It's the difference in driving impedance, "detector" source
vs passive-match "sink". An ionized charge track can present
a pretty low impedance (though you may find this dependent
on species dE/dx in material, and access resistance to the
"business end" of that material).
It's not uncommon for amplifiers to exhibit asymmetric slew
rates, especially if they have no, or loose, full power HF
distortion specs.
Perhaps the common emitter transistor presents a lower
impedance in both directions, or lower Cin to the point that
the track impulse is closer to its "in physics" shape. It seems
to be "soaking up less signal" going by the plot axes. Which
alone makes it a better choice.
But why not go for some common-emitter gain while you're
at it, if the transistor is that good (or, if you can find one
even better)?