PSA Below 9 ng/mL Triggers Unnecessary Biopsies

STANFORD, Calif—Attempts to reduce prostate cancer mortality by ordering biopsies at ever-lower prostate-specific antigen (PSA) levels are driving up health care costs and subjecting thousands of men to early prostatectomies they could safely defer for years, according to researchers at Stanford University Medical Center.

Now, a large study has correlated serum PSA levels with subsequent quantitative histology of the cancer in radical prostatectomies at 1 ng/mL intervals of preoperative PSA. Thomas A. Stamey, MD, and his colleagues found that PSA levels between 2 and 9 ng/mL provide no reliable predictive information about prostate cancer and are principally the result of benign prostatic hypertrophy (BPH). This study was published in the Journal of Urology (167:103-111, 2002).

The purpose of the study was to determine whether preoperative PSA levels correlated with any of a number of morphological variables known to affect cure rates from radical prostatectomy.

"We are overdiagnosing prostate cancer at a huge rate," said Dr. Stamey, professor of urology, Stanford University. He suggests that physicians may wish to avoid biopsy at PSA levels lower than 4 ng/mL. He thinks that biopsy can be delayed safely for a few years to confirm a PSA rise from 4 ng/mL into the 7 ng/mL range, and that prostatectomy can be delayed for years in men whose PSA is in the 2 to 4 ng/mL range and whose prostate is larger than 30 g.

Prostate cancer is a ubiquitous cancer of aging that starts in the 20s and increases in prevalence in every decade, from 8% in men age 20 to 30 to 80% in those age 70 to 80, Dr. Stamey said. "If you look for cancer with enough biopsies in men over age 50, you will often find it. However, few men die of the disease," he said. The SEER data show that the annual prostate cancer death rate in men over age 65 is 0.23% (226/100,000).

"We all thought there was a relationship between a PSA of 2 to 9 ng/mL and prostate cancer, but there is not. The cancers detected in biopsies of men at those levels is due to serendipity," he said.

Since 1983, Dr. Stamey and pathologist John E. McNeal, MD, have sectioned every prostate removed at Stanford into 3 mm step sections and quantified the percent of Gleason grade 4 or 5 tumors, volume of grade 4 or 5 tumors, volume of the largest tumor, percent of tumors smaller than 0.5 mL, percent with capsular penetration, percent with positive surgical margins, percent with positive lymph nodes, percent with positive seminal vesicles, percent with cancer confined to the organ, percent with transitional zone involvement, prostate weight, patient age, and percent of tumors that were stage T1c-T2c.

Statistician Iain M. Johnstone, PhD, helped cross reference the pathology data with preoperative PSA levels to answer two questions: What was the relationship between PSA and tumor pathology, and how well did PSA reflect the actual morphology of prostate cancer?

The study included 784 men with serum PSA levels of 2 to 22 ng/mL. The overall correlation between PSA and cancer morphology for the 579 men whose largest tumor was in the peripheral zone was 10% for all cancers and 3% for high-grade cancers, indicating that "almost no PSA produced by these peripheral zone cancers enters the serum," he said.

The morphological variables changed "at rates of doubtful medical significance" in patients with PSA below 7 to 9 ng/mL, and the correlation for these changes was never better than 15% even at PSA levels above 9 ng/mL. Large individual variation in cancer morphology was seen at all PSA levels.

Prostate Weight

Prostate weight increased by 21% for each doubling of PSA below 9 ng/mL but by only 4.8% for each doubling of PSA above that level. In essence, the larger the prostate, the lower the risk that an elevated PSA is due to cancer at serum PSA levels of 2 to 9 ng/mL.

Conversely, Dr. Stamey told ONI that physicians should have a heightened index of suspicion for cancer when PSA rises into the 2 to 7 ng/mL range and the prostate is smaller than 30 g.

"Preoperative serum PSA has a clinically useless relationship with cancer volume and grade in radical prostatectomy specimens, and a limited relationship with PSA cure rates [defined as serum PSA less than 0.07 ng/mL] at preoperative serum PSA levels of 2 to 9 ng/mL," he concluded. BPH was responsible for most PSA elevations below 9 ng/mL and constituted the primary cause of the overdiagnosis of prostate cancer.

Deferring Biopsy

According to Dr. Stamey, standard practice has been to biopsy patients with PSA levels between 4 and 10 ng/mL. More recently, some clinicians have begun to biopsy patients with PSA levels as low as 2 ng/mL. Dr. Stamey said that 10% of men over age 50 have PSA levels of 4 to 10 ng/mL and that dropping the criterion down to 2 ng/mL would add another 20% of the adult male population to the pool of biopsy candidates.

"We cannot afford that cost for so little benefit," he said. "I would not biopsy anybody with a PSA level below 10 ng/mL and a normal-feeling prostate. Many men with PSA in that range can safely defer surgery and retain an intact prostate for years before undergoing prostatectomy."

Another benefit from deferring biopsy until PSA rises to 7 ng/mL is that it should decrease the number of clinically insignificant cancers removed. This is an important benefit, he said. Researchers at Johns Hopkins have reported that 26% to 29% of radical prostatectomies at their institution contained potentially insignificant tumors, Dr. Stamey said.

"Prostate cancer and breast cancer are two diseases we thought we had great screening tests for: PSA and mammography. Now it appears neither is accurate for detecting cancers in the early, curable stages. That is a little humbling," Dr. Stamey said.

His laboratory is actively involved in efforts to discover the gene responsible for Gleason grade 4-5 cancer. "For every 10% increase in Gleason 4-5 cancer, we lose 10% curability in men treated by radical prostatectomy," he said.