In Pulmonary Hypertension, Digital Twins Could Speed Personalization


By creating digital models based on individual patients, clinicians might one day be better able to personalize care.

Since the dawn of medicine, practitioners have faced a frustrating problem: there is no guarantee a particular treatment will work on a particular patient. However, in the age of artificial intelligence (AI), some physicians believe technology may soon be able to sharply increase confidence levels when making treatment decisions.

One way of bringing about that certainty is the idea of “digital twins.” And if the concept works as hoped, patients with pulmonary hypertension (PH) might be among the first to benefit.

In the computer industry, a digital twin involves combining the specifications of a particular machine with AI and machine learning (ML) in order to create a replica that can be used to analyze, maintain, and optimize the original. Now, some are proposing the use of the concept in medicine. The theory is that by creating personalized “twins” of patients, clinicians might eventually be better able to understand how individual patients might respond to individual therapies.

Reinhard Laubenbacher, Ph.D.

Reinhard Laubenbacher, Ph.D.

Reinhard C. Laubenbacher, Ph.D., of the University of Florida, said while computer simulation and digital twin technology has become ubiquitous in many technology-focused industries, the same is not true in medicine.

“But there is now a lot of interest in medical digital twins that hopefully will translate into significant funding to jumpstart a concerted research effort,” he said.

Last month, he and colleagues published an article in the journal Pulmonary Circulation making the case for digital twins as a treatment tool in PH. On the one hand, it is a highly complex and heterogeneous condition. On the other hand, there is a vast array of actionable data on disease characteristics and patient outcomes. Both of those traits are ideal characteristics for ML and AI projects.

If all of that data can be leveraged to create individualized digital twins of particular patients, it might be possible to not only help optimize patient choices regarding existing therapies, but also to more quickly develop new therapies, they said.

Laubenbacher and colleagues noted how the data from decades of research related to bone‐morphogenetic protein receptor 2 (BMPR2) dysfunction in familial‐cases of pulmonary arterial hypertension helped lead to the development and approval of sotatercept (Winrevair).

“With eventual development of the digital twin system, one could imagine rapid development of individualized treatment based upon these granular databases, in combination with characteristics of known cellular signaling networks, organ‐level fibrotic and angiogenesis‐based physiologic assessment, and even development of in silico clinical trials to speed up drug development for patients, given the relatively rare yet complex and deadly nature of the disease,” Laubenbacher and colleagues wrote.

Andrew J. Bryant, M.D., a colleague of Laubenbacher’s and a co-author of the article, said the response to the paper has been “overwhelmingly positive.” However, while the concept has been well-received, he said there are issues related to implementation that still need to be worked out.

Related Content
© 2024 MJH Life Sciences

All rights reserved.