Original Research

Fully automated localization of prostate peripheral zone tumors on apparent diffusion coefficient map MR images using an ensemble learning method

Corresponding Author

Fatemeh Zabihollahy MASc

[email protected]

orcid.org/0000-0003-3362-1009

Department of Systems and Computer Engineering, Carleton University, Ottawa, Ontario, Canada

Address reprint requests to: F.Z., Department of Systems and Computer Engineering, Carleton University, 339 Riversedge Crescent, Ottawa, ON, Canada, K1V0Y6. E-mail: [email protected]Search for more papers by this author

Eranga Ukwatta PhD, PEng,

Eranga Ukwatta PhD, PEng

School of Engineering, University of Guelph, Guelph, Ontario, Canada

Search for more papers by this author

Satheesh Krishna MD,

Satheesh Krishna MD

Department of Medical Imaging, University of Toronto, Toronto, Ontario, Canada

Search for more papers by this author

Nicola Schieda MD,

Nicola Schieda MD

Department of Radiology, University of Ottawa, Ottawa, Ontario, Canada

Search for more papers by this author

Fatemeh Zabihollahy MASc,

Corresponding Author

Fatemeh Zabihollahy MASc

[email protected]

orcid.org/0000-0003-3362-1009

Department of Systems and Computer Engineering, Carleton University, Ottawa, Ontario, Canada

Eranga Ukwatta PhD, PEng,

Eranga Ukwatta PhD, PEng

School of Engineering, University of Guelph, Guelph, Ontario, Canada

Search for more papers by this author

Satheesh Krishna MD,

Satheesh Krishna MD

Department of Medical Imaging, University of Toronto, Toronto, Ontario, Canada

Search for more papers by this author

Nicola Schieda MD,

Nicola Schieda MD

Department of Radiology, University of Ottawa, Ottawa, Ontario, Canada

Search for more papers by this author

First published: 28 August 2019

https://doi.org/10.1002/jmri.26913

Citations: 13

Contract grant sponsor: Natural Sciences and Engineering Research Council of Canada (NSERC); Ontario Graduate Scholarship (OGS; to F.Z.).

Share a link

Email
Wechat
Bluesky

Abstract

Background

Accurate detection and localization of prostate cancer (PCa) in men undergoing prostate MRI is a fundamental step for future targeted prostate biopsies and treatment planning. Fully automated localization of peripheral zone (PZ) PCa using the apparent diffusion coefficient (ADC) map might be clinically useful.

Purpose/Hypothesis

To describe automated localization of PCa in the PZ on ADC map MR images using an ensemble U-Net-based model.

Study Type

Retrospective, case–control.

Population

In all, 226 patients (154 and 72 patients with and without clinically significant PZ PCa, respectively), training, and testing was performed using dataset images of 146 and 80 patients, respectively.

Field Strength

3T, ADC maps.

Sequence

ADC map.

Assessment

The ground truth was established by manual delineation of the prostate and prostate PZ tumors on ADC maps by dedicated radiologists using MRI-radical prostatectomy maps as a reference standard.

Statistical Tests: Performance of the ensemble model was evaluated using Dice similarity coefficient (DSC), sensitivity, and specificity metrics on a per-slice basis. Receiver operating characteristic (ROC) curve and area under the curve (AUC) were employed as well. The paired t-test was used to test the differences between the performances of constituent networks of the ensemble model.

Results

Our developed algorithm yielded DSC, sensitivity, and specificity of 86.72% ± 9.93%, 85.76% ± 23.33%, and 76.44% ± 23.70%, respectively (mean ± standard deviation) on 80 test cases consisting of 41 and 39 instances from patients with and without clinically significant tumors including 660 extracted 2D slices. AUC was reported as 0.779.

Data Conclusion

An ensemble U-Net-based approach can accurately detect and segment PCa in the PZ from ADC map MR prostate images.

