Self-reported testimonials are subject to selection bias, and are almost useless for identifying a causative link among multiple variables. It has nothing to do with the honesty of the people reporting the results. It simply isn't possible to separate out all the confounding variables with such a small sample size, for something as complicated as cancer.
There is nothing complicated about cancer. Cancer is the body's last line of defense of a compromised immune system.
Screw phase 4 trials. How about at least phase 2? Or even phase 1?
Before you can claim that garlic cures cancer, you should be able to answer most of these questions:
1. How many people have been cured by garlic?
2. How many have tried to use garlic as a cure, and failed?
3. What is the breakdown by dosage?
4. What is the breakdown by cancer type and stage?
5. Based on these results, how effective is garlic compared to other treatment options (chemo, radiation, etc.)?
We don't have the support of the medical establishment to study such things on a large scale. It takes a lot of money.
Carefully compiled statistics are not really needed before trying a natural treatment, since nothing is dangerous. American naturopathic doctors typically operate experimentally, and have a lot of leeway to try many things, since the medicines used are not at all harmful to the body. It became tribal knowledge from experiment to experience that certain herbs are good for certain ailments.
Traditional Chinese Medicine does have the support of the medical establishment in China, however, and Chinese Medical Journals are replete with such statistics. They are available if someone knowledgeable in Chinese were inclined to look for them. Traditional Chinese Medicine is a bit more structured, and visiting a Traditional Chinese Medicine doctor is a wise choice when faced with a serious ailment.