What really shields the rectum more than anything, and this is true of both primary and salvage radiation, is setting a zero margin at the rectum and tight dose constraints, including point dose constraints for the rectum. IGRT (image guidance) is important too. With primary radiation, fiducials (or transponders) can very accurately help aim the radiation (this is done continually throughout each session with SBRT for super-accurate beam localization).But with salvage radiation, IGRT is problematic. While some ROs seem able to use fiducials somehow (i.e., at UCSF), others say they don't hold in the loose fossa tissue and move around too much to be useful. They use soft tissue landmarks instea, at the start of every treatment with cone beam CT or stereoscopic Xrays. Some use bones as landmarks, which is a particularly bad idea (although easier to do). It's a bad idea because soft tissue is highly deformable and continually changes its shape and position with respect to bones. Those who site on bones will miss the proper targets and increase toxicity. The problem is compounded when there are two targets, as there are when both the pelvic lymph nodes and prostate bed are treated. The two targets change their shape and position with respect to each other, increasing the risk of misses. UCSF has an algorithm for dealing with this, others use judgment.
The single best aid is a full bladder. It anchors the tissue in the prostate bed and lifts most of the bladder away from the radiation target. A similar thing can be done in the rectum with an inflatable rectal balloon. It lifts most of the rectum away from the radiation field, although the anterior part is actually moved closer to it. (It's the approach they use with primary proton therapy). It also obviates the problem of gas passing through which changes the shape of everything.
I should rush to say that despite the anecdotal complaints you hear on HW (which I am not at all discounting), rectal side effects are typically mild and transient. They are rarely severe or long-lasting with today's very accurate linacs. For example, in the following study at UOP, there was
no late-term rectal toxicity (≥grade 2) among men getting 70 Gy of salvage prostate bed radiation.
www.redjournal.org/article/S0360-3016(11)00562-1/fulltext(BTW - late term rectal toxicity is similarly rare after primary radiation, which is why I consider SpaceOAR to be an expensive solution in search of a problem).