The cavitation damage to the engines' HP fuel pumps would seem to indicate incipient fuel starvation - or at least aeration - for whatever reason. Cavitation damage occurs when the pumps lose their continuing lubrication by the fuel itself. None of the in-tank debris (plastic etc) found would explain one or (particularly not) both engines becoming fuel-starved.
.
The possibilities remain prima facie limited to establishing the reasons why fuel (or trapped water) could be cooled to below freezing point (even if that point is 10 degrees below the specification freezing point for JET A1 of minus 47 deg C). The pointer would seem to be towards either water in the fuel becoming frozen or waxy/illiquid. There's another possibility however.
.
Think about the fact that certain "aerodynamically quiet" areas on an airframe can accumulate de-icing fluids and cause flight-control restrictions (see instances by Googling that term ["aerodynamically quiet"] and recent and historical AD's - particularly for BAe146 etc). In a similar fashion, icing will not form and accumulate uniformly over an airframe. It will first be seen in the corners of windshields, along windscreen wipers, wing/tail leading edges etc. Now think about what can happen in another fluid (i.e. not air, but fuel) that's contaminated with very cold water. What I'm getting at here is that some quiet (or stagnant) areas in an LP fuel system (tanks thru to HP pumps and filters) can support the formation of ice or waxy deposits due to higher (than normal/ambient) super-cooling in certain discrete areas. Result could be (a) large(ish) lump(s) of melting ice that will find its (their) way into a critical passage, filter, pipe or valve and temporarily cause constricted flow-rates. Because both sides of the aircraft mirror image each other, it's conceivable that identical "releases" of this icy lodgment could occur almost simultaneously - and then disappear (i.e. great quantities of water would not be required and wouldn't later be detectable anyway).
However why limit oneself to considering frozen water? Why not just think in terms of fuel itself being trapped and frozen in fluidically quiet areas and becoming supercooled/frozen. I'm guessing that residual fuel in the center tank/tail tank (and their collector tanks) might qualify and could later be induced into the system by both warming and the attitude changes inherent in both descent and configuring (gear and flap extension).
At what temp would the low fuel temp warning activate? On BA 777's the fuel freeze temp can be changed in the FMC but is set to -40 by default, the actual warning starts at 3 degrees above that value. So by default (but it can be changed) the low fuel temp warning would sound at minus 37.
However, as with all such system measurements, the validity and usefulness of the sensor data depends upon where the sensor is mounted. Temperature measurement systems are inherently different to pressure measurement. Pressure is exerted equally in all directions. Temperature can vary over a significant range within a large volumetric system, some of whose components/areas are temporarily isolated by dint of being virtually (but not really) empty. If a center tank's output to the wings ceases relatively early in the cruise attitude (because it's "virtually" empty), does that mean it cannot thereafter contribute some previously "baffled" slushy ice-dregs (merely by gravity feed perhaps) once the nose lowers to the descent attitude? It's also a liquid fact that smaller stagnant quantities (such as dregs) will freeze sooner (and also melt sooner) than larger quantities (i.e. in wing tanks) that are being subjected to throughput recirculation (of bypass fuel), engine feed and fuel/oil heat exchanging etc. Wing carried fuel is also heated to some extent by the wing's passage through the air (and by the sun on the upsun side). That's not the case with fuel carried (or baffle-lodged) in the center-section and tail-tanks.
Why would this have happened to BA038 and not other 777's previously? My guess is that there is a trigger temperature and that the period that G-YMMM spent in the super cold pool of their track-miles between the Urals and Eastern Scandinavia ("unusually cold but not exceptional" - UK Met Office) may have allowed such pockets of super-cooled fuel to form into ice. The time of day (sun/no sun) and chosen CRZ hts may have also helped format the process.
.