Rare Community Profiles is a Patient Worthy article series of long-form interviews featuring various stakeholders in the rare disease community, such as patients, their families, advocates, scientists, and more.

Volv’s Innovative Approach Harnesses the Power of AI for Rare Disease Detection

The EveryLife Foundation for Rare Diseases shared that, on average, individuals with rare diseases wait for 6.3 years for an accurate diagnosis. This diagnostic delay can have immense physical, mental, and emotional repercussions. In addition to being mistreated with the wrong therapies, individuals with rare diseases may feel isolated, alone, or like they aren’t believed by the healthcare system. Says Léon van Wouwe, the Clinical Innovation Director at Volv Global:

“The path to diagnosis is often a hard and arduous one for people in the rare disease community. It’s mindboggling – but also not. There are thousands of rare diseases, each of which expresses differently in individuals. Most healthcare providers are poorly equipped to recognize the subtle early symptoms of the many existing rare diseases. And how could they? Generalist healthcare providers are trained to deal with the masses and need to triage the people they see as effectively as possible. Specialist providers typically know a lot about specific diseases, but not more broadly. It’s not hard to see how people living rare, who often present with confusing and diffuse symptoms, are referred to different specialists over and over without ever ending up with the one person who can diagnose them correctly.”

But what if we could transform the diagnostic odyssey – and significantly reduce it in the process? That’s right: I want you to imagine a world where rare diseases are diagnosed in days, not decades. A world where emerging technologies like artificial intelligence (AI) and machine learning can sift through millions of data points, spotting patterns invisible to the human eye.

That is Volv Global’s mission – and one that van Wouwe supports wholeheartedly. Volv Global is a world-class leader in applying AI to healthcare with a focus on rare or difficult-to-diagnose diseases. Says van Wouwe:

“Rare diseases are very heterogenous in expression. It’s impossible for a doctor to always get it right. At Volv, we have conceptualized a space where machines can play a helpful role. With advances in machine learning and artificial intelligence, machines can crawl through data at scale much quicker and more efficiently than we can do as humans. We train our models on large-scale datasets and tune them specifically for rare disease expressions.”

Recently, Léon van Wouwe sat down with Patient Worthy to discuss the growing onset of artificial intelligence in healthcare, the importance of remembering the human dimension, and how Volv’s inTrigue methodology has already been used to improve diagnostics in Fabry disease and Pompe disease in the UK.

World Orphan Drug Congress: A First Look into Volv

I first came across Volv during the 2024 World Orphan Drug Congress in Boston, MA, where van Wouwe provided an overview of Volv’s AI algorithmic solutions. Currently, van Wouwe has over 20 years of experience in the healthcare industry at large. Volv provides the ideal foundation through which he can explore medical innovation and next-generation healthcare.

During the course of the conference, van Wouwe and I discussed the inTrigue methodology that Volv deploys, a topic we later expanded on during our interview.

Rare diseases affect millions of people worldwide. But, as described above, symptoms and manifestations often vary greatly. Physicians may not be aware of certain rare diseases, leaving themselves and patients feeling stumped. This, says van Wouwe, is where AI has the potential to make a difference:

“Primarily the problem that we’re trying to solve is that patients with a rare disease often don’t get diagnosed or get diagnosed super late. First and foremost, what we’re looking for is how we can address that. Can we learn more about these patients who typically don’t get diagnosed, characterize them better, and deploy the algorithm into the clinical system to flag clinicians for attention? If that doesn’t work, can we – from the highly complex algorithmic models – extract enough clinical insights to deploy for humans? I think it is indeed possible.”

Volv’s first answer to these questions is its inTrigue methodology. The website explains that:

“inTrigue is a new way of building computational models using machine learning to detect undiagnosed rare and orphan disease patients in population-scale databases such as electronic health records. It builds accurate, tailored algorithms for individual countries based on unique medical data structures and clinical diagnosis procedures to the disease in that country.”

This disease-agnostic methodology does not just create similarity searches based on diagnosed patients or combine existing data sets to create symptom lists or genome mutation libraries. It also doesn’t just search by genotype or phenotypes. So, what does it do? Says van Wouwe:

“Volv learns diverse and robust prediction models to characterize clusters of patients that go unnoticed and specify or map out where they’re being seen.”

inTrigue analyzes vast datasets including medical records, lab results, and genetic data to identify patterns that humans might miss. This prioritizes symptoms for physicians to pay attention to, provides more insight into the patient experience, finds patient cohorts, and mitigates risk. See how the different inTrigue components improve patient identification.

A Diagnostic Support Tool

Through inTrigue, Volv’s system can provide doctors with comprehensive summaries, flagging patients at risk of specific rare diseases. Additionally, it offers educational materials about these conditions, creating what van Wouwe calls a learning moment in time:

“We can point at where these patients are that are yet to be diagnosed to make it actionable. For example, we assisted with a project on acute hepatic porphyria where one question asked to key opinions leaders was where to find undiagnosed patients. The unanimous vote was gastroenterology or emergency wards. But our project showed that while 48% of typically unidentified patients had their primary interactions in gastroenterology, the remaining 52% mainly were seen by psychology, neurology or OB/GYNs. This discovery prompted the launch of educational materials in other areas to help physicians identify or recognize potentially undiagnosed patients.”

