Patients whose localized prostate cancer contained a loss of the phosphatase and tensin homolog gene and amplification of the v-myc myelocytomatosis viral oncogene homolog gene were 53 times as likely to die from the disease, compared with similar patients without the genetic changes.

The increased risk was seen despite similarly staged tumors and similar levels of prostate-specific antigen (PSA) at the time of diagnosis in analyses of 458 patients in multiple cohorts, according to Wennuan Liu, Ph.D., and his associates.

If the findings are confirmed, they could help clinicians decide which patients would best benefit from aggressive treatment with prostatectomy or radiation and which might be fine with conservative management.

The investigators first assayed DNA samples from 125 patients who underwent prostatectomy for localized prostate cancer at Johns Hopkins Hospital, in Baltimore between 1988 and 2004. They found 20 significant regions of chromosomal copy number alterations in the cancer cells, including four copy number alterations and the target genes within those alterations that had not been identified before (Cancer 2013 April 22 [doi:10.1002/cncr.27954]).

Alterations in the chromosomal copy numbers of the phosphatase and tensin homolog (PTEN) and v-myc myelocytomatosis viral oncogene homolog (MYC) genes, when occurring together, were associated with a 53-fold increased risk of death from prostate cancer in a multivariate analysis. That prognostic information was independent of information gleaned from pathologic stage, Gleason score, and initial PSA level, reported Dr. Liu of Wake Forest University’s Center for Cancer Genomics and Center for Genomics and Personalized Medicine Research, in Winston-Salem, N.C.

Independently, loss of PTEN was associated with a sevenfold increased risk of death from prostate cancer, while gain in MYC was associated with an eightfold increased risk of disease-related death. The prognostic information from the chromosomal copy number alterations was most relevant for patients whose tumors had a pathologic Gleason score of 7 or less.

Many of the patients in the initial cohort had aggressive prostate cancer, with a Gleason score of 8 or greater in 34% of patients, a pathologic stage of T3b or higher in 33%, and a pretreatment PSA level higher than 10 ng/mL in 44% of patients. Twenty-two patients in this cohort died of prostate cancer (18%); the rest are alive or died of other causes.

The investigators confirmed the relationship between fatal prostate cancer and copy number alterations of PTEN and MYC in analyses of three subsequent cohorts.

Among 103 patients who underwent prostatectomy at the Karolinska University Hospital in Sweden, most had less-aggressive prostate cancer, with a pathologic Gleason score of 7 or lower in 85% of patients, a pathologic stage less than T3 in 82%, and a pretreatment PSA no greater than 10 ng/mL in 64%. Four of these patients died of prostate cancer (4%). They also analyzed publicly available clinicopathologic and survival information for 216 patients treated at Memorial Sloan-Kettering Cancer Center in New York, 4 of whom died of prostate cancer (2%).

When the Karolinska and Memorial Sloan-Kettering cohorts were combined, only 1 of 201 patients who lacked the chromosomal copy number alterations in PTEN and MYC died of prostate cancer (0.5%), compared with death from prostate cancer in 6% of the remaining patients, whose tumors did have one or both of the prognostic markers, Dr. Liu reported.

Analyses of tumor samples obtained at autopsy for 14 separate patients seen at Johns Hopkins who died of progressive prostate cancer found chromosomal copy number alterations in one of the genes in 100% of patients, and alterations in both genes in 57%. In comparison, 58% of the original cohort of 125 patients treated for localized disease at Johns Hopkins had alterations in one of the genes, and 10% had alterations in both.

The findings are supported by previous in-vitro and animal studies, the investigators noted, but more studies are needed because of the relatively small number of deaths in the Karolinska and Memorial Sloan-Kettering cohorts.

The study was limited by the possibility of mutations influencing tumor phenotype or the selection of specific copy number alterations. The median clinical follow-up of 5-7 years may not be predictive of long-term mortality risk.

Further research into the lethal tumor phenotype’s effect on tumor evolution may lead to the design of therapies to interrupt disease progression, Dr. Liu noted.

Prostate cancer is the most common cancer among men. Approximately 28,000 U.S. men die from the cancer each year. Prostatectomy improves 15-year mortality rates modestly, compared with conservative management – 15% vs. 21%, respectively – and 1 in 7 patients who undergo prostatectomy relapse and die of prostate cancer. These statistics have propelled efforts to identify better prognostic markers and therapies.

The study was funded by the National Institutes of Health and the U.S. Department of Defense. Dr. Liu and his associates made no disclosures regarding potential conflicts of interest.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

DNA changes predict prostate cancer death