I read the article, thinking this was an incorrect claim in the summary. Nope, the article insists in several places that it was "undetectable" by the surgeons. Now, anyone who's played any online FPS knows that 50ms ping times are not only detectable, but are approaching unplayable because some punk kid that's only 10ms away from the server is always taking the head shots before you can even see him.
So I figured there has to be something else. The best hypothesis I could come up with is the current robotic surgery tools introduce their own lag such that the surgeons were unable to distinguish normal device response times from network latency. That, and the goals of a surgeon are completely different from an FPS shooter. A surgeon isn't trying to race anything or anyone - they don't have to shoot first. In a live operating theatre, they are methodical and cautious. It's not like there are sudden surprises that leap out at them that they have to instantly react to. Even a burst blood vessel takes a few moments to assess and plan a recovery from. So maybe if they're used to very slow approach, the latency doesn't impact them as much.
That 50ms lag in a game is noticeable because there are other people who are also experiencing latency and the game is doing things to compensate for everyone being at a slightly different point of reference. If you were just controlling a camera pointed out your window with a 100ms latency, your mind would not notice the latency between your motion with the hand and the results with your eye.