Autoprostate: Towards automated reporting of prostate mri for prostate cancer assessment using deep learning

Pritesh Mehta; Michela Antonelli; Saurabh Singh; Natalia Grondecka; Edward W. Johnston; Hashim U. Ahmed; Mark Emberton; Shonit Punwani; Sébastien Ourselin

doi:10.3390/cancers13236138

Autoprostate: Towards automated reporting of prostate mri for prostate cancer assessment using deep learning

Pritesh Mehta^*, Michela Antonelli, Saurabh Singh, Natalia Grondecka, Edward W. Johnston, Hashim U. Ahmed, Mark Emberton, Shonit Punwani, Sébastien Ourselin

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

15 Citations (Scopus)

Abstract

Multiparametric magnetic resonance imaging (mpMRI) of the prostate is used by radiologists to identify, score, and stage abnormalities that may correspond to clinically significant prostate cancer (CSPCa). Automatic assessment of prostate mpMRI using artificial intelligence algorithms may facilitate a reduction in missed cancers and unnecessary biopsies, an increase in inter-observer agreement between radiologists, and an improvement in reporting quality. In this work, we introduce AutoProstate, a deep learning-powered framework for automatic MRI-based prostate cancer assessment. AutoProstate comprises of three modules: Zone-Segmenter, CSPCaSegmenter, and Report-Generator. Zone-Segmenter segments the prostatic zones on T2-weighted imaging, CSPCa-Segmenter detects and segments CSPCa lesions using biparametric MRI, and ReportGenerator generates an automatic web-based report containing four sections: Patient Details, Prostate Size and PSA Density, Clinically Significant Lesion Candidates, and Findings Summary. In our experi-ment, AutoProstate was trained using the publicly available PROSTATEx dataset, and externally validated using the PICTURE dataset. Moreover, the performance of AutoProstate was compared to the performance of an experienced radiologist who prospectively read PICTURE dataset cases. In comparison to the radiologist, AutoProstate showed statistically significant improvements in prostate volume and prostate-specific antigen density estimation. Furthermore, AutoProstate matched the CSPCa lesion detection sensitivity of the radiologist, which is paramount, but produced more false positive detections.

Original language	English
Article number	6138
Journal	Cancers
Volume	13
Issue number	23
DOIs	https://doi.org/10.3390/cancers13236138
Publication status	E-pub ahead of print - 6 Dec 2021

Keywords

Automatic report
Computer-aided diagnosis
Convolutional neural network
Deep learning
Lesion classification
Lesion detection
Magnetic resonance imaging
Prostate cancer
Segmentation

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.3390/cancers13236138

Cite this

@article{a2b4e0ead77b451db9f2521d0df908f8,

title = "Autoprostate: Towards automated reporting of prostate mri for prostate cancer assessment using deep learning",

abstract = "Multiparametric magnetic resonance imaging (mpMRI) of the prostate is used by radiologists to identify, score, and stage abnormalities that may correspond to clinically significant prostate cancer (CSPCa). Automatic assessment of prostate mpMRI using artificial intelligence algorithms may facilitate a reduction in missed cancers and unnecessary biopsies, an increase in inter-observer agreement between radiologists, and an improvement in reporting quality. In this work, we introduce AutoProstate, a deep learning-powered framework for automatic MRI-based prostate cancer assessment. AutoProstate comprises of three modules: Zone-Segmenter, CSPCaSegmenter, and Report-Generator. Zone-Segmenter segments the prostatic zones on T2-weighted imaging, CSPCa-Segmenter detects and segments CSPCa lesions using biparametric MRI, and ReportGenerator generates an automatic web-based report containing four sections: Patient Details, Prostate Size and PSA Density, Clinically Significant Lesion Candidates, and Findings Summary. In our experi-ment, AutoProstate was trained using the publicly available PROSTATEx dataset, and externally validated using the PICTURE dataset. Moreover, the performance of AutoProstate was compared to the performance of an experienced radiologist who prospectively read PICTURE dataset cases. In comparison to the radiologist, AutoProstate showed statistically significant improvements in prostate volume and prostate-specific antigen density estimation. Furthermore, AutoProstate matched the CSPCa lesion detection sensitivity of the radiologist, which is paramount, but produced more false positive detections.",

keywords = "Automatic report, Computer-aided diagnosis, Convolutional neural network, Deep learning, Lesion classification, Lesion detection, Magnetic resonance imaging, Prostate cancer, Segmentation",

