Digital Twins: Virtual Patients Shaping the Future of Medicine

A New Horizon in Medicine

“Is it possible to create a digital copy of a patient's body and use it to foresee disease progression, treatment response or complications?”

Digital twin (DT) technology boldly answers this question with “Yes.” Having revolutionized sectors from aviation to automotive manufacturing, this approach is now preparing to break new ground in medicine. It is becoming a cornerstone of personalized medicine, early diagnosis and AI-supported clinical decisions.

What Is a Digital Twin?

A digital twin is a digitally modeled representation of a real physical system (a machine, a city or a human organ). In healthcare, digital twins aim to virtually recreate a patient's entire body, specific organs (such as the heart, lungs or liver), biological systems (immune, nervous, circulatory) or processes at the cellular level.

These virtual representations are more than 3D models based on imaging. They also incorporate genetic profiles, metabolism, behavior, health records, environmental exposures, social circumstances and wearable device data. A digital twin can therefore be dynamically updated and “learn” through AI and advanced data analytics.

How Do Personalized Virtual Models Work?

Digital twin models are broadly classified into 4 main functional types:

1. Descriptive: Statically represents the patient's current health status. Example: 3D organ models

2. Predictive: Simulates disease progression or treatment response

3. Prescriptive: Recommends which treatment will be more effective under which conditions

4. Generative: Estimates and fills in missing or hard-to-obtain data

The success of these functions depends on the diversity and accuracy of the twin's data. Clinical, biological, physiological and psychosocial information must be integrated, modeled and validated together.

How Are Digital Twins Created?

Digital twins typically combine three core components:

• Data Layer: Clinical (EPIC, PACS), genomic (NGS), behavioral (mHealth), environmental (social determinants) and physiological (wearable device) data

• Modeling Layer: Mathematical models, machine learning, systems biology approaches, physics-based simulations

• Communication Layer: Interfaces enabling data flow between the digital twin and physical entity (sensors, applications, data pipelines)

The continuous feedback loop among these layers keeps the twin updated in real time.

The image aboveillustrates a digital twin model built using microscopic and macroscopic data collected throughout an individual's life (from birth to death). Information from the physical entity (patient reports, omics, environmental and social factors, mobile health, etc.) remains continuously connected to the twin, enabling prediction of health trajectories, identification of optimal treatment strategies and even virtual clinical trials. By strengthening modeling, early intervention and decision support, this structure forms a foundation for personalized medicine.

A Systematic Summary of Pioneering Digital Twin Studies

The chart belowshows the growth in scientific publications on digital twins over the years. A particularly striking rise in healthcare publications is evident after 2020.

The table belowcompares current digital twin applications and research findings in healthcare and oncology. It presents the designs, target disease groups, data sources and main findings of five high-impact studies in accessible terms. This review aims to improve understanding of the potential for clinical decision support, personalized treatment and integration into healthcare systems.

Applications

Digital twins can model the natural course of diseases and support long-term care planning as well as simulate treatment effects. Before symptoms appear, particularly when risk factors for inherited diseases are present, they may help predict disease onset.

These models may offer strong evidence-based arguments to improve adherence to preventive care.

Digital twins enable not only treatment of existing diseases but also prediction and prevention before disease onset, together with personalized lifestyle recommendations.

Integration with Medical Devices

One advantage of digital twins is integration with medical devices such as wearables. These devices provide real-time or periodic health data, allowing continuous updates.

Digital twins can also support medical device design and optimize implant placement by simulating integration with the body and physiological effects in advance, enabling more precise, personalized interventions.

The Evolution of Digital Twins: 4 Stages of Development

The study classifies digital twins into four stages and explains which levels are used in medicine:

Supporting Clinical Decisions

Digital twins simplify complex clinical data, reducing clinicians' cognitive burden during decision-making. Simulation and clear visualization enable more reliable, faster decisions in uncertain clinical situations.

