Digital twins in healthcare: Revolutionizing personalized medical care

In episode 9 of The Good Doctor, two physicians use virtual reality to experiment with surgical approaches on a model of a patient’s heart. Over the course of the scene they try different options and see if the virtual patient survives. It’s a creative and visual example of the possibility of digital twins in healthcare. In the rapidly evolving landscape of healthcare technology, digital twins are emerging as a groundbreaking tool to improve patient care, advance medical research, and accelerate treatment strategies. 

The concept of a digital twin usually involves three ingredients: a physical entity in the real world, a digital representation in a virtual world, and information exchange between these two. In the healthcare context, a digital twin is typically a replica of a patient, a healthcare device, or even an entire healthcare system. It integrates a vast range of data, including patient health records, genetics, real-time sensor data, and environmental factors, to provide a more accurate, predictive, and personalized view of health and treatment. 

Digital twins allow physicians and organizations to apply “what if” scenarios to healthcare problems. While implementation of digital twins is still nascent, there are some compelling clinical applications for the technology. 

Four compelling clinical applications

Personalized cancer treatment planning

In cancer treatment, digital twins are offering new ways to personalize care, predict outcomes, and optimize treatment strategies. By creating a virtual replica of a cancer patient (or even specific tumors), digital twins allow clinicians to simulate the behavior of cancer cells, predict how tumors will respond to therapies, and tailor treatments for better efficacy and fewer side effects.

Once a digital twin is created, clinicians can simulate how chemotherapy, immunotherapy, targeted therapies, or radiation therapy affect tumors. Then doctors can predict which treatment options are most likely to be effective, avoiding trial-and-error.

In 2020, the US National Cancer Institute, and the US Department of Energy, initiated collaborative projects at the intersection of computing and research to explore the development and implementation of predictive cancer patient digital twins. They explored the use in pancreatic cancer, melanoma, and other cancers. These studies demonstrated the opportunities and limitations of the use of digital twins in oncology. 

Cardiovascular disease management

By creating a virtual replica of a patient’s cardiovascular system, digital twins can help healthcare providers better understand the patient’s specific anatomy, predict disease progression, and personalize treatment.

Dr. Amanda Randles’ research takes 3D images of a patient’s blood vessels, then simulates and forecasts their expected fluid dynamics. Doctors who use the software can not only measure pulse and blood pressure, but also understand the blood’s behavior inside the vessel and the potential for heart disease. According to an interview with Quanta Magazine, “A decade ago, Randles’ team could simulate blood flow for only about 30 heartbeats, but today they can foresee over 700,000 heartbeats (about a week’s worth). And because their models are interactive, doctors can also predict what will happen if they take measures such as prescribing medicine or implanting a stent.” This video shows how the tech works – super cool!

Chronic disease monitoring

For patients with complex chronic conditions like diabetes, digital twins provide continuous monitoring and predictive insights. These virtual models can track metabolic changes, predict potential complications, and suggest real-time adjustments to medication or lifestyle interventions, enabling more proactive and personalized healthcare management.

An Oct 2024 study published in Nature reported outcomes of a digital twin study of 1,853 participants with Type 2 diabetes. The researchers found that the digital twin intervention significantly improved glycemic control, reduced the need for anti-diabetic medications, and enhanced overall metabolic health in real individuals with T2D over the course of a year.  

Surgical planning and simulation

Surgeons are experimenting with digital twins to create precise, patient-specific surgical simulations. By generating virtual replicas of a patient’s anatomy, medical teams can practice complex procedures, identify potential challenges, and develop optimized surgical approaches before entering the operating room.

A team at Louis Pradel Hospital, led by Dr. Antoine Millon, has leveraged digital twin technology for cardiovascular surgery. Their partnership with PrediSurge enables them to choose the optimal stent device dimensions, anticipate difficulties and adjust the device design preoperatively, reducing time to intervention and risk from surgery. 

Challenges and opportunities for digital twins

Various challenges exist with bringing digital twins out of research and experimentation and into mainstream medicine and healthcare operations. Running these models requires large amounts of data and significant computing power. Dr. Randle shared that her models are running simulations with up to 580 million red blood cells with terabytes of data. These analyses can’t be conducted on your typical laptop, but would require the use of a supercomputer. 

While public databases are available to start creating digital twin models, most publicly available datasets are not multimodal and not longitudinal. In the same way that AI requires a constant stream of updated data to train models, digital twins predicting health outcomes compared to simulated patients require feeding as well. A federated data donation approach by multiple institutions who all have access would accelerate learning and implementation of digital twins. 

Even with the best of data, a virtual model of a specific human will never be perfect. A study at Georgetown University simulated the effect of two different pancreatic cancer treatments and adjusted for dose based upon resistance to the treatment. The models showed that patient survival doubled, however despite the one million digital twin “patients”, there was difficulty mapping actual humans to the simulation data. While this technology is still nascent, creating an individualized twin versus finding a similar one in population level data might lead to better outcomes. 

Finally, the need for shared data to advance innovation needs to be balanced with privacy and security. There is an inherent risk in having digital twin data tied to specific individuals. This data should be protected with the same standards as HIPAA. 

While challenges remain, the potential of digital twins to revolutionize healthcare is immense and fascinating. By transforming how we treat patient conditions, manage population health, and avoid adverse outcomes, digital twins are paving the way for a more responsive, individualized approach to medical science.

Katie D. McMillan, MPH is the CEO of Well Made Health, LLC, a business strategy consulting firm for health technology companies. She is also a curious researcher and writer focusing on digital health evidence, healthcare innovation, and women’s health. Katie can be reached at katie@wellmadehealth.com or LinkedIn.