I'm thinking that such a failure should also have caused a torque spike, which should have been recorded in the FADEC memory? After all, there would have to be a lot of power added to break the shaft, the governing system would have added power to maintain RRPM as that blade failed?
just throwing out thoughts.