Abstract
Objective
To train and to test for prostate zonal segmentation an existing algorithm already trained for whole-gland segmentation.
Methods
The algorithm, combining model-based and deep learning–based approaches, was trained for zonal segmentation using the NCI-ISBI-2013 dataset and 70 T2-weighted datasets acquired at an academic centre. Test datasets were randomly selected among examinations performed at this centre on one of two scanners (General Electric, 1.5 T; Philips, 3 T) not used for training. Automated segmentations were corrected by two independent radiologists. When segmentation was initiated outside the prostate, images were cropped and segmentation repeated. Factors influencing the algorithm’s mean Dice similarity coefficient (DSC) and its precision were assessed using beta regression.
Results
Eighty-two test datasets were selected; one was excluded. In 13/81 datasets, segmentation started outside the prostate, but zonal segmentation was possible after image cropping. Depending on the radiologist chosen as reference, algorithm’s median DSCs were 96.4/97.4%, 91.8/93.0% and 79.9/89.6% for whole-gland, central gland and anterior fibromuscular stroma (AFMS) segmentations, respectively. DSCs comparing radiologists’ delineations were 95.8%, 93.6% and 81.7%, respectively. For all segmentation tasks, the scanner used for imaging significantly influenced the mean DSC and its precision, and the mean DSC was significantly lower in cases with initial segmentation outside the prostate. For central gland segmentation, the mean DSC was also significantly lower in larger prostates. The radiologist chosen as reference had no significant impact, except for AFMS segmentation.
Conclusions
The algorithm performance fell within the range of inter-reader variability but remained significantly impacted by the scanner used for imaging.
Key Points
• Median Dice similarity coefficients obtained by the algorithm fell within human inter-reader variability for the three segmentation tasks (whole gland, central gland, anterior fibromuscular stroma).
• The scanner used for imaging significantly impacted the performance of the automated segmentation for the three segmentation tasks.
• The performance of the automated segmentation of the anterior fibromuscular stroma was highly variable across patients and showed also high variability across the two radiologists.
Similar content being viewed by others
Abbreviations
- ABD:
-
Average boundary distance
- AFMS:
-
Anterior fibromuscular stroma
- CG:
-
Central gland
- DNN:
-
Deep neural network
- DSC:
-
Dice similarity coefficient
- GE:
-
General Electric
- HD:
-
Hausdorff distance
- HD95:
-
95Th percentile of the Hausdorff distance
- MR:
-
Magnetic resonance
- MRI:
-
Magnetic resonance imaging
- PZ:
-
Peripheral zone
- R1:
-
Radiologist 1
- R2:
-
Radiologist 2
- RVD:
-
Relative volume difference
References
Pagniez MA, Kasivisvanathan V, Puech P, Drumez E, Villers A, Olivier J (2020) Predictive factors of missed clinically significant prostate cancers in men with negative magnetic resonance imaging: a systematic review and meta-analysis. J Urol 204:24–32
Lee DK, Sung DJ, Kim CS et al (2020) Three-dimensional convolutional neural network for prostate MRI segmentation and comparison of prostate volume measurements by use of artificial neural network and ellipsoid formula. AJR Am J Roentgenol 214:1229–1238
Penzkofer T, Padhani AR, Turkbey B et al (2021) ESUR/ESUI position paper: developing artificial intelligence for precision diagnosis of prostate cancer using magnetic resonance imaging. Eur Radiol. https://doi.org/10.1007/s00330-021-08021-6
Schelb P, Wang X, Radtke JP et al (2021) Simulated clinical deployment of fully automatic deep learning for clinical prostate MRI assessment. Eur Radiol 31:302–313
Turkbey B, Rosenkrantz AB, Haider MA et al (2019) Prostate imaging reporting and data system version 2.1: 2019 update of prostate imaging reporting and data system version 2. Eur Urol 76:340–351
Becker AS, Chaitanya K, Schawkat K et al (2019) Variability of manual segmentation of the prostate in axial T2-weighted MRI: a multi-reader study. Eur J Radiol 121:108716
Meyer A, Ghosh S, Schindele D et al (2021) Uncertainty-aware temporal self-learning (UATS): semi-supervised learning for segmentation of prostate zones and beyond. Artif Intell Med 116:102073
Montagne S, Hamzaoui D, Allera A et al (2021) Challenge of prostate MRI segmentation on T2-weighted images: inter-observer variability and impact of prostate morphology. Insights Imaging 12:71
Liu L, Tian Z, Zhang Z, Fei B (2016) Computer-aided detection of prostate cancer with MRI: technology and applications. Acad Radiol 23:1024–1046
Makni N, Iancu A, Colot O, Puech P, Mordon S, Betrouni N (2011) Zonal segmentation of prostate using multispectral magnetic resonance images. Med Phys 38:6093–6105
Yuan J, Ukwatta E, Qiu W et al (2013) Jointly segmenting prostate zones in 3D MRIs by globally optimized coupled level-sets. In: Heyden A, Kahl F, Olsson C, Oskarsson M, Tai XC, (eds) Lecture notes in computer science 8081. Springer, pp 12–25
