Repeatability of image features extracted from FET PET in application to post-surgical glioblastoma assessment

Barry, Nathaniel; Rowshanfarzad, Pejman; Francis, Roslyn J.; Nowak, Anna K.; Ebert, Martin A.

doi:10.1007/s13246-021-01049-4

Cited by 9 publications

(6 citation statements)

References 62 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…variability have been assessed using coe cient of variation (CV) [15,17], intra-class correlation coe cient (ICC) [16, [19][20][21][22], and Spearman's correlation coe cient [16,22].…”

Section: Discussionmentioning

confidence: 99%

“…In previous studies, Friedman test was applied for evaluation of ComBat performance [25][26][27], whereas CV and ICC were frequently applied to assess feature variance and inter-rater reliability and have been complemented by Spearman's rank correlation coe cient [15][16][17][19][20][21][22]. Friedman test quanti es signi cant differences in feature distributions considering the paired nature of feature values of each patient.…”

Section: Discussionmentioning

confidence: 99%

“…To properly translate radiomic models into clinical routine, they should be validated on large datasets that preferably include data from multiple centers and thus improve reproducibility and generalizability of radiomics analyses. However, reports have shown that features are sensitive to variations of several factors, including image acquisition, image reconstruction, tumor segmentation, as well as test-retest imaging [15][16][17][18][19][20][21][22]. Thus, it is essential to ensure the reproducibility and robustness of features in this regard.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Radiomic features derived from [18F]FET or [18F]GE-180 PET images of gliomas – does feature harmonization facilitate data pooling?

Zounek

Albert

Holzgreve

et al. 2022

Preprint

View full text Add to dashboard Cite

Purpose Large datasets are required to ensure reliable non-invasive glioma assessment with radiomics-based machine learning methods. This can often only be achieved by pooling images from different centers. Moreover, trained models should perform with high accuracy when applied to data from different centers. In this study, the impact of reconstruction settings and segmentation methods on radiomic features derived from amino acid and TSPO PET images of glioma patients was examined. Additionally, the ability to model and thus reduce feature differences was investigated. Methods [ 18 F]FET and [ 18 F]GE-180 PET data were acquired from 19 glioma patients. For each acquisition, 10 reconstruction settings and 9 segmentation methods were included to emulate multicentric data. Statistical robustness measures were calculated before and after ComBat harmonization. Differences between features due to setting variations were assessed using Friedman test, coefficient of variation (CV) and inter-rater reliability measures, including intraclass and Spearman’s rank correlation coefficients and Fleiss’ Kappa. Results According to Friedman analyses, most features (> 60%) showed significant differences. Yet, CV and inter-rater reliability measures indicated higher robustness. ComBat resulted in almost complete harmonization (> 87%) according to Friedman test and little to no improvement according to CV and inter-rater reliability measures. [ 18 F]GE-180 features displayed higher sensitivity to reconstruction settings than [ 18 F]FET features. Conclusions According to Friedman test, feature distributions could be successfully aligned using ComBat. However, depending on settings, changes in patient ranks were observed for some features and could not be eliminated by harmonization. Thus, for clinical utilization it is recommended to exclude affected features.

show abstract

“…variability have been assessed using coe cient of variation (CV) [15,17], intra-class correlation coe cient (ICC) [16, [19][20][21][22], and Spearman's correlation coe cient [16,22].…”

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Radiomic features derived from [18F]FET or [18F]GE-180 PET images of gliomas – does feature harmonization facilitate data pooling?