author = "Pritesh Mehta and Michela Antonelli and Saurabh Singh and Natalia Grondecka and Johnston, {Edward W.} and Ahmed, {Hashim U.} and Mark Emberton and Shonit Punwani and S{\'e}bastien Ourselin",

note = "Funding Information: P.M.?s research is supported by the Engineering and Physical Sciences Research Council (EPSRC) [EP/R512400/1]. P.M.?s work was additionally supported by the EPSRC-funded UCL Centre for Doctoral Training in Intelligent, Integrated Imaging in Healthcare (i4health) [EP/S021930/1]. M.A.?s research is supported by the Wellcome/EPSRC Centre for Medical Engineering King?s College London and by the London Medical Imaging and AI Centre for Value-Based Healthcare. H.U.A.?s research is supported by core funding from the UK?s National Institute of Health Research (NIHR) Imperial Biomedical Research Centre. HUA currently also receives funding from the Wellcome Trust, Medical Research Council (UK), Cancer Research UK, Prostate Cancer UK, The Urology Foundation, BMA Foundation, Imperial Health Charity, Sonacare Inc., Trod Medical and Sophiris Biocorp for trials in prostate cancer. M.E. and S.P. receive research support from the University College London/University College London Hospital (UCL/UCLH) Biomedical Research Centre. Funding Information: Acknowledgments: P.M.{\textquoteright}s research is supported by the Engineering and Physical Sciences Research Council (EPSRC) [EP/R512400/1]. P.M.{\textquoteright}s work was additionally supported by the EPSRC-funded UCL Centre for Doctoral Training in Intelligent, Integrated Imaging in Healthcare (i4health) [EP/S021930/1]. M.A.{\textquoteright}s research is supported by the Wellcome/EPSRC Centre for Medical Engineering King{\textquoteright}s College London and by the London Medical Imaging and AI Centre for Value-Based Healthcare. H.U.A.{\textquoteright}s research is supported by core funding from the UK{\textquoteright}s National Institute of Health Research (NIHR) Imperial Biomedical Research Centre. HUA currently also receives funding from the Wellcome Trust, Medical Research Council (UK), Cancer Research UK, Prostate Cancer UK, The Urology Foundation, BMA Foundation, Imperial Health Charity, Sonacare Inc., Trod Medical and Sophiris Biocorp for trials in prostate cancer. M.E. and S.P. receive research support from the University College London/University College London Hospital (UCL/UCLH) Biomedical Research Centre. Publisher Copyright: {\textcopyright} 2021 by the authors. Licensee MDPI, Basel, Switzerland.",

year = "2021",

month = dec,

day = "6",

doi = "10.3390/cancers13236138",

language = "English",

volume = "13",

journal = "Cancers",

issn = "2072-6694",

publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",

number = "23",

}

TY - JOUR

T1 - Autoprostate

T2 - Towards automated reporting of prostate mri for prostate cancer assessment using deep learning

AU - Mehta, Pritesh

AU - Antonelli, Michela

AU - Singh, Saurabh

AU - Grondecka, Natalia

AU - Johnston, Edward W.

AU - Ahmed, Hashim U.

AU - Emberton, Mark

AU - Punwani, Shonit

AU - Ourselin, Sébastien

N1 - Funding Information: P.M.?s research is supported by the Engineering and Physical Sciences Research Council (EPSRC) [EP/R512400/1]. P.M.?s work was additionally supported by the EPSRC-funded UCL Centre for Doctoral Training in Intelligent, Integrated Imaging in Healthcare (i4health) [EP/S021930/1]. M.A.?s research is supported by the Wellcome/EPSRC Centre for Medical Engineering King?s College London and by the London Medical Imaging and AI Centre for Value-Based Healthcare. H.U.A.?s research is supported by core funding from the UK?s National Institute of Health Research (NIHR) Imperial Biomedical Research Centre. HUA currently also receives funding from the Wellcome Trust, Medical Research Council (UK), Cancer Research UK, Prostate Cancer UK, The Urology Foundation, BMA Foundation, Imperial Health Charity, Sonacare Inc., Trod Medical and Sophiris Biocorp for trials in prostate cancer. M.E. and S.P. receive research support from the University College London/University College London Hospital (UCL/UCLH) Biomedical Research Centre. Funding Information: Acknowledgments: P.M.’s research is supported by the Engineering and Physical Sciences Research Council (EPSRC) [EP/R512400/1]. P.M.’s work was additionally supported by the EPSRC-funded UCL Centre for Doctoral Training in Intelligent, Integrated Imaging in Healthcare (i4health) [EP/S021930/1]. M.A.’s research is supported by the Wellcome/EPSRC Centre for Medical Engineering King’s College London and by the London Medical Imaging and AI Centre for Value-Based Healthcare. H.U.A.’s research is supported by core funding from the UK’s National Institute of Health Research (NIHR) Imperial Biomedical Research Centre. HUA currently also receives funding from the Wellcome Trust, Medical Research Council (UK), Cancer Research UK, Prostate Cancer UK, The Urology Foundation, BMA Foundation, Imperial Health Charity, Sonacare Inc., Trod Medical and Sophiris Biocorp for trials in prostate cancer. M.E. and S.P. receive research support from the University College London/University College London Hospital (UCL/UCLH) Biomedical Research Centre. Publisher Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland.

