A New Era Begins for Diagnostic Testing
From AI-powered insights to the rise of at-home testing, 2024 was a turning point for diagnostics.
Complete the form below to unlock access to ALL audio articles.
The diagnostics landscape has transformed over the last 12 months, fueled by advances in genomics, hyperscale biology and artificial intelligence (AI). As a result, at-home testing gained traction and the boundaries between research and clinical application blurred, innovation surged across disease areas from neurodegeneration to autoimmune conditions.
In this interview, Mara Aspinall, a leader in diagnostic innovation and policy and partner at Illumina Ventures, reflected on the defining moments of the last year and what lies ahead. From liquid biopsies and regulatory shifts to the expanding role of AI, Aspinall offered timely insights into how diagnostics are being reimagined.
Looking back over the last 12 months, what were the most defining moments or trends in the diagnostics industry?
Mara Aspinall (MA):
The diagnostics industry saw several major trends in 2024, punctuated by pivotal moments in technology and market adoption.
At a glance:
- Three foundational technological systems have become orders of magnitude more capable at lower costs - Genomics 3.0, hyperscale biology and AI
- As these advances converged, our ability to use less invasive samples to understand, diagnose and ultimately treat disease has been enhanced
- Perhaps the most notable is the widespread adoption of liquid biopsy tests for cancer screening, diagnosis and disease recurrence monitoring
- Further growth and acceptance of at-home testing and sampling
Genomics 1.0 was the identification of the human genome; Genomics 2.0 was the initial understanding of how mutations drive many diseases, both heritable and in cancer; Genomics 3.0 is a deepening understanding and integration of genomic tools from next-generation sequencing with transcriptomics (mRNA), epigenomics, proteomics and spatialomics used in clinical medicine.
Hyperscale biology is using these omics technologies to study biology in ever finer-grained detail (to the single-cell level) and to do so more broadly with greater throughput and scale. The multiomic data generated from both Genomics 3.0 and hyperscale biology is immensely powerful, but its complexity and volume make it challenging to isolate the most relevant clinical insights. Research data is no longer about thousands of data points but millions or billions of data points. And that’s why advances in AI and machine learning algorithms are so essential to allow us to process this data at two different scales: generating far more possibilities (e.g. therapeutic options) than the one-at-a-time decadal timeline permitted, while focusing down to differences in individual patients that define what will be an effective therapy for them.
Collectively, these advances have made it possible to analyze disease in humans with AI models in greater detail and discover novel diagnostic biomarkers, be they patterns in cell-free DNA fragments or the accumulation of an abnormal proteoform.
Last year saw many firsts in at-home diagnostic testing, beginning with those for COVID/flu, syphilis, home menstrual blood-based tests and the approval of the first over-the-counter home glucose meters for non-diabetics. The wearables that allow individuals and their physicians to track trends are now ubiquitous, and where the tests have to be done under professional supervision, we also saw the introduction of more patient at-home sample-taking technologies. This was most pronounced in sexually transmitted infection testing with self-collected vaginal and penile swabs. I believe that we will see more self-collection devices, including improved home blood collection systems, bridge the gap between the professional lab and the individual.
IE:
MA:
I believe the shift to at-home and over-the-counter testing is a fundamental shift toward decentralization of healthcare, which has changed and will continue to change the innovation pipeline. There are four reasons for this:
- The COVID-19 pandemic pushed much of the industry to focus on at-home test development and the creation of smaller and less expensive tests
- Many individuals are impatient to see their results as quickly as possible, and post-COVID-19 has shown that this is possible
- There is a critical shortage of skilled medical and laboratory staff creating delays in office and laboratory-based testing
- There is a simultaneous shift towards patients administering their own therapeutics at home – especially new is the patient administered injectable chronic disease treatments. Patients can administer their own treatments for conditions like psoriasis, Crohn's disease, neutropenia, obesity, asthma, migraine and more. Roughly 12% of the US population has used a GLP-1 antagonist. This is a significant shift in healthcare that’s taken place over the past decade. It’s natural, then, that the diagnostics paired with these tests would follow a similar trend
- Lastly, large segments of the population are concerned about their privacy. At-home tests, for those who can afford them, enable privacy.
