The Image Biomarker Standardization Initiative: Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights

Philip Whybra; Alex Zwanenburg; Vincent Andrearczyk; Roger Schaer; Aditya P. Apte; Alexandre Ayotte; Bhakti Baheti; Spyridon Bakas; Andrea Bettinelli; Ronald Boellaard; Luca Boldrini; Irène Buvat; Gary J.R. Cook; Florian Dietsche; Nicola Dinapoli; Hubert S. Gabryś; Vicky Goh; Matthias Guckenberger; Mathieu Hatt; Mahdi Hosseinzadeh; Aditi Iyer; Jacopo Lenkowicz; Mahdi A.L. Loutfi; Steffen Löck; Francesca Marturano; Olivier Morin; Christophe Nioche; Fanny Orlhac; Sarthak Pati; Arman Rahmim; Seyed Masoud Rezaeijo; Christopher G. Rookyard; Mohammad R. Salmanpour; Andreas Schindele; Isaac Shiri; Emiliano Spezi; Stephanie Tanadini-Lang; Florent Tixier; Taman Upadhaya; Vincenzo Valentini; Joost J.M. van Griethuysen; Fereshteh Yousefirizi; Habib Zaidi; Henning Müller; Martin Vallières; Adrien Depeursinge

doi:10.1148/radiol.231319

The Image Biomarker Standardization Initiative: Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights

Philip Whybra, Alex Zwanenburg^*, Vincent Andrearczyk, Roger Schaer, Aditya P. Apte, Alexandre Ayotte, Bhakti Baheti, Spyridon Bakas, Andrea Bettinelli, Ronald Boellaard, Luca Boldrini, Irène Buvat, Gary J.R. Cook, Florian Dietsche, Nicola Dinapoli, Hubert S. Gabryś, Vicky Goh, Matthias Guckenberger, Mathieu Hatt, Mahdi HosseinzadehAditi Iyer, Jacopo Lenkowicz, Mahdi A.L. Loutfi, Steffen Löck, Francesca Marturano, Olivier Morin, Christophe Nioche, Fanny Orlhac, Sarthak Pati, Arman Rahmim, Seyed Masoud Rezaeijo, Christopher G. Rookyard, Mohammad R. Salmanpour, Andreas Schindele, Isaac Shiri, Emiliano Spezi, Stephanie Tanadini-Lang, Florent Tixier, Taman Upadhaya, Vincenzo Valentini, Joost J.M. van Griethuysen, Fereshteh Yousefirizi, Habib Zaidi, Henning Müller, Martin Vallières, Adrien Depeursinge

^*Corresponding author for this work

Cancer Imaging

Research output: Contribution to journal › Review article › peer-review

61 Citations (Scopus)

Abstract

Filters are commonly used to enhance specific structures and patterns in images, such as vessels or peritumoral regions, to enable clinical insights beyond the visible image using radiomics. However, their lack of standardization restricts reproducibility and clinical translation of radiomics decision support tools. In this special report, teams of researchers who developed radiomics software participated in a three-phase study (September 2020 to December 2022) to establish a standardized set of filters. The first two phases focused on finding reference filtered images and reference feature values for commonly used convolutional filters: mean, Laplacian of Gaussian, Laws and Gabor kernels, separable and nonseparable wavelets (including decomposed forms), and Riesz transformations. In the first phase, 15 teams used digital phantoms to establish 33 reference filtered images of 36 filter configurations. In phase 2, 11 teams used a chest CT image to derive reference values for 323 of 396 features computed from filtered images using 22 filter and image processing configurations. Reference filtered images and feature values for Riesz transformations were not established. Reproducibility of standardized convolutional filters was validated on a public data set of multimodal imaging (CT, fluorodeoxyglucose PET, and T1-weighted MRI) in 51 patients with soft-tissue sarcoma. At validation, reproducibility of 486 features computed from filtered images using nine configurations × three imaging modalities was assessed using the lower bounds of 95% CIs of intraclass correlation coefficients. Out of 486 features, 458 were found to be reproducible across nine teams with lower bounds of 95% CIs of intraclass correlation coefficients greater than 0.75. In conclusion, eight filter types were standardized with reference filtered images and reference feature values for verifying and calibrating radiomics software packages. A web-based tool is available for compliance checking.

