"Surgery or radiation for high risk PCa" redux - a cautionary study example

Following the lively discussion on Pratoman's "Surgery or Radiation for G9-10?' recent study post, this one came in my "Uroalerts - Prostate Cancer Weekly" email feed.

Here's a link to the study, though I present it with some trepidation. In my humble opinion, this is only an illustrative example of the limits of retrospective studies, and how some researchers may put a particular spin on the results.

Comparative Effectiveness of Radical Prostatectomy Versus External Beam Radiation Therapy Plus Brachytherapy in Patients with High-risk Localized Prostate Cancer

This study is a good example of what I so often caution about, the need for carefully reading studies with strongly stated and sometimes provocative conclusions. This is a retrospective study, and actually shows only a very slight advantage for surgery and no suggestion at all of a range of error in the conclusion graphic.

In fact, even at 10 years of follow up, surgery (~84%) showed only about 4% better overall survival vs RT (~80%), surely within the margin of error. To me, that's showing equivalence. And yet, check out this conclusion from the article:

"Our results, showing that young and healthy men presenting with high-risk localized PCa had better survival with RP compared with EBRT + BT, call into question other recent research suggesting better or equivalent outcomes with EBRT + BT." (emphasis mine)

It's a retrospective study, so it has inherent biases, some of which they attempted to compensate with some specialized techniques. There were over 10 times as many surgery cases as radiation (FTA: "In total, 12 283 (88%) patients underwent RP versus 1702 (12%) patients who underwent EBRT + BT."). 91 months median follow up, so at least it's a good long-term look.

As an engineer I'd probably be more likely to conclude this showed the two treatment plans to be about equivalent, as a number of other studies have indicated. They're trying to say the opposite, and it looks to me like they are really overstating the value of their analysis.

Be careful reading studies. What the headline giveth, the fine print taketh away!

Certainly interesting because it was quite a large study compared to Pratoman's link with a similar conclusion - that the triple-play (BBT + ADT) and surgery+radiation have similar mortality in relatively young patients.