This is very broad as asked. I will answer with one example, there are others.
Immunotherapy gets the patient's own immune system to destroy the tumours rather than the treatment itself destroying them. This might involve making the immune system more active, or interfering with the way that tumours hide or shield themselves from the immune system. For example, Wikipedia says of Ipilimumab
T lymphocytes can recognize and destroy cancer cells. However, an inhibitory mechanism interrupts this destruction. Ipilimumab turns off this inhibitory mechanism and allows the lymphocytes to continue to destroy cancer cells.
(Lymphocytes are your own white blood cells, the same mechanism that cleans up infections in your body all the time.)
This article at cancer.gov summarizes some studies showing how Ipilimumab and other immunotherapies are changing outcomes for melanoma. You can do you own searches for other drugs and other cancers. For example Nivolumab is combined with "ipi" for melanoma. I have also seen ads on tv for Nivolumab for lung cancer.
How much longer will patients live? That depends on the cancer. Melanoma has dismal outcomes if it's Stage IV - median survival of about 6 months, less than 10% surviving 5 years - and chemotherapy doesn't help survival. (One article on how bad stage IV is with traditional treatments.) So there's plenty of room for improvement. The studies in the article above include some of that improvement. It's hard to find much on long-term odds, because they haven't been using these drugs long - ten years at most.
Why not use it everywhere? Well they literally do not know what cancers it works for (presumably not all of them) or precisely what kind of survival improvements to expect. It's a no-brainer to use it with something that has survival times of just a few months, but should you use it with something that has survival times of 2 or 3 years? Because there is a cost. First, a dollar-cost: $200,000 a year is typical. And second, the side effects are very bad. They can be fatal. And when they're not fatal, they're hospitalize-you-and-major-steroids bad, and take-a-compensating-medication-for-the-rest-of-your-life bad. The immune system can over-react causing inflammation and swelling that might damage or destroy your thyroid, pancreas, adrenal gland, or whatnot, or give you colitis. Understanding the side effects is a big part of the current studies. One study describing how people went from months of survival to years mentions "55% of patients will have high-grade adverse events" so widespread use will be held back by that.
Here's a study I'm very familiar with. It compares two dose regimens of two different immunotherapy drugs for melanoma. These regimens have already been shown to have the same response rate; the study compares the side effects. It's normal for some patients to have to stop taking the treatment because the side effects are intolerable. And a similar study (same combination of drugs, same disease) reported results where over 60% of patients getting the combination were still alive after 2 years. 65% is so much better than 10%!
There are other immunotherapy drugs, but in general you see:
- the treatment doesn't attack the cancer; rather it enables the patients own immune system to do so
- the improvements in survival are astonishing. People living years instead of months, and possibly decades instead of months. People becoming immediately "better" with no symptoms instead of lingering with constant side effects of chemo and radiation, or surgeries many times a year
- the long term outcomes are not well known at all
- the side effects are very dangerous when they happen
- the cost is high
From a personal point of view, yesterday I said to my doctor "it feels like being around when insulin or antibiotics were first being used" and he agreed, yes it does. But that includes the part where you don't know for sure how things will turn out.