This is not going to be easy, the vocal is all over the stereo-field so using a centre-channel extractor unsurprisingly didn't work well.
Except if the vocal track is fairly unprocessed and clean, its actually very difficult to remove it from the recording.
You could try using a spectral editor like iZotope RX, and remove it manually, or some clever automation with EQ or even perhaps a multi-band compressor.
It's usually less hassle to try to get the authors to send you the instrumental mix.
edit: I forgot, Audition has a built-in spectral editor (albeit less powerful than RX), so you could try with that.