Random forest-based modelling to detect biomarkers for prostate cancer progression

Abstract

BackgroundThe clinical course of prostate cancer (PCa) is highly variable, demanding an individualized approach to therapy. Overtreatment of indolent PCa cases, which likely do not progress to aggressive stages, may be associated with severe side effects and considerable costs. These could be avoided by utilizing robust prognostic markers to guide treatment decisions.ResultsWe present a random forest-based classification model to predict aggressive behaviour of prostate cancer. DNA methylation changes between PCa cases with good or poor prognosis (discovery cohort with n = 70) were used as input. DNA was extracted from formalin-fixed tumour tissue, and genome-wide DNA methylation differences between both groups were assessed using Illumina HumanMethylation450 arrays. For the random forest-based modelling, the discovery cohort was randomly split into a training (80%) and a test set (20%). Our methylation-based classifier demonstrated excellent performance in discriminating prognosis subgroups in the test set (Kaplan-Meier survival analyses with log-rank p value < 0.0001). The area under the receiver operating characteristic curve (AUC) for the sensitivity analysis was 95%. Using the ICGC cohort of early- and late-onset prostate cancer (n = 222) and the TCGA PRAD cohort (n = 477) for external validation, AUCs for sensitivity analyses were 77.1% and 68.7%, respectively. Cancer progression-related DNA hypomethylation was frequently located in ‘partially methylated domains’ (PMDs)—large-scale genomic areas with progressive loss of DNA methylation linked to mitotic cell division. We selected several candidate genes with differential methylation in gene promoter regions for additional validation at the protein expression level by immunohistochemistry in > 12,000 tissue micro-arrayed PCa cases. Loss of ZIC2 protein expression was associated with poor prognosis and correlated with significantly shorter time to biochemical recurrence. The prognostic value of ZIC2 proved to be independent from established clinicopathological variables including Gleason grade, tumour stage, nodal stage and prostate-specific-antigen.ConclusionsOur results highlight the prognostic relevance of methylation loss in PMD regions, as well as of several candidate genes not previously associated with PCa progression. Our robust and externally validated PCa classification model either directly or via protein expression analyses of the identified top-ranked candidate genes will support the clinical management of prostate cancer.

Advances in Prognostic Methylation Biomarkers for Prostate Cancer

“…The resulting aggressiveness classifier consisted of 25 CpG sites (including in NXPH2 , TRIB1 and PCDHA1-PCDHA8 ), and was successfully validated in the TCGA cohort ( n = 351) through accurate prediction of lymph node metastases and invasive pathological stage T3 tumours [ 136 ]. Finally, the most recent HM450K study which aligns with our criteria of prognostic discovery, used random-forest-based modelling to identify markers that could differentiate between good prognosis, defined as organ-confined disease (pT2) and no BCR for at least 5 years ( n = 35), and poor prognosis, defined as systemic metastatic disease with recurrence within 3 years ( n = 35) [ 137 ]. A DNA methylation-based classifier consisting of 598 sites was developed, validating in two independent cohorts of patients with publicly available methylation data, based on the same selection criteria (ICGC cohort n = 63, TCGA cohort: n = 84) [ 137 ].…”

“…Finally, the most recent HM450K study which aligns with our criteria of prognostic discovery, used random-forest-based modelling to identify markers that could differentiate between good prognosis, defined as organ-confined disease (pT2) and no BCR for at least 5 years ( n = 35), and poor prognosis, defined as systemic metastatic disease with recurrence within 3 years ( n = 35) [ 137 ]. A DNA methylation-based classifier consisting of 598 sites was developed, validating in two independent cohorts of patients with publicly available methylation data, based on the same selection criteria (ICGC cohort n = 63, TCGA cohort: n = 84) [ 137 ]. Further analyses highlighted the independent prognostic value of a gene overlapping one of the 598 sites, with immunostaining analysis reporting significant association between loss of ZIC2 protein expression and poorer prognosis (adjusted for GS, pathological T- stage, nodal stage and PSA) [ 137 ].…”

“…A DNA methylation-based classifier consisting of 598 sites was developed, validating in two independent cohorts of patients with publicly available methylation data, based on the same selection criteria (ICGC cohort n = 63, TCGA cohort: n = 84) [ 137 ]. Further analyses highlighted the independent prognostic value of a gene overlapping one of the 598 sites, with immunostaining analysis reporting significant association between loss of ZIC2 protein expression and poorer prognosis (adjusted for GS, pathological T- stage, nodal stage and PSA) [ 137 ].…”

Advances in Prognostic Methylation Biomarkers for Prostate Cancer

“…The resulting aggressiveness classifier consisted of 25 CpG sites (including in NXPH2, TRIB1 and PCDHA1-PCDHA8), and was successfully validated in the TCGA cohort (n = 351) through accurate prediction of lymph node metastases and invasive pathological stage T3 tumours [136]. Finally, the most recent HM450K study which aligns with our criteria of prognostic discovery, used random-forest based modelling to identify markers that could differentiate between good prognosis, defined as organconfined disease (pT2) and no BCR for at least 5 years (n = 35), and poor prognosis, defined as systemic metastatic disease with recurrence within 3 years (n = 35) [137]. A DNA methylation-based classifier consisting of 598 sites was developed, validating in two independent cohorts of patients with publicly available methylation data, based on the same selection criteria (ICGC cohort n = 63, TCGA cohort: n = 84) [137].…”

Characterization of Dysregulated lncRNA-Associated ceRNA Network Reveals Novel lncRNAs With ceRNA Activity as Epigenetic Diagnostic Biomarkers for Osteoporosis Risk

“…To identify potential lncRNA biomarkers with ceRNA activity, a random forest approach and leave-one-out cross-validation (LOOCV) were used to select optimal lncRNAs biomarkers using the R package "randomForest" and out-of-bag (OOB) error, which measure the performance of the model on the training set (Lv et al, 2019;Tan et al, 2019). The OOB error will produce an unbiased estimate for the classification error, while the bagging method will decrease the chance of overfitting (Toth et al, 2019). Then, a random forest-based classifier was built using the optimal lncRNA biomarkers, and a receiver operating characteristic (ROC) curve and the area under ROC curve (AUC) was used to measure the diagnostic performance of the lncRNA classifier (Lai et al, 2019).…”