Author: Dr. Vijay Rajagopal, A/PROF Biomedical Engineering from University of Melbourne
Co-Authors: Marcus Watson, Liam Murray and Joshua Chung
This article is featured in the second edition of MEMKO’s Academic Exchange, our quarterly newsletter sharing insights, research and thought leadership. Read the full edition here.
Cardiovascular disease (CVD) is the leading cause of death worldwide. Although advances in pharmaceuticals, surgical procedures, and implant technologies have reduced mortality rates, CVD continues to burden the healthcare system; in 2021-2022, $14.3B of the Australian healthcare budget was attributed to CVD[1].
Current CVD treatments are suboptimal. Pharmaceutical interventions mostly provide symptomatic relief, surgical procedures are invasive and not universally accessible, while implants (i.e. artificial hearts, valve replacements) are only temporary solutions. Resolving these limitations is critical to managing the growing cost of healthcare in Australia. Therefore, innovative approaches are being sought to screen for CVD earlier, identify effective drug targets in a time- and cost-efficient manner, and to assess the effectiveness of surgical interventions. An emerging tool for addressing these bottlenecks are virtual counterparts to patient hearts, ‘Digital Twins’.
Figure 1: Digital twin models of the heart are being created at all scales of biology – from the subcellular to the whole organ. The vision is to enable patient-specific modelling of cardiac function to assist precision medicine and improve patient care. This figure was created with biorender.
The heart is often compared to a mechanical pump, with its two atrial and two ventricular chambers creating pressure gradients that drive blood through the lungs to oxygenate organs and tissues. Yet, unlike a mechanical pump, the heart can dynamically vary its force and rate to meet demand, with short-term adaptations occurring over seconds to minutes. For example, during a sprint, the heart beats harder and faster to supply skeletal muscles with more oxygen.
Similarly, the heart can adapt to more sustained changes in demand, such as during pregnancy, endurance training, and in age-related diseases like hypertension. In these cases, the heart grows to accommodate greater blood volumes and generate stronger contractions. These adaptations are influenced by an interplay of factors involving the heart, vasculature, and nervous system.
Naturally, the complex regulation of heart function by interdependent processes makes it difficult to predict treatment outcomes or observe functional state. This problem is complicated further by limited clinical measures. However, it is precisely these sophisticated adaptations of the heart – and their malfunction in disease – that motivate the development of ‘Digital Twins’. Digital twins that biophysically encode the physiology of the heartbeat and its long-term adaptation mechanisms will enable healthcare providers to precisely track and predict patient risks and inform treatment strategies.
Since the advent of desktop computers in the 1970s, engineering scientists and physiologists have collaborated in pursuit of creating individual-specific digital twin models of the heart. These digital twins stretch beyond reconstructions of an individual’s cardiac anatomy, embedding algorithms that simulate cell- and tissue-level biophysical processes, from subcellular ion transport to ventricular mechanics. Modern workflows create patient-specific digital twins from medical data, including imaging and diagnostic test results[2].
We can now reproduce short-term adaptations like the “flight or fight” acute stress response. However, precise mechanisms that control long-term adaptation involving heart growth remain less understood, representing the next frontier in cardiac digital twin research.
Heart growth arises from the enlargement of individual muscle cells, the fundamental force-producing unit of each heartbeat. This cellular growth is regulated by an intricate network of mechanical and chemical interactions taking place between the cell’s internal and external environment, including extracellular matrix composition, circulating stress hormones, and local signalling activity[3]. At the Cell Systems and Mechanobiology research group at the University of Melbourne, we develop digital twin models alongside live-cell and molecular biology experiments to uncover how these factors collectively influence muscle cell function and growth.
Recent work by our PhD student, Liam Murray, reveals new details about the forces muscle cells generate during each heartbeat. Since the 1950s, mathematical models of muscle cell contraction assumed that these cells only generate unaxial forces. However, by combining high-resolution electron microscopy, artificial intelligence-based image processing, and finite element modelling, Liam demonstrated that non-parallel intracellular muscle fibres generate shear forces which may implicate intracellular signalling pathways that limit or enhance heart growth.
Further, our group is developing digital twins of the muscle cell nucleus, the compartment inside the cell housing DNA required for the transcription of proteins that fundamentally drive heart growth. We are also developing models of intracellular metabolism[4] as metabolic reaction networks regulate the energy required for muscle cells to contract and grow. Our goal is to incorporate new cell-level insights of mechanics, calcium signalling, metabolism, and structural remodelling into existing digital twin modelling platforms to better predict patient responses to treatments than what is currently possible.
In addition to advancing digital twin cardiac technology at the cellular level, we are developing new collaborations with clinical colleagues in the Parkville precinct to integrate digital twin technologies in patient diagnosis and prognosis workflows in the hospital. At Austin Health, our digital twin research capabilities are being applied to develop CT-image analysis workflows to help detect early signs of aortic dissection. Cardiac digital twins have the potential to make significant positive impact in intensive care where patient cardiovascular function is monitored real-time for the duration of their care. Any adverse cardiovascular events (AE) in this setting can leave patients with further complications and may even be fatal (~25% of critically ill patients experience at least one AE during their ICU stay, 12% of AEs were life-threatening and 2% resulted in death).
We aim to apply our digital twin modelling capabilities to develop new algorithms that can predict risk of AEs from real-time monitoring data. Reducing ICU cardiovascular adverse events by just 10% could save Australia approximately $22.9 million AUD annually, and nearly $2 billion AUD globally. A 50% reduction would yield even greater savings approximately $114.7 million AUD for Australia and $9.9 billion AUD worldwide—highlighting the immense economic value of improving ICU Cardiac safety. This collaborative endeavour not only allows insights from patient care to refine our research but also enables their translation to clinical impact.
For further details about our research and for potential collaborative opportunities, please contact A/Prof Vijay Rajagopal of the Department of Biomedical Engineering at the University of Melbourne: vijay.rajagopal@unimelb.edu.au
[1] A. I. of H. and Welfare;, “Heart, stroke and vascular disease: Australian facts,” AIHW, Canberra, 2025. [Online]. Available: https://www.aihw.gov.au/reports/heart-stroke-vascular-diseases/hsvd-facts
[2] S. A. Niederer, J. Lumens, and N. A. Trayanova, “Computational models in cardiology,” Nat. Rev. Cardiol., vol. 16, no. 2, Art. no. 2, Feb. 2019, doi: 10.1038/s41569-018-0104-y.
[3] J. Chung et al., “Calcium-dependent regulation of physiological vs pathological cardiomyocyte hypertrophy,” Biochim. Biophys. Acta BBA – Mol. Cell Res., vol. 1872, no. 8, p. 120046, Dec. 2025, doi: 10.1016/j.bbamcr.2025.120046.
[4] S. Ghosh et al., “Effects of altered cellular ultrastructure on energy metabolism in diabetic cardiomyopathy: an in silico study,” Philos. Trans. R. Soc. B Biol. Sci., vol. 377, no. 1864, p. 20210323, Nov. 2022, doi: 10.1098/rstb.2021.0323.