Largest Medical Cannabis & Cancer Study Ever: Can These Findings Push Rescheduling? ft Ryan Castle

Listen Now

Ryan Castle: How do we use this information to change the medical establishment's view of how they approach medical cannabis? Because that's dangerous — if there's this much consensus and people are still not embracing it, then they're leaving tools on the table that could be used to improve people's lives. Bryan Fields: What's up guys? Welcome back to another episode of The Dime. I'm Bryan Fields, and with me as always is Kellen Finney. This week we've got a very special guest, Ryan Castle, research director at Whole Health Oncology Institute. Ryan, thanks for taking the time. How are you doing today? Ryan Castle: I'm very happy to be here. Thank you. Bryan Fields: Excited to have you here, Kellen. How are you doing? Kellen Finney: I'm doing well, I'm doing well. Really excited to talk to Ryan. How are you doing today, Bryan? Bryan Fields: Yeah, I'm stoked. We're gonna talk about a little medical cannabis and maybe a little AI stuff — dive into some of the elements that people have perceptions on that might not actually be correct. But before we get into that, Ryan, it'll be really helpful if you can give a quick background on yourself and how you found your way to the cannabis space. Ryan Castle: Yeah, so I actually found my way to the cannabis space trying to prove that cannabis doesn't work. My background is as a public health researcher, so I do consulting work — people hire me to find out what the facts and the truth are behind different controversial ideas. That's what I specialize in, really digging into hard numbers. When somebody says, "Well, it's complicated," I'm the person who tries to uncomplicate it. I was hired by an organization to investigate medical cannabis, specifically cannabis and oncology — its use in cancer. I had them sign paperwork up and down saying, "You may not like what I have to say, because I'm going to tell you exactly what the science says, no matter what." They signed off on it, and I started doing the research. My specialty is in public health applications and systems dynamics, meaning we look at as big a collection of data as we can get. I ended up putting together the largest set of studies on medical cannabis that had ever been done, and to my surprise discovered it works much more than anybody thought — to the point where I came back to the folks who hired me and said, "I was 100% wrong on this. The evidence changed my mind, and I'd like to work with you guys on a more permanent basis going forward." That's where I am today. Bryan Fields: A couple of high-level questions. What was your perception going in — that some of these perceptions people had just weren't data-accurate? And when you say "works," was there a specific segment you were looking at — pain, nausea — or just anyone who's used cannabis for medicinal purposes in general? Ryan Castle: The study initially looked to capture medical cannabis across all genres. We have background data for virtually every major symptom and condition out there — over 10,000 papers, over 800,000 data points, and the numbers get crazy at the levels of analytics. But specifically, we focused in on cancer, both in terms of symptoms and the cancer itself. When you ask what I meant by "it works" — my question going in was whether placebo and user assumptions were going to be a very large part of this, which is honestly the case with the majority of pharmacological interventions; most drugs the FDA approves are pretty close to placebo in efficacy. But what we found with cannabis was that not only were the effect sizes way higher than placebo in a lot of studies, but these rates of efficacy showed up for pain, nausea, fatigue, and sleep. It also showed up for tumor size, tumor growth, rates of remission, and rates of the cancer returning afterward. So it wasn't just treating symptoms — it was treating the disease itself. Our study specifically tried to look at this huge collection of studies and find the broad conclusions each one reached. For example, if there were 300 studies discussing cannabis and cancer apoptosis — the cell death of cancer cells while leaving healthy cells intact — we wanted to know what percentage reported a conclusion, and of those, what percentage found cannabis effective versus ineffective or inconclusive. With such a large data set, we're not reporting that 17 patients succeeded in their treatment goals — we're reporting the high-level rates of success across the scientific field, the points of consensus, and the points of disagreement, based on sheer math and machine learning. Bryan Fields: Taking it one step back — you went in skeptical, and at some point that skepticism started to waver. Do you remember exactly when that moment was, and can you tell us about it? Ryan Castle: Yeah, absolutely, I know exactly what it was, because I ended up having to eat some crow in front of my employers. To be clear, I've never had anything against medical cannabis, but I am a scientist — an ideological skeptic. I don't believe anything until I have a reason to, and then once I see the evidence, I'll believe it. I went in thinking maybe it'll be effective for a couple of things, but there's no way it's effective for everything I'd heard it listed for. It was originally going to be a relatively small review, but out of the first 15 papers I looked at, 14 said cannabis was effective, and at least five reported direct effects on cancer cells. That's not something you get with most pharmaceutical treatments. I went back and said, "I don't feel comfortable with this, it's a shocking finding for a controversial topic — let's look bigger." We did, and found pretty much the same thing. We kept going bigger until we did the biggest meta-analysis on medical cannabis ever done, because I couldn't get