Legal hypothetical about duty-to-act aside, even in the OP’s scenario nobody knows that the treatment is a miracle cure until they’ve done most of the clinical testing necessary for FDA approval. Not even the evil Mustache Twirling big pharma.
At best, after many millions of dollars invested in early stage preclinical research, they’ll know (say) that the drug effectively inhibits some biochemical process that’s necessary for cancer growth, and it can kill cancer cells in a dish. But that’s a very low bar, just barely qualifying as “promising”. There are many thousands of possible treatments that have crossed
After another several million spent on late stage preclinical research, they’ll know that the drug effectively treats a few different animal models of cancer, and it is not toxic to animals or human cells in a dish. Now the drug qualifies as a pretty good candidate for clinical testing. The average success rate for “pretty good candidates” is 10%. Maybe, to be really generous, we could posit that the preclinical evidence is so positive and overwhelming that we could say that this drug has a whopping 20% or 30% chance of being a miracle cure. At this point nobody knows whether the treatment will be safe or effective for actual patients.
Now Mustache Twirling Pharma has to decide whether they want to invest tens of millions more into early stage clinical research. After Phase I trials, they’ll know that the miracle treatment doesn’t harm healthy people, but they still don’t know that it is a true “miracle cure”. The first hints that they have a “miracle cure” will be after Phase II trials, which are small scale trials in patient populations. Usually these are too small to provide definitive evidence of efficacy. However, if the drug is a “miracle cure” it’ll be pretty obvious after one of ten cancer patients in the control group has a partial remission, but ten of ten patients in the treatment group have full remissions. At last, Mustache Twirling Pharma is fairly certain (but not completely sure!) that they have a miracle cure. That they want to suppress for some unfathomable reason*. And as far as the FDA has concerned, they have not provided sufficient proof that the drug is safe and effective.
All this testing has taken the better part of a decade, so the remaining patent life is maybe 10 or 15 years.
Going back to Little Nemo’s hypothetical:
Could you sue someone for suppressing a promising but unproven treatment that perhaps has a 1% chance of actually working? A 20% chance? A 90% chance, but for a treatment that has not been officially approved by the FDA?
*Again suppression makes no economic sense at this point. The more miraculous the results in the Phase 2 trials, the cheaper the Phase 3 trial (a bigger effect means you need a smaller sample size to convincingly demonstrate the effect), and the better the justification for charging high prices.