As before, the "promise" of the therapy is the chance it will give you a significant benefit. The type and size of benefit may depend on what the goals of the therapy are. It may be to cure an advanced cancer, but may also (and possibly more realistically) be to reduce pain and anxiety and to improve the quality of life.
In this scheme you estimate the promise of the therapy by considering all available evidence which supports or undermines the therapy as well as how plausible the therapy is to begin with. The quality of your decision making will depend on the accuracy of your estimates.
Remember that in addition to positive studies and cases, there can also be negative studies and cases, and that proponents may fail to cite the negative evidence and opponents may fail to cite the positive evidence. You will have to search to find both the positive and negative evidence.
Once you find the evidence you have to weigh the positive and negative together to form your best guess as to the chance the therapy actually works. One negative study should not automatically doom an alternative therapy - I can cite completely negative studies of IL-2, the conventional immunotherapy which saved me. But you do have to consider the quality of the studies. If the negative evidence is of a higher quality - more carefully done studies - than the positive evidence it counts correspondingly more. If on the other hand, the positive evidence is more or at least as impressive as the negative that weighs in favor of the therapy. For many alternative therapies, there is only fragmentary positive evidence, with no truly rigorous studies of any kind.