The US doesn't use PANS-OPS. It uses TERPS - their own set of standards & procedures so PANS-OPS doesn't apply. There are differences.
ANGEE is just a waypoint from one of the IFR routes. It, DIDDY and VRB are commonly used as IFR waypoints into KFPR by ATC (local knowledge and not really important other than giving common navigation point between the IFR charts & the approach plate).You'll normally be radar vectored to final or to whatever IAF you request.
So, the issue is having arrived at ANGEE.....what next? Most commonly vectors to final (or IAF if you want to waste time). However, if for some reason, you want or need to do a completely pilot interpreted arrival, what necessary info is available on the plate to get from ANGEE to the NDB? Each waypoint has BRG/DIST from the NDB so that lets one fix position at the waypoint although there's no necessity to arrive via the waypoint. You would track to the NDB quite conventionally from whereever you are. The MSA gives a minimum altitude within 25nm so you mustn't descend below that unless you're being vectored.
Option #1: Once over the NDB and still above the MSA use whatever manoeuvering you find convenient to arrive back over the the NDB to track outbound on the approach. It could be a standard sector entry but needn't be - you're above the MSA remember.
Option #2: Join the charted hold conventionally while remaining above the MSA then when on the outbound leg of the hold 'merge' with the approach's outbound leg.
Once established outbound commence descent to not below 2000'. Commence the procedure turn sufficiently in advance to *ensure* you remain with 10nm of the aid. It's up to you to determine how to do that but historically most people would work out a time based on their groundspeed - typically a couple of minutes. Now days use the GPS...
Both methods are accepted and much of a muchness in time & turns. I lean towards option #1 because I don't have to dick around with formal hold entries compared to #2, I like an accurate as possible start to the leg and I can commence the descent sooner, but in reality I prefer not to muck around and nearly always get vectors to as short a final as I can arrange.