Well, let's take a look at his trends:
What we see in this first chart is that he's getting hit hard. That LD rate has jumped through the roof. An LD rate in the mid 20s is high. The good news here, though, is that the LD rate jump came from his FB rate and not from his extraordinarily high GB rate (compared to his historical norms). If a guy has to choose between converting FBs or GBs to LDs, he should almost always take FBs. That lowers the chance of HRs, and the continuance of GBs will potentially lead to double plays that can get him out of trouble.
On a red, yellow, green warning scale, I'd give this first chart a yellow.

Next, we're going to take a look at his BABIP. Insomuch as he was getting lucky during his hot streak, he is getting equally "unlucky" now. However, knowing what we know about his LD rate from the last chart, I would say that this BABIP, while high, makes some degree of sense. If you're hit pitches are getting clobbered nearly 25% of the time, those guys are probably going to get on base more often than that.
I'd give this chart another yellow rating on the scale.

Next, let's look at some of his important peripherals to see if there is any skill degradation or reversion going on here.
As we see below, his K rate was higher during his streak, but it wasn't ridiculous - and it's only slightly reverted during this cold streak. What is disturbing, though, is his increase in walks. Granted, his walk rate was not only minuscule by his own standard, it was ridiculous by every standard. So, this increase was expected, even if it is unwelcome.
I would rate these charts a yellow, too.


Now let's take one last look at his game lines. Over the last 4 games you summarized, 2 were in hitter-friendly parks (Great American & Arlington). How'd he do in those two games? 9.28 ERA, 10 K, 3 BB, 16 FB (3 HR), 16 GB, 8 LD. How'd he do in the other two games? 4.26 ERA, 11 K, 5 BB, 10 FB (0 HR), 22 GB, 10 LD. During those games, his strike-to-ball ratio didn't really move too much from their numbers earlier in the season, which is a good sign.
At this point in my analysis, I'm ready to say that this guy is just leveling out and that he will still provide value, even if there will be good days and bad days mixed in. But then I noticed his pitch counts this year:
108, 102, 109, 120, 97, 103, 117, 96, 113, 101, 112
He's only thrown less than 100 pitches in 2 games this year, and one of them was when he got clobbered by the Reds. Even though he threw 200+ innings in '05 and '06, this is a guy that only pitched in 97 innings last year. My first "red" warning sign is that possibly he's getting worn out. If his pitches are still making it over the plate, and if they're just getting hit better, either hitters have figured him out, or he's lost some movement/speed on his fastball. Possibly, he's having location issues with his fastball. Hitters did comment that during his streak, his fastball was ridiculous not only in its movement but also in it lethal accuracy.
So what to do with Lee? Well, you can't bench the guy. That just makes no sense. I'd say the days of plenty are looking like they've come to an end. If you can trade him for a top-tier pitcher right now, I would do it in a heartbeat. If not, hold on to him and keep an eye on that LD rate. If it continues to stay up, or if his FB rate returns with an increase in HR/FB, I'd say that's a pretty serious warning sign. If the LD rate normalizes, then this was just a spate of him getting some hot hitters, and his numbers should go back to being contributory in nature.