In late July, I was in the middle of an epic bike ride through the Alps, going over all the major cols: Galibier, Telegraph, Alpe du Huez, Col D’Iseran. I was a few days in and, despite being fully trained for the expedition, I was just not feeling the normal energy and exhilaration I usually get from being out in nature. My energy levels were the lowest I had felt in a long time. When I did a body inventory, I felt a lump the size of a mouse in my groin. I bailed from the trip and flew back to the UK. I couldn’t get my NHS GP to see me immediately, so I went private. After some pushing and prodding, I finally got an MRI scheduled which confirmed the tumor as malignant: stage 1 follicular lymphoma (FL).

Of all cancer diagnoses, this is hardly on the severe side. But like most cancer patients the C word was intimidating to me nonetheless. Standard of care in the UK for FL stage I is either “watch and wait” or local “involved site radiotherapy” (ISRT). Because of the proximity to the groin and hence important organs, ISRT was the recommended standard of care therapy.

Having some familiarity with some more advanced cancer treatment science from my board position at HealthTree, I felt that there should be better treatments available for this. And in fact there have been successful trials of experimental treatments such as this one on rituximab plus involved field radiotherapy with impressive five year progression-free survival (PFS) rates relateive to standard of care radiotherapy alone. When I discussed this with my oncologist he coldly informed me that he wasn’t willing to discuss finding me trials.

Thus I began my odyssey through the many worldwide trials databases. The difficulty with searching sites like clinicaltrials.gov and the WHO’s ICTRP is that they don’t have structured eligibility criteria. Instead you can just do a broad search for your disease and stage and then you must read the unstructured prose each and every trial to understand each of their participation criteria in order to figure out if the trial is a fit. This is in incredibly time-consuming and error-prone task. From my work with HealthTree I know that it takes their trained clinicians 10 hours per patient to find them appropriate trials.

As a result of these obstacles, only 2% of cancer patients sign up for clinical trials. 86 percent of trials are delayed for years because of such low participation. Patients are missing out on possible cures. Researchers can’t complete their studies. And the science sits still. This status quo needs to change. With my background building companies and products based on AI Large Language Models (LLMs), I felt called to help solve this problem.

So I decided to build a better approach, that we now call CancerBot. We gather all of the world’s clinical trials (in 14 different databases and counting) and use Large Language Models to extract the structured participation criteria. Patients can sign up with just their email (without providing their personal identity) and provide only a minimum of information such as their disease, stage and previous therapy. CancerBot then provides a list of trials that the patient is eligible for. No information is provided to anyone else including trial researchers, unless you choose to apply to a specific trial.

Some Follicular Lymphoma Trials That I’m Eligible For

CancerBot also provides a list of trials that are potential matches that are quite close. For each of the potential matching trials, we suggest which diagnostics, lab values or answers can be provided to make the patient fully eligible for the trial. Once the patient sees a trial of interest (based on the the specific treatment, the location or other factors) they can verify quickly whether they are a fit for the by looking the the AI-extracted trial criteria. Clicking Apply then sends all the patient data (with the participation criteria match highlighted prominently) to the researcher.

Right now, CancerBot supports follicular lymphoma and multiple myeloma. We will be adding support for other diseases one by one. For those patients still uncertain of their willingness to participate in trials, we will also be providing data-backed suggestions for standard of care treatments for the many cancers where the standard of care has many options.

If you are a cancer patient who is interested in pursuing better than decades old standard of care treatments, but are finding the trial searching as difficult as I was, I would encourage you to sign up for free today.