PY - 2021/12/6

Y1 - 2021/12/6

N2 - Multiparametric magnetic resonance imaging (mpMRI) of the prostate is used by radiologists to identify, score, and stage abnormalities that may correspond to clinically significant prostate cancer (CSPCa). Automatic assessment of prostate mpMRI using artificial intelligence algorithms may facilitate a reduction in missed cancers and unnecessary biopsies, an increase in inter-observer agreement between radiologists, and an improvement in reporting quality. In this work, we introduce AutoProstate, a deep learning-powered framework for automatic MRI-based prostate cancer assessment. AutoProstate comprises of three modules: Zone-Segmenter, CSPCaSegmenter, and Report-Generator. Zone-Segmenter segments the prostatic zones on T2-weighted imaging, CSPCa-Segmenter detects and segments CSPCa lesions using biparametric MRI, and ReportGenerator generates an automatic web-based report containing four sections: Patient Details, Prostate Size and PSA Density, Clinically Significant Lesion Candidates, and Findings Summary. In our experi-ment, AutoProstate was trained using the publicly available PROSTATEx dataset, and externally validated using the PICTURE dataset. Moreover, the performance of AutoProstate was compared to the performance of an experienced radiologist who prospectively read PICTURE dataset cases. In comparison to the radiologist, AutoProstate showed statistically significant improvements in prostate volume and prostate-specific antigen density estimation. Furthermore, AutoProstate matched the CSPCa lesion detection sensitivity of the radiologist, which is paramount, but produced more false positive detections.

AB - Multiparametric magnetic resonance imaging (mpMRI) of the prostate is used by radiologists to identify, score, and stage abnormalities that may correspond to clinically significant prostate cancer (CSPCa). Automatic assessment of prostate mpMRI using artificial intelligence algorithms may facilitate a reduction in missed cancers and unnecessary biopsies, an increase in inter-observer agreement between radiologists, and an improvement in reporting quality. In this work, we introduce AutoProstate, a deep learning-powered framework for automatic MRI-based prostate cancer assessment. AutoProstate comprises of three modules: Zone-Segmenter, CSPCaSegmenter, and Report-Generator. Zone-Segmenter segments the prostatic zones on T2-weighted imaging, CSPCa-Segmenter detects and segments CSPCa lesions using biparametric MRI, and ReportGenerator generates an automatic web-based report containing four sections: Patient Details, Prostate Size and PSA Density, Clinically Significant Lesion Candidates, and Findings Summary. In our experi-ment, AutoProstate was trained using the publicly available PROSTATEx dataset, and externally validated using the PICTURE dataset. Moreover, the performance of AutoProstate was compared to the performance of an experienced radiologist who prospectively read PICTURE dataset cases. In comparison to the radiologist, AutoProstate showed statistically significant improvements in prostate volume and prostate-specific antigen density estimation. Furthermore, AutoProstate matched the CSPCa lesion detection sensitivity of the radiologist, which is paramount, but produced more false positive detections.

KW - Automatic report

KW - Computer-aided diagnosis

KW - Convolutional neural network

KW - Deep learning

KW - Lesion classification

KW - Lesion detection

KW - Magnetic resonance imaging

KW - Prostate cancer

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=85120623152&partnerID=8YFLogxK

U2 - 10.3390/cancers13236138

DO - 10.3390/cancers13236138

M3 - Article

AN - SCOPUS:85120623152

SN - 2072-6694

VL - 13

JO - Cancers

JF - Cancers

IS - 23

M1 - 6138

ER -

Autoprostate: Towards automated reporting of prostate mri for prostate cancer assessment using deep learning

Abstract

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this