A comparative study of survival models for breast cancer prognostication revisited: the benefits of multi-gene models.

BACKGROUND: The development of clinical -omic biomarkers for predicting patient prognosis has mostly focused on multi-gene models. However, several studies have described significant weaknesses of multi-gene biomarkers. Indeed, some high-profile reports have even indicated that multi-gene biomarkers fail to consistently outperform simple single-gene ones. Given the continual improvements in -omics technologies and the availability of larger, better-powered datasets, we revisited this "single-gene hypothesis" using new techniques and datasets. RESULTS: By deeply sampling the population of available gene sets, we compare the intrinsic properties of single-gene biomarkers to multi-gene biomarkers in twelve different partitions of a large breast cancer meta-dataset. We show that simple multi-gene models consistently outperformed single-gene biomarkers in all twelve partitions. We found 270 multi-gene biomarkers (one per ~11,111 sampled) that always made better predictions than the best single-gene model. CONCLUSIONS: The single-gene hypothesis for breast cancer does not appear to retain its validity in the face of improved statistical models, lower-noise genomic technology and better-powered patient cohorts. These results highlight that it is critical to revisit older hypotheses in the light of newer techniques and datasets.

Valection: design optimization for validation and verification studies.

BACKGROUND: Platform-specific error profiles necessitate confirmatory studies where predictions made on data generated using one technology are additionally verified by processing the same samples on an orthogonal technology. However, verifying all predictions can be costly and redundant, and testing a subset of findings is often used to estimate the true error profile. RESULTS: To determine how to create subsets of predictions for validation that maximize accuracy of global error profile inference, we developed Valection, a software program that implements multiple strategies for the selection of verification candidates. We evaluated these selection strategies on one simulated and two experimental datasets. CONCLUSIONS: Valection is implemented in multiple programming languages, available at: http://labs.oicr.on.ca/boutros-lab/software/valection.

Germline contamination and leakage in whole genome somatic single nucleotide variant detection.

BACKGROUND: The clinical sequencing of cancer genomes to personalize therapy is becoming routine across the world. However, concerns over patient re-identification from these data lead to questions about how tightly access should be controlled. It is not thought to be possible to re-identify patients from somatic variant data. However, somatic variant detection pipelines can mistakenly identify germline variants as somatic ones, a process called "germline leakage". The rate of germline leakage across different somatic variant detection pipelines is not well-understood, and it is uncertain whether or not somatic variant calls should be considered re-identifiable. To fill this gap, we quantified germline leakage across 259 sets of whole-genome somatic single nucleotide variant (SNVs) predictions made by 21 teams as part of the ICGC-TCGA DREAM Somatic Mutation Calling Challenge. RESULTS: The median somatic SNV prediction set contained 4325 somatic SNVs and leaked one germline polymorphism. The level of germline leakage was inversely correlated with somatic SNV prediction accuracy and positively correlated with the amount of infiltrating normal cells. The specific germline variants leaked differed by tumour and algorithm. To aid in quantitation and correction of leakage, we created a tool, called GermlineFilter, for use in public-facing somatic SNV databases. CONCLUSIONS: The potential for patient re-identification from leaked germline variants in somatic SNV predictions has led to divergent open data access policies, based on different assessments of the risks. Indeed, a single, well-publicized re-identification event could reshape public perceptions of the values of genomic data sharing. We find that modern somatic SNV prediction pipelines have low germline-leakage rates, which can be further reduced, especially for cloud-sharing, using pre-filtering software.

NanoStringNormCNV: pre-processing of NanoString CNV data.

Summary: The NanoString System is a well-established technology for measuring RNA and DNA abundance. Although it can estimate copy number variation, relatively few tools support analysis of these data. To address this gap, we created NanoStringNormCNV, an R package for pre-processing and copy number variant calling from NanoString data. This package implements algorithms for pre-processing, quality-control, normalization and copy number variation detection. A series of reporting and data visualization methods support exploratory analyses. To demonstrate its utility, we apply it to a new dataset of 96 genes profiled on 41 prostate tumour and 24 matched normal samples. Availability and implementation: NanoStringNormCNV is implemented in R and is freely available at http://labs.oicr.on.ca/boutros-lab/software/nanostringnormcnv. Contact: paul.boutros@oicr.on.ca. Supplementary information: Supplementary data are available at Bioinformatics online.

Germline Mutations in the Kallikrein 6 Region and Predisposition for Aggressive Prostate Cancer.

