Pharma companies need a lot of data for approval. Melanoma is a common cancer for which much data can be rapidly generated. With a 25% success rate, which is low relative to proven therapies, some as high as 99% cure, it would be unethical to use this drug for all cancers and without proven methods being used first.
Docs have the option to use a drug “off-label” for other cancers, but most would not want to risk the lawsuits that are likely in such scenarios. Medicine is about stats and ethics; rolling the dice is not looked favorably upon by the profession, the FDA, or the people who like to file lawsuits.
Just to update—10 years ago my wife was stage 4 brain cancer when detected. I was paying 900 a day for an experimental drug for a few weeks at my request. Luckily, the drug companies had a program to assist those making less than 75000. That helped but she shortly was switched to a more conventional regime.
It was not to be, but what was learned from her situation was said to be helpful, and that brave charitable fact gave her great comfort at the end.