any larger a collection to make myself feel comfortable. We got every applicable study we could find — over 10,000 studies — and finally I came back and said, "Once we've reviewed everything that exists, I think I have to be satisfied." Bryan Fields: This study is published, right? You guys published it — it's available on the NIH website? Ryan Castle: Yes. Bryan Fields: Can you walk our listeners through exactly what that meta-analysis process looked like? You're talking about studies — whether on individuals or cell lines — with some sort of positive outcome. Is that what you're pulling out of each study to build the data set? Ryan Castle: Yeah, absolutely. There are several tiers of evidence. There are reviews of existing work, clinical trials where a treatment is actively given and tested, and double-blind studies testing against placebo. All of these are one-off collections of data — sometimes comparable, but usually standalone. The highest level of analysis in research are systematic reviews and meta-analyses, which is what we did. Instead of conducting individual research with 10, 20, or 100 people, we analyzed the findings of every one of those studies. Our estimate is that all told, roughly 50,000 patients were captured across our analysis, which we used to determine the overall direction science is taking — because people always say you can find a study for anything, and that can be true. Systematic reviews and meta-analyses exist to look at all of them and say, yes, you found one study supporting an opinion, but there are 99 studies supporting the other opinion, and that's where the evidence lies. We tried to see, out of 10,000 studies, how many actually supported cannabis for a given symptom or disease versus how many said it shouldn't be used. We did not exclude any study from either side — critical or supportive — and we built custom machine-learning systems to identify which studies to include and what their conclusions were, based on keywords and language, so we could keep our own bias out of it. It's pure statistical analysis, so we can say this is a close approximation of what the medical and scientific community is showing through research. Kellen Finney: As a non-scientist, I'll probably ask a bad question, but going in there are so many varying factors — gender, weight, formulation, how something feels. How are all of those factors accounted for so you can feel comfortable saying medical cannabis is good? I think that becomes the challenge — there's the endocannabinoid system, all these variables. Where does it stop, and where do you say, okay, we feel good with these factors for these conclusions? Ryan Castle: That's one of the biggest hurdles. We started with nearly 40,000 data sets and got down to 10,000 by excluding everything that didn't have clear-cut, analyzable data. There's a large collection of studies that say something like, "We examined three college-age males and found they all love cannabis, so it looks like everybody loves cannabis." Obviously there are methodological problems there — but also, technically true based on that observation. That's where meta-analysis comes in. I don't think the people who wrote those small studies acted in bad faith — they reported the facts they found. But the facts one small group finds aren't usually indicative of the entire field. You have to look at the bigger picture to find where the actual scientific consensus is moving. In this case, the overall consensus was overwhelmingly in support of medical cannabis. Kellen Finney: The factors are endless — I think about my dad, who sometimes uses cannabis for pain, and his 1-to-10 pain scale is subjective. Sometimes the edible helps, sometimes it doesn't. For you to come in skeptical and end up feeling overwhelmingly positive is a substantial finding — you must have felt that throughout the process, thinking the data says this but still questioning if it's accurate enough. Ryan Castle: Yeah, and that came down to my belief that the more extraordinary the claim, the more extraordinary the evidence required. This went against a lot of institutional advice — a lot of large-scale organizations, which also receive federal funding, so there are caveats to their motivations, say they don't sign off on medical cannabis. But you can't just look at "my buddy's uncle used cannabis and it really worked for him" as evidence. Personal stories aren't the same as evidence unless you work them into structured patient-reported outcome surveys, which most casual conversations don't do. That's part of why we felt this study was necessary — so people don't have to rely on cherry-picked studies or hearsay, and can instead look at what the overall scientific consensus says. Consensus is a dangerous word in science, and I don't blame people for getting upset about that term, but when you've looked at as big a collection as we did and found a consensus for cannabis 31 times greater than the studies opposing it, that's not a number you see in just about any other pharmaceutical intervention outside of maybe penicillin. Bryan Fields: Wow — timely, too, given the recent news. [laughter] Is there other nuance from the meta-analysis — other than the 31-times-greater support — across different form factors? People get stuck debating strains, dosage, or whether it's treating symptoms of chemotherapy versus actually causing apoptosis. Were there other data points that helped wash away some of those debates? Ryan Castle: Yeah, absolutely. We broke this down into over 200 individual data points, from overall chemotherapy down to individual symptoms — appetite issues, nausea, pain, sleeplessness — and then into biomarkers like tumor size, cancer growth, remission, tumor necrosis factor, inflammatory markers, and recurrence. We tried to include enough that someone could look at a glance and see, for example, a 