Level of Evidence: 4

Technical Efficacy: Stage 1

J. Magn. Reson. Imaging 2020;51:1223–1234.

References

1Canadian Cancer Statistics 2017. Toronto, ON: Canadian Cancer Society; 2017. Available at: cancer.ca/Canadian-CancerStatistics-2017-EN.pdf (accessed [2019/07/15]).
Google Scholar
2De Rooij M, Hamoen EHJ, Fütterer JJ, Barentsz JO, Rovers MM. Accuracy of multiparametric MRI for prostate cancer detection: A meta-analysis. Am J Roentgenol 2014; 202:343–351.
10.2214/AJR.13.11046
PubMed Web of Science® Google Scholar
3Ahmed HU, El-Shater Bosaily A, Brown LC, et al. Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): A paired validating confirmatory study. Lancet 2017; 389:815–822.
10.1016/S0140-6736(16)32401-1
Web of Science® Google Scholar
4Kasivisvanathan V, Rannikko AS, Borghi M, et al. MRI-targeted or standard biopsy for prostate-cancer diagnosis. N Engl J Med 2018; 378:1767–1777.
10.1056/NEJMoa1801993
PubMed Web of Science® Google Scholar
5Quon JS, Moosavi B, Khanna M, Flood TA, Lim CS, Schieda N. False positive and false negative diagnoses of prostate cancer at multi-parametric prostate MRI in active surveillance. Insights Imaging 2015; 6:449–463.
10.1007/s13244-015-0411-3
PubMed Web of Science® Google Scholar
6Mussi TC, Yamauchi FI, Tridente CF, et al. Interobserver agreement and positivity of PI-RADS version 2 among radiologists with different levels of experience. Acad Radiol 2018; 26:1017–1022.
PubMed Web of Science® Google Scholar
7Prostate Imaging Reporting and Data System Version 2.1: 2019 Update of Prostate Imaging Reporting and Data System Version 2. Eur Urol 2019; 76:340–351. https://doi.org/10.1016/j.eururo.2019.02.033.
Google Scholar
8Gaur S, Harmon S, Rosenblum L, et al. Can apparent diffusion coefficient values assist PI-RADS version 2 DWI scoring? A correlation study using the pi-radsv2 and international society of urological pathology systems. Am J Roentgenol 2018;211:W33-W41.
10.2214/AJR.17.18702
PubMed Web of Science® Google Scholar
9Gaur S, Lay N, Harmon SA, et al. Can computer-aided diagnosis assist in the identification of prostate cancer on prostate MRI? A multi-center, multi-reader investigation. Oncotarget 2018;9:33804–33817.
10.18632/oncotarget.26100
PubMed Google Scholar
10Lemaître G, Martí R, Freixenet J, Vilanova JC, Walker PM, Meriaudeau F. Computer-aided detection and diagnosis for prostate cancer based on mono and multi-parametric MRI: A review. Comput Biol Med 2015; 60:8–31.
10.1016/j.compbiomed.2015.02.009
PubMed Web of Science® Google Scholar
11Wang S, Burtt K, Turkbey B, Choyke P, Summers RM. Computer aided-diagnosis of prostate cancer on multiparametric MRI: A technical review of current research. Biomed Res Int 2014; 2014:789561–11.
10.1155/2014/789561
Web of Science® Google Scholar
12Niu X ke, Chen X hui, Chen Z fan, Chen L, Li J, Peng T. Diagnostic performance of biparametric MRI for detection of prostate cancer: A systematic review and meta-analysis. Am J Roentgenol 2018; 211:369–378.
Web of Science® Google Scholar
13Giger ML, Chan HP, Boone J. Anniversary paper: History and status of CAD and quantitative image analysis: The role of medical physics and AAPM. Med Phys 2008; 35:5799–5820.
10.1118/1.3013555
Web of Science® Google Scholar
14Rokach L. Ensemble-based classifiers. Artif Intell Rev 2010;33:1–39.
10.1007/s10462-009-9124-7
Web of Science® Google Scholar
