PHILADELPHIA—After treatment for prostate cancer, biochemical failure should be defined as any increase of at least 2 ng/mL from the lowest PSA reading. That was the conclusion of a report about a new classification system presented here at the ASTRO Annual Meeting.
The new system—known as the Phoenix or Nadir+2 definition—offers a more robust early determinant of patient outcome, in contrast to the standard ASTRO definition of three consecutive rises in PSA after treatment, said the chief investigator, Matthew C. Abramowitz, MD, Chief Resident of the Department of Radiation Oncology at Fox Chase Cancer Center.
In a study comparing the two classification systems, the Phoenix definition better predicted for distant metastases, prostate cancer-death, and overall mortality, he reported. “The Phoenix definition enables us to better predict which patients with PSA failure will be a clinical failure. Since men with prostate cancer can live for years, having an early and robust endpoint is crucially important.”
While “PSA failure doesn't dictate treatment per se, you can tell a patient that once he reaches this threshold, you would like to consider salvage therapy,” he added.
Problems with ASTRO Definition
There are several problems with the ASTRO definition of biochemical failure, according to Dr. Abramowitz.
“The ASTRO definition predicts clinical failure and prostate-cancer specific survival among patients treated with radiation alone. But if they get hormone therapy or implants, there can be problems. When a man comes off hormone therapy, for example, PSA naturally rises. He would be considered a treatment failure under the ASTRO definition, when in fact he is not.”
Additionally, the ASTRO definition incorporates backdating, the time of failure being defined as the midway point between nadir and the time of the third rise, Dr. Abramowitz said. ''That causes artificial flattening of survival curves, which makes things look than they are.
“The Nadir+2 definition avoids these problems. It is associated with less misclassification following withdrawal of androgen deprivation and has better sensitivity and specificity for clinical failure.”
The analysis included 1,831 patients with pathologically diagnosed prostate cancer treated at Fox Chase between 1987 and 2001. Post-prostatectomy patients, as well as patients with node-positive disease or evidence of metastasis at presentation, were excluded.
The median age of the patients was 69, and their average pretreatment PSA level was 7.1 ng/mL. A total of 400 patients (22%) were given initial hormone therapy.
Of the total, 716 patients (41%) had Stage T1 disease, 857 (49%) had Stage T2 disease, and 171 (10%) had Stage T3/T4. The Gleason score was 2 to 6 points in 1,244 patients (68%), 7 in 442 patients (24%), and 8 to 10 in 145 patients (8%).
A total of 81 patients were treated with conventional radiation therapy, and 1,750 men received three-dimensional conformal radiation; the dose ranged from 60 Gy to 79 Gy.
In general, patients with Stage T1 to T2a,b disease and a Gleason score of less than 7 received treatment to the prostate alone. Patients with Stage T2c to T4 disease or a Gleason score of 7 to 10 were given a dose of 46 to 50 Gy to the prostate and periprostatic tissues, followed by a boost to the prostate and seminal vesicles.
Using the Phoenix definition, 389 (21%) men had a biochemical failure. In contrast, 460 (25%) patients were considered in biochemical failure when the ASTRO definition was applied.
At a median follow-up of 71 months, 84 patients (4.6%) developed distant metastases, 48 (3%) died of prostate cancer, and 404 (22%) died of any cause.
Multivariate analysis showed that the Phoenix definition of biochemical failure was a significant predictor of overall mortality, cause-specific mortality, and distant metastases.
Specifically, men considered in biochemical failure using the Phoenix definition were twice as likely to die of any cause, 307 times more likely to die of prostate cancer, and 173 times more likely to develop distant metastases than those not in biochemical failure.
In contrast, men considered in biochemical failure using the ASTRO definition were 26 times more likely to die of prostate cancer and 63 times more likely to develop distant metastases than those not in biochemical failure.
Still, there was no significant association with overall mortality, the study showed.
The analyses were adjusted for T stage, initial hormone treatment, Gleason score, radiation dose, last pretreatment PSA level, and age. Biochemical failure was treated as a time-dependent covariate because the hazard of death or a given recurrence for a patient at the time of biochemical failure is no longer the same as that before the failure, Dr. Abramowitz said.
ASTRO Consensus Statement Supports Use of the Phoenix Definition
Dr. Abramowitz noted that an ASTRO consensus statement, released this year, supports use of the Phoenix definition (Int J Radiat Oncol Biol Phys 2006;65:965–974).
Also, many radiation oncologists at the ASTRO Annual Meeting appeared to be in favor of the change.
However, there were some reservations, mainly due to the difficulty in comparing findings of new studies using the Phoenix definition with those from older studies that utilized the older definition.
Phillip M. Devlin, MD, Assistant Professor of Radiation Oncology at Harvard Medical School and Director of Brachytherapy at Brigham and Women's Hospital, said, “This is a compelling change that needs to be played out in large national databases. For now, I think you can expect us to continue having this conversation.”