Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility Study

<b>Background/Objectives:</b> Tenosynovitis is a common feature of psoriatic arthritis (PsA) and is typically assessed using semi-quantitative magnetic resonance imaging (MRI) scoring. However, visual scoring s variability. This study evaluates a fully automated, deep-learning approach f...

Full description

Saved in:
Bibliographic Details
Main Authors: Saeed Arbabi, Vahid Arbabi, Lorenzo Costa, Iris ten Katen, Simon C. Mastbergen, Peter R. Seevinck, Pim A. de Jong, Harrie Weinans, Mylène P. Jansen, Wouter Foppen
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Diagnostics
Subjects:
Online Access:https://www.mdpi.com/2075-4418/15/12/1469
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849432022602219520
author Saeed Arbabi
Vahid Arbabi
Lorenzo Costa
Iris ten Katen
Simon C. Mastbergen
Peter R. Seevinck
Pim A. de Jong
Harrie Weinans
Mylène P. Jansen
Wouter Foppen
author_facet Saeed Arbabi
Vahid Arbabi
Lorenzo Costa
Iris ten Katen
Simon C. Mastbergen
Peter R. Seevinck
Pim A. de Jong
Harrie Weinans
Mylène P. Jansen
Wouter Foppen
author_sort Saeed Arbabi
collection DOAJ
description <b>Background/Objectives:</b> Tenosynovitis is a common feature of psoriatic arthritis (PsA) and is typically assessed using semi-quantitative magnetic resonance imaging (MRI) scoring. However, visual scoring s variability. This study evaluates a fully automated, deep-learning approach for ankle tenosynovitis segmentation and volume-based quantification from MRI in psoriatic arthritis (PsA) patients. <b>Methods:</b> We analyzed 364 ankle 3T MRI scans from 71 PsA patients. Four tenosynovitis pathologies were manually scored and used to create ground truth segmentations through a human–machine workflow. For each pathology, 30 annotated scans were used to train a deep-learning segmentation model based on the nnUNet framework, and 20 scans were used for testing, ensuring patient-level disjoint sets. Model performance was evaluated using Dice scores. Volumetric pathology measurements from test scans were compared to radiologist scores using Spearman correlation. Additionally, 218 serial MRI pairs were assessed to analyze the relationship between changes in pathology volume and changes in visual scores. <b>Results:</b> The segmentation model achieved promising performance on the test set, with mean Dice scores ranging from 0.84 to 0.92. Pathology volumes correlated with visual scores across all test MRIs (Spearman ρ = 0.52–0.62). Volume-based quantification captured changes in inflammation over time and identified subtle progression not reflected in semi-quantitative scores. <b>Conclusions:</b> Our automated segmentation tool enables fast and accurate quantification of ankle tenosynovitis in PsA patients. It may enhance sensitivity to disease progression and complement visual scoring through continuous, volume-based metrics.
format Article
id doaj-art-e4c21dfbbc71463fa9910d5f5c33a13b
institution Kabale University
issn 2075-4418
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Diagnostics
spelling doaj-art-e4c21dfbbc71463fa9910d5f5c33a13b2025-08-20T03:27:28ZengMDPI AGDiagnostics2075-44182025-06-011512146910.3390/diagnostics15121469Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility StudySaeed Arbabi0Vahid Arbabi1Lorenzo Costa2Iris ten Katen3Simon C. Mastbergen4Peter R. Seevinck5Pim A. de Jong6Harrie Weinans7Mylène P. Jansen8Wouter Foppen9Image Sciences Institute, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsDepartment of Orthopedics, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsImage Sciences Institute, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsDepartment of Radiology, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsDepartment of Rheumatology & Clinical Immunology, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsImage Sciences Institute, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsDepartment of Radiology, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsDepartment of Orthopedics, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsDepartment of Rheumatology & Clinical Immunology, University Medical Center Utrecht, 3584 CX Utrecht, The NetherlandsDepartment of Radiology, University Medical Center Utrecht, 3584 CX Utrecht, The Netherlands<b>Background/Objectives:</b> Tenosynovitis is a common feature of psoriatic arthritis (PsA) and is typically assessed using semi-quantitative magnetic resonance imaging (MRI) scoring. However, visual scoring s variability. This study evaluates a fully automated, deep-learning approach for ankle tenosynovitis segmentation and volume-based quantification from MRI in psoriatic arthritis (PsA) patients. <b>Methods:</b> We analyzed 364 ankle 3T MRI scans from 71 PsA patients. Four tenosynovitis pathologies were manually scored and used to create ground truth segmentations through a human–machine workflow. For each pathology, 30 annotated scans were used to train a deep-learning segmentation model based on the nnUNet framework, and 20 scans were used for testing, ensuring patient-level disjoint sets. Model performance was evaluated using Dice scores. Volumetric pathology measurements from test scans were compared to radiologist scores using Spearman correlation. Additionally, 218 serial MRI pairs were assessed to analyze the relationship between changes in pathology volume and changes in visual scores. <b>Results:</b> The segmentation model achieved promising performance on the test set, with mean Dice scores ranging from 0.84 to 0.92. Pathology volumes correlated with visual scores across all test MRIs (Spearman ρ = 0.52–0.62). Volume-based quantification captured changes in inflammation over time and identified subtle progression not reflected in semi-quantitative scores. <b>Conclusions:</b> Our automated segmentation tool enables fast and accurate quantification of ankle tenosynovitis in PsA patients. It may enhance sensitivity to disease progression and complement visual scoring through continuous, volume-based metrics.https://www.mdpi.com/2075-4418/15/12/1469tenosynovitisMRIdeep learning
spellingShingle Saeed Arbabi
Vahid Arbabi
Lorenzo Costa
Iris ten Katen
Simon C. Mastbergen
Peter R. Seevinck
Pim A. de Jong
Harrie Weinans
Mylène P. Jansen
Wouter Foppen
Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility Study
Diagnostics
tenosynovitis
MRI
deep learning
title Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility Study
title_full Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility Study
title_fullStr Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility Study
title_full_unstemmed Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility Study
title_short Deep Learning Based Automatic Ankle Tenosynovitis Quantification from MRI in Patients with Psoriatic Arthritis: A Feasibility Study
title_sort deep learning based automatic ankle tenosynovitis quantification from mri in patients with psoriatic arthritis a feasibility study
topic tenosynovitis
MRI
deep learning
url https://www.mdpi.com/2075-4418/15/12/1469
work_keys_str_mv AT saeedarbabi deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT vahidarbabi deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT lorenzocosta deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT iristenkaten deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT simoncmastbergen deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT peterrseevinck deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT pimadejong deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT harrieweinans deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT mylenepjansen deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy
AT wouterfoppen deeplearningbasedautomaticankletenosynovitisquantificationfrommriinpatientswithpsoriaticarthritisafeasibilitystudy