Litjens G, Debats O, Van de Ven W, Karssemeijer N, Huisman H (2012) A pattern recognition approach to zonal segmentation of the prostate on MRI. In: Ayache N, Delingette H, Golland P, Mori K (eds) Lecture notes in computer science 7511. Springer, pp 413–420
Alvarez C, Martinez F, Romero E (2017) A multiresolution prostate representation for automatic segmentation in magnetic resonance images. Med Phys 44:1312–1323
Shahedi M, Cool DW, Bauman GS, Bastian-Jordan M, Fenster A, Ward AD (2017) Accuracy validation of an automated method for prostate segmentation in magnetic resonance imaging. J Digit Imaging 30:782–795
Chilali O, Puech P, Lakroum S, Diaf M, Mordon S, Betrouni N (2016) Gland and zonal segmentation of prostate on T2W MR images. J Digit Imaging 29:730–736
Guo Y, Gao Y, Shen D (2016) Deformable MR prostate segmentation via deep feature learning and sparse patch matching. IEEE Trans Med Imaging 35:1077–1089
Cheng R, Roth HR, Lay N et al (2017) Automatic magnetic resonance prostate segmentation by deep learning with holistically nested networks. J Med Imaging (Bellingham) 4:041302
Jia H, Xia Y, Song Y, Cai W, Fulham M, Feng DD (2018) Atals registration and ensemble deep convolutional neural network-based prostate segmentation using magnetic resonance imaging. Neurocomputing 275:1358–1369
To MNN, Vu DQ, Turkbey B, Choyke PL, Kwak JT (2018) Deep dense multi-path neural network for prostate segmentation in magnetic resonance imaging. Int J Comput Assist Radiol Surg 13:1687–1696
Tian Z, Liu L, Zhang Z, Fei B (2018) PSNet: prostate segmentation on MRI based on a convolutional neural network. J Med Imaging (Bellingham) 5:021208
Yan K, Wang X, Kim J, Khadra M, Fulham M, Feng D (2019) A propagation-DNN: deep combination learning of multi-level features for MR prostate segmentation. Comput Methods Programs Biomed 170:11–21
Zhou W, Tao X, Wei Z, Lin L (2020) Automatic segmentation of 3D prostate MR images with iterative localization refinement. Digit Signal Proc 98:102649
Qiu W, Yuan J, Ukwatta E, Sun Y, Rajchl M, Fenster A (2014) Dual optimization based prostate zonal segmentation in 3D MR images. Med Image Anal 18:660–673
Zhu Y, Wei R, Gao G et al (2019) Fully automatic segmentation on prostate MR images based on cascaded fully convolution network. J Magn Reson Imaging 49:1149–1156
Cheng R, Lay N, Roth HR et al (2019) Fully automated prostate whole gland and central gland segmentation on MRI using holistically nested networks with short connections. J Med Imaging (Bellingham) 6:024007
Zabihollahy F, Schieda N, Krishna Jeyaraj S, Ukwatta E (2019) Automated segmentation of prostate zonal anatomy on T2-weighted (T2W) and apparent diffusion coefficient (ADC) map MR images using U-Nets. Med Phys 46:3078–3090
Yin Y, Fotin SV, Periaswany S et al (2012) Fully automated prostate segmentation in 3D MR based on normalized gradient fields cross-correlation initialization and LOGISMOS refinement. In: Haynor DR, Ourselin S, (eds) Proceedings of SPIE pp 83106
Yin Y, Fotin SV, Periaswany S et al (2012) Fully automated prostate central gland segmentation in MR images: a LOGISMOS based approach. In: Haynor DR, Ourselin S, (eds) Proceedings of SPIE, pp 83143B
Rundo L, Han C, Nagano Y et al (2019) USE-Net: incorporating squeeze-and-excitation blocks into U-Net for prostate zonal segmentation of multi-institutional MRI datasets. Neurocomputing 365:31–43
Liu Y, Yang G, S.A. M et al (2019) Automatic prostate zonal segmentation using fully convolutional netwok with feature pyramid attention. arXiv 1911.00127v1 [eess.iv]
Bardis M, Houshyar R, Chantaduly C et al (2021) Segmentation of the prostate transition zone and peripheral zone on MR images with deep learning. Radiol Imaging Cancer 3:e200024
Cuocolo R, Comelli A, Stefano A et al (2021) Deep learning whole-gland and zonal prostate segmentation on a public MRI dataset. J Magn Reson Imaging. https://doi.org/10.1002/jmri.27585
Aldoj N, Biavati F, Michallek F, Stober S, Dewey M (2020) Automatic prostate and prostate zones segmentation of magnetic resonance images using DenseNet-like U-net. Sci Rep 10:14315
Liu Y, Yang G, Hosseiny M et al (2020) Exploring uncertainty measures in Bayesian deep attentive neural networks for prostate zonal segmentation. IEEE Access 8:151817–151828
Khan Z, Yahya N, Alsaih K, Ali SSA, Meriaudeau F (2020) Evaluation of deep neural networks for semantic segmentation of prostate in T2W MRI. Sensors (Basel) 20:3183
Nai YH, Teo BW, Tan NL et al (2020) Evaluation of multimodal algorithms for the segmentation of multiparametric MRI prostate images. Comput Math Methods Med 2020:8861035
Litjens G, Toth R, van de Ven W et al (2014) Evaluation of prostate segmentation algorithms for MRI: the PROMISE12 challenge. Med Image Anal 18:359–373
