Sugar, spice, and everything nice- that’s what little girls are made of. Digital hearts, AI, and non-invasive sensors- that’s what digital patients are made of.
Perhaps one of the most famous examples of pairing technology, the precursor to digital twin technology, was saving the Apollo 13 crew in 1970. One of the oxygen tanks on board exploded two days into the mission and NASA was tasked with solving the technical malfunctions from 200,000 miles away to save their 3 astronauts [1]. Without the analogue model of Apollo 13 on earth, this rescue mission may not have been successful. Upon carbon dioxide rising to life-threatening levels in Apollo 13’s lunar module, NASA engineers created an improvised air purifier and instructed the astronauts on how to build it with materials inside the spacecraft using their model. Simultaneously, astronauts on earth ran simulations to test procedures on getting the crew back on earth alive and well, which they did successfully 4 days later [1]. Since 1970, analogue models have been replaced with digital models, but the concept remains the same: twin models enable the diagnosis of issues, status monitoring, and solution testing remotely.
Fast forward to 2018 and digital twins were named one of the top ten strategic technology trends by Gartner [1]. Partner digital twins with artificial intelligence (AI) and you get a model which allows you to analyse systems remotely in real-time, identify potential problems before they arise, test new products in virtual environments before building them, and allows for timely repair and replacements of critical components [1, 2]. The time has come where these concepts that have previously been used mainly to improve manufacturing processes, are now being applied to create digital replications of living things where data is seamlessly transmitted between the physical and virtual worlds [3].
Representing people in this way has led to the term digital patient or digital avatar [3]. Digital patients will take precision medicine to the next level, allowing healthcare to be tailored to anticipate the responses of individuals [1]. This can lead to not only better resolutions when defining the health of patients but can also change the perception of the model of a ‘healthy patient’. In the past ‘healthy’ has been viewed as the absence of disease symptoms, but now we can define what ‘healthy’ really is by comparing ‘healthy’ patients to the rest of the population [1].
Digital patients may be complied by integrating different measurements of a person over time to build a digital model of a body part, i.e. an organ, and eventually leading up to a fully integrated model of their anatomy and physiology. The eventual goal would be to assemble a personalised model of a patient over their lifetime which is updated with every scan or exam, including behavioural and genetic data [2].
These models of patients with all their organ and cellular functions combined with AI could lead to predicting how effective treatments will be for a specific patient in the future. You would be able to predict weeks or months in advance which patients will get ill, how each patient would react to therapy and ultimately who would benefit most [4]. It has the potential to revolutionise medicine as we know it by predicting outcomes, helping doctors make more precise diagnoses, and avoiding unnecessary surgeries. This would also lead to saving tens of thousands of dollars in cut costs [4].
While we are not at the point yet of fully-fledged digital patients, digital twin technology has demonstrated promising applications in one organ: the heart. It is difficult for clinicians to reconstruct and interpret the anatomy of a patient’s heart from a set of 2D images. Based on population data, these models also do not capture the unique characteristics of each heart, but a digital twin does [4]. A cardiologist at Heidelberg University Hospital is testing Siemens Healthineers’ digital heart software and has used the digital heart of a patient with congestive heart failure. He ran simulations to test whether their pacemaker could keep the patient alive before attempting surgery [4]. A 6-year trial is taking place doing just that on 100 digital hearts to test whether predictions match up with the actual outcomes. If results are promising, this system will be tested in a larger, multi-centre trial as the next step in getting the software approved by regulators for commercial use [4].
Another example of where this technology is being exercised is at the Swiss Federal Laboratories for Materials Science and Technology. A complex multi-physical skin model is under development which has been assembled with data from non-invasive sensor systems attached to human skin [5]. This together with a linked-up computer model are being used to simulate human physiological responses to drugs and medicine, i.e. changes in skin temperature or perspiration rate, which are being analysed to enable precise and personalised dosage of active agents. This is being used to reduce the dosage of, for example, pain killers, to such an extent that the patient gets exactly the right amount for them [5].
It will be interesting to see how these projects develop and which others will arise as digital twin technology and AI is adopted more and more in healthcare. Maybe we will all have a digital twin sooner than we think.
Mia Georgiou
References
- Van Houten, H., The rise of the digital twin: how healthcare can benefit. 2020.
- Van Houten, H., How a virtual heart could save your real one. 2020.
- El Saddik, A. Digital Twins: The Convergence of Multimedia Technologies – IEEE Journals & Magazine. of Publication: 03 August 2018; Available from: https://ieeexplore.ieee.org/document/8424832.
- Copley, C., Medtech firms get personal with digital twins. 2018.
- @HospiMedica, Digital Avatar Helps Plan Personalized Therapies. 2019.
