Prostate Cancer: Improving the Flow of Research

Data analysis algorithm for the development of extracellular miRNA-based diagnostic systems for prostate cancer

Urine of prostate cancer (PCa) carries miRNAs originated from prostate cancer cells as a part of both nucleoprotein complexes and cell-secreted extracellular vesicles. The analysis of such miRNA-markers in urine can be a convenient option for PCa screening. The aims of this study were to reveal miRNA–markers of PCa in urine and design a robust and precise diagnostic test, based on miRNA expression analysis. The expression analysis of the 84 miRNAs in paired urine extracellular vesicles (EVs) and cell free urine supernatant samples from healthy donors, patients with benign and malignant prostate tumours was done using miRCURY LNA miRNA qPCR Panels (Exiqon, Denmark). Sets of miRNAs differentially expressed between the donor groups were found in urine EVs and urine supernatant. Diagnostically significant miRNAs were selected and algorithm of data analysis, based on expression data on 24-miRNA in urine and obtained using 17 analytical systems, was designed. The developed algorithm of data analysis describes a series of steps necessary to define cut-off values and sequentially analyze miRNA expression data according to the cut-offs to facilitate classification of subjects in case/control groups and allows to detect PCa patients with 97.5% accuracy.

Prostate cancer detection from multi-institution multiparametric MRIs using deep convolutional neural networks

Multiparametric magnetic resonance imaging (mpMRI) of the prostate aids in early diagnosis of prostate cancer, but is difficult to interpret and subject to interreader variability. Our objective is to generate probability maps, overlaid on original mpMRI images to help radiologists identify where a cancer is suspected as a computer-aided diagnostic (CAD). We optimized the holistically nested edge detection (HED) deep convolutional neural network. Our dataset contains T2, apparent diffusion coefficient, and high b -value images from 186 patients across six institutions worldwide: 92 with an endorectal coil (ERC) and 94 without. Ground-truth was based on tumor segmentations manually drawn by expert radiologists based on histologic evidence of cancer. The training set consisted of 120 patients and the validation set and test set included 19 and 47, respectively. Slice-level probability maps are evaluated at the lesion level of analysis. The best model: HED using 5 × 5 convolutional kernels, batch normalization, and optimized using Adam. This CAD performed significantly better ( p < 0.001 ) in the peripheral zone ( AUC = 0.94 ± 0.01 ) than the transition zone. It outperforms a previous CAD from our group in a head-to-head comparison on the same ERC-only test cases ( AUC = 0.97 ± 0.01 ; p < 0.001 ). Our CAD establishes a state-of-the-art performance for predicting prostate cancer lesions on mpMRIs.