AMSTERDAM — Experts at HLTH Europe 2024 told conference participants that "digital biomarkers" are on the horizon.
Industry experts explained at a session on June 19 that these relatively new technologies leverage existing tools, like the ECG, alongside the latest wearable digital sensors.
Questions remain, however, about their clinical relevance or practicality for physicians.
Manuel Marina Breysse, MD, a cardiologist and CEO of Idoven, a company that uses artificial intelligence to analyze ECGs and detect cardiac patterns, believes they have much potential.
"If you look at history, every single time there is a new device, or a new way to measure something, that allows us to change everything," he told conference participants.
Blood tests, x-rays, CT, and even the microscope are examples of former game-changing technologies, he explained.
"For me, digital biomarkers are the same. It's a way to measure something that has never been measured before. The beauty of that is we are dealing with big data [and] super-computational power." This is giving medicine "a completely new way" to approach the diagnosis and management of a disease, he said.
Digitizing Biomarkers: What Does It Mean?
Medscape Medical News spoke with experts to examine how far the field of digital biomarkers has progressed and the challenges it currently faces.
Jakob Nikolas Kather, MD, professor of clinical artificial intelligence at Technische Universität Dresden, Dresden, Germany, explained that biomarkers are objectively measured indicators of biological or pathogenic processes or of pharmacologic responses to an intervention.
Although a bit of a stretch, he continued, the outputs of wearable technologies, like smartwatches, are sometimes included in this very broad definition as they can be used to monitor treatment effects.
Conventional biomarkers can also be digitized, explained Łukasz Kołtowski, MD, first chair at the Department of Cardiology at the Medical University of Warsaw, Warsaw, Poland.
An example is the ECG, which started out as an analog system producing traces on a piece of paper and is now moving over to digital, a journey also taking place with blood pressure and blood glucose level measurements.
From Concept to Clinic: An Ongoing Challenge
Thomas Berger, MD, professor and chair of the Comprehensive Center for Clinical Neurosciences and Mental Health at the Medical University of Vienna, Vienna, Austria, said there are a lot of "fancy tools and devices" coming in the market, but "one of the major problems, and also major demands, is that the validation and reliability of those tools needs to be…evidence based."
Kołtowski is conducting a review that has found a "long tail" of digital technologies that have not been validated in a population of patients independent of the population in which they were developed and tested.
He estimates that only around 1% apps, devices, and other technologies have been properly validated in this way, while the remaining 99% have been produced and released without any formal testing from a research point of view.
If a technology is externally validated, Kołtowski said, the next step on the road to the clinic is for it to be recommended in a European guideline. It's only after that when bigger players will get involved to push it out in the market, he explained, which is a journey that can take years. That is even before addressing the issue of national healthcare system reimbursement, he said.
But even if a novel technology does overcome these hurdles and has been shown, for example, to improve the quality of life or reduce mortality or hospital admissions, Kołtowski said, that doesn't mean it will automatically get adopted.
The reality is that it is "critical that the stakeholders in the system are able to use it" and incorporate it into their practice.
Berger explained many clinicians are concerned that digital biomarker tools will ultimately create more work for them, rather than ease their workflows, due to information overload.
Many novel technologies also face the problem of not fitting within preexisting hospital systems.
Some companies have tried to circumnavigate this by encouraging clinicians to install apps on their smartphone, for example, an app has been developed that analyzes ECG photographs to aid diagnosis, explained Kołtowski.
But this requires a credit card subscription or pay-per-use. Kołtowski said physicians do not use their private credit cards in clinical settings and are often not even allowed to use their private smartphones in the hospital to start with.
The Future
Despite the challenges, digital biomarker technologies are making strides in some specialties, such as the numerous examples of artificial intelligence being used to triage images in radiology. Kołtowski added that cardiology is also strongly adopting digital solutions.
This is not so much the case in oncology, however, where currently used biomarkers are "very simple and nondigital," explained Kather.
The hope, he said, is that future digital biomarkers can be used to simplify and improve the reliability of genome sequencing and immunohistochemistry applications.
An example is how we measure programmed death ligand 1 in some cancers to determine if the tumor expresses the antigen sufficiently for immunotherapy to be prescribed.
Currently, this is subjectively quantified by a human in a way that is "not very reproducible," Kather explained.
Computer-based image analyses could more reliably and reproducibly assess tumor staining, which would not only improve patient selection for immunotherapy but also "make clinical trials easier because if you have a more quantitative biomarker, you have…less uncertainty in your clinical trial recruitment."
"These are the types of systems that we can very soon expect in the clinic."
Neurology could also be a field soon benefiting from these technologies.
A Swiss company has developed a "digital loft" that uses sensors and video cameras to examine an individual's neurologic performance over 24 hours, for example, if they are at a risk for falling. This provides a much broader picture than a single-point consultation.
Given current workforce challenges, Berger said, it's currently not possible to monitor people for that long, but blankets can be equipped with sensors that set off an alarm if a patient who should not stand up on their own starts moving. That way nurses can be alerted to rush to a patient within a very short timeframe.
While it could prevent adverse outcomes in the hospital, "you can also transfer this in a home setting," Berger noted, adding: "Nevertheless, of course, it's far away from routine use."
Moving forward, challenges relating to privacy, equity, and environmental sustainability will need to be carefully balanced against cost and healthcare priorities for these new technologies to be applied, concluded Dylan Powell, lecturer in public health and innovation at the University of Stirling, Stirling, Scotland, in a recent News & Views piece for npj Digital Medicine.