Article
The Reproducibility of Deep Learning-Based Segmentation of
the Prostate Gland and Zones on T2-Weighted MR Images
Mohammed R. S. Sunoqrot
1,
* , Kirsten M. Selnæs
1,2
, Elise Sandsmark
2
, Sverre Langørgen
2
,
Helena Bertilsson
3,4
, Tone F. Bathen
1,2
and Mattijs Elschot
1,2
Citation: Sunoqrot, M.R.S.; Selnæs,
K.M.; Sandsmark, E.; Langørgen, S.;
Bertilsson, H.; Bathen, T.F.; Elschot, M.
The Reproducibility of Deep
Learning-Based Segmentation of the
Prostate Gland and Zones on
T2-Weighted MR Images. Diagnostics
2021, 11, 1690. https://doi.org/
10.3390/diagnostics11091690
Academic Editors: Keun Ho Ryu and
Nipon Theera-Umpon
Received: 10 August 2021
Accepted: 15 September 2021
Published: 16 September 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
1
Department of Circulation and Medical Imaging, NTNU—Norwegian University of Science and Technology,
7030 Trondheim, Norway; kirsten.margrete.selnes@stolav.no (K.M.S.); tone.f.bathen@ntnu.no (T.F.B.);
mattijs.elschot@ntnu.no (M.E.)
2
Department of Radiology and Nuclear Medicine, St. Olavs Hospital, Trondheim University Hospital,
7030 Trondheim, Norway; elise.sandsmark@stolav.no (E.S.); sverre.langorgen@stolav.no (S.L.)
3
Department of Cancer Research and Molecular Medicine, NTNU—Norwegian University of Science and
Technology, 7030 Trondheim, Norway; helena.bertilsson@ntnu.no
4
Department of Urology, St. Olavs Hospital, Trondheim University Hospital, 7030 Trondheim, Norway
* Correspondence: mohammed.sunoqrot@ntnu.no
Abstract:
Volume of interest segmentation is an essential step in computer-aided detection and
diagnosis (CAD) systems. Deep learning (DL)-based methods provide good performance for prostate
segmentation, but little is known about the reproducibility of these methods. In this work, an in-
house collected dataset from 244 patients was used to investigate the intra-patient reproducibility of
14 shape features for DL-based segmentation methods of the whole prostate gland (WP), peripheral
zone (PZ), and the remaining prostate zones (non-PZ) on T2-weighted (T2W) magnetic resonance
(MR) images compared to manual segmentations. The DL-based segmentation was performed
using three different convolutional neural networks (CNNs): V-Net, nnU-Net-2D, and nnU-Net-3D.
The two-way random, single score intra-class correlation coefficient (ICC) was used to measure
the inter-scan reproducibility of each feature for each CNN and the manual segmentation. We
found that the reproducibility of the investigated methods is comparable to manual for all CNNs
(14/14 features), except for V-Net in PZ (7/14 features). The ICC score for segmentation volume
was found to be 0.888, 0.607, 0.819, and 0.903 in PZ; 0.988, 0.967, 0.986, and 0.983 in non-PZ;
0.982, 0.975, 0.973, and 0.984 in WP for manual, V-Net, nnU-Net-2D, and nnU-Net-3D, respec-
tively. The results of this work show the feasibility of embedding DL-based segmentation in CAD
systems, based on multiple T2W MR scans of the prostate, which is an important step towards the
clinical implementation.
Keywords: prostate; segmentation; deep learning; MRI; computer-aided diagnosis
1. Introduction
Prostate cancer is the most detected cancer in men and the second most common cause
of cancer related death for men worldwide [
1
]. An early diagnosis of prostate cancer is
essential for a better disease management [
2
]. Following reasonable suspicion of prostate
cancer, based on elevated prostate-specific antigen (PSA) levels in blood and a digital rectal
examination (DRE), the patient, in many countries, is likely to be referred to a pre-biopsy
magnetic resonance imaging (MRI) to guide the collection of biopsies [
3
]. To improve
the diagnostic process, the use of multi-parametric MRI (mpMRI) has been established
through international guidelines [
4
–
6
]. Additionally, mpMRI has been employed in active
surveillance programs to follow up the patients with indolent lesions [
7
], prostate cancer
risk calculators [
8
], and treatment response monitoring [
6
,
9
]. Currently, the mpMR images
are interpreted qualitatively by a radiologist, which is a tedious, time-consuming [
10
],
and reader opinion-dependent [
11
,
12
] process. The resulting vulnerability to inter and
Diagnostics 2021, 11, 1690. https://doi.org/10.3390/diagnostics11091690 https://www.mdpi.com/journal/diagnostics
