Gene-expression signature may signify risk for recurrence, metastasis in prostate cancer

The report has been issued online in PNAS Early Edition.

"Radical prostatectomy is the standard of care for men whose cancer is advanced but confined to the prostate gland, but we know that the factors we use to determine which patients need radiation therapy after surgery are inadequate," says W. Scott McDougal, MD, of the MGH Department of Urology, corresponding author of the report. "The treatments available to our patients can have significant impact on their quality of life, so a better way to know which patients with localized cancer need additional therapy after surgery and which require no additional treatment is a significant unmet need."

Gene expression signatures indicating patient prognosis and sometimes the most appropriate treatment have been incorporated into care for breast cancer and other tumors. Studies looking for such markers in prostate cancer have had variable results, and their potential usefulness to guide treatment has not been determined. For the current study the research team -- led by Chin-Lee Wu, MD, PhD, of the MGH Department of Pathology -- examined samples of malignant tissue from around 200 prostate cancer patients who had radical prostatectomies at the MGH between 1993 and 1995, analyzing the expression patterns of more than 1,500 genes associated with prostate cancer in earlier studies. With the results of that analysis, they developed a 32-gene index to reflect the likelihood that a patient's tumor would recur, signified by detectable levels of prostate-specific antigen (PSA) after the gland had been remove, or spread.

To validate the usefulness of the index, they used it to analyze tissue samples from a different group of almost 300 patients who had their prostates removed in 1996 and 1997, comparing the index with currently used prognostic factors -- such as PSA levels, physical examination, and a tumor's microscopic appearance -- to see how accurately each predicted the actual incidence of tumor recurrence or metastasis during the 10 years after surgery. The expression-based index proved to be the most accurate method. Among those it designated as high-risk, the actual incidence of tumor recurrence was 47 percent and of metastasis, 14 percent. Among those classified as intermediate risk, actual recurrence was 22 percent, and metastasis occurred in 2 percent. No recurrence or metastasis were seen in patients classified as low-risk by the gene-expression index.

To get a sense of whether the index could help determine risk at the time of diagnosis, the researchers used it to assess presurgical needle biopsy samples from 79 patients in the validation group. The risk assignment based on biopsy results closely matched the assessment based on surgically removed tissue, and the prognostic ability of the index was better than that of other pathological information available at the time a biopsy was taken. Because the current report is based on study of patients treated at a single institution, the authors note, it requires confirmation in larger, multi-institutional studies.

"A more accurate prognosis at the time of diagnosis could give patients and their physicians much more confidence in choosing a definitive therapy or pursuing active surveillance for those at low risk, which could reduce over-treatment, a critical issue in disease management," says lead author Wu, an associate professor of Pathology at Harvard Medical School. McDougal is the the Kerr Professor of Urology, at HMS.