65% greater support for cannabis helping with chemotherapy than against it. Bryan Fields: When you say support, let's clarify — does that mean 65% of people in the study felt benefits versus not? Ryan Castle: Technically we're not looking at individual people within each study, for the exact reason that there are so many complicating factors. We're looking at the conclusions of each individual study. So out of those 10,000 studies, approximately 65% that discussed chemotherapy said cannabis helped chemotherapy symptoms in their study — whether that study had 5 people or 500. When we get into the nitty-gritty, like TNF-alpha, a biomarker showing how disregulated someone's inflammatory system is, the consensus was overwhelmingly in support of cannabis, even after applying statistical filters to exclude studies that don't meet a significance threshold. It was unparalleled in areas like inflammation and pain symptoms. Bryan Fields: Is it consistent across all forms of cancer? Ryan Castle: Not all of the results came back uniform. There were some instances where cannabis didn't show a strong improvement — clinician-reported remission, for example, where the doctor follows up and reports whether the patient is cancer-free or the cancer returned. We didn't find a relationship there that we could determine, which was interesting given everything else we saw, like shrinking tumor sizes and reduced cancer rates. That led to a follow-up project: reaching out directly to patients, because remission relies on the doctor and patient reconnecting. We started a project called the Patient Reported Outcome Hub, where we gather patient-reported outcomes longitudinally — their experiences, the cannabis used, the disease, the symptoms. We found a large percentage of patients who reported having cancer and using medical cannabis said they went into remission earlier than anticipated, and when asked if they followed up with their doctor, the vast majority said no, because their doctor didn't approve of them using medical cannabis. That's a confounder — a possible explanation for why that one metric showed the opposite trend of everything else: patients for whom it worked well may have been disillusioned with the medical system's approach to cannabis and just didn't report it. We're exploring how to use this to change the medical establishment's approach, because if there's this much consensus and it's still not embraced, people are leaving tools on the table that could improve lives. Bryan Fields: That's probably the most powerful part so far — this is a tool that could help people, and if doctors' role is to explore all the options, how do we tighten the pipes between doctors and research so new facts change minds? How do we shift that narrative in front of a doctor? Ryan Castle: If you want to reach a group of people, you have to speak the language they understand. If I moved to Spain without speaking Spanish, I wouldn't get far explaining medical cannabis to a random person. If I'm talking to a researcher or medical director using personal anecdotes, they'll nod politely and ignore it. You have to speak science, data, hard numbers, statistics, efficacy rates, odds ratios — the terms they look for. That's one area where there hasn't been as concerted a push in medical cannabis, and we're trying to remedy that. Research papers like this — we're grateful for the traction it got, around 100,000 downloads by scientific and medical organizations and over 500 million reads worldwide, enough attention that I think it'll make an impact. We're also building follow-up tools like the Patient Reported Outcome Hub, so doctors can survey their own patient groups and track symptoms day by day. And we're building a custom AI suite to answer these questions for laypeople or clinicians, giving them vetted, complete data on the science behind medical cannabis. Bryan Fields: Before we get into the AI stuff, I want you to put your skeptical hat back on. Is there a gap in the study — an area where, if you were a skeptic, you'd say this isn't enough? Ryan Castle: Absolutely — that's the process I go through; as soon as I share the paper with my co-authors, everybody rips it to shreds. If I were being critical, I'd say we don't have effect sizes — not just whether something had an effect, but how much. Without that, it's hard to estimate the impact of implementing it broadly. I'd also attack the statistics — someone could reassess this with a completely different set of metrics. We ran our data through two completely different types of analysis to check they lined up, and they did, but a determined critic could probably still find a way to poke at it statistically. Finally, someone could attack the people involved, saying I'm employed by an organization that published the paper and want to make them happy. My answer to all of that is that this is an open-access paper — we deliberately published all the data, methodology, stats, and raw numbers publicly so anyone can try to debunk us. We want the best science available. Bryan Fields: I appreciate that transparency — people hear new facts that conflict with what they currently believe, and their first instinct is to reject it. From a governmental standpoint, if this is just the beginning and we put real funding and resources behind it, opening up all the resources you might have been constrained on, how much bigger could the pool be to address some of those remaining doubts? Ryan Castle: That's what we're hoping for — that policymakers see the consensus and the evidence and decide to support it, since the evidence is solidly in favor and there's almost no evidence of harm. The last time there was this much scientific and medical support behind a treatment that hadn't been widely implemented was right before the abolishment of polio. That required a dramatic step of huge participation