Convergence: Industry veterans share lessons learned from jumping CDx hurdles

Editor's note: This article was updated with quotes from EMA's Antonella Baron.

PITTSBURGH — Ambitions for precision medicine may be lofty, but behind the scenes are some tough challenges for developing companion diagnostics (CDx) that need to be overcome for the market to thrive. Industry experts brainstormed solutions during an 8 October session at RAPS Convergence 2025.
 
Precision medicine promises to deliver the right care to the right patient at the right time, guided by biomarkers. That’s a tall order, but one with a lot of potential.
 
Growing use of biomarkers across therapeutic areas is driving the global CDx market, which is expected to increase from $9.38 billion in 2024 to $1.75 billion by 2035, expanding at a CAGR of 11.75%, according to a recent report from Vantage Market Research. Use has been particularly strong in oncology, with 70% of drug approvals now including a CDx.
 
“It's no secret that precision medicine has slowly been changing the landscape of healthcare over the last 20 years or so,” said moderator Maritza Ward, RA/QA advisor at Pearl Pathways. “It's replacing the one size fits all approach to medicine with a more tailored approach, where biomarkers are leveraged to identify patients who may benefit most from a specific therapeutic and also identify patients who may be more susceptible to toxic effects.”
 
But the challenges in development and commercialization are not trivial, including how to design analytical and clinical validations when we're targeting small patient subpopulations, Ward added.
 
With any assay used in a clinical trial, there is rigorous validation testing for things like accuracy, precision, dynamic range and so on, said Doug Rains, chief scientific officer at Quantigen Biosciences, which provides diagnostic laboratory and contract research organization (CRO) services. For validation, samples that are biomarker-positive are collected from real-world patients – that’s generally the expectation of the US Food and Drug Administration (FDA).
 
“This becomes problematic, though, where we have a rare biomarker, because clinical samples tend to get prioritized for clinical studies,” Rains added. “They're not really sent to the lab as a top priority, and so as a lab, we're kind of left in the lurch.”
 
Consequently, Rains said they always look for some level of regulatory flexibility in communications with FDA.
 
“It's important for them to understand what limitations the lab's facing, and what are the ways that we can overcome those, in terms of getting alternate sources of samples,” Rains explained.
 
Options include getting samples from related clinical trials, taking samples from a different tumor type with similar mutations or using contrived (artificially created) samples.
 
“But I think it's really important, however you're doing it, to come up with a plan,” Rains said. “And be very honest with the FDA about what the certain limitations might be of taking that approach.”
 
Developers may need to bring in consultants to develop these rationales for alternative samples and negotiate with FDA upfront, advised Christine Gathers, senior consultant at À propos Regulatory Consulting Services and former big pharma executive. They may also want to run a disease history study in advance of a pivotal trial to understand more about the patient population and collect samples for use in analytical studies, added Gathers, whose experience includes leading Eli Lilly’s companion diagnostic and biomarker qualification efforts.
 
Diversity matters
 
Gathers also addressed diversity in clinical validation studies for conditions with small patient populations. From a global perspective, the prevalence of a biomarker could be very different in the US, Europe, or Japan, and that has to be considered when you are developing your regulatory strategy, she advised.
 
“I think this is where we're lacking as an industry, especially for understanding the performance of an assay and making sure that we cover the right enrollment across different ethnic backgrounds, to be sure that we're being able to identify if there are issues with using the assay and get the right drug to the right patient,” Gathers explained.
 
Another challenge when working with small patient populations is finding a suitable cut-off for immunohistochemistry (IHC) assays or immunoassays for pivotal trial eligibility, especially with expedited programs aimed at supporting accelerated approvals, Gathers said. You don’t want to include patients unlikely to get any benefit from the test drug, but at the same time you don’t want to exclude those patients who will, Gathers said.
 
It's important to collect as much information as you can in Phase 1, Gathers advised. Developers should look at the literature and try to understand what that prevalence might be, and a disease history study may be very appropriate to try and collect additional information and establish the right IHC cut-off, she said.

In the EU, regulatory bodies request and outline the type of evidence they need on the analytical and clinical performance of the test that is used to select patients, said Antonella Baron, senior product lead and companion diagnostic topic lead at the European Medicines Agency (EMA), Netherlands, who participated live remotely.

“The cut-off selection really can enlarge or restrict the population, and therefore the indication,” Baron said. “It’s important to have enough information.”
 
Managing prescreening challenges
 
Panelists also addressed challenges in prescreening with assays to enrich precision medicine trials with appropriate patients for a test therapy. Companies often use local laboratory developed tests (LDTs), which are not FDA-reviewed – not the assay that will become the marketed companion diagnostic – to speed enrollment.
 
“There's a lot of drive to get the study going quickly, to get patients enrolled faster, to get the drug to market quicker. And sometimes that comes at the cost of the CDx planning,” said Amanda Jensen, PhD, companion diagnostic development lead at Genentech.
 
When it comes time to file for approval, sponsors may be asked by regulators to perform a bridging study to show the LDT used in a pivotal trial and the CDx that will be marketed are equivalent.
 
“If you don't have enough concordance, that can be really problematic, and you can actually put your development at risk,” Jensen said. “Theoretically, FDA could withhold your drug approval for the CDx.”
 
From a technical standpoint, obviously not all assays are created equal, Rains commented. With some technologies, for example next generation sequencing, the risk is lower – most labs tend to get comparable data.
 
“But when you look at something like immunohistochemistry, the variability is astounding – sometimes shocking – and that's where you could really run into some trouble,” Rains added.
 
For bridging studies, sponsors will need to re-test samples from screened patients, including patients who were negative and screened out. If you're not using the CDx for study enrollment, you could find yourself in a situation where you don't actually have access to negative samples for the bridging studies, Rains said.
 
Sponsors have to really think about trying to find a means of developing an analytically validated assay that may fall under an investigative device exemption (IDE), Gathers advised.
 
“All of this should be discussed up front with the FDA,” Gathers said.

In the EU, things work a bit differently than the US – there are two development pathways, one for the medicinal product and one for the diagnostic, explained Baron. At a specific time point during the clinical study of the medicinal product, sponsors are required to assess performance of the test that might become the companion diagnostic. 

“There should be an alignment between the clinical study and the performance study for the companion diagnostic,” said Baron. “This is one point of contact where there really needs to be concordance and agreement between the manufacturer of the diagnostic and the pharma.” 
 
Early birds
 
Planning ahead is also crucial for successful CDx co-development programs involving pharma and in vitro diagnostic (IVD) partners. Pharma companies and the diagnostic partners often are making assumptions about each other and not really speaking the same language, resulting in slipped timelines, for example when drug data are ready but the local validation for the diagnostic hasn’t been completed, Jensen said. Having a robust face-to-face dialogue at the very beginning of kickoff, complete with nitty gritty details, can help, she suggested.
 
“Really having a deep dive into the project plan between both the pharma partner and the IVD partner is absolutely critical,” Jensen said.
 
In addition to scientific plans, sponsors must be mindful of the legal agreements underpinning CDx co-development.
 
“Contracting takes so long, and the fact is that a lot of partners won't start until those contracts are signed,” Jensen advised. “Don’t forget about contracting. Talk to your legal team and try to get them to understand that.”
 
RAPS Convergence