Original language	English
Article number	231319
Journal	Radiology
Volume	310
Issue number	2
DOIs	https://doi.org/10.1148/radiol.231319
Publication status	Published - Feb 2024

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1148/radiol.231319

Cite this

Whybra, P., Zwanenburg, A., Andrearczyk, V., Schaer, R., Apte, A. P., Ayotte, A., Baheti, B., Bakas, S., Bettinelli, A., Boellaard, R., Boldrini, L., Buvat, I., Cook, G. J. R., Dietsche, F., Dinapoli, N., Gabryś, H. S., Goh, V., Guckenberger, M., Hatt, M., ... Depeursinge, A. (2024). The Image Biomarker Standardization Initiative: Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights. Radiology, 310(2), Article 231319. https://doi.org/10.1148/radiol.231319

@article{177ca682363b4f6cb55f501361f0ac2c,

title = "The Image Biomarker Standardization Initiative: Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights",

abstract = "Filters are commonly used to enhance specific structures and patterns in images, such as vessels or peritumoral regions, to enable clinical insights beyond the visible image using radiomics. However, their lack of standardization restricts reproducibility and clinical translation of radiomics decision support tools. In this special report, teams of researchers who developed radiomics software participated in a three-phase study (September 2020 to December 2022) to establish a standardized set of filters. The first two phases focused on finding reference filtered images and reference feature values for commonly used convolutional filters: mean, Laplacian of Gaussian, Laws and Gabor kernels, separable and nonseparable wavelets (including decomposed forms), and Riesz transformations. In the first phase, 15 teams used digital phantoms to establish 33 reference filtered images of 36 filter configurations. In phase 2, 11 teams used a chest CT image to derive reference values for 323 of 396 features computed from filtered images using 22 filter and image processing configurations. Reference filtered images and feature values for Riesz transformations were not established. Reproducibility of standardized convolutional filters was validated on a public data set of multimodal imaging (CT, fluorodeoxyglucose PET, and T1-weighted MRI) in 51 patients with soft-tissue sarcoma. At validation, reproducibility of 486 features computed from filtered images using nine configurations × three imaging modalities was assessed using the lower bounds of 95% CIs of intraclass correlation coefficients. Out of 486 features, 458 were found to be reproducible across nine teams with lower bounds of 95% CIs of intraclass correlation coefficients greater than 0.75. In conclusion, eight filter types were standardized with reference filtered images and reference feature values for verifying and calibrating radiomics software packages. A web-based tool is available for compliance checking.",

author = "Philip Whybra and Alex Zwanenburg and Vincent Andrearczyk and Roger Schaer and Apte, {Aditya P.} and Alexandre Ayotte and Bhakti Baheti and Spyridon Bakas and Andrea Bettinelli and Ronald Boellaard and Luca Boldrini and Ir{\`e}ne Buvat and Cook, {Gary J.R.} and Florian Dietsche and Nicola Dinapoli and Gabry{\'s}, {Hubert S.} and Vicky Goh and Matthias Guckenberger and Mathieu Hatt and Mahdi Hosseinzadeh and Aditi Iyer and Jacopo Lenkowicz and Loutfi, {Mahdi A.L.} and Steffen L{\"o}ck and Francesca Marturano and Olivier Morin and Christophe Nioche and Fanny Orlhac and Sarthak Pati and Arman Rahmim and Rezaeijo, {Seyed Masoud} and Rookyard, {Christopher G.} and Salmanpour, {Mohammad R.} and Andreas Schindele and Isaac Shiri and Emiliano Spezi and Stephanie Tanadini-Lang and Florent Tixier and Taman Upadhaya and Vincenzo Valentini and {van Griethuysen}, {Joost J.M.} and Fereshteh Yousefirizi and Habib Zaidi and Henning M{\"u}ller and Martin Valli{\`e}res and Adrien Depeursinge",

note = "Publisher Copyright: {\textcopyright} RSNA, 2024.",

year = "2024",

month = feb,

doi = "10.1148/radiol.231319",

language = "English",

volume = "310",

journal = "Radiology",

issn = "0033-8419",

publisher = "Radiological Society of North America",

number = "2",

}

Whybra, P, Zwanenburg, A, Andrearczyk, V, Schaer, R, Apte, AP, Ayotte, A, Baheti, B, Bakas, S, Bettinelli, A, Boellaard, R, Boldrini, L, Buvat, I, Cook, GJR, Dietsche, F, Dinapoli, N, Gabryś, HS, Goh, V, Guckenberger, M, Hatt, M, Hosseinzadeh, M, Iyer, A, Lenkowicz, J, Loutfi, MAL, Löck, S, Marturano, F, Morin, O, Nioche, C, Orlhac, F, Pati, S, Rahmim, A, Rezaeijo, SM, Rookyard, CG, Salmanpour, MR, Schindele, A, Shiri, I, Spezi, E, Tanadini-Lang, S, Tixier, F, Upadhaya, T, Valentini, V, van Griethuysen, JJM, Yousefirizi, F, Zaidi, H, Müller, H, Vallières, M & Depeursinge, A 2024, 'The Image Biomarker Standardization Initiative: Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights', Radiology, vol. 310, no. 2, 231319. https://doi.org/10.1148/radiol.231319

TY - JOUR

T1 - The Image Biomarker Standardization Initiative

T2 - Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights

AU - Whybra, Philip

AU - Zwanenburg, Alex

AU - Andrearczyk, Vincent

AU - Schaer, Roger

AU - Apte, Aditya P.

