Three separate professional research teams searched for clinical trials on my behalf. My oncologist’s team. The genetic testing company. An outside institution. None of them found the trial I ended up in — the one at the National Institutes of Health that made me the first human to receive TCR-T cell therapy for my specific type of solid tumor.
I found it myself. And I had zero medical background.
That single fact should tell you everything about how the clinical trial search system works — and why you cannot rely on anyone else to do it for you.
Why Trials Matter More Than You Think
When you’re first diagnosed, clinical trials feel abstract. You’re busy processing the shock, starting chemo, learning words you never wanted to know. Trials sound like something for later — a last resort when standard treatment fails.
That’s backwards. The best time to start researching trials is now, even if you don’t need one yet. Keep them in your back pocket. Treatment plans change. What doesn’t qualify you today might be exactly what you need six months from now.
My oncologist actually encouraged me to do this research on my own — to have options ready “in my back pocket” should I need them later. That advice turned out to be one of the most important things anyone ever told me.
How the System Actually Works (And Why It Fails You)
Here’s what I learned the hard way: the system isn’t designed to find the best trial for you. It’s designed to find the most obvious matches.
My Tempus genetic report listed three trials. None were highly relevant to my mutations. My primary care provider’s research team found the same results. An outside institution found those plus a couple of long shots that weren’t even targeted to my specific mutation profile — more generalized approaches.
Meanwhile, my own searching — done on my couch with no medical degree — turned up highly specific results and trials none of them had located. Including the NIH trial.
Being honest, there’s a real satisfaction in knowing you’ve influenced your own outcome in a way no professional had managed to do for you. And here’s what matters most: I had absolutely no medical background and even less interest in it before cancer forced me to learn. If I could refine my searches to that level, so can you.
Step by Step: How I Did It
1. Get your FULL genetic report.
Request the complete file from the testing lab — Tempus, Foundation Medicine, whoever ran your panel. Not the curated clinical summary your oncologist shows you. That report is a subset, limited to the “most targetable” findings at the time. The broader dataset contains variants that might be relevant to trial eligibility or prognosis that no one flagged for you.
I had a buried variant that subtly favored long-term immune memory — a slight predisposition toward forming CD8 memory T cells. That would have been useful context early on when thinking about immunotherapy responsiveness. And it ended up being directly relevant to the response I had in the NIH trial.
2. Extract your mutation list.
Identify every reported variant, not just the “most targetable.” If the report is dense and technical, use an AI tool to translate the annotations into plain language and flag anything potentially relevant.
3. Search ClinicalTrials.gov yourself.
Enter each mutation one at a time AND your cancer type. Filter by “Recruiting” status. For me, three mutations stood out: TP53 Y220C and BRAF among them. I searched them together, then one by one, and always came back with new possibilities.
4. Cross-reference with AI tools.
Here’s an example prompt I used:
“Search for currently recruiting clinical trials related to the mutation TP53 Y220C. List them by relevance, trial phase, and recruiting status. For each trial, include the registry link, key inclusion criteria, major locations, and the primary endpoint. Cite primary sources with dates. Give a confidence score 0–10. State what you could not verify.”
More on AI tools below — they’ve evolved so fast that what I used mid-journey is already obsolete.
5. Search mutations individually AND in combination.
Some trials target single mutations. Others target pathway-level vulnerabilities. Cast a wide net.
6. Compare multiple search sources.
Your oncologist’s list. Your hospital’s research team. Outside institutions. AND your own search. The overlap between all four of my sources was embarrassingly small. Each found different things. None found the best thing.
7. Contact trial coordinators directly.
Most are surprisingly responsive to phone calls and emails once you track them down. Don’t be intimidated. You’re not wasting their time — enrollment is literally their job.
8. Check eligibility requirements carefully.
Some trials require measurable disease above a threshold. Some require washout periods from prior therapy. I actually had to let my tumors grow back to measurable size just to qualify for the NIH trial — they’d paused my treatment for washout, and ironically, I needed the cancer to come back enough to meet the enrollment criteria.
9. Keep a “back pocket” list.
Even trials you don’t currently qualify for may become relevant as your treatment evolves.
10. Recheck every 2–4 weeks.
New trials open constantly. Set AI alerts if possible.