Applications and Regulatory Considerations

Digital twin technologies show promise in neurology, cardiology, oncology and other fields. In Spain, for example, they are used to study recovery from psychosis and support cardiac arrhythmia diagnosis. Despite rapid development, the technology remains at an early stage, with regulatory and technical challenges. In the longer term, however, it has substantial potential for safe virtual experiments and evaluation of multiple treatment options.

Notable Digital Twin Applications

• SIMULIA Living Heart: An approved digital twin modeling cardiac function, developed through an FDA-Dassault partnership.

• FEops HEARTguide: An AI-supported digital twin for planning cardiac procedures such as TAVI (approved in the EU and USA).

• Project BreathEasy: A lung digital twin predicting respiratory support requirements in COVID-19.

• MindBank Ai and Babylon: Wellness-focused applications tracking lifestyle and stress levels.

The Digital Twin Revolution in Clinical Trials

In-silico clinical trials can evaluate drug and treatment efficacy in advance using digital simulations instead of real patients. With FDA and EMA support, indications for drugs such as palbociclib and alectinib have been expanded through digital twin-based studies. VICTRE successfully performed preliminary mammography device evaluations using synthetic breast images.

Personalized Oncology Applications

• Oropharyngeal cancer: A patient-physician digital twin modelused deep Q-learning and simulation to compare which chemoradiotherapy strategy (sequential or concurrent) was more effective in oropharyngeal squamous cell carcinoma. This study of approximately 536 patients (87% accuracy) showed that digital twin-guided sequential or concurrent models could be optimized for both survival and toxicity.

• Triple-Negative Breast Cancer (TNBC): Nature npj Digital Medicine (2025) According to the article, a biological mathematical model based on MRI data from 105 TNBC patients in the ARTEMIS trial created digital twins to predict pathological complete response (pCR). The model had 0.82% AUC performance. These twins optimized individual neoadjuvant chemotherapy across 128 dose-frequency combinations, increasing pCR rates by approximately 21–25%.

• Non-Small Cell Lung Cancer (NSCLC): Large datasets (25,000+ lesions) and predictive modeling have been used to optimize NSCLC treatment. These studies compared efficacy through digital twin simulations, particularly for pembrolizumab and chemotherapy options. Academic work in this area is supported by comprehensive reviews such as “From virtual to reality: innovative practices of digital twins in tumor therapy.”

An Advanced Application from Spain

The Spanish National Cancer Research Centre (CNIO) runs the “High-Definition Oncology in Women's Cancers” project, applying digital twin technology to women with advanced metastatic cancer. Broad data—including tumor metabolism, gene and protein expression, physiological parameters and psychological status—are integrated to create patient-specific twins. An important innovation is estimating biological age to assess whether disease accelerates or slows biological aging, then adjusting treatment intensity and combinations.

Clinical Validation and Data Infrastructure

Clinical validation requires comparison with real-world data to prove that digital twins add value beyond existing clinical methods. Harmonized, interoperable data infrastructure across countries is critical for global implementation.

Regulatory and Ethical Framework

Key issues such as data ownership, validation of clinical benefit and data equity still lack a clear regulatory framework. Although EU (GDPR) and US (HIPAA) laws exist, comprehensive digital twin-specific regulations remain under development.

Global Consortia

• SDTC (Sweden): Personalized drug screening using single-cell RNA sequencing.

• DIGIPREDICT (EU): Research focused on early intervention in cardiovascular disease.

• DT4H.org: An interdisciplinary collaboration platform for diseases such as sepsis, diabetes and lung cancer.

Future Projections

• Real-time whole-body digital twins through exascale computing.

• Digital twin applications in VR-based clinical training.

• Secure data-sharing platforms using blockchain technology.

Conclusion

Digital twins have the potential to create a paradigm shift in personalized medicine and clinical decision support. By integrating dynamic data and improving real-world outcomes, they enhance healthcare effectiveness, increase the accuracy of clinical decisions and transform health economics.