2021
DOI: 10.1002/jmri.27595
|View full text |Cite
|
Sign up to set email alerts
|

Performance of Deep Learning and Genitourinary Radiologists in Detection of Prostate Cancer Using 3‐T Multiparametric Magnetic Resonance Imaging

Abstract: Background: Several deep learning-based techniques have been developed for prostate cancer (PCa) detection using multi-parametric MRI (mpMRI), but few of them have been rigorously evaluated relative to radiologists' performance or whole-mount histopathology (WMHP).Purpose: To compare the performance of a previously proposed deep learning algorithm, FocalNet, and expert radiologists in the detection of PCa on mpMRI with WMHP as the reference.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
33
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
7
1

Relationship

0
8

Authors

Journals

citations
Cited by 27 publications
(33 citation statements)
references
References 26 publications
0
33
0
Order By: Relevance
“…Three out of 12 DL studies (25%) [ 37 – 39 ] that underwent quality screening using CLAIM failed at least three pre-identified mandatory criteria, with 2/12 [ 40 , 41 ] failing two, and 2/12 [ 42 , 43 ] failing just one criterion. Four of the seven rejected papers (57%) [ 37 – 39 , 43 ] did not describe data processing steps in sufficient detail (Q9), 4/7 [ 38 40 , 42 ] did not explain the exact method of selecting the final model (Q26), and 3/7 [ 38 , 40 , 41 ] failed to provide enough details on training approach (Q25). Following the subsequent full CLAIM assessment of the remaining five papers, we found that none of them reported the following items: selection of data subsets (Q10), robustness or sensitivity analysis (Q30), validation or testing on external data (Q32), and failure analysis of incorrectly classified cases (Q37).…”
Section: Resultsmentioning
confidence: 99%
“…Three out of 12 DL studies (25%) [ 37 – 39 ] that underwent quality screening using CLAIM failed at least three pre-identified mandatory criteria, with 2/12 [ 40 , 41 ] failing two, and 2/12 [ 42 , 43 ] failing just one criterion. Four of the seven rejected papers (57%) [ 37 – 39 , 43 ] did not describe data processing steps in sufficient detail (Q9), 4/7 [ 38 40 , 42 ] did not explain the exact method of selecting the final model (Q26), and 3/7 [ 38 , 40 , 41 ] failed to provide enough details on training approach (Q25). Following the subsequent full CLAIM assessment of the remaining five papers, we found that none of them reported the following items: selection of data subsets (Q10), robustness or sensitivity analysis (Q30), validation or testing on external data (Q32), and failure analysis of incorrectly classified cases (Q37).…”
Section: Resultsmentioning
confidence: 99%
“…Algorithm training and testing were performed retrospectively using cross-validation without an independent test set and application to new cases. In a recent study, Cao et al 42 trained a FocalNet to automatically detect prostate lesions with a confidence score achieving comparable results to experienced radiologists only in a high-sensitivity or high-specificity setting. Most recently, Li et al 40 trained a fully automatic system for lesion detection and classification and demonstrated an improvement in radiologist performance with AI support; however, the 2 readers possessed comparably low sensitivity of 59% and 79%, respectively, before AI review.…”
Section: Discussionmentioning
confidence: 99%
“…target Gleason score, growth patterns, surface area, volume) need to be accurately established to correctly rate a given algorithm's underlying dataset relative to the natural history of the targeted prostate cancer. These ground truth data are difficult to compile, as digital annotations of prostate MRI scans are rarely performed in clinical practice, and histopathologic correlation with whole gland sampling [30] for MRI verification does not happen in the clinical routine. For example, men with negative scans may not undergo biopsy and men with positive scans may only undergo targeted biopsies.…”
Section: Data Science Prerequisitesmentioning
confidence: 99%