Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Sep:13435:570-579.
doi: 10.1007/978-3-031-16443-9_55. Epub 2022 Sep 16.

Atlas-based Semantic Segmentation of Prostate Zones

Jiazhen Zhang  1 , Rajesh Venkataraman  2 , Lawrence H Staib  1   3   4 , John A Onofrey  1   5   4
Affiliations

Atlas-based Semantic Segmentation of Prostate Zones

Jiazhen Zhang et al. Med Image Comput Comput Assist Interv. 2022 Sep.

Abstract

Segmentation of the prostate into specific anatomical zones is important for radiological assessment of prostate cancer in magnetic resonance imaging (MRI). Of particular interest is segmenting the prostate into two regions of interest: the central gland (CG) and peripheral zone (PZ). In this paper, we propose to integrate an anatomical atlas of prostate zone shape into a deep learning semantic segmentation framework to segment the CG and PZ in T2-weighted MRI. Our approach incorporates anatomical information in the form of a probabilistic prostate zone atlas and utilizes a dynamically controlled hyperparameter to combine the atlas with the semantic segmentation result. In addition to providing significantly improved segmentation performance, this hyperparameter is capable of being dynamically adjusted during the inference stage to provide users with a mechanism to refine the segmentation. We validate our approach using an external test dataset and demonstrate Dice similarity coefficient values (mean±SD) of 0.91±0.05 for the CG and 0.77±0.16 for the PZ that significantly improves upon the baseline segmentation results without the atlas. All code is publicly available on GitHub: https://github.com/OnofreyLab/prostate_atlas_segm_miccai2022.

Keywords: MRI; deep learning; image segmentation; probabilistic atlas; prostate.

PubMed Disclaimer

Figures

Fig. 1.
Fig. 1.. Atlas-based semantic segmentation framework.
Our proposed framework consists of four modules: (1) pre-processing to transform imaging data from the native space to the atlas space; (2) a deep neural network to segment the CG with the aid of a probabilistic CG shape atlas LAtlas; (3) a hyperparameter λ to dynamically adjust the atlas weight; and (4) post-processing to transform segmentation results from the atlas space back to the native space.
Fig. 2.
Fig. 2.. Quantitative results.
Dice overlap segmentation results comparing the ground truth, the baseline results, and atlas-based results. Boxplots show median and inter-quartile range and outliers. * and ** indicate statistically significant results (p <0.001) for PZ and CG, respectively, compared to baseline with WG mask results.
Fig. 3.
Fig. 3.. Dynamic segmentation with λ for one subject.
CG (orange) and PZ (blue) segmentation results comparing the proposed method using various λ values for a single subject to the baseline method with WG mask. The top row shows the axial views of segmentation results superimposed on T2W images. The middle and bottom rows show segmentation results superimposed on ground-truth labels (gray: PZ, white: CG) in axial and sagittal views, respectively. Images are displayed in atlas space.
Fig. 4.
Fig. 4.. Segmentation results for five subjects.
CG (orange) and PZ (blue) segmentation results using our proposed method (λ = 0.4) for 5 subjects overlayed on T2W images. Results show slices from the middle of prostate gland in axial views (top and middle rows) and sagittal view (bottom row). Images are displayed in atlas space.
See this image and copyright information in PMC

References

    1. Armato SG, Huisman H, Drukker K, Hadjiiski L, Kirby JS, Petrick N, Redmond G, Giger ML, Cha K, Mamonov A, et al.: Prostatex challenges for computerized classification of prostate lesions from multiparametric magnetic resonance images. Journal of Medical Imaging 5(4), 044501 (2018) - PMC - PubMed
    1. Bohlender S, Oksuz I, Mukhopadhyay A: A survey on shape-constraint deep learning for medical image segmentation. arXiv preprint arXiv:2101.07721 (2021) - PubMed
    1. Cuocolo R, Comelli A, Stefano A, Benfante V, Dahiya N, Stanzione A, Castaldo A, De Lucia DR, Yezzi A, Imbriaco M: Deep Learning Whole-Gland and Zonal Prostate Segmentation on a Public MRI Dataset. Journal of Magnetic Resonance Imaging p. jmri.27585 (2 2021) - PubMed
    1. Hoeks CMA, Barentsz JO, Hambrock T, Yakar D, Somford DM, Heijmink SWTPJ, Scheenen TWJ, Vos PC, Huisman H, van Oort IM, Witjes JA, Heerschap A, Futterer JJ: Prostate Cancer: Multiparametric MR Imaging for Detection, Localization, and Staging. Radiology 261(1), 46–66 (2011) - PubMed
    1. Hoopes A, Hoffmann M, Fischl B, Guttag J, Dalca AV: Hypermorph: amortized hyperparameter learning for image registration. In: International Conference on Information Processing in Medical Imaging. pp. 3–17. Springer; (2021)

LinkOut - more resources