An assay which uses reverse-transcriptase polymerase chain reaction (PCR) allows the detection of small numbers of cells expressing prostate-specific antigen (PSA) and may be useful in preoperatively identifying patients with early metastatic spread of disease, reported Mitchell C. Benson, MD, at the First Sonoma Conference on Prostate Cancer. Dr. Benson is Professor of Urology at Columbia University College of Physicians and Surgeons, and Director of Urologic Oncology at the Columbia Presbyterian Medical Center.
Were using RT-PCR for PSA, to detect extremely small quantities of cancer cells beyond the limits of the primary organ...And the current sensitivity of these assays is that we can detect about one cell in a million, Dr. Benson said.
Potential Applications of PCR
Potential applications of PCR in prostate cancer would be preoperative assessment for metastatic disease, and postoperative monitoring, Dr. Benson stated. However, for postoperative monitoring, you can view PCR as a hyper-hypersensitive PSA assaybut it only buys you about 3 months in terms of earlier detection. So, I dont think that theres any role for PCR in postoperative monitoring in terms of its ability to detect recurrent disease earlier.
He does, however, see an important role for PCR in preop assessment. The rationale for molecular prognosis is that we can detect small numbers of cells. Its clearly more sensitive than radiological screening.
Archival Lymph Node Study
Dr. Benson reviewed data from an archival lymph node study on a radical prostatectomy series at Boston University. All of the patients had negative lymph nodes at surgery.
The study showed that 44%, or 16 of the 36 patients who had no pathologic evidence of metastatic disease, had PCR-demonstrable metastatic disease in their lymph nodes, Dr. Benson said. And thats very, very impressive. Among the 16 patients who had a positive PCR result, 14 had a PSA failure within 5 years.
To test the usefulness of PCR in identifying metastatic disease at earlier stages, the assay was applied to a cohort of more than 200 patients, all of them candidates for radical prostatectomy based on current diagnostic modalities that predicted prostate cancer localized to the prostate gland. The mean PSA was 8.6.
Blood was drawn from the patients within 4 weeks of surgery and assays performed in a single laboratory for consistency. Among the results:
- Of the 2 patients with positive lymph nodes, PCR detected 1. Dr. Benson noted that only a single blood assay was performed and said that studies still need to be done to determine the number of PCR assays needed to be confident of results.
- Of 15 patients with positive seminal vesicles, 10 were shown by PCR to be positive for cancer cells. So, we have a 66% or 67% sensitivity in that setting, Dr. Benson said.
- Of 43 patients with positive surg-ical margins, 28about two-thirdswere positive for PSA via PCR prior to surgery.
- Of 119 patients with organ-confined disease, 107 were PCR negative. But it would be wrong to classify the 12 who tested positive on PCR as false-positives, Dr. Benson said, because PCR is being used for molecular prognosis, not molecular staging.
What were saying here, Dr. Benson explained, is that in 107 patients, on a single blood assay, we could not detect that these patients prostate cancer had shown any sign of spreading beyond the prostate gland.
- Of the 140 patients who were PCR negative, 121 had cancer that was confined to the specimen. I think this is really the key, Dr. Benson stated. In other words, if we only operated on people who were PCR negative, we could dramatically increase the percentage of patients in whom we would have long-term surgical cures. Thats taking 30% of the patients who fall into our favorable category, who we think are good surgical candidates, and saying theyre not really good surgical candidates.
Stratifying by Serum PSA
Stratifying by serum PSA, there were 23 patients with a PSA less than 4, with 7 of those being PCR positive, and 2 of the 7 having adverse pathologic findings. But, to date, none of these patients are PSA failures, Dr. Benson noted, adding that these are patients who would be in a good prognostic category, but that the length of follow-up was short, with a mean of about 18 months.
Of 129 patients with PSA between 4 and 10, 32 were PCR positive. Half of these had adverse pathologic findings and 1 was a PSA failure. Clearly people in the higher PSA group may be having some higher volume cancer and more aggressive disease, Dr. Benson said.
Of 48 patients with a PSA greater than 10, 21 were PCR positive, with 90% of them having adverse pathologic findings, and 9 of the 21 have already demonstrated PSA failure in short follow-up.
Although the breakdown into categories of PSA scores seems arbitrary, it seems to hold in terms of our ability to stratify patients, Dr. Benson observed. He admitted that the greater than 10 PSA category allows for a very wide range. One of the things thats dangerous about greater than 10, he said, is that 100 is greater than 10, and 11 is greater than 10. So theres a lot of variability in greater than 10.
Predictive Values Are High
The predictive value for a positive PCR, for some type of adverse event--meaning either adverse pathology or PSA failure--right now is 80%, Dr. Benson summarized. And the negative predictive value of a negative PSA is 76%.
If we look at collected literature, we can look at some controls, he said. If one looks at people without prostate cancer, 99.5% of them are PCR negative. And the 1 patient who was positive came from an assay where they really pushed the number of cycles of the PCR to the limit.
The true value of PCR in patients with prostate cancer, Dr. Benson concluded, will be determined by the clinical course of these 200 patients.