Abstract
Purpose:
Diagnostic prostate biopsies are subject to sampling bias. We hypothesize that quantitative imaging with multiparametric (MP)‐MRI can more accurately direct targeted biopsies to index lesions associated with highest risk clinical and genomic features.
Methods:
Regionally distinct prostate habitats were delineated on MP‐MRI (T2‐weighted, perfusion and diffusion imaging). Directed biopsies were performed on 17 habitats from 6 patients using MRI‐ultrasound fusion. Biopsy location was characterized with 52 radiographic features. Transcriptome‐wide analysis of 1.4 million RNA probes was performed on RNA from each habitat. Genomics features with insignificant expression values (<0.25) and interquartile range <0.5 were filtered, leaving total of 212 genes. Correlation between imaging features, genes and a 22 feature genomic classifier (GC), developed as a prognostic assay for metastasis after radical prostatectomy was investigated.
Results:
High quality genomic data was derived from 17 (100%) biopsies. Using the 212 ‘unbiased’ genes, the samples clustered by patient origin in unsupervised analysis. When only prostate cancer related genomic features were used, hierarchical clustering revealed samples clustered by needle‐biopsy Gleason score (GS). Similarly, principal component analysis of the imaging features, found the primary source of variance segregated the samples into high (≥7) and low (6) GS. Pearson's correlation analysis of genes with significant expression showed two main patterns of gene expression clustering prostate peripheral and transitional zone MRI features. Two‐way hierarchical clustering of GC with radiomics features resulted in the expected groupings of high and low expressed genes in this metastasis signature.
Conclusions:
MP‐MRI‐targeted diagnostic biopsies can potentially improve risk stratification by directing pathological and genomic analysis to clinically significant index lesions. As determinant lesions are more reliably identified, targeting with radiotherapy should improve outcome. This is the first demonstration of a link between quantitative imaging features (radiomics) with genomic features in MRI‐directed prostate biopsies.
The research was supported by NIH‐ NCI R01 CA 189295 and R01 CA 189295; E Davicioni is partial owner of GenomeDx Biosciences, Inc. M Takhar, N Erho, L Lam, C Buerki and E Davicioni are current employees at GenomeDx Biosciences, Inc.