Background: There is a need for markers that can specifically identify individuals at increased risk of harboring aggressive forms of prostate cancer (PCa). Methods: We surveyed the Kallikrein ( KLK ) region ( KLK 1-15) for single-nucleotide polymorphisms (SNPs) associated with aggressive PCa (Gleason Score ≥ 8) in 1858 PCa patients. Discovery cohorts (Swiss arm of the European Randomized Study of Screening for PCa, n = 379; Toronto, Canada, n = 540) and a validation cohort (Prostate, Lung, Colorectal and Ovarian [PLCO] screening trial, n = 939) were analyzed. Fine-mapping within the KLK region was carried out by genotyping and imputation in the discovery cohort, whereas PLCO data were provided through database of Genotypes and Phenotypes ( dbGaP ). The influence of SNPs of interest on biochemical-free survival was evaluated in a cohort of localized PCa from the International Cancer Genome Consortium (ICGC; n = 130) analyzed with next-generation sequencing. Single- and multi-SNP association studies, as well as haplotype analyses, were performed. All statistical tests were two-sided. Results: Several SNPs in very strong linkage disequilibrium in the KLK 6 region and located within the same haplotype (rs113640578, rs79324425, rs11666929, rs28384475, rs3810287), identified individuals at increased risk of aggressive PCa in both discovery (odds ratio [OR] = 3.51-3.64, 95% confidence interval [CI] = 2.01 to 6.36, P = 1.0x10 -5 -8.4x10 -6 ) and validation (OR = 1.89-1.96, 95% CI = 0.99 to 3.71, P = .04-.05) cohorts. The overall test of haplotype association was highly statistically significant in each cohort ( P = 3.5x10 -4 and .006, respectively) and in the three data sets combined ( P = 2.3x10 -5 ). These germline SNPs independently predicted relapse in the ICGC cohort (hazard ratio = 3.15, 95% CI = 1.57 to 6.34, P = .001). Conclusions: Our fine-mapping study has identified novel loci in the KLK 6 region strongly associated with aggressive PCa.

Modulation of long noncoding RNAs by risk SNPs underlying genetic predispositions to prostate cancer.

Long noncoding RNAs (lncRNAs) represent an attractive class of candidates to mediate cancer risk. Through integrative analysis of the lncRNA transcriptome with genomic data and SNP data from prostate cancer genome-wide association studies (GWAS), we identified 45 candidate lncRNAs associated with risk to prostate cancer. We further evaluated the mechanism underlying the top hit, PCAT1, and found that a risk-associated variant at rs7463708 increases binding of ONECUT2, a novel androgen receptor (AR)-interacting transcription factor, at a distal enhancer that loops to the PCAT1 promoter, resulting in upregulation of PCAT1 upon prolonged androgen treatment. In addition, PCAT1 interacts with AR and LSD1 and is required for their recruitment to the enhancers of GNMT and DHCR24, two androgen late-response genes implicated in prostate cancer development and progression. PCAT1 promotes prostate cancer cell proliferation and tumor growth in vitro and in vivo. These findings suggest that modulating lncRNA expression is an important mechanism for risk-associated SNPs in promoting prostate transformation.

Tumour genomic and microenvironmental heterogeneity for integrated prediction of 5-year biochemical recurrence of prostate cancer: a retrospective cohort study.

BACKGROUND: Clinical prognostic groupings for localised prostate cancers are imprecise, with 30-50% of patients recurring after image-guided radiotherapy or radical prostatectomy. We aimed to test combined genomic and microenvironmental indices in prostate cancer to improve risk stratification and complement clinical prognostic factors. METHODS: We used DNA-based indices alone or in combination with intra-prostatic hypoxia measurements to develop four prognostic indices in 126 low-risk to intermediate-risk patients (Toronto cohort) who will receive image-guided radiotherapy. We validated these indices in two independent cohorts of 154 (Memorial Sloan Kettering Cancer Center cohort [MSKCC] cohort) and 117 (Cambridge cohort) radical prostatectomy specimens from low-risk to high-risk patients. We applied unsupervised and supervised machine learning techniques to the copy-number profiles of 126 pre-image-guided radiotherapy diagnostic biopsies to develop prognostic signatures. Our primary endpoint was the development of a set of prognostic measures capable of stratifying patients for risk of biochemical relapse 5 years after primary treatment. FINDINGS: Biochemical relapse was associated with indices of tumour hypoxia, genomic instability, and genomic subtypes based on multivariate analyses. We identified four genomic subtypes for prostate cancer, which had different 5-year biochemical relapse-free survival. Genomic instability is prognostic for relapse in both image-guided radiotherapy (multivariate analysis hazard ratio [HR] 4·5 [95% CI 2·1-9·8]; p=0·00013; area under the receiver operator curve [AUC] 0·70 [95% CI 0·65-0·76]) and radical prostatectomy (4·0 [1·6-9·7]; p=0·0024; AUC 0·57 [0·52-0·61]) patients with prostate cancer, and its effect is magnified by intratumoral hypoxia (3·8 [1·2-12]; p=0·019; AUC 0·67 [0·61-0·73]). A novel 100-loci DNA signature accurately classified treatment outcome in the MSKCC low-risk to intermediate-risk cohort (multivariate analysis HR 6·1 [95% CI 2·0-19]; p=0·0015; AUC 0·74 [95% CI 0·65-0·83]). In the independent MSKCC and Cambridge cohorts, this signature identified low-risk to high-risk patients who were most likely to fail treatment within 18 months (combined cohorts multivariate analysis HR 2·9 [95% CI 1·4-6·0]; p=0·0039; AUC 0·68 [95% CI 0·63-0·73]), and was better at predicting biochemical relapse than 23 previously published RNA signatures. INTERPRETATION: This is the first study of cancer outcome to integrate DNA-based and microenvironment-based failure indices to predict patient outcome. Patients exhibiting these aggressive features after biopsy should be entered into treatment intensification trials. FUNDING: Movember Foundation, Prostate Cancer Canada, Ontario Institute for Cancer Research, Canadian Institute for Health Research, NIHR Cambridge Biomedical Research Centre, The University of Cambridge, Cancer Research UK, Cambridge Cancer Charity, Prostate Cancer UK, Hutchison Whampoa Limited, Terry Fox Research Institute, Princess Margaret Cancer Centre Foundation, PMH-Radiation Medicine Program Academic Enrichment Fund, Motorcycle Ride for Dad (Durham), Canadian Cancer Society.

