When it comes to imaging biomarkers in drug development, few experts have seen—and shaped—the field as much as Brad Wyman. With a career spanning over two decades across pharma, biotech, and medical devices, Brad has been on the front lines of innovation, witnessing firsthand how imaging tools have evolved from basic structural snapshots to powerful predictors of clinical outcomes.

In this interview, Brad shares his thoughts on the future of imaging biomarkers, the role of artificial intelligence (AI), and the common challenges and opportunities facing companies trying to do more with their imaging data.

Imaging Then and Now: The Evolution of Biomarkers

Brad takes us back to 1998, just as computer-aided diagnosis (CAD) tools were gaining traction in mammography. But the road to adoption was anything but smooth. “It was fascinating at the time—the resistance to such technology,” he recalls. “Experts resistant to accept that a computer-aided tool could be of any benefit.”

While regulators were careful to keep the radiologist at the center of interpretation, early CAD systems sometimes performed better without human intervention. But those results were often buried in fine print, lest they present a direct challenge to the human experts.

Flash forward two decades: AI is here and it’s directly challenging the experts and transforming imaging across the industry.

“There are still pockets of resistance,” Brad says. “But the utility is undeniable; and that it will replace the human component in a lot of what we do.”

Automation and Real-Time Feedback

The power of AI extends far beyond interpretation. Brad is particularly excited about how AI can improve imaging before analysis even begins, during acquisition. “There’s adherence to protocols, quality control, checking anatomy coverage, making sure the patient didn’t move—it’s all still mostly done manually,” he explains. “The ability to do it automatically, and in real time, means we could correct issues while the patient is still on the table.”

For modalities such as CT, where patient recall is complicated by radiation exposure, this kind of immediate feedback could be a game-changer by ensuring consistent quality imaging at every timepoint.

Regulatory Lag and the Validation Gap

One of the most persistent challenges in imaging is not just proving a biomarker works—but getting it accepted by regulators. Brad highlights the example of RAMRIS, an MRI-based scoring system for rheumatoid arthritis.

“It’s been widely studied. There are committees, standardizations, longitudinal and cross-sectional data. But the regulatory agencies still won’t accept it for a structural claim.” As a result, many sponsors have abandoned the tool altogether—despite the scientific support behind it. “Why bother with the time and expense,” he says, “if it doesn’t yield any commercial or regulatory benefit?”

At Bracken, we help clients navigate these complexities, bridging the gap between scientific innovation and regulatory readiness.

Go/No-Go Decision-Making: Getting it Right

With a strong background in consulting, Brad is often brought in to advise on early-phase go/no-go decisions. For him, it all starts with asking the right questions—before the trial begins. “If you measure this biomarker, what does it tell you that’s going to shape your decision? Has it been validated in a similar population or drug class?”

He’s also quick to distinguish between different types of biomarkers. Some are one-sided; they can clearly tell you when to stop. Others are two-sided: they might suggest both when to go forward or to halt development. Knowing the difference early can save time, money, and effort—especially when the stakes are high.

Trial Design: Pitfalls to Avoid

In Brad’s view, one of the most common missteps in trial design is what he calls “getting trapped in history.”

“Some people will be resistant to technology because they've had a bad experience,” Brad explains. “Sometimes when you dissect it, you realize it was just a poorly controlled study and quality just was horrendous because they didn’t put controls into place. Now they refuse to ever use that biomarker again.” This kind of legacy bias can limit innovation and cause teams to abandon potentially useful tools too early.

Another issue Brad flags is the misapplication of exploratory biomarkers. While these tools can offer exciting insights in early-phase research, they’re often not yet validated enough to support regulatory claims or to serve as definitive efficacy endpoints in later-stage trials.

“It’s a mistake to take an exploratory biomarker and try to make it a primary or secondary endpoint when it’s not ready for prime time.” Brad emphasizes the importance of knowing where a biomarker is in its lifecycle—and ensuring that sponsors apply it appropriately based on its validation status. At Bracken, we work with clients to evaluate their biomarker strategies critically and match them to the right study phase, indication, and regulatory context.

Looking Ahead: AI and the Promise of Discovery 

What’s next for imaging and AI? Brad is cautiously optimistic. After decades of seeing technologies evolve slowly—and sometimes fall short of their initial promise—he sees AI as a genuine inflection point in the field.

“I really think we may be ushering in a new age of discovery, with AI as the basis,” he says. “Hopefully, it will help us revitalize drug discovery and find new treatments.” But he’s also seen how early hype doesn’t always translate into immediate impact. When the human genome was first sequenced in 1998, it was widely heralded as the beginning of a golden age in medicine. In reality, it took years—decades, even—for that promise to translate into actionable therapies and 25+ years later have still have fallen short of the initial promise.

 “Hopefully AI will move faster. It can process massive amounts of data, make connections, and go deeper than any human could.”

Imaging, in particular, stands to benefit. AI can analyze thousands of high-resolution images, spot patterns across patient populations, and zoom in on subtle changes that even expert human readers might miss.

“Most people don’t have the attention span to really thoroughly examine stacks of images and scale in at the level they want to,” Brad notes. “AI is not limited by fatigue. It can certainly do that deeper dive and not only find things of finer and finer significance, but then start to understand the relevant meaning.”

For companies seeking to integrate these tools into their programs, Bracken offers the strategic, operational, and scientific expertise to help turn that promise into measurable progress.

Advice for Early-Career Scientists 

For those entering the field, Brad’s advice is clear: “Be data-driven. Get comfortable with AI. Even if you don’t master statistics, surround yourself with people who have.”

But perhaps most importantly, don’t lose sight of the clinical context. “Understand the disease. Understand the biology behind what you’re seeing in the image. Imaging is a tool—but it needs to be connected to the study goals, the drug mechanism, the endpoints and particularly the patient and their wellbeing.”

At Bracken, we help our clients do exactly that. Whether it's biomarker selection, regulatory positioning, or early clinical design, we bring clarity to complex decisions—so you can move forward with confidence. Contact us today about working with Brad and Bracken’s team of medical imaging experts.