Brosch T, Peters J, Groth A, Stehle T, Weese J (2018) Deep learning-based boundary detection for model-based segmentation with application to MR prostate segmentation. In: Frangi AF, Schnabel JA, Davatzikos C, Alberola-Lopez C, Fichtinger G (eds) Lecture notes in computer science 11073. Springer, pp 515–522
Brosch T, Peters J, Groth A, Weber FM, Weese J (2021) Model-based segmentation using neural network-based boundary detectors: application to prostate and heart segmentation in MR images. Machine Learning with Applications 6:100078
Heimann T, Meinzer HP (2009) Statistical shape models for 3D medical image segmentation: a review. Med Image Anal 13:543–563
Bratan F, Niaf E, Melodelima C et al (2013) Influence of imaging and histological factors on prostate cancer detection and localisation on multiparametric MRI: a prospective study. Eur Radiol 23:2019–2029
Farahani K, Bloch N, Madabhushi A et al (2013) NCI-ISBI 2013 challenge - automated segmentation of prostate structures. Available via https://wiki.cancerimagingarchive.net/display/Public/NCI-ISBI+2013+Challenge+-+Automated+Segmentation+of+Prostate+Structures. Accessed 22 Sept 2020
Cribari-Neto F, Zeileis A (2010) Beta regression in R. J Stat Softw 34:1–24
de Rooij M, Israel B, Tummers M et al (2020) ESUR/ESUI consensus statements on multi-parametric MRI for the detection of clinically significant prostate cancer: quality requirements for image acquisition, interpretation and radiologists’ training. Eur Radiol 30:5404–5416
Bouye S, Potiron E, Puech P, Leroy X, Lemaitre L, Villers A (2009) Transition zone and anterior stromal prostate cancers: zone of origin and intraprostatic patterns of spread at histopathology. Prostate 69:105–113
Heijmink SW, Futterer JJ, Hambrock T et al (2007) Prostate cancer: body-array versus endorectal coil MR imaging at 3 T–comparison of image quality, localization, and staging performance. Radiology 244:184–195
Heijmink SW, Scheenen TW, van Lin EN et al (2009) Changes in prostate shape and volume and their implications for radiotherapy after introduction of endorectal balloon as determined by MRI at 3T. Int J Radiat Oncol Biol Phys 73:1446–1453
Funding
This research project was sponsored and funded by the Hospices Civils de Lyon and performed under the framework of the collaboration between the Hospices Civils de Lyon and Philips that is part of the GOPI public contract whose holder is Philips.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Guarantor
The scientific guarantor of this publication is Prof. Olivier Rouvière.
Conflict of interest
Four co-authors (AV, AG, MaR, JW) are Philips employees. No author has financial conflict of interest regarding the products assessed in this study.
Statistics and biometry
Two of the authors have significant statistical expertise.
Informed consent
Training cases were selected from a prospectively maintained database (CLARA-P database). The creation of this database was approved by the Comité de Protection des Personnes Sud-Est IV and included patients provided written informed consent for the use of their MR images and histological results for research purposes.
Test cases were randomly selected among all the prostate MRIs in our PACS. According to the French law, retrospective analysis of these routine cases was approved by our Ethics Committee. In addition, all patients whose MRI was selected as a test case received a letter explaining the study and giving them the possibility to withdraw.
Ethical approval
Institutional Review Board approval was obtained.
Study subjects or cohorts overlap
Some of the training MRIs (included in the CLARA-P database) have been used in other studies:
Bratan F et al., Influence of imaging and histological factors on prostate cancer detection and localisation on multiparametric MRI: a prospective study. Eur Radiol 2013; 23:2019.
Vaché T et al., Characterization of prostate lesions as benign or malignant at multiparametric MR imaging: comparison of three scoring systems in patients treated with radical prostatectomy. Radiology 2014; 272:446.
Niaf E et al., Prostate focal peripheral zone lesions: characterization at multiparametric MR imaging–influence of a computer-aided diagnosis system. Radiology 2014; 271:761.
Bratan F et al., How accurate is multiparametric MR imaging in evaluation of prostate cancer volume? Radiology 2015; 275:144.
Hoang Dinh A et al., Quantitative analysis of prostate multiparametric MR images for detection of aggressive prostate cancer in the peripheral zone: a multiple imager study Radiology 2016; 280:117.
None of these studies assessed automated prostate zonal segmentation.
Methodology
• retrospective
• diagnostic or prognostic study
• performed at one institution
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Rouvière, O., Moldovan, P.C., Vlachomitrou, A. et al. Combined model-based and deep learning-based automated 3D zonal segmentation of the prostate on T2-weighted MR images: clinical evaluation. Eur Radiol 32, 3248–3259 (2022). https://doi.org/10.1007/s00330-021-08408-5
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00330-021-08408-5