Here is a sentence that is entirely true and almost entirely useless: this compound has been shown to interfere with microtubule formation in cancer cells.
It is true. It is published. It sounds like it means something enormous. And on its own, it tells you close to nothing about whether the compound helps a human being with cancer.
The distance between "we know how it could work" and "we know that it does work" is the single most misunderstood concept in the repurposed-drug conversation. It has a name in the research world — the translational gap — and understanding it is the difference between reading this literature clearly and being misled by it, in either direction.
What a mechanism actually is
A mechanism is a plausible story about how a substance could affect a disease. Antiparasitic compounds have mechanisms that sound, on paper, encouraging for oncology. Some interfere with microtubules — the internal scaffolding cells use to divide. Cancer is a disease of uncontrolled division. Interfering with division sounds like it should interfere with cancer.
This reasoning is not wrong. It is incomplete, and the incompleteness is where everything happens.
Why dish results do not transfer
In vitro work — cells in a dish — is where most mechanistic findings originate. It is also the least predictive setting in all of biology, for reasons that are structural rather than incidental:
- Concentration. The dose applied to a dish is often far higher than anything achievable in human blood at a tolerable dose. A compound that kills cells at a concentration your bloodstream cannot reach is not a treatment.
- No body in the way. A dish has no liver metabolising the compound, no kidneys clearing it, no gut deciding how much gets absorbed in the first place. Many substances that work beautifully in glass never reach the tumour in a person.
- No tumour microenvironment. A real tumour is not a lawn of identical cells. It is a structure with poor blood supply, oxygen gradients, immune interactions, and dense supporting tissue that drugs struggle to penetrate.
- No selection pressure. Real cancers evolve. Dish cultures, over the timescale of an experiment, largely do not.
This is why "kills cancer cells in vitro" is a genuinely low bar. Enormous numbers of substances clear it — including many that would kill the patient long before the tumour.
Why mice are not small humans
Animal models are a real advance: there is a body, a metabolism, a circulatory system. And they still fail to predict human results with striking regularity. The attrition rate from promising animal data to successful human trials is famously brutal — the large majority of compounds that look good in mice do not survive human testing.
The reasons include species differences in metabolism, tumour models that are induced rather than spontaneous, dosing that does not scale cleanly across species, and the simple fact that a mouse tumour and a human cancer are different biological objects.
The gap in one sentence
Mechanism tells you a thing is possible. Only a controlled human trial tells you it is true.
Every step of the chain — plausible mechanism, in-vitro activity, animal efficacy, human safety, human efficacy — is a filter, and each one removes most of what came before it. A compound sitting three steps down the chain is not "nearly proven." It is at a stage where most candidates historically fail.
Why this cuts both ways
It would be easy to read the above as a debunking, and that would be a misreading. The translational gap is not an argument that repurposed compounds cannot work. It is an argument about what we currently know.
Two claims are both unsupported by the evidence:
- "The mechanism is compelling, therefore it works." This skips every filter. Mechanism is a hypothesis, not a result.
- "There's no Phase III trial, therefore it's nonsense." This confuses unproven with disproven. The absence of a completed trial is not evidence of failure — it is evidence that the trial has not been done. Repurposed generic compounds face a structural funding problem precisely because no one owns a patent worth financing a Phase III trial to defend.
Both errors mistake the current state of evidence for a verdict. The honest position is genuinely uncomfortable: the mechanistic and preclinical literature is real, and it has not been carried across the gap. That is not a satisfying answer to anyone. It is simply the accurate one.
What to do with this
When you encounter a claim about a repurposed compound, ask where on the chain the evidence sits. If the answer is "a dish," treat it as a hypothesis. If it is "mice," treat it as a stronger hypothesis. If it is "a case report," treat it as a reason someone should run a trial. If it is "a completed randomised controlled trial," you are looking at something categorically different — and you will find, if you look, that this is the step the antiparasitic-oncology literature has not yet reached.
Knowing precisely where the evidence stops is not pessimism. It is the only foundation on which a real conversation with your oncologist can be built.
Disclaimer: This article is for educational purposes only and is not medical advice. It does not diagnose, prescribe, or recommend any treatment. No human efficacy has been established for these compounds in cancer treatment, and none is approved by any regulator for that purpose. Consult a qualified oncologist about your care.