Your circuitry is waaaayyy to simple to work reliably! You need a protection circuitry against ESD at the inputs, you need line filtering, you need a Schmitt trigger input and you need a reliable reset chip for the micro. You also need some current limiting for the lamp to avoid huge inrush currents and you need some sort of slew rate control to prevent unsane feedback from the lamp switchings into the switch line.

All these things you omit and then you wonder that your circuit will work only with certain cables. Believe me, it has nothing to do with the sort of cable but with the total lack of bandwidth limiting! You give a 10m long wire directly to the input of a monflop to reset the micro? Hallelujah, this cannot work!!
By the way, your "schematic" is wrong, too.

I would recommend the following circuit as a start:


All other diodes are 1N4148.


Use a cable with two twisted pairs and a common shield. Connect the shield to ground, and don't use it as a current return!
At all inputs you see suited transzorbs to absorb ESD events. Diodes in series to the supply lines prevent ESD events from contaminating the supply voltages by injecting currents into them.
Caps nearby decouple the supplies from the current spikes, when the switch is pressed or the lamp is turned-on.
The switch line is low pass filtered by 100k/100n RC-combination. Diodes to the supply rails protect the Schmitt trigger input.
LED is driven by output of first Schmitt trigger. This allows a higher current to be driven through the LED, which is needed to better notice the LED flahing.
Output of second Schmitt trigger goes to /PBRST input of MAX1232, which is a much better reset solution, than a 74LS123 monoflop!
Lamp is driven by a slew rate limited constant current source, which keeps the noise caused by turning-on the lamp as minimal as possible, which is of fundamental interest, when routing all the signals via same cable!!
BD137 is chosen to allow permanent short circuit of lamp, without resulting in a damaged transistor.

Kai