An interesting exercise:
1. Review some of the anectodal news reports and internet postings of people who have used high potency cannabis oil to treat their cancer. A common thread emerges: The importance of titrating the dose quickly to establish the tolerance necessary to use very high doses. There's no way to know how much THC, CBD, etc is present in these products but the patients generally describe symptoms you would see with extremely high doses of THC.
2. Search pubmed for GPR55, CB1, CB2, heteromization, and cancer. You will find a number of recent mechanistic studies which describe the biphasic action of cannabinoid agonists on cancer cells that express GPR55-CB heteromers. At low concentrations, THC (but not CBD) seems to act via GPR55 associated pathways with oncogenic downstream effects. Conversely, THC triggers Cannabinoid receptor mediated apoptosis in these same cells when present in higher concentrations. The effect of upward titrating cannabis oil doses in treatment of Ph+ T-Cell-ALL in a human being has been documented in one case report in "Case Reports in Oncology" published by Karger.
Put simply, patients using cannabis or THC for palliative reasons only use small doses, while the complementary/herbal medicine community seems to think that very high doses are necessary if the patient is trying to treat their cancer (induce remission). Moreover, the patients who claim cannabis oil helped to "cure" their cancer invariably report they used very high doses. These cases all predated the discovery of THC's biphasic action on cancer cells. In other words, there is now a pharmacologic explanation for a phenomenon that has been reported within the alternative medicine community. Could all be coincidence, but definitely suggests to me that high dose cannabis as a potential chemotherapy drug needs to be studied in randomized trials.
There is also something called "the entourage effect" which may be unique to eicosanoid signalling (and hence endocannabinoid signaling). In basic terms, the presence of various endocannabinoid analogues at one receptor site seem to influence it's action. While it hasn't been studied yet, one might postulate that the 60 different cannabinoids in Cannabis help shift THC signaling toward activation of CB1 and CB2 apoptotic pathways and away from GPR-55 mediated proliferation.