" approach to personalized prevention. They also help empower women by providing visual, evidence-based explanations of their delivery options and risks.
How important is interdisciplinary collaboration between clinicians, engineers, computer scientists, and industry partners in advancing this field?
Realizing the VHT is beyond the capabilities of any single organization or discipline; it requires an inclusive ecosystem of clinicians, engineers, computer scientists, regulators, and industry partners. This collaboration is essential to bridge the gap between academic research and clinical deployment, ensuring models are both scientifically robust and practically useful for patients.
How can projects such as PELVITRACK contribute to the broader vision of Virtual Human Twins in healthcare?
PELVITRACK aims to develop the first-ever real-time solution to predict pelvic floor tearing during delivery. It embodies the broader VHT vision by integrating multi-modal data, combining mechanical biomarkers with Digital Twins,to transform a reactive diagnostic process into a personalized, preventive maternal care tool.
What is the most exciting breakthrough you expect to see in computational childbirth modelling over the next decade?
Despite significant progress in computational childbirth modelling, many important physiological mechanisms remain to be incorporated into our simulations. Current models lack detailed representations of vascularization, innervation, and tissue remodelling. In most cases, the foetal head trajectory is prescribed rather than emerging naturally from the simulation itself. While separate models exist that simulate foetal expulsion from the uterus, these have not yet been fully integrated with models that capture the complete passage of the foetus through the birth canal.
One of the most exciting developments I expect over the next decade is the convergence of these individual modelling approaches into comprehensive, patient-specific Virtual Human Twins. In childbirth, we are effectively dealing with at least two interconnected virtual twins—the mother and the foetus—whose mechanical, physiological, and biological interactions determine the outcome of delivery. In the future, these coupled digital representations could evolve into highly realistic simulations capable of predicting labour progression, assessing risks, and supporting personalized clinical decision-making. The prospect becomes even more fascinating when considering multiple pregnancies, where three or more interacting virtual twins would need to be modelled simultaneously.
At the European level, substantial investments are being made in VHT technologies, and dedicated VHT Platform is currently under development. As these infrastructures mature, I expect computational childbirth models to become an integral component of the broader VHT ecosystem. This integration will enable richer multimodal representations by combining medical imaging, biomechanics, physiology, genomics, and longitudinal clinical data into unified patient-specific models.
Artificial intelligence will play a central role in this evolution by accelerating model personalization, uncertainty quantification, and real-time clinical predictions. Looking further ahead, Quantum Computing (QC) may provide additional opportunities for tackling some of the most computationally demanding challenges. Future quantum and hybrid quantum-classical algorithms could accelerate large-scale biomechanical simulations, parameter estimation, inverse problems, and the training of complex AI models that underpin Virtual Human Twins. While QC is unlikely to replace classical High-Performance Computing (HPC) in the near term, it may become a powerful complementary technology for enabling processing of huge datasets and exploring complex physiological systems.