Ice around the edges?
We are told that the proportion of water in the fuel was too low for significant quantities of ice crystals to form in the bulk fuel in the tanks or in the bore of the pipework. But could some water have frozen out on the insides of the pipes themselves, in much the same way that water from fairly dry air builds up ice deposits over time on the cold plates of domestic fridges and freezers?
Any ice deposited in that way would probably be fairly soft and easily detached to clog up the FOHE by increased fuel flow as thrust was increased.