Saltar para: Post [1], Comentários [2], Pesquisa e Arquivos [3]


Quinta-feira, 20.08.15

Molecular Profiles Early Detection for Prostate Cancer


In a first-ever study, findings identified five gene sets that can predict clinical outcome in independent patient cohorts with prostate cancer, according to researchers in England.1  

Combining gene copy number and gene expression data has yielded a new 100-gene risk-stratification signature for prostate cancer that appears to outperform prostate specific antigen (PSA) and Gleason scores, and might allow early detection of the most aggressive cases.

Whereas previous studies had used either gene expression or copy number data to develop molecular risk-stratification signatures for prostate cancer, this study's authors assessed each approach and then tried something new: they combined gene copy-number and expression data.

“Combining these data provides a more powerful tool to predict outcome following surgery,” lead study author Helen Ross-Adams, PhD, told Cancer Therapy Advisor. Dr. Ross-Adams is at the Cancer Research UK Cambridge Institute at the University of Cambridge, in Cambridge, United Kingdom.

The study “is one of the most comprehensive looks at the integration of copy number and transcriptomic data in a large cohort of patients with localized prostate cancer,” noted Sumanta Kumar Pal, an ASCO Expert and an assistant professor at the Department of Medical Oncology & Therapeutics Research at City of Hope Comprehensive Cancer, in Duarte, CA.

Using healthy blood and tissue samples and prostate tumor tissue samples from discovery and validation cohorts of 125 and 103 men, respectively, the researchers searched tumor-genetics data predictive of biochemical relapse. The study confirmed prostate tumor development associations for MYC amplification, PCA3, and AMACR overexpression, and the loss of MSMB expression, and for six genes already known to be involved in prostate cancer: MAP3K7, MELK, RCBTB2, ELAC2, TPD52, and ZBTB4.

But surprisingly, the combined-data analysis also discovered new prostate cancer associations for 94 other genes—genes whose associations with prostate cancer progression “would not have been detected using either transcript or copy number data alone,” the study authors noted.

The METABRIC study2 used the same integrative method to show that breast cancer represents at least 10 distinct molecular malignancies, Dr. Ross-Adams noted.

Gene expression and copy number alterations for a total of 100 prostate cancer-associated genes allowed differentiation of five separate prognostic subgroups that consistently predicted biochemical relapse.1

The resulting genetic assay “outperforms established clinical predictors of poor prognosis like PSA and Gleason score” and the authors report that they found in the validation cohort that the 100-gene signature showed significant power to separate out a poor prognosis patient group with quicker time to recurrence, from the low-risk cohorts, with much lower chance for recurrence and slower disease progression.

The signature involved different molecular processes (such as nucleic acid processing and protein phosphorylation) than previously published gene signatures that are based on cell-cycle and lipid-metabolism genes, the researchers noted.

They next compared the 100-gene signature's prognostic performance with that of previously published prostate cancer gene signatures and the OncoType Dx Prostate Cancer assay.

“Our 100-gene signature outperformed all other gene sets in identifying patients with early time to biochemical relapse” (P=0.0001), they reported.

“We've developed a tool that will help clinicians identify the most aggressive prostate cancers, with the highest risk of relapsing after surgery,” Dr. Ross-Adams explained.

That suggests their gene signature might indeed emerge “as an effective tool to outperform previously published gene signatures, or the conventional PSA or histology risk markers,” as noted by Joseph Ragaz, MD, FRCP, MRCS-LRCP, senior medical oncologist and clinical professor at the University of British Columbia's School of Population and Public Health, in Vancouver, Canada, and an expert on cancer outcome research including cancer risk stratification.

The new study is “just a first step,” Dr. Ross-Adams is quick to acknowledge. “We're quite a way away from this being offered routinely in clinics. The temptation is to run before we can walk, but these findings need to be confirmed in larger clinical trials.”


RELATED: Prostate Cancer May Impact Relationship Satisfaction Among Couples

Using more genes “certainly may improve the performance of a classifier, but it is also possible that it could add more ‘noise' to the signal,” cautioned Dr. Pal.

But the study was based on biochemical relapse – elevation in blood PSA – rather than documented metastasis, Drs. Pal and Ragaz each warned.

“Other platforms have also shown the ability to predict clinical metastases, arguably a more important endpoint in patients with prostate cancer,” Dr. Pal said.

“Biochemical recurrence, the resurgence of PSA after definitive treatment, can be a long and indolent disease state. Furthermore, if you look carefully at the Kaplan-Meier curves [in the new study], you can see that there is considerable overlap in clinical outcome between the risk groups defined by this algorithm.”

Because her team's patient cohorts had no deaths attributable to prostate cancer and few metastases, however, that was not possible for this study, Dr. Ross-Adams noted.

“Our primary cohort is maturing and our data are publically available, so this could certainly be revisited in the future,” she added.

“We have also made matched tissue microarrays (TMAs) for these patients, with the idea being that we can offer this as a resource for further translational research to address these sorts of questions.”

But her team's results represent a promising advance with potentially important clinical implications, Dr. Ross-Adams said.

“At the moment, treatment decisions for prostate cancer are made on the basis of tumor histology, PSA, and tumor staging, which are relatively ‘blunt' instruments—it's not uncommon for patients to respond differently than expected, based on these parameters.”

Better predicting which tumors will progress aggressively would allow life-saving, early interventions, and Dr. Ross-Adams and her colleagues hope their proposed 100-gene signature may help do just that, as an additional tool used alongside tumor stage, histology, and PSA, to better determine prognosis, identify the most dangerous cases, and to plan treatments accordingly.

If this signature proves itself in larger clinical trials, it will join OncotypeDx and other gene signature tests. But using these tools in guidelines is still a matter of contention, or at least considered premature or “experimental.”

“There is no shortage of molecular predictors for clinical outcome in prostate cancer,” Dr. Pal noted. “The real question is what to do with them. Patients often face a difficult decision between surgery, radiation, or active surveillance when they are diagnosed with prostate cancer. To my knowledge, none of the genomic tools that have been introduced to date definitely help arbitrate this decision.”

Ragaz said that the “real problem is a long-term validation, and also lack of studies testing their predictive worth”—the ability of the test to associate a genetic signature with response to a therapy, be it surgery, chemotherapy, radiation or hormones.


RELATED: A Decade of Dramatic Change in the Treatment of Prostate Cancer

The study used surgically-excised tumor tissue. “The next step is to see whether these findings would be as useful in identifying aggressive cases using tissue from diagnostic biopsies,” Dr. Ross-Adams noted.

One concern is that such biopsies will not prove to be representative of all cancer cells in a tumor, because of tumors' genetic heterogeneity. But template biopsies that use MRI to identify the best target regions, might help, she said.


  1. Ross-Adams H, Lamb AD, Dunning MJ, et al. Integration of copy number and transcriptomics provides risk stratification in prostate cancer: a discovery and validation cohort study. EBioMedicine. 2015. [epub ahead of print]. doi: 10.1016/j.ebiom.2015.07.017.
  2. Curtis C, Shah SP, Chin SF, et al. The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups. Nature. 2012;486:346-352. 

Autoria e outros dados (tags, etc)

por cyto às 12:05

1 comentário

De protein expression a 29.09.2015 às 03:52

Looking forward to the further progress!

Comentar post

Mais sobre mim

foto do autor

Subscrever por e-mail

A subscrição é anónima e gera, no máximo, um e-mail por dia.


Pesquisar no Blog  


Agosto 2015