Prediction of malignancy by a radiomic signature from contrast agent-free diffusion MRI in suspicious breast lesions found on screening mammography.

• Verfasst von: Bickelhaupt, Sebastian [VerfasserIn]
• Verfasst von: Vollmuth, Philipp [VerfasserIn]
• Titel: Prediction of malignancy by a radiomic signature from contrast agent-free diffusion MRI in suspicious breast lesions found on screening mammography.
• Verf.angabe: Sebastian Bickelhaupt, MD, Daniel Paech, MD, Philipp Kickingereder, MD, Franziska Steudle, Wolfgang Lederer, MD, Heidi Daniel, MD, Michael Götz, PhD, Nils Gählert, Diana Tichy, PhD, Manuel Wiesenfarth, PhD, Frederik B. Laun, PhD, Klaus H. Maier‐Hein, PhD, Heinz-Peter Schlemmer, MD, PhD, and David Bonekamp, MD
• Jahr: 2017
• Umfang: 13 S.
• Teil: volume:46
• Teil: year:2017
• Teil: number:2
• Teil: pages:604-616
• Teil: extent:13
• Fussnoten: First published: 02 February 2017 ; Gesehen am 29.06.2018
• Titel Quelle: Enthalten in: Journal of magnetic resonance imaging
• Ort Quelle: New York, NY : Wiley-Liss, 1991
• Jahr Quelle: 2017
• Band/Heft Quelle: 46(2017), 2, Seite 604-616
• ISSN Quelle: 1522-2586
• DOI: doi:10.1002/jmri.25606
• Datenträger: Online-Ressource
• Sprache: eng
• Sach-SW: apparent diffusion coefficient
• Sach-SW: diffusion-weighted imaging with background suppression
• Sach-SW: DWIBS
• Sach-SW: magnetic resonance
• Sach-SW: mammography
• Sach-SW: radiomics
• K10plus-PPN: 1577032306

Abstract

Purpose To assess radiomics as a tool to determine how well lesions found suspicious on breast cancer screening X-ray mammography can be categorized into malignant and benign with unenhanced magnetic resonance (MR) mammography with diffusion-weighted imaging and T2-weighted sequences. Materials and Methods From an asymptomatic screening cohort, 50 women with mammographically suspicious findings were examined with contrast-enhanced breast MRI (ceMRI) at 1.5T. Out of this protocol an unenhanced, abbreviated diffusion-weighted imaging protocol (ueMRI) including T2-weighted, (T2w), diffusion-weighted imaging (DWI), and DWI with background suppression (DWIBS) sequences and corresponding apparent diffusion coefficient (ADC) maps were extracted. From ueMRI-derived radiomic features, three Lasso-supervised machine-learning classifiers were constructed and compared with the clinical performance of a highly experienced radiologist: 1) univariate mean ADC model, 2) unconstrained radiomic model, 3) constrained radiomic model with mandatory inclusion of mean ADC. Results The unconstrained and constrained radiomic classifiers consisted of 11 parameters each and achieved differentiation of malignant from benign lesions with a .632 + bootstrap receiver operating characteristics (ROC) area under the curve (AUC) of 84.2%/85.1%, compared to 77.4% for mean ADC and 95.9%/95.9% for the experienced radiologist using ceMRI/ueMRI. Conclusion In this pilot study we identified two ueMRI radiomics classifiers that performed well in the differentiation of malignant from benign lesions and achieved higher performance than the mean ADC parameter alone. Classification was lower than the almost perfect performance of a highly experienced breast radiologist. The potential of radiomics to provide a training-independent diagnostic decision tool is indicated. A performance reaching the human expert would be highly desirable and based on our results is considered possible when the concept is extended in larger cohorts with further development and validation of the technique. Level of Evidence: 1 Technical Efficacy: Stage 2