What Finding That Trial Actually Felt Like
The truth is, I was terrified on the drive to NIH. I’d done my homework on this trial, maybe too much. Two different chemotherapies designed to obliterate the immune system. A list longer than my arm of potential side effects. A one-in-ten chance of ending up in the ICU. A minimum of three to four weeks in a hospital bed.
I never imagined I’d actually need this trial — it was just a fall-back option I’d kept in my pocket. But now here I was: eligible, accepted, and on my way to one of the most selective studies in the country. Both grateful and terrified.
A year earlier I’d been a typical middle-aged guy running a typical business. Work every day. Dinner every night. Life on repeat. Then, in what felt like a blink, I was outside Washington, D.C., inside a federal hospital, surrounded by some of the brightest medical minds in the world — about to receive an infusion no human had ever received for my type of cancer.
The experience was brutal. More than a month inpatient. Fevers every night. Weakness so deep that a walk to Starbucks one floor down nearly put me on the ground. But it was also profound. When Dr. Y told my wife, “This TCR-T therapy is many times more potent than any vaccine — and I’ve been working on those vaccines for two decades” — that single exchange shifted something inside me. What had been pure anxiety became something stranger: stress braided with excitement.
AI Changed Everything — And It’s Still Changing
I need to be honest about something: when I started my trial search, AI didn’t exist for me. Not in any useful way. I was doing this the old-fashioned way — one website, one paper, one email at a time. Hours on ClinicalTrials.gov, manually cross-referencing mutations, reading dense research papers I barely understood, and slowly teaching myself a language I never asked to learn.
Then, right in the middle of my cancer journey, the AI wave hit. ChatGPT. Then the flood of tools that followed. And suddenly, searches that used to take me days could be done in minutes. I could paste a dense genetic report into a prompt and get plain-language explanations of variants I’d been struggling to decode for months.
But here’s what’s wild: even that version of AI — the one that felt revolutionary to me at the time — is now ancient history. The tools available today are ten levels beyond where I started using them. The speed of this evolution is staggering, and if you’re reading this even six months after I wrote it, the landscape has probably shifted again.
Some of what’s emerged recently:
- DeepSomatic (Google/DeepMind, 2025) — an open-source AI built specifically to identify cancer-causing mutations in tumor DNA. It catches variants that traditional methods miss, especially insertions and deletions. It’s a lab tool, not something you use at home, but it means the genetic reports you’re receiving are getting more accurate as institutions adopt it.
- TrialGPT (NIH, 2025) — an AI that matches patient profiles to clinical trial eligibility criteria. In testing, it performed comparably to human clinicians at determining whether a patient met specific trial requirements. This is closer to what you’d actually use: feed it your profile, get matched to trials.
- AlphaGenome (Google DeepMind, 2025) — can link specific genetic variants to disease mechanisms, including non-coding mutations that standard testing often ignores entirely.
The point isn’t to memorize these tool names — they’ll be replaced by something better before the ink dries. The point is this: AI is becoming your most powerful research partner, and it’s getting better faster than most oncologists can keep up with. Use it. But use it deliberately.
Treat AI as a tool, not an authority. Ask for sources and dates. Cross-check important claims. Follow the links and read the originals yourself. These systems produce confident answers even when the underlying information is incomplete or wrong. The skill now isn’t searching — it’s verification.
I started this fight with nothing but Google and stubbornness. If I’d had what’s available today, I might have found that NIH trial in a weekend instead of months. You have access to tools I couldn’t have imagined when I was diagnosed. Use every one of them.
The Point
Sometimes the difference between what’s available and what’s actually possible is just how far you’re willing to look. Every step in this process — from pulling your full genetic report to cold-calling a trial coordinator — is something you can do. You don’t need a medical degree. You need persistence.
Don’t assume that because your oncologist hasn’t mentioned a trial, it doesn’t exist. Don’t assume that because one research team searched and came up empty, there’s nothing out there. Keep searching. Keep asking. Keep refining your queries.
That’s how I found the door that no one else had opened for me. And if I could do it, so can you.
Keep reading:
→ You Are Not a Statistic
→ TCR-T Immunotherapy: What It’s Like to Be the First Patient
→ Scanxiety: How to Survive the Wait for Scan Results
The CRC Roadmap 
Leave a comment