In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioning

Abstract Despite academic success, radiomics-based machine learning algorithms have not reached clinical practice, partially due to limited repeatability/reproducibility. To address this issue, this work aims to identify a stable subset of radiomics features in prostate MRI for radiomics modelling....

Full description

Saved in:
Bibliographic Details
Main Authors: Kevin Sun Zhang, Christian Jan Oliver Neelsen, Markus Wennmann, Thomas Hielscher, Balint Kovacs, Philip Alexander Glemser, Magdalena Görtz, Albrecht Stenzinger, Klaus H. Maier-Hein, Johannes Huber, Heinz-Peter Schlemmer, David Bonekamp
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-09989-7
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849766738438127616
author Kevin Sun Zhang
Christian Jan Oliver Neelsen
Markus Wennmann
Thomas Hielscher
Balint Kovacs
Philip Alexander Glemser
Magdalena Görtz
Albrecht Stenzinger
Klaus H. Maier-Hein
Johannes Huber
Heinz-Peter Schlemmer
David Bonekamp
author_facet Kevin Sun Zhang
Christian Jan Oliver Neelsen
Markus Wennmann
Thomas Hielscher
Balint Kovacs
Philip Alexander Glemser
Magdalena Görtz
Albrecht Stenzinger
Klaus H. Maier-Hein
Johannes Huber
Heinz-Peter Schlemmer
David Bonekamp
author_sort Kevin Sun Zhang
collection DOAJ
description Abstract Despite academic success, radiomics-based machine learning algorithms have not reached clinical practice, partially due to limited repeatability/reproducibility. To address this issue, this work aims to identify a stable subset of radiomics features in prostate MRI for radiomics modelling. A prospective study was conducted in 43 patients who received a clinical MRI examination and a research exam with repetition of T2-weighted and two different diffusion-weighted imaging (DWI) sequences with repositioning in between. Radiomics feature (RF) extraction was performed from MRI segmentations accounting for intra-rater and inter-rater effects, and three different image normalization methods were compared. Stability of RFs was assessed using the concordance correlation coefficient (CCC) for different comparisons: rater effects, inter-scan (before and after repositioning) and inter-sequence (between the two diffusion-weighted sequences) variability. In total, only 64 out of 321 (~ 20%) extracted features demonstrated stability, defined as CCC ≥ 0.75 in all settings (5 high-b value, 7 ADC- and 52 T2-derived features). For DWI, primarily intensity-based features proved stable with no shape feature passing the CCC threshold. T2-weighted images possessed the largest number of stable features with multiple shape (7), intensity-based (7) and texture features (28). Z-score normalization for high-b value images and muscle-normalization for T2-weighted images were identified as suitable.
format Article
id doaj-art-6324e3a231f748aab83d97a5c6732eb0
institution DOAJ
issn 2045-2322
language English
publishDate 2025-08-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-6324e3a231f748aab83d97a5c6732eb02025-08-20T03:04:29ZengNature PortfolioScientific Reports2045-23222025-08-0115111410.1038/s41598-025-09989-7In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioningKevin Sun Zhang0Christian Jan Oliver Neelsen1Markus Wennmann2Thomas Hielscher3Balint Kovacs4Philip Alexander Glemser5Magdalena Görtz6Albrecht Stenzinger7Klaus H. Maier-Hein8Johannes Huber9Heinz-Peter Schlemmer10David Bonekamp11German Cancer Research Center (DKFZ), Division of RadiologyGerman Cancer Research Center (DKFZ), Division of RadiologyGerman Cancer Research Center (DKFZ), Division of RadiologyGerman Cancer Research Center (DKFZ), Division of BiostatisticsGerman Cancer Research Center (DKFZ), Division of Medical Image ComputingGerman Cancer Research Center (DKFZ), Division of RadiologyDepartment of Urology, Heidelberg University HospitalInstitute of Pathology, Heidelberg University HospitalGerman Cancer Research Center (DKFZ), Division of Medical Image ComputingDepartment of Urology, Heidelberg University HospitalGerman Cancer Research Center (DKFZ), Division of RadiologyGerman Cancer Research Center (DKFZ), Division of RadiologyAbstract Despite academic success, radiomics-based machine learning algorithms have not reached clinical practice, partially due to limited repeatability/reproducibility. To address this issue, this work aims to identify a stable subset of radiomics features in prostate MRI for radiomics modelling. A prospective study was conducted in 43 patients who received a clinical MRI examination and a research exam with repetition of T2-weighted and two different diffusion-weighted imaging (DWI) sequences with repositioning in between. Radiomics feature (RF) extraction was performed from MRI segmentations accounting for intra-rater and inter-rater effects, and three different image normalization methods were compared. Stability of RFs was assessed using the concordance correlation coefficient (CCC) for different comparisons: rater effects, inter-scan (before and after repositioning) and inter-sequence (between the two diffusion-weighted sequences) variability. In total, only 64 out of 321 (~ 20%) extracted features demonstrated stability, defined as CCC ≥ 0.75 in all settings (5 high-b value, 7 ADC- and 52 T2-derived features). For DWI, primarily intensity-based features proved stable with no shape feature passing the CCC threshold. T2-weighted images possessed the largest number of stable features with multiple shape (7), intensity-based (7) and texture features (28). Z-score normalization for high-b value images and muscle-normalization for T2-weighted images were identified as suitable.https://doi.org/10.1038/s41598-025-09989-7ProstateMagnetic resonance imagingReproducibility of resultsObserver variationRadiomics
spellingShingle Kevin Sun Zhang
Christian Jan Oliver Neelsen
Markus Wennmann
Thomas Hielscher
Balint Kovacs
Philip Alexander Glemser
Magdalena Görtz
Albrecht Stenzinger
Klaus H. Maier-Hein
Johannes Huber
Heinz-Peter Schlemmer
David Bonekamp
In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioning
Scientific Reports
Prostate
Magnetic resonance imaging
Reproducibility of results
Observer variation
Radiomics
title In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioning
title_full In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioning
title_fullStr In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioning
title_full_unstemmed In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioning
title_short In vivo variability of MRI radiomics features in prostate lesions assessed by a test-retest study with repositioning
title_sort in vivo variability of mri radiomics features in prostate lesions assessed by a test retest study with repositioning
topic Prostate
Magnetic resonance imaging
Reproducibility of results
Observer variation
Radiomics
url https://doi.org/10.1038/s41598-025-09989-7
work_keys_str_mv AT kevinsunzhang invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT christianjanoliverneelsen invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT markuswennmann invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT thomashielscher invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT balintkovacs invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT philipalexanderglemser invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT magdalenagortz invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT albrechtstenzinger invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT klaushmaierhein invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT johanneshuber invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT heinzpeterschlemmer invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning
AT davidbonekamp invivovariabilityofmriradiomicsfeaturesinprostatelesionsassessedbyatestreteststudywithrepositioning