Digital twins in medicine

Abstract

Medical digital twins, which are potentially vital for personalized medicine, have become a recent focus in medical research. Here we present an overview of the state of the art in medical digital twin development, especially in oncology and cardiology, where it is most advanced. We discuss major challenges, such as data integration and privacy, and provide an outlook on future advancements. Emphasizing the importance of this technology in healthcare, we highlight the potential for substantial improvements in patient-specific treatments and diagnostics.

Figure 1.. Applications for medical digital twins.

1) Keeping healthy patients healthy. For a given patient (in yellow), a safe cholesterol level is determined, using genetic information, family history, and other data. The yellow line indicates the trend of the patient’s cholesterol levels over time, if untreated. Yellow boxes represent measurements. The patient’s digital twin (blue), on the other hand, forecasts the trajectory and recommends periodic preventive interventions (blue arrows), resulting in cholesterol levels following the blue curve. 2) Restoring health in ill patients. Upon admission to the ICU the patient (green) is being evaluated and receives initial treatment. A computer algorithm personalizes an appropriate computational disease model, together with information from a database of reference patients to recommend optimal interventions. As more repeated measurements are taken from the patient, the reference population is refined, the model is re-calibrated to the patient at later time points, and the recommendations for optimal treatment are refined. The cones represent the likely trajectory of the infection as determined by the digital twin. With time and larger number of patient datapoints the uncertainty in the predictions decreases (the cone becomes narrower) and subsequent patient timepoints fall closer to the center of the previous prediction cone. The improvement in the parameter ensemble that describes the patient is reflected the corresponding virtual cohort that describes the patient at each timepoint, which is depicted as increasingly containing more females and green subjects, like the patient being treated. 3) Development of novel therapeutics. Currently, clinical trials typically involve the use of animals and patient cohorts; left panel. With the advent of MDTs, it will be possible to reduce the number of animals used in preclinical trials, and to optimize patient trials using virtual patients. They can be used to screen large numbers of drug targets, drug candidates and to perform initial optimization studies using large numbers of patient MDTs and virtual patients; middle panel. Optimal drug regimes, doses and combinations can also be inferred by MDT prior to administering drugs to patients, thus minimizing side effects; right panel.