Once the orbiter was committed to entry by executing the deorbit burn, nothing could have been done. The entry trajectory was tightly constrained for structural, thermal, and aerodynamic reasons. There was no way to abort an entry and return to orbit. This answer on Aviation Stack Exchange gives a good overview of the shuttle entry trajectory constraints.
The entry could be waved off at any time up to the deorbit burn; this was done many times for weather reasons.
The only exception would have been that if the damage had been somehow minor enough to keep the vehicle from breaking up or losing control, but major enough to prevent reaching a runway, the crew could have bailed out if stable gliding flight was possible. (This was not the case for the actual accident).
For what could have been done prior to entry, see my answer to this question:
What would NASA have done if they knew Columbia was catastrophically damaged?
The document referenced in this answer discusses some entry trajectory modification that might have been considered had the damage been assessed before entry; the consensus is that none can be stated to have any certain appreciable benefit, and both were discussed only in the context of flying the entry with the on-orbit wing repair completed.
4.7 ADDITIONAL ENTRY OPTIONS – THE “CAIN REPORT ” NASA Flight Director Leroy Cain presented the report from the “Entry Options Tiger
Team” to the Orbiter Vehicle Engineering Working Group (OVEWG) on
April 22. This report was a very complete analysis of the results of
jettisoning most of the payload bay cargo and coldsoaking the wing.
Although this report looked at options within the certified entry
design envelope, the options presented required some very difficult
EVA tasks like cutting power and fluid cables, cutting through a
tunnel, and large mass handling. This study does not assess the
feasibility of these tasks, but it simply notes that whatever jettison
tasks that could be accomplished in any remaining time during the
two “repair” EVAs would be performed, as this would decrease the entry
heating by a small amount. As there is a very large uncertainty band in the thermal analysis of a wing leading edge repair, it
is sufficient to say that jettison of equipment would have occurred
during any remaining EVA time, and this may have helped the overall
total heat load.
4.8 UNCERTIFIED OPTIONS - INCREASED ANGLE OF ATTACK /LOW DRAG PROFILE The Entry Options Tiger Team was requested to look at certified options only. The only uncertified entry flight design options
that could significantly reduce the wing leading edge temperature
would be to change guidance to fly a lower drag profile during entry
or to raise the angle of attack (alpha) to a reference of 45 degrees,
vice the standard 40 degrees. However, it should be noted that while
flying either one of these entry profiles would reduce heating on the
leading edge, the heat load would increase on another part of the TPS
structure. A simplified analysis that does not account for heating
effects due to boundary layer tripping from a damaged area shows that
a wing leading edge peak temperature could be decreased from a
reference of 2,900 degrees F to 2,578 degrees F. This would be
considered as an additional tool in attempting to maintain the spar
structural integrity. It should be noted that changing the reference
alpha would require a significant software patch to entry guidance
From the CAIB Report, "STS-107 Inflight Options Assessment"
.
