Survival among patients with castration-resistant prostate cancer is highly variable, and there is a scarcity of prognostic markers that could be used to stratify patients in clinical trials, improve treatment, and accelerate drug development. Two recent studies reported in Lancet Oncology have shown that whole-blood RNA profiling can sort patients with prostate cancer into distinct risk groups.
In one study, Ross and colleagues from Dana-Farber Cancer Institute, Harvard Medical School, and GeneNews in Boston, Statistical Innovations in Belmont, Massachusetts, and Mount Sinai School of Medicine, Memorial-Sloan Kettering Cancer Center (MSKCC), and Weill Cornell Medical College in New York, investigated a whole-blood RNA transcript-based model as a prognostic biomarker in castration-resistant prostate cancer. Peripheral blood was prospectively collected from 62 men with castration-resistant prostate cancer on various treatment regimens at the Dana-Farber Cancer Institute (August 2006–June 2008, training set) and from 140 patients with castration-resistant prostate cancer at MSKCC (August 2006–February 2009, validation set).
Both of the cohorts were heterogeneous with respect to patient characteristics and treatment exposures. A panel of 168 inflammation-related and prostate cancer–related genes was assessed with optimized quantitative reverse transcriptase polymerase chain reaction testing to determine potential associations with survival.
A six-gene model (ABL2, SEMA4D, ITGAL, C1QA, TIMP1, and CDKN1A) distinguished a low-risk group of patients with median survival of more than 34.9 months (median survival not reached) and a high-risk group with a median survival of 7.8 months (P < .0001). The prognostic utility of the model was validated in the independent cohort, although the difference in survival was less dramatic than in the training cohort (9.2 vs 18.5 months [hazard ratio = 2.3, 95% confidence interval = 1.39–3.70]), potentially reflecting model overfitting. In the validation cohort, the model had greater prognostic power than a five-variable clinicopathologic model (P = .0067).
In another study, Olmos and colleagues from Royal Marsden NHS Foundation Trust and Institute of Cancer Research in Sutton and London, Spanish National Cancer Research Centre in Madrid, Sidney Kimmel Center for Prostate and Urologic Cancers, MSKCC, and Weill Cornell College of Medicine in New York, Beatson West of Scotland Cancer Centre and University of Glasgow in Glasgow, and Johns Hopkins University in Baltimore, used whole-blood mRNA expression arrays to identify patients with metastatic castration-resistant prostate cancer with poorer outcome.
In a test set cohort consisting of 64 patients with metastatic castration-resistant disease and 30 patients on watchful waiting from Royal Marsden Hospital NHS Foundation Trust and Beatson West of Scotland Cancer Centre, profiling sorted patients into four molecular subgroups, with one of these groups containing only patients with metastatic castration-resistant disease. Patients with castration-resistant prostate cancer in this subgroup had significantly poorer overall survival compared with patients with castration-resistant disease in other subgroups (10.7 vs 25.6 months, P < .0001).
A nine-gene signature (TERF2IP, TMCC2, GABARAPL2, SNCA, RIOK3, TFDP1, SLC4A1, HMBS, and STOM) classified patients into this subgroup with a very low percentage of misclassification (1.2%). In the validation set cohort consisting of 70 patents with progressive metastatic castration-resistant disease from MSKCC, the nine-gene signature was significantly associated with worse overall survival (9.2 vs 21.6 months, P = .001). The signature was an independent prognostic factor on multivariable analysis in both the test and validation cohorts.
These findings with two distinct gene signatures suggest that whole-blood gene profiling can identify higher- and lower-risk patients with castration-resistant prostate cancer. As stated by Dr. Karina Sørensen in an editorial accompanying the two studies, “The biological relevance of these prognostic signatures, which are the first of their kind, is largely unknown. Whole-blood expression profiles presumably reflect changes in peripheral blood mononuclear cells, because [circulating tumor cells] are extremely rare. The [nine-gene] signature correlated with increased immature erythroid cells and decreased lymphoid immune response, while the six-gene signature associated with impaired immunity of the myeloid lineage. The signatures represent molecular snapshots of distinct host immune functions, relevant for patient survival. However, further investigation into the underlying biological mechanisms is needed … [and] future studies should clarify if any of the 15 signature genes are functionally relevant, or are surrogate markers for other processes.”
Ross RW, et al: Lancet Oncol. October 8, 2012 (early release online).
Olmos D, et al: Lancet Oncol. October 8, 2012 (early release online).
Sørensen KD: Lancet Oncol. October 8, 2012 (early release online).