Keywords:
Tissue characterisation, Neoplasia, Cancer, Segmentation, Computer Applications-Detection, diagnosis, Computer Applications-3D, MR-Diffusion/Perfusion, MR, Image manipulation / Reconstruction, Pelvis, Genital / Reproductive system male, Computer applications
Authors:
D. grimaldi1, S. Monti2, C. Cavaliere2, V. Brancato1, L. Palumbo1, M. Aiello2, G. Di Costanzo1, M. Salvatore1, A. Ragozzino1; 1Napoli/IT, 2Naples/IT
DOI:
10.26044/ecr2019/C-3388
Aims and objectives
Prostate cancer (PCa) is one of the most common malignant neoplasms and is the second leading cause of cancer-related death among older males in Western countries (1).
The early detection and grading of the PCa are fundamental in patient management and in the evaluation of long-term survival.
In 2015,
the Prostate Imaging and Reporting and Data System version 2 (PI-RADS v2) was described to promote standardized reporting criteria for the interpretation of mp-MRI scans in the assessment of PCa,
including clinical indications for prostate mp-MRI,
minimal and optimal imaging acquisition protocols,
and a structured category assessment system (2)(3)(14).
Encouraging results have been reported in the literature on the role of PI-RADS v2 in the diagnosis and characterization of PCa (4)(5)(6)(7).
However,
the variability of the inter-operator evaluation remains significant,
even in PI-RADS v2 (8)(9).
Radiomics,
the extraction of multiple quantitative imaging features from medical images,
has attracted much attention(10)(11).
With automatic feature extraction algorithms,
imaging data can be converted to high- dimensional mineable data and provide valuable information for assessing the diagnosis and prognosis of various disorders (12 (13)(16).
Aim of this study is to investigate the usefulness of radiomics to disentangle prostate pathology,
and mainly alterations scored as doubt (PIRADS 3),
using bioptic findings as gold standard.