This is a real can of worms question.
The answer is, not in the simple way that would seem to be the most obvious.
The problem is that cues in a timeline group are offset by their pre-wait from the timeline group, as soon as the timeline group is fired the pre-wait runs on a separate clock that has no understanding of what the group is doing.
Take an example of a timeline group containing an audio file with slices at 2 & 4 seconds, with an infinite loop. a slice at 7s with an infinite loop, and a MIDI cue placed on the timeline view against the audio waveform at 5 seconds.

Although the audio file won't reach 5 seconds during the first loop the MIDI cue will still fire at 5 seconds. In the timeline group you will actually see two cursors.
What is confusing is that the timeline group itself never reaches 5secs as the action time of the group follows the audio cue. But the 5 second pre wait on the MIDI cue isn't actually being fired off the elapsed action time of the timeline, but off it's own prewait timer and so fires at 5secs after the group cue is triggered, regardless of any slicing or looping.
There are 2 work arounds.
1, You could use duplicate audio cues, and set the start and end times of each cue to create sections and whole cue loops which correspond to where you would have sliced your audio.
You then set your MIDI Cues on the timeline relative to the start time of each chunk of audio.
There are several problems with this approach.
You have to calculate the time of any MIDI cues relative to the start time of the chunk of audio you are playing in that cue.
Gapless playback at the devamp point is not guaranteed because QLab is now treating each section as a separate audio file.
If the MIDI cue is in a looping section of audio, it will only fire once.
2. You could make a multichannel audio file with your audio file and LTC and slice and loop it as before.
You loop back the LTC track on your audio interface and set the cue list to use LTC.
You trigger your lighting cues off Incoming timecode.
The disadvantages are:
It's messy and complex
You can't see the waveform to spot points in the audio cue. Any audio detail is obliterated by the LTC waveform, as you can't select which channels you see in the waveform.
The Advantages are:
Midi cues fire at the exact point in the audio, which is not guaranteed in any other method.
A Midi Cue in a looping section will fire repeatedly every time that the audio and time code value is looped through.
All the above is easier to demonstrate than explain so attached are links to a video and workspace demonstrating the 3 methods.

Although the example uses MIDI cues for the LX cues, they could just as easily be OSC or internal QLab lighting cues.
You can download themeless here:
The files aren't that big but exceed the limits for this google group hence I couldn't attach to the post.
Mic