It'd have several desirable properties: 1. You'd have a flat smooth surface that you could apply a liner to, instead of a surface that undulates and (probably)

It'd have several desirable properties: 1. You'd have a flat smooth surface that you could apply a liner to, instead of a surface that undulates and (probably) has fissures. 2. The diameter would also be constant, so that you could design your liner to fit based on known dimensions. Curves would also be of a known radius. 3. Also, our tunnel would go exactly where we want, and not wander around based on some ancient flow pattern. Or form oxbows where it loops back on itself. 4. So far as a structural liner, if you look at some of the tunnels under rivers where there is the danger of water intrusion, to my knowledge none of them have ever used a natural formation. Brunel used a brick structural liner under the Thames, and modern TBMs use interlocking concrete panels with rubber gaskets.