I’ve been thinking about this for a while, and I want to share something I’ve noticed both in my own practice and from my time in industry.
Most clinical decision support tools in oncology are built around the idea that what oncologists need is help remembering what the guidelines say. I understand why. Guidelines are the most visible, most structured part of clinical decision-making, and they’re relatively easy to encode. But in my experience, that’s not where oncologists get stuck.
We know the guidelines. We know the major trials. That’s not the hard part.
The conversations that don’t have a clean answer
The hard part shows up in clinic, usually several times a week, and it tends to look something like this.
A patient comes in with early breast cancer. She’s genomically high risk, clinically lower risk. We know the treatment landscape: chemotherapy, endocrine therapy, CDK4/6 inhibition. But she doesn’t want chemotherapy. No one does. So the conversation shifts to: how much does she actually lose by skipping the chemo and going with CDK4/6 and endocrine therapy instead? And if she does proceed with chemo, is adding CDK4/6 on top worth it for her specifically, or does the incremental benefit get lost in the noise of her individual risk profile? These are the questions that matter to her, and they’re the questions I’m least well-equipped to answer precisely. I can give her a directional answer based on the trial data and my clinical experience, but I can’t give her a number she can hold onto. And the further we move from the standard of care path, the less confident I feel in my counselling. Not because I don’t know the data, but because the data wasn’t designed to answer the question she’s actually asking.
Or a patient eligible for a kidney cancer drug. The trial that supports it enrolled a wide swath of patients, some very high risk, some not so much. “Intermediate risk” is a category that can mean a lot of different things. My patient falls somewhere in that range, and what I’m really trying to figure out is: what is the realistic ceiling on benefit for someone at their specific risk level within that population? Because the median benefit in the trial is not their expected benefit. And once I have a sense of that, the question becomes whether that benefit is enough to justify what treatment actually asks of them. The time, the side effects, the cost, the disruption to their life. That is a conversation I want to have with real numbers, and most of the time I’m working with far less than that.
Or the question that comes up more often than people might expect: what happens if we can’t complete the full treatment course? If a patient can only tolerate a portion of the planned cycles, what does the evidence say about expected outcomes from a partial course? Most guidelines are silent on this. Most tools don’t address it at all.
These conversations are unsatisfying in a way that I think is worth acknowledging. I’m a well-trained, experienced oncologist with access to the literature, and I’m still often giving patients answers that amount to “your situation is probably better or worse than average in this trial, so we might expect a bit more or less benefit than the headline number.” That’s not good enough, and I think we can do better.
What this means for clinical decision support investment
Here is something I think the industry underestimates. When a drug gets approved and has strong trial data, there is often an assumption that uptake across the eligible patient population will follow. The data is there, the indication is clear, the guidelines will reflect it. Oncologists will use it.
In practice, that’s not what happens. The addressable population ends up smaller than expected. Not because oncologists are ignoring the drug, but because the questions I described above don’t have good answers yet. Patients ask about the tradeoffs. Oncologists hesitate at the edges of the indication. The grey areas create friction, and that friction quietly compresses the market.
What I haven’t seen much of is meaningful investment in clinical decision support at launch or before it. The assumption seems to be that approval and data are sufficient. But approval tells you the drug works in the trial population. It doesn’t tell an oncologist how to counsel the patient in front of them who doesn’t fit that population cleanly, or how to quantify what they lose by modifying the standard regimen, or whether the expected benefit at their specific risk level is enough to justify the full treatment burden.
That gap is where I think there is a real opportunity. Investing in tools that help oncologists navigate those specific questions, at the point of the decision, could meaningfully expand how a drug gets used in the real world. Not by pushing guidelines, but by helping clinicians answer the questions their patients are already asking.
That is what we are working on with Precision Path, and it is the kind of collaboration I find genuinely interesting to think through with companies at the right stage of development. If any of this resonates with what you are seeing in your own programs, I would enjoy the conversation.
Dr. Henry Conter is a Medical Oncologist and Hematologist at William Osler Health System and the founder of Kesis & Sisters. He trained in Medical Oncology at MD Anderson Cancer Center and spent six years at Hoffmann-La Roche in progressively senior roles spanning oncology clinical development, portfolio strategy, and medical and regulatory affairs, across both national and global functions.