Together, these pressures enabled by innovation have led to a steady increase in at-home diagnostics. The success of COVID tests encouraged a broadening of the innovation pipeline. The challenge ahead is reimbursement and integration into medical practice. Most home tests are not reimbursed by health insurance companies. In addition, many physicians do not trust the process and integrity of at-home tests as patients conduct them themselves.
But, like many first waves of innovation, I believe that acceptance and adoption will increase steadily in the next decade.
IE:
MA:
Without a doubt, we’ve seen tremendous diagnostic momentum for neurodegenerative diseases, including Alzheimer's disease, Parkinson's disease and other neurodegenerative conditions.
This gain in momentum is largely driven by the convergence of Genomics 3.0, specifically advances in proteomics technology that are able to generate highly sensitive data on a cost-effective basis. The field has also advanced in understanding the role of the immune system in a disease’s natural history. This will be the foundation of the next generation of tests.
IE:
MA:
The recent regulatory changes in the US have been focused on rolling back oversight and reducing regulatory barriers to the market. In the short term, I have no doubt that this will make the product development for labs and diagnostics companies easier. It puts fewer regulatory requirements on new products and should allow faster-to-market development pipelines.
In the long term, however, I am concerned that the lack of US Food and Drug Administration (FDA) oversight for LDTs will hurt the industry because it permits remaining diagnostic skepticism, making it more difficult for physicians to trust the adoption of a new test and reducing payers’ willingness to pay for that test. Without formal FDA oversight and uniform regulatory standards, the accuracy of tests will continue to be raised as a point of contention.
In Europe, the delayed rollout of diagnostics-related regulations will help companies prepare for future regulation.
IE:
MA:
I would like policy changes to increase the level of knowledge about diagnostics for physicians and all healthcare professionals. Medical schools are under enormous pressure to redesign their curricula as fast as the practice of medicine is changing. As part of this, I’d like to see diagnostics more fully integrated into both medical school curricula and continuing medical education requirements in every state, making diagnostics a part of physician and nursing board exams and requalification exams. There is currently too little emphasis in healthcare professional education on testing technologies and how to use tests to better guide clinical care.
While not a policy change, we need a nomenclature change. “Diagnostics” implies that it is a domain focusing on diagnosis alone, but our industry is so much more – it is the foundation of sustaining high-quality health, providing critical information on health status and selecting disease treatments when needed. I believe that the “Diagnostics” industry should be renamed the “Testing” industry. Calling it “Testing” helps make it clearer that we drive every aspect of health management from screening to risk assessment, to therapeutic choice, to monitoring and, of course, diagnosis.
IE:
MA:
The most significant growth will be in three areas. Number one is in our ability to diagnose neurodegenerative and autoimmune disorders early enough to start making a difference in patient outcomes. Except for a very few unfortunate individuals who have inherited damaging mutations from their parents, these diseases are more about transcriptomics, epigenomics and the resulting proteomics. Over the next few years, continued research into the natural history of these conditions will lead to better and more timely diagnostics and, ultimately, more effective therapeutics.
Number two is the use of so-called “alternative sample types”. In the past, the lower sensitivity of diagnostic technologies meant that samples had to be at source, i.e., where the definitive changes are most extreme – e.g., tissue biopsies for breast and prostate cancer; spinal column puncture for cerebral fluid; bone marrow for blood disorders. These traditional diagnostic methods are more invasive and higher cost. Today’s technologies are sensitive enough to find the “needle in the haystack,” e.g., of pathological proteins, DNA, etc. in very low concentrations to be found in peripheral blood. And it doesn’t stop there – urine, breath and sweat are other “haystacks” in which we are increasingly able to find disease “needles”.
AI and machine learning made their early engagement in healthcare by analyzing already digital data better than humans, e.g., imaging and electrocardiology. These days, it is hard to find any diagnostic technology that does not (or at least claims) to contain “AI inside”. Beyond this, what is now emerging is AI’s ability to find patterns not otherwise apparent. Examining and integrating large volumes of disparate data (e.g., multiomics) to find effective biomarkers from such untraditional aspects such as gait, speech and writing patterns. This momentum will certainly continue, even if it is not clear exactly where this is all headed, given AI’s extraordinary potential against a background of worrisome deficiencies (e.g., diverse/dynamic clinical backgrounds, out-of-bounds responses, errors of omission and inclusion).
