MILAN, Italy—Genomic profiling could help reduce the risk of overtreatment in primary prostate cancer (PC) by identifying patients and healthy individuals whose genes put them at greatest risk of developing dangerous disseminated disease, said Norman J. Maitland, PhD, Professor of Molecular Biology and Director of the Cancer Research Unit at York University, U.K. He was speaking at the 2016 European Multidisciplinary Meeting on Urological Cancers.
He interpreted data from several recent studies—investigating the influence gene mutations had on prostate tumor development—that infer gene arrays could soon distinguish patients with dangerous tumors from those whose disease does not need aggressive management.
“[Currently], we don't know which [patients] are going to die of prostate cancer and which ones can be left untreated for the rest of their lives,” he said. “Genetics is one tool to allow us to understand that.”
Healthy Population at High Risk
One of the studies Maitland discussed identified 23 susceptibility loci for PC that “could pick out the top 1 percent of the risk distribution” among healthy individuals—who had nearly a five-fold higher risk of developing primary PC than average (Nat Genet 2013;45(4):385-391,391e1-2).
Improved “bioinformatics” were needed, said Maitland, to assess multiple genetic indicators of risk in each patient and to interpret genome-wide associations of PC-specific survival so as to recognize lethal forms of PC from among many less-threatening types.
Maitland illustrated the difficulty of genetic prediction by discussing data from a genome-wide survival analysis of cause-specific death among 24,023 patients with PC (in whom there were 3,513 disease-specific deaths) that found no significant association between genetic variants and prostate cancer survival (Cancer Epidemiol Biomarkers Prev 2015;24(11):1796-1800).
Because of tumor heterogeneity Maitland warned that biopsy was not always good enough for prediction—mutations could be different depending on which part of the tumor sampled (Nature Genetics 2015;47:367-372).
In addition, targeted therapies might not work because of the multiplicity of potential targets and the likelihood that some of them would elude the best endeavors of the targeteers.
There was also another hurdle to overcome—a cancer “genomic halo” that could limit the effectiveness of local therapy for PC, Maitland explained. Cancer cells in the tumor radiated their cancer-causing influence into surrounding tissues—in a process he described as a “field effect” or “field cancerization”—through which cancer-causing mutations present in the tumor were also found beyond tumor margins.
“Around [the tumor] there appears to be a field of cells, which are activated to become cancer,” he noted. The logical conclusion was that focal therapies—like radical prostatectomy—could only be curative if affected surrounding cells were ablated along with the tumor.
Maitland described PC tumor samples as not very different from healthy tissues.
“The gene complement in many prostate cancers is more than 99 percent identical to the original normal tissue. It's just that a critical gene becomes mutated and that triggers the whole cancer event.”
But Maitland was encouraged by the fact that research had yielded a “catalogue of mutations likely to occur in prostate cancer” that were already reflected in panels currently being tested in research on the genomics of prostate cancer—even though PC was not such a rich source of mutations as most other common cancers.
“There are many mutations [in PC], but they are a fraction of the number that are present in lung cancer, melanoma, and in almost any other solid cancer including breast cancer. Prostate cancer seems to be mutagenically quite separate from the other cancers.”
In PC, thousands of cells of a tumor needed to be sequenced, he said. “But it might be the cell that looks normal that's the one to go on to become cancer.”
To make matters even more complicated, Maitland noted that not all gene mutations detected by arrays were active. Some mutated genes—active early in the tumor's development—could have been switched off at later stages since they were no longer required.
“Identifying them will be important in deciding how a cancer starts, but useless if you designed a treatment against them,” he said.
“Every time you treat a cancer you are selecting for survivors. With prostate cancer, the first drive in this [process] is either to mutate the androgen receptor gene or to amplify it,” Maitland continued. “This comes about because the drugs are dampening down the androgen response critical in the growth of prostate cancer and the cell needs to find a new way to respond to these minimal levels of androgen.”
In addition to mutations, PC also had gene rearrangements that seemed to be diagnostic, according to Maitland. “But they're diagnostic for good prognosis rather than bad prognosis, so they're not ideal tools.”
When he was asked if genomic profiling could become a powerful tool in attacking prostate cancer, Maitland suggested genetic “mapping” could eventually be viewed as more useful for clinical planning than measuring the size and shape of a tumor.
“If you could image the genetics of the cancer rather than [just] its appearance, then you could extend focal therapy to as much of the field as you can get,” he concluded.
Peter M. Goodwin is a contributing writer.