15Ronneberger O, Fischer P, Brox T. U-Net: Convolutional networks for biomedical image segmentation. In: Lect Notes Comput Sci (including Subser Lect Notes Artif Intell Lect Notes Bioinformatics); 2015;9351:234–241.
Google Scholar
16Purysko AS, Rosenkrantz AB, Barentsz JO, Weinreb JC, Macura KJ. PI-RADS Version 2: A pictorial update. RadioGraphics 2016;36:1354–1372.
10.1148/rg.2016150234
Web of Science® Google Scholar
17Bhavsar A, Verma S. Anatomic imaging of the prostate. Biomed Res Int 2014;2014:728539–728547.
10.1155/2014/728539
PubMed Web of Science® Google Scholar
18Weinreb JC, Barentsz JO, Choyke PL, et al. PI-RADS Prostate Imaging — Reporting and Data System: 2015, Version 2. Eur Urol 2016;69:16–40.
10.1016/j.eururo.2015.08.052
Web of Science® Google Scholar
19Zabihollahy F, Schieda N, Krishna Jeyaraj S, Ukwatta E. Automated segmentation of prostate zonal anatomy on T2-weighted (T2W) and apparent diffusion coefficient (ADC) map MR images using U-Nets. Med Phys 2019;46:3078–3090.
10.1002/mp.13550
Web of Science® Google Scholar
20Bratan F, Melodelima C, Souchon R, et al. How accurate is multiparametric MR imaging in evaluation of prostate cancer volume? Radiology 2014;275:144–54.
Web of Science® Google Scholar
21Krishna S, Lim CS, McInnes MDF, et al. Evaluation of MRI for diagnosis of extraprostatic extension in prostate cancer. J Magn Reson Imaging 2018;47:176–185.
10.1002/jmri.25729
Web of Science® Google Scholar
22Milletari F, Navab N, Ahmadi SA; V-Net: Fully convolutional neural networks for volumetric medical image segmentation. In: Proc – 2016 4th Int Conf 3D Vision, 3DV 2016;2016:565–571. https://doi.org/10.1109/3DV.2016.79.
Google Scholar
23Zeiler MD. ADADELTA: An adaptive learning rate method. 2012; arXiv:1212.5701
Google Scholar
24Barentsz JO, Weinreb JC, Verma S, et al. Synopsis of the PI-RADS v2 guidelines for multiparametric prostate magnetic resonance imaging and recommendations for use. Eur Urol 2016;69:41–49.
10.1016/j.eururo.2015.08.038
Web of Science® Google Scholar
25Rosenkrantz AB, Ream JM, Nolan P, Rusinek H, Deng FM, Taneja SS. Prostate cancer: Utility of whole-lesion apparent diffusion coefficient metrics for prediction of biochemical recurrence after radical prostatectomy. Am J Roentgenol 2015;205:1208–1214.
10.2214/AJR.15.14482
Web of Science® Google Scholar
26Kim TH, Kim CK, Park BK, et al. Relationship between Gleason score and apparent diffusion coefficients of diffusion-weighted magnetic resonance imaging in prostate cancer patients. Can Urol Assoc J 2016;10:E377–E382.
10.5489/cuaj.3896
Web of Science® Google Scholar
27Chen M, Dang HD, Wang JY, et al. Prostate cancer detection: Comparison of T2-weighted imaging, diffusion-weighted imaging, proton magnetic resonance spectroscopic imaging, and the three techniques combined. Acta Radiol 2008;49:602–10.
10.1080/02841850802004983
Web of Science® Google Scholar
28Miao H, Fukatsu H, Ishigaki T. Prostate cancer detection with 3-T MRI: Comparison of diffusion-weighted and T2-weighted imaging. Eur J Radiol 2007;61:297–302.
10.1016/j.ejrad.2006.10.002
PubMed Web of Science® Google Scholar
29Scialpi M, Rondoni V, Aisa MC, et al. Is contrast enhancement needed for diagnostic prostate MRI? Transl Androl Urol 2017;6:499–509.
10.21037/tau.2017.05.31
PubMed Web of Science® Google Scholar
30Song Y, Zhang YD, Yan X, et al. Computer-aided diagnosis of prostate cancer using a deep convolutional neural network from multiparametric MRI. J Magn Reson Imaging 2018;48:1570–1577.
10.1002/jmri.26047
Web of Science® Google Scholar
