With any of the the existing LimeSDR boards, I would be ~99% sure that no is the answer.
But if what you are asking is could the LMS7002M chip be treated as 4 differential baseband direct sampling ADC’s, where more than three quarters of the LMS chip is sitting idle. If I look at the block diagram, my answer would be with a lot of work possibly, maybe.
It would be a very inefficient use of hardware, unless all the signals of interest were between DC and 80 MHz, it may still work out comparable, or even lower, in price to four 160 MSPS 12-bit ADC’s. At a guess all of the TSP would also need to be reconfigured to, do nothing no processing at all and, just pass through raw I Q samples for channel 0 and 1. But in theory you could always use an analogue bandpass filter(s) and undersample to access signals slightly higher up in frequency.
As far as I know none of the existing LimeSDR hardware expose the differential input pins (adcin_ip_1,adcin_in_1; adcin_qp_1,adcin_qn_1; adcin_qp_2,adcin_qn_2; adcin_ip_2,adcin_in_2) on the LMS7002M chip that would allow you to attempt oddball ideas. Probably because if you screw up, even once, you would irreparably damage the LMS chip. I honestly have no idea if there is ESD protection on those pins, but usually there is not.