I have heard that a 'vitamin B17' can cure cancer, but that the medical industry never talks about it, since making it legal would cause them loss of billions. But I have never found a reliable report on whether B17 really works or not. There is a question on 'vitamin B17' deficiency as a cause of cancer (see this post), however, not on 'vitamin B17' as a possible medication for cancer. What are the proposed mechanisms? And is there evidence that 'vitamin B17' really works in curing cancer?
Laetrile/Amygdalin has been claimed to be a suitable treatment for 'cancer' (which is a summary term for an extremely heterogeneous class of diseases). Even though laetrile/amygdalin in these claims is often called a vitamin, it in fact is not a vitamin as the molecule is not essential for the metabolism (Greenberg (1980)).
The claimed effects of laetrile/amygdalin are not only based on supplementing the molecule to balance a putative deficiency (with the false assumption that it is in fact necessary), though. Both other putative mechanisms are based on reactions that form HCN/cyanide: the enzyme $\beta$-Glucosidase breaks down laetrile/amygdalin and this reaction releases HCN. The argument now is that tumour cells have higher levels of $\beta$-Glucosidase which leads to accumulation of cyanide in tumour cells - with the effect of killing the tumour cells selectively. This would be great... if healthy and tumour cells actually had different levels of the enzyme. Unfortunately, they seem to have similarly low levels of $\beta$-Glucosidase (Greenberg (1980)). Even more problematic, Greenberg (1980) also reports elevated levels of $\beta$-Glucosidase in the liver and small intestine ... and you would not really want to release cyanide there.
There is a second mechanism, though: The enzyme $\beta$-Glucuronidase is claimed to be at higher levels in cancer tissue than in healthy tissues (leading to the same desirable effect described above). But again, there does not seem to be a difference in enzyme levels between healthy and cancer tissues (Dorr and Paxinos (1977)), and even more: $\beta$-Glucuronidase does not even break down the molecule as laetrile/amygdalin has a glucoside bond - so there is not even any cyanide release in that case (Holzbecher et al. (1984)).
Additionally, there is no clinical evidence that laetrile/amygdalin is an appropriate treatment for any form of cancer. This topic has been reviewed substantially (see Milazzo et al. (2006) for a review and this fairly recent Cochrane review by Milazzo and Horneber (2015)) and none of the studies included and analysed indicate any positive effect of laetrile/amygdalin. On the contrary, the in vivo cyanide formation after laetrile/amygdalin uptake makes the treatment very dangerous as it can lead to cyanide poisoning (and death, eventually) - especially after oral uptake due to higher $\beta$-Glucosidase concentrations in the digestive system. The authors of the Cochrane review therefore conclude that
The claims that laetrile or amygdalin have beneficial effects for cancer patients are not currently supported by sound clinical data. There is a considerable risk of serious adverse effects from cyanide poisoning after laetrile or amygdalin, especially after oral ingestion. The risk–benefit balance of laetrile or amygdalin as a treatment for cancer is therefore unambiguously negative.