Incorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation
Digital mammography has seen an explosion in the number of radiomic features used for risk-assessment modeling. However, having more features is not necessarily beneficial, as some features may be overly sensitive to imaging physics (contrast, noise, and image sharpness). To measure the effects of i...
Guardado en:
| Autores principales: | , , , , , , , , , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
MDPI AG
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/e71709f14ea1497aa9a63b8d583d6da1 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| id |
oai:doaj.org-article:e71709f14ea1497aa9a63b8d583d6da1 |
|---|---|
| record_format |
dspace |
| spelling |
oai:doaj.org-article:e71709f14ea1497aa9a63b8d583d6da12021-11-11T15:33:44ZIncorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation10.3390/cancers132154972072-6694https://doaj.org/article/e71709f14ea1497aa9a63b8d583d6da12021-11-01T00:00:00Zhttps://www.mdpi.com/2072-6694/13/21/5497https://doaj.org/toc/2072-6694Digital mammography has seen an explosion in the number of radiomic features used for risk-assessment modeling. However, having more features is not necessarily beneficial, as some features may be overly sensitive to imaging physics (contrast, noise, and image sharpness). To measure the effects of imaging physics, we analyzed the feature variation across imaging acquisition settings (kV, mAs) using an anthropomorphic phantom. We also analyzed the intra-woman variation (IWV), a measure of how much a feature varies between breasts with similar parenchymal patterns—a woman’s left and right breasts. From 341 features, we identified “robust” features that minimized the effects of imaging physics and IWV. We also investigated whether robust features offered better case-control classification in an independent data set of 575 images, all with an overall BI-RADS<sup>®</sup> assessment of 1 (negative) or 2 (benign); 115 images (cases) were of women who developed cancer at least one year after that screening image, matched to 460 controls. We modeled cancer occurrence via logistic regression, using cross-validated area under the receiver-operating-characteristic curve (AUC) to measure model performance. Models using features from the most-robust quartile of features yielded an AUC = 0.59, versus 0.54 for the least-robust, with <i>p</i> < 0.005 for the difference among the quartiles.Raymond J. AcciavattiEric A. CohenOmid Haji MaghsoudiAimilia GastouniotiLauren PantaloneMeng-Kang HsiehEmily F. ConantChristopher G. ScottStacey J. WinhamKarla KerlikowskeCeline VachonAndrew D. A. MaidmentDespina KontosMDPI AGarticleradiomicsdigital mammographyrobustnessfeature selectionanthropomorphic phantomcase-control analysisNeoplasms. Tumors. Oncology. Including cancer and carcinogensRC254-282ENCancers, Vol 13, Iss 5497, p 5497 (2021) |
| institution |
DOAJ |
| collection |
DOAJ |
| language |
EN |
| topic |
radiomics digital mammography robustness feature selection anthropomorphic phantom case-control analysis Neoplasms. Tumors. Oncology. Including cancer and carcinogens RC254-282 |
| spellingShingle |
radiomics digital mammography robustness feature selection anthropomorphic phantom case-control analysis Neoplasms. Tumors. Oncology. Including cancer and carcinogens RC254-282 Raymond J. Acciavatti Eric A. Cohen Omid Haji Maghsoudi Aimilia Gastounioti Lauren Pantalone Meng-Kang Hsieh Emily F. Conant Christopher G. Scott Stacey J. Winham Karla Kerlikowske Celine Vachon Andrew D. A. Maidment Despina Kontos Incorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation |
| description |
Digital mammography has seen an explosion in the number of radiomic features used for risk-assessment modeling. However, having more features is not necessarily beneficial, as some features may be overly sensitive to imaging physics (contrast, noise, and image sharpness). To measure the effects of imaging physics, we analyzed the feature variation across imaging acquisition settings (kV, mAs) using an anthropomorphic phantom. We also analyzed the intra-woman variation (IWV), a measure of how much a feature varies between breasts with similar parenchymal patterns—a woman’s left and right breasts. From 341 features, we identified “robust” features that minimized the effects of imaging physics and IWV. We also investigated whether robust features offered better case-control classification in an independent data set of 575 images, all with an overall BI-RADS<sup>®</sup> assessment of 1 (negative) or 2 (benign); 115 images (cases) were of women who developed cancer at least one year after that screening image, matched to 460 controls. We modeled cancer occurrence via logistic regression, using cross-validated area under the receiver-operating-characteristic curve (AUC) to measure model performance. Models using features from the most-robust quartile of features yielded an AUC = 0.59, versus 0.54 for the least-robust, with <i>p</i> < 0.005 for the difference among the quartiles. |
| format |
article |
| author |
Raymond J. Acciavatti Eric A. Cohen Omid Haji Maghsoudi Aimilia Gastounioti Lauren Pantalone Meng-Kang Hsieh Emily F. Conant Christopher G. Scott Stacey J. Winham Karla Kerlikowske Celine Vachon Andrew D. A. Maidment Despina Kontos |
| author_facet |
Raymond J. Acciavatti Eric A. Cohen Omid Haji Maghsoudi Aimilia Gastounioti Lauren Pantalone Meng-Kang Hsieh Emily F. Conant Christopher G. Scott Stacey J. Winham Karla Kerlikowske Celine Vachon Andrew D. A. Maidment Despina Kontos |
| author_sort |
Raymond J. Acciavatti |
| title |
Incorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation |
| title_short |
Incorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation |
| title_full |
Incorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation |
| title_fullStr |
Incorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation |
| title_full_unstemmed |
Incorporating Robustness to Imaging Physics into Radiomic Feature Selection for Breast Cancer Risk Estimation |
| title_sort |
incorporating robustness to imaging physics into radiomic feature selection for breast cancer risk estimation |
| publisher |
MDPI AG |
| publishDate |
2021 |
| url |
https://doaj.org/article/e71709f14ea1497aa9a63b8d583d6da1 |
| work_keys_str_mv |
AT raymondjacciavatti incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT ericacohen incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT omidhajimaghsoudi incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT aimiliagastounioti incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT laurenpantalone incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT mengkanghsieh incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT emilyfconant incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT christophergscott incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT staceyjwinham incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT karlakerlikowske incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT celinevachon incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT andrewdamaidment incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation AT despinakontos incorporatingrobustnesstoimagingphysicsintoradiomicfeatureselectionforbreastcancerriskestimation |
| _version_ |
1718435215744958464 |