New findings on prostate cancer may enable doctors to make better diagnoses and prognoses for patients and provide novel directions for therapies, according to a report from The Cancer Genome Atlas (TCGA) Network.
Investigators published the in-depth analysis of 333 prostate cancer tumors online Nov. 5 in Cell. TCGA is jointly supported and managed by the National Human Genome Research Institute and the National Cancer Institute, both parts of the National Institutes of Health.
While 90 percent of prostate cancers are now identified as clinically localized tumors, once diagnosed, these cancers tend to have a heterogeneous and unpredictable course of progression, ranging from slow-growing to fatal disease.
“We have identified seven clearly defined subtypes of prostate cancer based on genetic alterations,” said Massimo Loda, MD, director of the Center for Molecular Oncologic Pathology at Dana-Farber/Brigham and Women’s Cancer Center, and a co-principal director of the study.
Loda added, “Interestingly, it is clear that there is substantial diversity within each of these subtypes. We can now put this critical information into clinical and pathological context and, by developing biomarkers for the different genetic variants, use them to guide therapeutic options.”
"Until now, we haven't had a reliable way of predicting the way a primary prostate cancer will act by looking at the genome," said Chris Sander, PhD, principal investigator and chair of the computational biology program at Memorial Sloan Kettering Cancer Center. "The TCGA study gives us much more information about the spectrum of alterations in tumors and can help us predict the development of the disease. This will also inform the design of new clinical trials."
According to the American Cancer Society, prostate cancer will be newly diagnosed in more than 220,000 men in the United States in 2015, making it the second most common cancer affecting men and the second leading cause of death from cancer in men. Most prostate cancers are detected early while still confined to the prostate, a walnut-sized gland located below the bladder. While most cases remain harmless - benign - for decades, other subtypes of prostate cancers can be aggressive, and spread to other parts of the body (metastasize), making them extremely difficult to treat. It is currently difficult for healthcare providers to distinguish which cancers will remain harmless and which will metastasize.
The scientists studied five aspects of the prostate tumors:
- The number and kinds of genetic mutations.
- Gene fusions (when genes attach to each other or otherwise combine).
- The number of copies of DNA segments (abnormal differences in the cell's number of copies of DNA segments can contribute to cancer).
- Gene activity, including when genes are turned on or off, and how much activity is seen.
- DNA methylation (methyl chemical groups are added to many places on a cell's DNA and act like on/off switches for a gene). Mistakes in DNA methylation can turn genes on or off at the wrong time and contribute to cancer.
Of the seven subtypes, the investigators found that four are characterized by gene fusions, while the other three are defined by mutations in the SPOP, FOXA1 and IDH1 genes. The subtypes with SPOP and FOXA1 mutations share several genomic characteristics, suggesting that mutations in these genes cause similar disruptions in the cell to bring about cancer.
Investigators discovered that mutations in the IDH1 gene are similar to those found in leukemia and brain cancer. Such a cancer subtype could be a candidate for a "basket" clinical trial, which tests for similar mutations across cancer types. The goal of this type of trial would be to personalize a patient's treatment based on the mutations, not on the anatomical location of the cancer.
Sander said that while previous studies have examined gene copy number or the number and kinds of genetic mutations or the level of gene activity in prostate tumors, the TCGA study is the first to comprehensively and systematically examine many different types of data together on a large scale.