Checking PSA levels every three to six months after radiotherapy, they found about 12% of the 964 patients had such a PSA bounce.

Although Pisters worried that a PSA bounce meant the cancer was more likely to recur, he found just the opposite. Five years after treatment, about 82% of men whose PSA bounced were free of signs of cancer, compared to about 58% of those whose PSA did not bounce.

A bounce usually began with less than a one-point rise (less than 1 ng/ml), because larger rises were less likely to drop back to their starting points, Pisters said.

Most bounces happened within two years of treatment; later rises were less likely to be part of a bounce, Pisters said. In their study, 80% of men had their bounce within two years and by two and one-half years, that number rose to almost 90%.

Unfortunately, the researchers found no way to predict whether rising PSA would later fall back to its starting point, becoming part of a bounce, said Pisters.

But more study may help scientists be able to make such predictions, which could be very useful to patients and to doctors planning their treatment, Pisters noted.

It's understandable why those with a bounce had less chance of recurrence. Men whose first rise after treatment didn't go back down to pre-treatment levels by definition didn't have a bounce, and may have been having a recurrence, Pisters said.

But the new definition better describes what a real bounce is and helps highlight that not all rises are recurrences, said Pisters.

The study is reassuring, and shows patients shouldn't "overinterpret" small rises in PSA by assuming they signal the return of cancer, said Timothy D. Gilligan, MD, a genitourinary oncologist at Dana-Farber Cancer Institute in Boston.

"We need to be careful about how we interpret PSA right after radiotherapy, so we don’t jump in with unnecessary testing or procedures prematurely, before it's clear that rising PSA means residual disease rather than just reaction to treatment," said Gilligan.