Another pilot study set out to explore Volv’s diagnostic model for Fabry disease and Pompe disease in the UK. For ten years, diagnostic rates for these two diseases have been relatively similar. In 2023, after implementing the algorithm, the diagnostic rate improved by 50%. The pilot study, though limited in scope, pointed GPs to over 160 individuals likely to have either Fabry disease or Pompe disease. Van Wouwe shares:

“If we look in literature, the typical uplift using an algorithmic approach is 2-5%. 50% is unheard of. While we are waiting to validate the data, it seems this approach was highly specific and successful.”

So far, inTrigue is available in England, Germany, France and the Netherlands as well as in the United States. At the same time, Volv is working on ways to offer the solution in countries like Canada, Australia and others also. Learn more about how Volv is deploying its technologies to care for rare.

The Real-World Impact of Volv Technologies

Volv’s approach also has another benefit: the revelation and understanding of systemic healthcare biases that may influence diagnosis rates.

Van Wouwe shared an example from the company’s work in alpha-1 antitrypsin deficiency. This disease is often seen in people of Nordic or Northern European ancestry. When a client wanted to understand whether a racial or ethnic bias existed in the diagnostic process, Volv dove in.

The company’s analysis found that diagnosed cases were underrepresented in Hispanic and Black communities. However, when Volv’s inTrigue predicted undiagnosed cases:

“The racial distribution evened out and the predicted undiagnosed patients showed more of a racial and ethnic spread. This suggests that unconscious biases may play a role in diagnostic rates, but working with AI tooling can address these biases.”

Through this study, Volv has detected that there may be approximately 57,000 undiagnosed individuals across the United States. This not only puts a burden on the healthcare system but leaves 57,000 people without answers – or adequate healthcare.

Volv’s objective, data-driven insights could promote more equitable healthcare outcomes across diverse populations. Says van Wouwe:

“We have been talking to the Alpha-1 Foundation. We want to see where the insights we generate would be helpful for patient advocacy groups and organizations.”

Diversifying the Approach

While Volv’s initial focus has been in the rare disease space, the company’s technology has broader applications. Their approach works for any condition with a complex diagnostic journey.

For example, Volv built the inAdvance algorithm to improve early detection and diagnosis of diseases like cardiomyopathies. As van Wouwe explains:

“The diagnostic delay is quite sad, really. If you catch cardiomyopathies early, there are newer classes of cardioprotective drugs that are good at keeping it at bay or even delaying its symptoms. This can prevent some of the tough outcomes associated with cardiomyopathies. But these diseases are often diagnosed quite late. Our models can demonstrably recommend a patient for diagnostic testing, between two and five years earlier than current clinical practice.”

Another algorithm – inFlow – could improve care in potentially life-or-death emergent situations. This prognostic support tool can differentiate between patients who are more quickly progressing towards severe outcomes and those with a slower-progressing disease course where more conservative treatments are sufficient. In HCM, says van Wouwe:

“Seventeen percent of people with cardiomyopathies are considered ‘fast progressors.’ It’s quite important to identify who they are as they should be monitored more closely to prevent serious cardiovascular complications.”

inFlow’s relevancy also extends into the oncology space – where Volv is expanding into. Shares van Wouwe:

“Treatment success for cancer strongly correlates to when treatment is started. Most cancer patients are only diagnosed in advanced stages. I’m very committed to improving early-stage cancer detection and improving survival odds.”

This capability could transform cancer care, helping doctors tailor treatment plans more effectively and make crucial decisions about adjuvant therapy.

Quantifying AI’s impact in healthcare is complex, but indicators for Volv are positive. In addition to cardiomyopathies and oncology, Volv plans to move into indications like diabetes, chronic kidney disease, and metabolic dysfunction-associated steatohepatitis (MASH) in the future. Van Wouwe sees this move as essential to reducing the burden and cost to the healthcare system while simultaneously improving patient quality-of-life.

Addressing the Challenges of AI in Healthcare Settings

As Volv continues to refine and expand its AI capabilities, the company faces several challenges: ensuring that healthcare systems can effectively act on AI-generated insights, building trust in AI, data privacy and security. Volv adheres to strict data protection regulations and transparency on how its system works. The company also emphasizes that Volv is a support tool, not a replacement for clinical judgment. Says van Wouwe:

“At Volv, we think addressing the human dimension in healthcare is extremely important. Technology is the tool that helps you solve the problem, but we’re using this tool in a setting defined by people. Our tools help patients. We help clinicians. And our tools are deployed in the healthcare setting, which is nothing without the people. You can have the best tools, but if people don’t know how to work with the tools or trust the tools, it’s not going to work. You have to be mindful of the impact and focus on the people. How do we deploy this technology in an actionable way without blowing up the system? That’s what we want to do.”

This philosophy shapes Volv’s approach to technological implementation; as the company moves forward, they plan to work with healthcare providers and patient communities to smoothly integrate AI technology into existing workflows. Volv’s balanced approach positions the company as a key player in the emerging field of AI assisted medical diagnostics.