In order to produce a personalized anti-cancer RNA vaccine, you need to
a) do a biopsy of the tumor and some healthy tissue, which is fairly cheap,
b) do genetic analysis of those samples to find the differences. This used to be prohibitively expensive, but now is fairly cheap too, the price went down six (!) orders of magnitude since 2000,
c) ask specialized software to find the optimal protein set to express. This software can be used for all the patients around, so cost of development will be split among them,
d) synthetize the correct messenger RNA and embed it in lipid nanoparticles. Not cheap, but not prohibitively expensive either, this is basically a biological equivalent of a printing job,
e) check the resulting drug and send it to the clinic where the patient is treated.
This all together is a much more straightforward and automatized process than extracting tumor-invading lymphocytes from a patient and babying them in vitro with a lot of attention from human scientists and use of fairly expensive reagents. The price tag for this process per se should be at least two orders of magnitude lower.
Then you must add the original development and testing costs, amortized equipment cost and profit. This is where the total price tag may balloon.