AI to shorten the rare disease diagnosis

More than 10,000 rare diseases are known today, yet the path to diagnosis remains one of medicine’s most complex challenges. Symptoms are often non-specific, allowing rare conditions to hide in plain sight, especially when physicians may encounter them perhaps once or twice in a career, if at all.

For patients and families, this complexity can mean years of uncertainty and, in some cases, irreversible consequences. It takes an average of five years to receive a diagnosis, involving over seven specialists and several misdiagnoses. Even then, fewer than 10% of known rare diseases have an approved treatment, underscoring the scale of unmet need.¹

Today, advances in artificial intelligence (AI) are beginning to change that trajectory. Alexion is applying these technologies, through both original innovations and strategic partnerships, to help shorten the diagnostic odyssey and bring new therapies to patients faster.

Watch the video to hear more from Alexion’s R&D leaders about promising areas of diagnostic innovation that are bringing new hope to millions of people impacted by rare diseases.

Applying AI to help identify rare diseases earlier

For people living with a rare disease, early diagnosis is critical to improving outcomes. Since the late 1960s, newborn screening has played a vital role in enabling identification of genetic disorders and early therapeutic intervention. Yet today, current newborn screening captures only a small fraction of rare diseases. More than 10,000 rare diseases are known today. Yet recommended newborn screening panels cover only about 40 core conditions leaving the vast majority of rare genetic diseases undetectable at birth.^2,3

To help close the diagnostic gap for the thousands of rare diseases beyond conventional screening panels, Alexion is collaborating with Rady Children’s Institute for Genomic Medicine to pioneer an approach to early detection as a founding member of a program called BeginNGS® (Begin Newborn Genomic Sequencing) that implements rapid whole-genome sequencing to identify rare genetic conditions before symptoms start. Research has already shown that this approach can deliver diagnoses for rare genetic diseases in just days, potentially leading to faster treatments, better health outcomes and lower healthcare costs.⁴

“Because roughly 80% of rare diseases have a genetic basis, I see genomic approaches like whole genome sequencing and creating AI algorithms that can help identify these patients as one of the greatest opportunities in diagnostics.”⁵ -Tom Defay, Deputy Head Diagnostics Strategy & Development at Alexion, AstraZeneca Rare Disease.

Not all genetic changes may lead to disease, however. To decrease the risk of confounding false alarms, Alexion is implementing AI to analyse large, multi-ancestry genetic databases to distinguish disease-causing changes from benign ones, which can reduce false alarms by 97%, bringing peace of mind to patients and families.⁶

While newborn screening is an important tool for early detection, many individuals living with rare diseases are not diagnosed until later in life. To continue to overcome barriers to diagnosis, it is critical to find new ways to identify undiagnosed patients and enable earlier treatment intervention.

Improving the rare disease diagnostic journey

Alexion’s commitment to close diagnostic gaps in rare disease extend well beyond newborn screening. For many patients, non-specific symptoms and inattentional blindness mean rare diseases hide in plain sight, delaying diagnosis for years. As Chathuri Daluwatte, Head of AI Diagnostics at Alexion, AstraZeneca Rare Disease, notes: “AI is the biggest changemaker that can happen to rare diseases in terms of diagnostics.”

Combined with clinical expertise, AI-based approaches can detect subtle signatures earlier that cannot be detected by the human eye and help uncover hidden patterns in clinical data that could assist healthcare providers in diagnosis and routine care. Through collaborations and partnerships, Alexion is exploring AI-based tools designed to support detection of patients who may warrant further evaluation for certain rare diseases, including cardiac amyloidosis and HPP.

Cardiac amyloidosis: As part of an ongoing collaboration, we are pursuing the development of an AI software as a medical device to support detection of subtle features associated with cardiac amyloidosis in routine echocardiograms aiming to facilitate early detection.

Hypophosphatasia (HPP): We are in an ongoing collaboration to advance the development of an AI-based clinical decision support software that analyses electronic medical records to connect data points that may be consistent with HPP. The approach is intended to evaluate a patient’s medical records to help detect patterns that warrant further assessment of HPP.

These technologies are being explored for their potential to support earlier detection of rare diseases prompting further diagnostic evaluation for patients, helping to streamline the diagnostic journey.

Looking Towards the Future

Artificial intelligence is redefining what is possible in rare disease – from early detection and faster diagnosis to treatment management. At Alexion, these innovations are helping shorten the diagnostic odyssey and expand treatment possibilities, moving us closer to a future where patients receive answers faster and can access therapies sooner.