Zounek

Albert

Holzgreve

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…It remains to be investigated if, for example, deep learning-based image harmonisation can be used to correct for both the batch effects and patient ranks simultaneously. Further frequently employed statistical measures for evaluating the robustness of radiomic features include the coefficient of variation (CV) and the intra-class correlation coefficient (ICC) 41 74 75 . This huge variety of statistical measures and their implications shows how important the correct choice of statistical metrics is for the assessment and optimisation of harmonisation techniques 33 71 .…”

Section: Data Preparation and Harmonisation For Multi-centre Applicat...mentioning

confidence: 99%

Enhancing Interoperability and Harmonisation of Nuclear Medicine Image Data and Associated Clinical Data

Fuchs,

Kaiser,

Müller

et al. 2023

Nuklearmedizin

View full text Add to dashboard Cite

Nuclear imaging techniques such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) in combination with computed tomography (CT) are established imaging modalities in clinical practice, particularly for oncological problems. Due to a multitude of manufacturers, different measurement protocols, local demographic or clinical workflow variations as well as various available reconstruction and analysis software, very heterogeneous datasets are generated. This review article examines the current state of interoperability and harmonisation of image data and related clinical data in the field of nuclear medicine. Various approaches and standards to improve data compatibility and integration are discussed. These include, for example, structured clinical history, standardisation of image acquisition and reconstruction as well as standardised preparation of image data for evaluation. Approaches to improve data acquisition, storage and analysis will be presented. Furthermore, approaches are presented to prepare the datasets in such a way that they become usable for projects applying artificial intelligence (AI) (machine learning, deep learning, etc.). This review article concludes with an outlook on future developments and trends related to AI in nuclear medicine, including a brief research of commercial solutions.

show abstract

“…Orlhac et al [25][26][27] used Friedman test and the equivalent Wilcoxon test to validate the ComBat harmonization method. Differences between scanners or reconstruction algorithms and test-retest variability have been assessed using coe cient of variation (CV) [15,17], intra-class correlation coe cient (ICC) [16, [19][20][21][22], and Spearman's correlation coe cient [16,22].…”

Section: Statistical Analysis ( )mentioning

confidence: 99%

Radiomic features derived from [18F]FET or [18F]GE-180 PET images of gliomas – does feature harmonization facilitate data pooling?

Zounek¹,

Albert²,

Holzgreve³

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: Large datasets are required to ensure reliable non-invasive glioma assessment with radiomics-based machine learning methods. This can often only be achieved by pooling images from different centers. Moreover, trained models should perform with high accuracy when applied to data from different centers. In this study, the impact of reconstruction settings and segmentation methods on radiomic features derived from amino acid and TSPO PET images of glioma patients was examined. Additionally, the ability to model and thus reduce feature differences was investigated. [18F]FET and [18F]GE-180 PET data were acquired from 19 glioma patients. For each acquisition, 10 reconstruction settings and 9 segmentation methods were included to emulate multicentric data. Statistical robustness measures were calculated before and after ComBat harmonization. Differences between features due to setting variations were assessed using Friedman test, coefficient of variation (CV) and inter-rater reliability measures, including intraclass and Spearman’s rank correlation coefficients and Fleiss’ Kappa.Results: According to Friedman analyses, most features (> 60%) showed significant differences. Yet, CV and inter-rater reliability measures indicated higher robustness. ComBat resulted in almost complete harmonization (> 87%) according to Friedman test and little to no improvement according to CV and inter-rater reliability measures. [18F]GE-180 features displayed higher sensitivity to reconstruction settings than [18F]FET features.Conclusions: According to Friedman test, feature distributions could be successfully aligned using ComBat. However, depending on settings, changes in patient ranks were observed for some features and could not be eliminated by harmonization. Thus, for clinical utilization it is recommended to exclude affected features.

show abstract

Repeatability of image features extracted from FET PET in application to post-surgical glioblastoma assessment

Cited by 9 publications

References 62 publications

Radiomic features derived from [18F]FET or [18F]GE-180 PET images of gliomas – does feature harmonization facilitate data pooling?

Radiomic features derived from [18F]FET or [18F]GE-180 PET images of gliomas – does feature harmonization facilitate data pooling?

Enhancing Interoperability and Harmonisation of Nuclear Medicine Image Data and Associated Clinical Data

Radiomic features derived from [18F]FET or [18F]GE-180 PET images of gliomas – does feature harmonization facilitate data pooling?

Contact Info

Product

Resources

About