5/8/2012

The detection of certain gene abnormalities in the blood or tumor tissue of prostate cancer patients can predict the likelihood of relapse, according to researchers at the University of Pittsburgh School of Medicine. Their findings, published online today in The American Journal of Pathology, show that the test also can indicate how aggressive or mild the relapse will be.

Currently, prostate-specific antigen, or PSA, blood levels are tested to monitor the status of prostate tumors, said senior investigator Jian-Hua Luo, M.D., Ph.D., associate professor, Department of Pathology, University of Pittsburgh School of Medicine. But measuring copy number variation (CNV), which is the deletion or increased redundancy of areas of DNA within chromosomes, in the tumor, neighboring tissues or blood better reflects the state of the cancer.

“Our method will allow us to determine at the time of, or after biopsy or prostate removal, whether the cancer is likely to come back and, if so, how aggressively,” he said. “It promises to more accurately predict the progression of the disease.”

For the study, the researchers analyzed the genomes of 238 samples obtained from men whose prostate glands were surgically removed; 104 prostate tumor samples; 85 blood samples from prostate cancer patients, and 49 samples of disease-free prostate tissues neighboring the tumors.

A third of the samples were from patients whose cancer had recurred and whose PSA level had doubled in less than four months, which is associated with lethal prostate cancer; a third from patients with disease recurrence with a slowly increasing PSA level that doubled in more than 15 months; and a third with no relapse more than five years after surgery. The researchers also examined an additional 25 samples from prostate cancer patients to validate their findings.

They found that deletion and increased redundancy of DNA occurred in all chromosomes in prostate cancer samples. Some of these changes occurred with high frequency. Deletion and increased redundancy of DNA also occurred in benign neighboring tissue and blood samples of the cancer patients.

Gene-specific tumor CNV could correctly predict 73 percent of cases that had relapsed and 75 percent of cases in which PSA levels rapidly doubled. The CNV model from disease-free neighboring tissue correctly predicted 67 percent of cases for relapse and 77 percent of cases for short PSA doubling time. A specific tumor CNV from blood could correctly predict 81 percent of relapse cases and 69 percent of the cases for short PSA doubling time.

The consistency of the associations across specimen types suggest that CNV analysis could reliably indicate what the likelihood of recurrence is either at the time of biopsy of a suspicious mass, when the tumor is removed, and in post-treatment blood monitoring, and could help doctors decide early in the disease process whether an individual’s cancer warrants additional therapy, Dr. Luo said.

Co-authors of the paper include Pitt’s Yan P. Yu, M.D., Ph.D., Baoguo Ren, M.D., William LaFramboise, Ph.D., and George Michalopoulos , M.D., Ph.D., all of the Department of Pathology; Chi Song, Ph.D., and George Tseng, Sc.D., both of the Department of Biostatistics, Pitt Graduate School of Public Health; and Joel Nelson, M.D., of the Department of Urology.

The study was funded by grant RO1CA098249 from the National Cancer Institute, part of the National Institutes of Health, and the University of Pittsburgh Cancer Institute.