across the entire country, but it saved millions of lives. I feel like we're in a similar place now — there's an untold number of lives that could be impacted by making this medication more freely available. Bryan Fields: Let's play — you have a magic wand, money and ethics aren't an issue. What study would you conduct next? Ryan Castle: I'd want to do a health impact assessment — a large-scale endeavor that looks at every layer of consequences of a change. I did one for a small intervention that improved heart rate variability by 5-10%, and we mapped out the ripple effects: decreased heart attack risk, fewer medical bills, less burden on families and employers, fewer bankruptcies, increased quality of life. Over 10 years, we projected saving 20,000 lives, improving 300,000 people's lives, and saving a cumulative $18 million for that collection of people. For something like medical cannabis, which affects cancer rates, cancer development and recovery, and inflammation — the number one predictor of chronic disease — a large-scale rollout's health impact assessment would be mind-blowing: millions of lives, billions of dollars, and a very different world, especially in savings from not relying on less efficient, more expensive treatments like opioids. Bryan Fields: The skeptic in me wonders if large opioid companies wouldn't want that revenue stream to go away, unless they had an opportunity to get in front of it. Ryan Castle: I have to be careful about any given person or group's motivations, but I'll simply observe that when the DEA opened its discussion about rescheduling cannabis, a huge swath of studies suddenly popped up claiming cannabis caused everything from spontaneous heart attacks to cancer, many with pretty bonkers methodology — stuff that wouldn't have made the cut in our analysis. One study concluded cannabis causes heart disease and heart attacks based on a telephone survey of about 100 heart attack survivors, nearly 80% of whom said they used medical cannabis — without asking whether they started using it after the heart attack to deal with recovery pain, or long before. When you trace the funding behind a lot of those studies, you find a certain group of interests I'll just leave at that. Bryan Fields: The chain's pretty clear — you trace it back and go, huh, that's the group that funded that, that doesn't totally surprise me. Ryan Castle: Right, and it actually makes sense. One of the areas where we got pushback in private was around pain — we compared cannabis to other pain treatments and found comparable efficacy to leading opioids, but with a huge reduction in side effects. Cannabis's side effects were feeling a bit out of it; opioid side effects included life-threatening bowel obstructions and crippling addiction. We also found patients who used cannabis alongside opioids were dramatically more likely to quit using opioids than those who used opioids alone — cannabis functioned as an anti-gateway drug, helping people get off harder drugs during treatment. We got some polite conversations asking, "Are you sure you want to put that in the paper?" Bryan Fields: What about rescheduling — have you been contacted, or submitted this information? This seems like exactly the kind of research the DEA would want when assessing medical evidence. Ryan Castle: Exactly, and we're really proud of this — we worked with the Cannabis Coalition to submit our evidence to the DEA during the public comment period for rescheduling, breaking down efficacy, danger, addictiveness, and prevalence. We're proud to have submitted the largest collection of evidence in support of cannabis during that discussion. At the time, they said to expect rescheduling to Schedule III within a month or two, allowing it to be researched scientifically and used medically. That obviously didn't happen — there was a change in DEA leadership and things went a different direction. But the evidence is logged in the record; anyone can go to the DEA's rescheduling comment section and find it, test it, see if it makes sense. We hope people will, because I've looked at this data a hundred different ways and can't find any way to conclude anything other than that this should be medically available. Bryan Fields: Powerful stuff. Last question — tell us about the Cancer Playbook. Ryan Castle: Cancer Playbook is a nonprofit project where the goal is to give cancer patients the opportunity to share their story and get the benefit of large-scale data — not just this study, but a huge collection of patient-reported outcomes and scientific studies across just about every facet of oncology, whole-health treatment, and medical cannabis. Someone logs onto cancerplaybook.org, takes a survey about who they are, their symptoms, their disease, and it kicks back a cutting-edge breakdown — which types of cannabis have been shown most effective for their disease, which terpenes are effective for a particular symptom, success rates for patients who did or didn't use medical cannabis, and patient stories. It's not medical advice — we're putting out third-party evidence and patient stories to help people make their own decisions about their cancer journey, 100% free, with scientific citations and statistics behind every statement, so people can take it to their doctor and ask if it holds up. Bryan Fields: Awesome. For our listeners who want to get in touch and learn more, where else can they find you? Ryan Castle: There's cancerplaybook.org, where I run research operations, and wholehealthoncology.com, the organization where I'm president and where we do the scientific operations and high-level clinical work. Cancer Playbook is where we interact directly with patients. Bryan Fields: Awesome, we'll link it all up in the show notes. Thanks for taking the time — this was a lot of fun. Ryan Castle: Thank you, I really appreciate it. Take care.