AU - Ayotte, Alexandre

AU - Baheti, Bhakti

AU - Bakas, Spyridon

AU - Bettinelli, Andrea

AU - Boellaard, Ronald

AU - Boldrini, Luca

AU - Buvat, Irène

AU - Cook, Gary J.R.

AU - Dietsche, Florian

AU - Dinapoli, Nicola

AU - Gabryś, Hubert S.

AU - Goh, Vicky

AU - Guckenberger, Matthias

AU - Hatt, Mathieu

AU - Hosseinzadeh, Mahdi

AU - Iyer, Aditi

AU - Lenkowicz, Jacopo

AU - Loutfi, Mahdi A.L.

AU - Löck, Steffen

AU - Marturano, Francesca

AU - Morin, Olivier

AU - Nioche, Christophe

AU - Orlhac, Fanny

AU - Pati, Sarthak

AU - Rahmim, Arman

AU - Rezaeijo, Seyed Masoud

AU - Rookyard, Christopher G.

AU - Salmanpour, Mohammad R.

AU - Schindele, Andreas

AU - Shiri, Isaac

AU - Spezi, Emiliano

AU - Tanadini-Lang, Stephanie

AU - Tixier, Florent

AU - Upadhaya, Taman

AU - Valentini, Vincenzo

AU - van Griethuysen, Joost J.M.

AU - Yousefirizi, Fereshteh

AU - Zaidi, Habib

AU - Müller, Henning

AU - Vallières, Martin

AU - Depeursinge, Adrien

PY - 2024/2

Y1 - 2024/2

N2 - Filters are commonly used to enhance specific structures and patterns in images, such as vessels or peritumoral regions, to enable clinical insights beyond the visible image using radiomics. However, their lack of standardization restricts reproducibility and clinical translation of radiomics decision support tools. In this special report, teams of researchers who developed radiomics software participated in a three-phase study (September 2020 to December 2022) to establish a standardized set of filters. The first two phases focused on finding reference filtered images and reference feature values for commonly used convolutional filters: mean, Laplacian of Gaussian, Laws and Gabor kernels, separable and nonseparable wavelets (including decomposed forms), and Riesz transformations. In the first phase, 15 teams used digital phantoms to establish 33 reference filtered images of 36 filter configurations. In phase 2, 11 teams used a chest CT image to derive reference values for 323 of 396 features computed from filtered images using 22 filter and image processing configurations. Reference filtered images and feature values for Riesz transformations were not established. Reproducibility of standardized convolutional filters was validated on a public data set of multimodal imaging (CT, fluorodeoxyglucose PET, and T1-weighted MRI) in 51 patients with soft-tissue sarcoma. At validation, reproducibility of 486 features computed from filtered images using nine configurations × three imaging modalities was assessed using the lower bounds of 95% CIs of intraclass correlation coefficients. Out of 486 features, 458 were found to be reproducible across nine teams with lower bounds of 95% CIs of intraclass correlation coefficients greater than 0.75. In conclusion, eight filter types were standardized with reference filtered images and reference feature values for verifying and calibrating radiomics software packages. A web-based tool is available for compliance checking.

AB - Filters are commonly used to enhance specific structures and patterns in images, such as vessels or peritumoral regions, to enable clinical insights beyond the visible image using radiomics. However, their lack of standardization restricts reproducibility and clinical translation of radiomics decision support tools. In this special report, teams of researchers who developed radiomics software participated in a three-phase study (September 2020 to December 2022) to establish a standardized set of filters. The first two phases focused on finding reference filtered images and reference feature values for commonly used convolutional filters: mean, Laplacian of Gaussian, Laws and Gabor kernels, separable and nonseparable wavelets (including decomposed forms), and Riesz transformations. In the first phase, 15 teams used digital phantoms to establish 33 reference filtered images of 36 filter configurations. In phase 2, 11 teams used a chest CT image to derive reference values for 323 of 396 features computed from filtered images using 22 filter and image processing configurations. Reference filtered images and feature values for Riesz transformations were not established. Reproducibility of standardized convolutional filters was validated on a public data set of multimodal imaging (CT, fluorodeoxyglucose PET, and T1-weighted MRI) in 51 patients with soft-tissue sarcoma. At validation, reproducibility of 486 features computed from filtered images using nine configurations × three imaging modalities was assessed using the lower bounds of 95% CIs of intraclass correlation coefficients. Out of 486 features, 458 were found to be reproducible across nine teams with lower bounds of 95% CIs of intraclass correlation coefficients greater than 0.75. In conclusion, eight filter types were standardized with reference filtered images and reference feature values for verifying and calibrating radiomics software packages. A web-based tool is available for compliance checking.

UR - http://www.scopus.com/inward/record.url?scp=85186751738&partnerID=8YFLogxK

U2 - 10.1148/radiol.231319

DO - 10.1148/radiol.231319

M3 - Review article

AN - SCOPUS:85186751738

SN - 0033-8419

VL - 310

JO - Radiology

JF - Radiology

IS - 2

M1 - 231319

ER -

The Image Biomarker Standardization Initiative: Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this