If the fuel metering device was giving a “false” reading, somewhat bellow true fuel flow, when the A/T commanded an increase of thrust it would have increased initially to the requested value but then the excess of fuel in the burners would made it drop dramatically but not entirely. Thrust decays, FADEC opens fuel valves even further to compensate, but instead of correcting it worsens the problem.
Too far off the mark?
Or was the excess of fuel in the burners in this scenario a mix of fuel, ice and water?
Thawing Ice, perhaps shedding from the compressor stages as the engines revved up from approach idle to a new target thrust setting, quenching or partially blocking the annular combustion chambers on both engines? In this case the anticipated EGT increase with reduced gas flow may have been negligible due to the cooling effect of water, hence no RPM increase but instead (partial) rollback to somewhere above idle?
Green-dot