Open AccessEvaluation of an Automated Analysis Tool for Prostate Cancer Prediction Using Multiparametric Magnetic Resonance Imaging
Author(s) -
Matthias Roethke,
Timur H. Kuru,
Maya Mueller-Wolf,
Erik Agterhuis,
Christopher Edler,
Markus Hohenfellner,
Heinz Peter Schlemmer,
Boris Hadaschik
Publication year - 2016
Publication title -
plos one
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.99
H-Index - 332
ISSN - 1932-6203
DOI - 10.1371/journal.pone.0159803
Subject(s) - prostate cancer , magnetic resonance imaging , medicine , malignancy , prostate , radiology , multiparametric mri , cancer , pathology
Objective To evaluate the diagnostic performance of an automated analysis tool for the assessment of prostate cancer based on multiparametric magnetic resonance imaging (mpMRI) of the prostate. Methods A fully automated analysis tool was used for a retrospective analysis of mpMRI sets (T2-weighted, T1-weighted dynamic contrast-enhanced, and diffusion-weighted sequences). The software provided a malignancy prediction value for each image pixel, defined as Malignancy Attention Index (MAI) that can be depicted as a colour map overlay on the original images. The malignancy maps were compared to histopathology derived from a combination of MRI-targeted and systematic transperineal MRI/TRUS-fusion biopsies. Results In total, mpMRI data of 45 patients were evaluated. With a sensitivity of 85.7% (with 95% CI of 65.4–95.0), a specificity of 87.5% (with 95% CI of 69.0–95.7) and a diagnostic accuracy of 86.7% (with 95% CI of 73.8–93.8) for detection of prostate cancer, the automated analysis results corresponded well with the reported diagnostic accuracies by human readers based on the PI-RADS system in the current literature. Conclusion The study revealed comparable diagnostic accuracies for the detection of prostate cancer of a user-independent MAI-based automated analysis tool and PI-RADS-scoring-based human reader analysis of mpMRI. Thus, the analysis tool could serve as a detection support system for less experienced readers. The results of the study also suggest the potential of MAI-based analysis for advanced lesion assessments, such as cancer extent and staging prediction.