Extending a load cell cable is routine, and the usual advice is to use six-wire cells and keep the cable short. That advice is directionally right but mixes up three separate effects, only one of which is fixed by the sense leads.
Excitation drop
The bridge draws current through the cable, so the voltage at the bridge is lower than at the instrument. With a 350 ohm bridge and 10 V excitation the loop current is about 28 mA; 50 m of 0.5 mm² cable adds roughly 1.7 ohm per conductor, dropping about 0.1 V. That is a 1% loss of sensitivity, which a six-wire connection corrects completely and a four-wire cell cannot.
Thermal EMF and noise
Dissimilar metals at junctions generate microvolt-level offsets that drift with temperature; twisted pair and a stable junction temperature reduce it. Noise pickup is a shielding problem: connect the shield at one end only, normally the instrument, to avoid a ground loop. Neither issue is improved by the sense pair.
Practical limits
Beyond a few hundred metres, cable capacitance and the indicator input filter interact and the reading becomes sluggish before it becomes inaccurate. At that distance, move the amplifier close to the cell and transmit a 4-20 mA or digital signal instead of extending the bridge wiring.
Kevin Huang