Deep learning algorithm performs similarly to radiologists in the assessment of prostate volume on MRI
Thimansson, Erick LU ; Bengtsson, J. LU
; Baubeta, E.
LU
; Engman, J.
LU
; Flondell-Sité, D.
LU
; Bjartell, A.
LU
and Zackrisson, S.
LU
(2023)
In European Radiology
33.
p.2519-2528
- Abstract
- Objectives
Prostate volume (PV) in combination with prostate specific antigen (PSA) yields PSA density which is an increasingly important biomarker. Calculating PV from MRI is a time-consuming, radiologist-dependent task. The aim of this study was to assess whether a deep learning algorithm can replace PI-RADS 2.1 based ellipsoid formula (EF) for calculating PV.
Methods
Eight different measures of PV were retrospectively collected for each of 124 patients who underwent radical prostatectomy and preoperative MRI of the prostate (multicenter and multi-scanner MRI’s 1.5 and 3 T). Agreement between volumes obtained from the deep learning algorithm (PVDL) and ellipsoid formula by two radiologists (PVEF1 and PVEF2) was... (More) - Objectives
Prostate volume (PV) in combination with prostate specific antigen (PSA) yields PSA density which is an increasingly important biomarker. Calculating PV from MRI is a time-consuming, radiologist-dependent task. The aim of this study was to assess whether a deep learning algorithm can replace PI-RADS 2.1 based ellipsoid formula (EF) for calculating PV.
Methods
Eight different measures of PV were retrospectively collected for each of 124 patients who underwent radical prostatectomy and preoperative MRI of the prostate (multicenter and multi-scanner MRI’s 1.5 and 3 T). Agreement between volumes obtained from the deep learning algorithm (PVDL) and ellipsoid formula by two radiologists (PVEF1 and PVEF2) was evaluated against the reference standard PV obtained by manual planimetry by an expert radiologist (PVMPE). A sensitivity analysis was performed using a prostatectomy specimen as the reference standard. Inter-reader agreement was evaluated between the radiologists using the ellipsoid formula and between the expert and inexperienced radiologists performing manual planimetry.
Results
PVDL showed better agreement and precision than PVEF1 and PVEF2 using the reference standard PVMPE (mean difference [95% limits of agreement] PVDL: −0.33 [−10.80; 10.14], PVEF1: −3.83 [−19.55; 11.89], PVEF2: −3.05 [−18.55; 12.45]) or the PV determined based on specimen weight (PVDL: −4.22 [−22.52; 14.07], PVEF1: −7.89 [−30.50; 14.73], PVEF2: −6.97 [−30.13; 16.18]). Inter-reader agreement was excellent between the two experienced radiologists using the ellipsoid formula and was good between expert and inexperienced radiologists performing manual planimetry.
Conclusion
Deep learning algorithm performs similarly to radiologists in the assessment of prostate volume on MRI.
Key Points
• A commercially available deep learning algorithm performs similarly to radiologists in the assessment of prostate volume on MRI.
• The deep-learning algorithm was previously untrained on this heterogenous multicenter day-to-day practice MRI data set. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/708d6f8d-37fd-48e2-b8fd-38728ee5bf74
- author
- Thimansson, Erick
LU
; Bengtsson, J.
LU
; Baubeta, E.
LU
; Engman, J.
LU
; Flondell-Sité, D.
LU
; Bjartell, A.
LU
and Zackrisson, S.
LU
- organization
- publishing date
- 2023
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Magnetic resonance imaging, Prostate neoplasms, Deep learning, Prostate-specific antigen
- in
- European Radiology
- volume
- 33
- pages
- 2519 - 2528
- publisher
- Springer
- external identifiers
-
- pmid:36371606
- scopus:85141694135
- ISSN
- 0938-7994
- DOI
- 10.1007/s00330-022-09239-8
- language
- English
- LU publication?
- yes
- id
- 708d6f8d-37fd-48e2-b8fd-38728ee5bf74
- date added to LUP
- 2022-11-14 10:55:58
- date last changed
- 2024-05-16 14:11:19
@article{708d6f8d-37fd-48e2-b8fd-38728ee5bf74,
abstract = {{Objectives<br/><br/>Prostate volume (PV) in combination with prostate specific antigen (PSA) yields PSA density which is an increasingly important biomarker. Calculating PV from MRI is a time-consuming, radiologist-dependent task. The aim of this study was to assess whether a deep learning algorithm can replace PI-RADS 2.1 based ellipsoid formula (EF) for calculating PV.<br/>Methods<br/><br/>Eight different measures of PV were retrospectively collected for each of 124 patients who underwent radical prostatectomy and preoperative MRI of the prostate (multicenter and multi-scanner MRI’s 1.5 and 3 T). Agreement between volumes obtained from the deep learning algorithm (PVDL) and ellipsoid formula by two radiologists (PVEF1 and PVEF2) was evaluated against the reference standard PV obtained by manual planimetry by an expert radiologist (PVMPE). A sensitivity analysis was performed using a prostatectomy specimen as the reference standard. Inter-reader agreement was evaluated between the radiologists using the ellipsoid formula and between the expert and inexperienced radiologists performing manual planimetry.<br/>Results<br/><br/>PVDL showed better agreement and precision than PVEF1 and PVEF2 using the reference standard PVMPE (mean difference [95% limits of agreement] PVDL: −0.33 [−10.80; 10.14], PVEF1: −3.83 [−19.55; 11.89], PVEF2: −3.05 [−18.55; 12.45]) or the PV determined based on specimen weight (PVDL: −4.22 [−22.52; 14.07], PVEF1: −7.89 [−30.50; 14.73], PVEF2: −6.97 [−30.13; 16.18]). Inter-reader agreement was excellent between the two experienced radiologists using the ellipsoid formula and was good between expert and inexperienced radiologists performing manual planimetry.<br/>Conclusion<br/><br/>Deep learning algorithm performs similarly to radiologists in the assessment of prostate volume on MRI.<br/>Key Points<br/><br/>• A commercially available deep learning algorithm performs similarly to radiologists in the assessment of prostate volume on MRI.<br/><br/>• The deep-learning algorithm was previously untrained on this heterogenous multicenter day-to-day practice MRI data set.}},
author = {{Thimansson, Erick and Bengtsson, J. and Baubeta, E. and Engman, J. and Flondell-Sité, D. and Bjartell, A. and Zackrisson, S.}},
issn = {{0938-7994}},
keywords = {{Magnetic resonance imaging; Prostate neoplasms; Deep learning; Prostate-specific antigen}},
language = {{eng}},
pages = {{2519--2528}},
publisher = {{Springer}},
series = {{European Radiology}},
title = {{Deep learning algorithm performs similarly to radiologists in the assessment of prostate volume on MRI}},
url = {{http://dx.doi.org/10.1007/s00330-022-09239-8}},
doi = {{10.1007/s00330-022-09239-8}},
volume = {{33}},
year = {{2023}},
}