| Summary: | A critical problem in the clinical management of prostate cancer is that it shows high intra- and inter-tumoural heterogeneity. As a result, accurate prediction of individual cancer behaviour is not achievable at the time of diagnosis, leading to substantial overtreatment. It remains an enigma that, in contrast to other cancers, no molecular biomarkers which define robust subtypes of prostate cancer with distinct clinical outcomes have been discovered. In the first part of this study, using data from exon microarrays, we developed a novel method that can identify transcriptional alterations within genes. The alterations might be the result of chromosomal rearrangements, such as translocations, and deletions, or of other abnormalities, such as read-through transcription and alternative transcriptional initiation sites. Using data from two independent datasets we identify several candidate alterations that are constantly correlated with the biochemical failure or that are linked to the development of metastasis. In the second part of the study we illustrate the application of an unsupervised Bayesian procedure, which identifies a subtype of the disease in five prostate cancer transcriptome datasets. Cancers assigned to this subtype (designated DESNT cancers) are characterized by low expression of a core set of 45 genes. For the four datasets with linked PSA failure data following prostatectomy, patients with DESNT cancer exhibited poor outcome relative to other patients (p = 2.65 ・ 10−5, p = 4.28 ・ 10−5, p = 2.98 ・ 10−8 and p = 1.22 ・ 10−3). The DESNT cancers are not linked with the presence of any particular class of genetic mutation, including ETS gene status. However, the methylation analysis reveals a possible role of epigenetic changes in the generation of the DESNT subtype. Our results demonstrate the existence of a novel poor prognosis category of human prostate cancer and will assist in the targeting of therapy, helping avoid treatment-associated morbidity in men with indolent disease.
|
|---|
