Improved PCa outcomes may be statistical artifact

December 1, 2005

Farmington, CT--Before epidemiologists and urologists pat themselves on the back for the improvements seen in outcomes following definitive treatment for prostate cancer, they may want to consider data assimilated by Peter C. Albertsen, MD, professor and chief of the division of urology, University of Connecticut Health Center, Farmington.

Farmington, CT-Before epidemiologists and urologists pat themselves on the back for the improvements seen in outcomes following definitive treatment for prostate cancer, they may want to consider data assimilated by Peter C. Albertsen, MD, professor and chief of the division of urology, University of Connecticut Health Center, Farmington.

He supports this observation with a study titled "Prostate cancer and the Will Rogers phenomenon," which was presented at the AUA annual meeting and was subsequently published in the Journal of the National Cancer Institute (2005; 97:1248-53).

Translating the Will Rogers phenomenon to prostate cancer epidemiology means that, although Gleason score-specific prostate cancer mortality rates appear to have fallen over time, some of the apparent improvement may be illusory: It may be attributable to the shift in tissue grading.

Reevaluating readings

To arrive at this observation, Dr. Albertsen and his colleagues isolated 1,858 histology slides from men 75 years of age or younger who had been diagnosed with prostate cancer between 1990 and 1992. The slides were reinterpreted by a referee pathologist blinded to the original readings, clinical baseline information, and outcomes. This pathologist's findings were supported by two other pathologists who analyzed 10% of the slides under similar circumstances.

The upward shifts in grading were dramatic. Dr. Albertsen reported that 1,028 (55%) of the 1,858 specimens were upgraded, compared with 251 that were downgraded. Roughly one-third (31%, or 579 slides) retained the same grade on reinterpretation.

The team then compared cause-specific survival of the patients according to the original or initial grade with that of patients classified under the contemporary grade. The cause-specific survival curve of the patients with contemporary grades was consistently better than the cause-specific survival of patients who were assigned that same grade in the 1990s. This improved survival shift was apparent at every Gleason grade.

The team reported that their results demonstrate that contemporary Gleason score readings produce an apparent statistically significant improvement when clinical outcomes are compared with outcomes associated with historic Gleason scores.

What is the difference?

"I don't think the Gleason system has changed, but its application may have," Dr. Albertsen said. "Factors such as the widespread application of PSA testing, transrectal ultrasonography, and the spring-loaded biopsy gun coincide with this shift."

He explained that the spring-loaded biopsy gun acquires much smaller tissue specimens than earlier technologies and procedures.

"Pathologists today have much less material to work with and are more comfortable calling prostate cancer at a minimum of Gleason 3. That drives the first number. If they are dealing with a microscopic focus of disease, they do not really have a second pattern. The results, therefore, become a Gleason 3 plus 3. That is where we are seeing many of the sixes we see now," he explained.

Interpretation is the key

The study shows that the Gleason score shift seen in the past decade has much more to do with modern interpretation of biopsy specimens than with selective identification of more aggressive tumors by current PSA screening practices, he said.

This is more than a curious epidemiologic observation, according to Dr. Albertsen, who suggested that if new and evolving technologies and procedures are going to be evaluated by comparing the results they generate with the results generated by past procedures, the Gleason shift must be factored in to produce accurate evaluations.