NBN gain is predictive for adverse outcome following image-guided radiotherapy for localized prostate cancer.

Despite the use of clinical prognostic factors (PSA, T-category and Gleason score), 20-60% of localized prostate cancers (PCa) fail primary local treatment. Herein, we determined the prognostic importance of main sensors of the DNA damage response (DDR): MRE11A, RAD50, NBN, ATM, ATR and PRKDC. We studied copy number alterations in DDR genes in localized PCa treated with image-guided radiotherapy (IGRT; n=139) versus radical prostatectomy (RadP; n=154). In both cohorts, NBN gains were the most frequent genomic alteration (14.4 and 11% of cases, respectively), and were associated with overall tumour genomic instability (p<0.0001). NBN gains were the only significant predictor of 5yrs biochemical relapse-free rate (bRFR) following IGRT (46% versus 77%; p=0.00067). On multivariate analysis, NBN gain remained a significant independent predictor of bRFR after adjusting for known clinical prognostic variables (HR=3.28, 95% CI 1.56-6.89, Wald p-value=0.0017). No DDR-sensing gene was prognostic in the RadP cohort. In vitro studies correlated NBN gene overexpression with PCa cells radioresistance. In conclusion, NBN gain predicts for decreased bRFR in IGRT, but not in RadP patients. If validated independently, Nibrin gains may be the first PCa predictive biomarker to facilitate local treatment decisions using precision medicine approaches with surgery or radiotherapy.

Identifying molecular features that distinguish fluvastatin-sensitive breast tumor cells.

Statins, routinely used to treat hypercholesterolemia, selectively induce apoptosis in some tumor cells by inhibiting the mevalonate pathway. Recent clinical studies suggest that a subset of breast tumors is particularly susceptible to lipophilic statins, such as fluvastatin. To quickly advance statins as effective anticancer agents for breast cancer treatment, it is critical to identify the molecular features defining this sensitive subset. We have therefore characterized fluvastatin sensitivity by MTT assay in a panel of 19 breast cell lines that reflect the molecular diversity of breast cancer, and have evaluated the association of sensitivity with several clinicopathological and molecular features. A wide range of fluvastatin sensitivity was observed across breast tumor cell lines, with fluvastatin triggering cell death in a subset of sensitive cell lines. Fluvastatin sensitivity was associated with an estrogen receptor alpha (ERα)-negative, basal-like tumor subtype, features that can be scored with routine and/or strong preclinical diagnostics. To ascertain additional candidate sensitivity-associated molecular features, we mined publicly available gene expression datasets, identifying genes encoding regulators of mevalonate production, non-sterol lipid homeostasis, and global cellular metabolism, including the oncogene MYC. Further exploration of this data allowed us to generate a 10-gene mRNA abundance signature predictive of fluvastatin sensitivity, which showed preliminary validation in an independent set of breast tumor cell lines. Here, we have therefore identified several candidate predictors of sensitivity to fluvastatin treatment in breast cancer, which warrant further preclinical and clinical evaluation.

Epigenetic markers of prostate cancer in plasma circulating DNA.

Epigenetic differences are a common feature of many diseases, including cancer, and disease-associated changes have even been detected in bodily fluids. DNA modification studies in circulating DNA (cirDNA) may lead to the development of specific non-invasive biomarkers. To test this hypothesis, we investigated cirDNA modifications in prostate cancer patients with locally confined disease (n = 19), in patients with benign prostate hyperplasias (n = 20) and in men without any known prostate disease (n = 20). This initial discovery screen identified 39 disease-associated changes in cirDNA modification, and seven of these were validated using the sodium bisulfite-based mapping of modified cytosines in both the discovery cohort and an independent 38-patient validation cohort. In particular, we showed that the DNA modification of regions adjacent to the gene encoding ring finger protein 219 distinguished prostate cancer from benign hyperplasias with good sensitivity (61%) and specificity (71%). We also showed that repetitive sequences detected in this study were meaningful, as they indicated a highly statistically significant loss of DNA at the pericentromeric region of chromosome 10 in prostate cancer patients (p = 1.8 × 10(-6)). Based on these strong univariate results, we applied machine-learning techniques to develop a multi-locus biomarker that correctly distinguished prostate cancer samples from unaffected controls with 72% accuracy. Lastly, we used systems biology techniques to integrate our data with publicly available DNA modification and transcriptomic data from primary prostate tumors, thereby prioritizing genes for further studies. These data suggest that cirDNA epigenomics are promising source for non-invasive biomarkers.