31Tsehay YK, Lay NS, Roth HR, et al. Convolutional neural network based deep-learning architecture for prostate cancer detection on multiparametric magnetic resonance images. In: Med Imaging 2017 Comput Diagnosis; 2017;10134.
Google Scholar
32Yang X, Liu C, Wang Z, et al. Co-trained convolutional neural networks for automated detection of prostate cancer in multi-parametric MRI. Med Image Anal 2017;42:212–227.
10.1016/j.media.2017.08.006
Web of Science® Google Scholar
33Khalvati F, Salmanpour A, Rahnamayan S, Haider MA, Tizhoosh HR. Sequential registration-based segmentation of the prostate gland in MR image volumes. J Digit Imaging 2016;29:254–263.
10.1007/s10278-015-9844-y
PubMed Web of Science® Google Scholar
34Dietterich TG. Ensemble methods in machine learning. 2007. LNCS, 1857:1–15.
Google Scholar
35Krogh A, Vedelsby J. Neural network ensembles, cross validation, and active learning. NIPS 7, 1995;7:231–238.
Google Scholar
36Breiman L. Bagging predictors. Mach Learn 2004;24:123–140.
Google Scholar
37Zenobi G, Cunningham P. Using diversity in preparing ensembles of classifiers based on different feature subsets to minimize generalization error. 2001 LNCS, 2167:576–587.
10.1007/3-540-44795-4_49
Google Scholar
38Melville P, Mooney RJ. Constructing diverse classifier ensembles using artificial training examples. In: IJCAI Int Jt Conf Artif Intell; 2003;18:505–510.
Google Scholar
39Brown G, Wyatt J, Harris R, Yao X. Diversity creation methods: A survey and categorisation. Inf Fusion 2005;6:5–20.
10.1016/j.inffus.2004.04.004
Google Scholar
40Tong G, Li Y, Chen H, Zhang Q, Jiang H. Improved U-NET network for pulmonary nodules segmentation. Optik (Stuttg) 2018;174:460–469.
10.1016/j.ijleo.2018.08.086
PubMed Web of Science® Google Scholar
41Fabijañska A. Segmentation of corneal endothelium images using a U-Net-based convolutional neural network. Artif Intell Med 2018;88:1–13.
PubMed Web of Science® Google Scholar
42Barentsz J. The role of (fast) bi-parametric MRi versus multi-parametric MRI and TRUS-biopsy for detecting clinically significant prostate cancer in biopsy naïve men with elevated PSA. EAU, Barcelona, Spain; 2019;34.
Google Scholar
43Woo S, Suh CH, Kim SY, Cho JY, Kim SH. Head-to-head comparison between high- and standard-b-value DWI for detecting prostate cancer: A systematic review and meta-analysis. Am J Roentgenol 2018;210:91–100.
10.2214/AJR.17.18480
Web of Science® Google Scholar
44Rosenkrantz AB, Kim S, Campbell N, Gaing B, Deng FM, Taneja SS. Transition zone prostate cancer: Revisiting the role of multiparametric MRI at 3 T. Am J Roentgenol 2015;204:W266–W272.
10.2214/AJR.14.12955
Web of Science® Google Scholar

Citing Literature

Volume51, Issue4

April 2020

Pages 1223-1234

Fully automated localization of prostate peripheral zone tumors on apparent diffusion coefficient map MR images using an ensemble learning method

Abstract

Background

Purpose/Hypothesis

Study Type

Population

Field Strength

Sequence

Assessment

Results

Data Conclusion

References

Citing Literature

References

Information

About Wiley Online Library

Help & Support

Opportunities

Connect with Wiley

Fully automated localization of prostate peripheral zone tumors on apparent diffusion coefficient map MR images using an ensemble learning method

Abstract

Background

Purpose/Hypothesis

Study Type

Population

Field Strength

Sequence

Assessment

Results

Data Conclusion

References

Citing Literature

References

Related

Information