Every case should be looked at individually, since details will vary, but as merrillwalt has found, if you do the calculations, you will probably find in many cases that such a bundle can indeed be lifted without a cradle.
Using the parallel axis theorem, it can be found that the stiffness of the bundle is very large. So, as long as the tube to tubesheet joint is adequate to transfer the tensile loads from the lifting operation (I have not done calculations to see if there would be a combination where it would not be - I would think if the joint meets Code requirements, it will be quite strong enough for the loads during lifting and rotating from horizontal to vertical.
We have a few vertical bundles at our site, one very large, with about 7000 tubes 3/4" diameter. We have sucessfully lifted all of these bundles vertically out of the shell, rotated to horizontal, cleaned and inspected, then lifted back to vertical and reinstalled without any damage to the bundle. In no case have I calculated any need for a cradle.
Our bundles do have floating heads at the bottom end, with lugs or trunnions which are used for the tailing crane. So long as the sling is not put around the tube bundle itself, no damage will occur to the tubes.
I agree a cradle could be beneficial for shipping puroposes, but if it is not needed for lifting, then I would recommend removing it before lifting. Trying to remove it after getting the bundle to vertical could be inviting an accident.
Test shells are a completely different matter - such a shell may be very useful for leak testing once the bundle is removed from its "normal" shell, but again, unless calculations prove it is necessary for lifting, I would not use it during the lifitng operation.