A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRI
Computer-aided detection and diagnosis (CAD) systems have the potential to improve robustness and efficiency compared to traditional radiological reading of magnetic resonance imaging (MRI). Fully automated segmentation of the prostate is a crucial step of CAD for prostate cancer, but visual inspect...
Main Authors: | , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-09-01
|
Series: | Diagnostics |
Subjects: | |
Online Access: | https://www.mdpi.com/2075-4418/10/9/714 |
id |
doaj-0cf7d7df410e405b92bf59f478f3b2e4 |
---|---|
record_format |
Article |
spelling |
doaj-0cf7d7df410e405b92bf59f478f3b2e42020-11-25T03:02:41ZengMDPI AGDiagnostics2075-44182020-09-011071471410.3390/diagnostics10090714A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRIMohammed R. S. Sunoqrot0Kirsten M. Selnæs1Elise Sandsmark2Gabriel A. Nketiah3Olmo Zavala-Romero4Radka Stoyanova5Tone F. Bathen6Mattijs Elschot7Department of Circulation and Medical Imaging, NTNU—Norwegian University of Science and Technology, 7030 Trondheim, NorwayDepartment of Circulation and Medical Imaging, NTNU—Norwegian University of Science and Technology, 7030 Trondheim, NorwayDepartment of Radiology and Nuclear Medicine, St. Olavs Hospital, Trondheim University Hospital, 7030 Trondheim, NorwayDepartment of Circulation and Medical Imaging, NTNU—Norwegian University of Science and Technology, 7030 Trondheim, NorwayDepartment of Radiation Oncology, University of Miami Miller School of Medicine, Miami, FL 33136, USADepartment of Radiation Oncology, University of Miami Miller School of Medicine, Miami, FL 33136, USADepartment of Circulation and Medical Imaging, NTNU—Norwegian University of Science and Technology, 7030 Trondheim, NorwayDepartment of Circulation and Medical Imaging, NTNU—Norwegian University of Science and Technology, 7030 Trondheim, NorwayComputer-aided detection and diagnosis (CAD) systems have the potential to improve robustness and efficiency compared to traditional radiological reading of magnetic resonance imaging (MRI). Fully automated segmentation of the prostate is a crucial step of CAD for prostate cancer, but visual inspection is still required to detect poorly segmented cases. The aim of this work was therefore to establish a fully automated quality control (QC) system for prostate segmentation based on T2-weighted MRI. Four different deep learning-based segmentation methods were used to segment the prostate for 585 patients. First order, shape and textural radiomics features were extracted from the segmented prostate masks. A reference quality score (QS) was calculated for each automated segmentation in comparison to a manual segmentation. A least absolute shrinkage and selection operator (LASSO) was trained and optimized on a randomly assigned training dataset (N = 1756, 439 cases from each segmentation method) to build a generalizable linear regression model based on the radiomics features that best estimated the reference QS. Subsequently, the model was used to estimate the QSs for an independent testing dataset (N = 584, 146 cases from each segmentation method). The mean ± standard deviation absolute error between the estimated and reference QSs was 5.47 ± 6.33 on a scale from 0 to 100. In addition, we found a strong correlation between the estimated and reference QSs (rho = 0.70). In conclusion, we developed an automated QC system that may be helpful for evaluating the quality of automated prostate segmentations.https://www.mdpi.com/2075-4418/10/9/714prostatesegmentationdeep learningradiomicsquality controlcomputer-aided detection and diagnosis |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Mohammed R. S. Sunoqrot Kirsten M. Selnæs Elise Sandsmark Gabriel A. Nketiah Olmo Zavala-Romero Radka Stoyanova Tone F. Bathen Mattijs Elschot |
spellingShingle |
Mohammed R. S. Sunoqrot Kirsten M. Selnæs Elise Sandsmark Gabriel A. Nketiah Olmo Zavala-Romero Radka Stoyanova Tone F. Bathen Mattijs Elschot A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRI Diagnostics prostate segmentation deep learning radiomics quality control computer-aided detection and diagnosis |
author_facet |
Mohammed R. S. Sunoqrot Kirsten M. Selnæs Elise Sandsmark Gabriel A. Nketiah Olmo Zavala-Romero Radka Stoyanova Tone F. Bathen Mattijs Elschot |
author_sort |
Mohammed R. S. Sunoqrot |
title |
A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRI |
title_short |
A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRI |
title_full |
A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRI |
title_fullStr |
A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRI |
title_full_unstemmed |
A Quality Control System for Automated Prostate Segmentation on T2-Weighted MRI |
title_sort |
quality control system for automated prostate segmentation on t2-weighted mri |
publisher |
MDPI AG |
series |
Diagnostics |
issn |
2075-4418 |
publishDate |
2020-09-01 |
description |
Computer-aided detection and diagnosis (CAD) systems have the potential to improve robustness and efficiency compared to traditional radiological reading of magnetic resonance imaging (MRI). Fully automated segmentation of the prostate is a crucial step of CAD for prostate cancer, but visual inspection is still required to detect poorly segmented cases. The aim of this work was therefore to establish a fully automated quality control (QC) system for prostate segmentation based on T2-weighted MRI. Four different deep learning-based segmentation methods were used to segment the prostate for 585 patients. First order, shape and textural radiomics features were extracted from the segmented prostate masks. A reference quality score (QS) was calculated for each automated segmentation in comparison to a manual segmentation. A least absolute shrinkage and selection operator (LASSO) was trained and optimized on a randomly assigned training dataset (N = 1756, 439 cases from each segmentation method) to build a generalizable linear regression model based on the radiomics features that best estimated the reference QS. Subsequently, the model was used to estimate the QSs for an independent testing dataset (N = 584, 146 cases from each segmentation method). The mean ± standard deviation absolute error between the estimated and reference QSs was 5.47 ± 6.33 on a scale from 0 to 100. In addition, we found a strong correlation between the estimated and reference QSs (rho = 0.70). In conclusion, we developed an automated QC system that may be helpful for evaluating the quality of automated prostate segmentations. |
topic |
prostate segmentation deep learning radiomics quality control computer-aided detection and diagnosis |
url |
https://www.mdpi.com/2075-4418/10/9/714 |
work_keys_str_mv |
AT mohammedrssunoqrot aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT kirstenmselnæs aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT elisesandsmark aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT gabrielanketiah aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT olmozavalaromero aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT radkastoyanova aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT tonefbathen aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT mattijselschot aqualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT mohammedrssunoqrot qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT kirstenmselnæs qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT elisesandsmark qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT gabrielanketiah qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT olmozavalaromero qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT radkastoyanova qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT tonefbathen qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri AT mattijselschot qualitycontrolsystemforautomatedprostatesegmentationont2weightedmri |
_version_ |
1724688928070434816 |