In my view, the first question is whether fusion can reasonably be expected to occur in a given asymptomatic patient with a migrated cage. If the answer to this question is Yes for that particular patient, then it may be justifiable to follow the patient clinically and radiologically. If the answer is No or uncertain, I would rather opt to remove and replace the cage. If revision surgery is necessary anyways, e.g. for pedicle screw failure, I believe adequate attempts should be made to correct the cage within the same procedure. The answer also depends on what exactly "migration" means, e.g. vertical sink-in or horizontal displacement. In the first scenario, I would be rather reluctant towards removal/replacement in an asymptomatic patient, whereas posterior dural compression or anterior vessel compromise would probably be reasons for revision surgery even in the absence of symtoms. Certainly, the outline above can only be a rough sketch, and a thorough risk-benefit assessment must always guide the treatment decision.
In my opinion, fusion process is important. if there is a migration with well developed fusion, it is not necessary to remove the cage. You can take a breath. The main problem is if there is no fusion. This means that there is a instability. Nearly all migration occur into the spinal canal for a while after the operation. If someone try to remove it from the posterior, there is high probability risk of neurologic damage due to developed fibrosis. The best way try to anterior root and remove the cage anteriorly. This procedure also provide to surgeon a new era for good fusion.
The issues to be considered are: 1. Is the patient symptomatic due to the migrated cage or is he likely to become so down the line? 2. What are the risks and benefits of a revision surgery to remove and or replace the cage? 3. Does one have the skills and the support to perform such a surgery?. Once you have answered these questions the answer to your problem presents itself.
I agree with the colleagues' opinion above. The main point is the fusion. If the fusion is complete and the patient is asymptomatic, you better leave the cage in situ. In any other situation you should remove it and try to replace it with another one.
I agree completely with my colleagues that fusion is the answer if there is fusion and patient is asymptomatic no need to interfere . but if the patient doesn't have fusion I prefer to go from anterior to remove it even if the patient is asymptomatic .