Deep learning for predicting future lesion emergence in high-risk breast MRI screening: a feasibility study.
Breast cancer
Deep learning
Generative adversarial network
High-risk women
Magnetic resonance imaging
Journal
European radiology experimental
ISSN: 2509-9280
Titre abrégé: Eur Radiol Exp
Pays: England
ID NLM: 101721752
Informations de publication
Date de publication:
07 06 2023
07 06 2023
Historique:
received:
25
01
2023
accepted:
04
04
2023
medline:
8
6
2023
pubmed:
7
6
2023
entrez:
6
6
2023
Statut:
epublish
Résumé
International societies have issued guidelines for high-risk breast cancer (BC) screening, recommending contrast-enhanced magnetic resonance imaging (CE-MRI) of the breast as a supplemental diagnostic tool. In our study, we tested the applicability of deep learning-based anomaly detection to identify anomalous changes in negative breast CE-MRI screens associated with future lesion emergence. In this prospective study, we trained a generative adversarial network on dynamic CE-MRI of 33 high-risk women who participated in a screening program but did not develop BC. We defined an anomaly score as the deviation of an observed CE-MRI scan from the model of normal breast tissue variability. We evaluated the anomaly score's association with future lesion emergence on the level of local image patches (104,531 normal patches, 455 patches of future lesion location) and entire CE-MRI exams (21 normal, 20 with future lesion). Associations were analyzed by receiver operating characteristic (ROC) curves on the patch level and logistic regression on the examination level. The local anomaly score on image patches was a good predictor for future lesion emergence (area under the ROC curve 0.804). An exam-level summary score was significantly associated with the emergence of lesions at any location at a later time point (p = 0.045). Breast cancer lesions are associated with anomalous appearance changes in breast CE-MRI occurring before the lesion emerges in high-risk women. These early image signatures are detectable and may be a basis for adjusting individual BC risk and personalized screening. Anomalies in screening MRI preceding lesion emergence in women at high-risk of breast cancer may inform individualized screening and intervention strategies. • Breast lesions are associated with preceding anomalies in CE-MRI of high-risk women. • Deep learning-based anomaly detection can help to adjust risk assessment for future lesions. • An appearance anomaly score may be used for adjusting screening interval times.
Sections du résumé
BACKGROUND
International societies have issued guidelines for high-risk breast cancer (BC) screening, recommending contrast-enhanced magnetic resonance imaging (CE-MRI) of the breast as a supplemental diagnostic tool. In our study, we tested the applicability of deep learning-based anomaly detection to identify anomalous changes in negative breast CE-MRI screens associated with future lesion emergence.
METHODS
In this prospective study, we trained a generative adversarial network on dynamic CE-MRI of 33 high-risk women who participated in a screening program but did not develop BC. We defined an anomaly score as the deviation of an observed CE-MRI scan from the model of normal breast tissue variability. We evaluated the anomaly score's association with future lesion emergence on the level of local image patches (104,531 normal patches, 455 patches of future lesion location) and entire CE-MRI exams (21 normal, 20 with future lesion). Associations were analyzed by receiver operating characteristic (ROC) curves on the patch level and logistic regression on the examination level.
RESULTS
The local anomaly score on image patches was a good predictor for future lesion emergence (area under the ROC curve 0.804). An exam-level summary score was significantly associated with the emergence of lesions at any location at a later time point (p = 0.045).
CONCLUSIONS
Breast cancer lesions are associated with anomalous appearance changes in breast CE-MRI occurring before the lesion emerges in high-risk women. These early image signatures are detectable and may be a basis for adjusting individual BC risk and personalized screening.
RELEVANCE STATEMENT
Anomalies in screening MRI preceding lesion emergence in women at high-risk of breast cancer may inform individualized screening and intervention strategies.
KEY POINTS
• Breast lesions are associated with preceding anomalies in CE-MRI of high-risk women. • Deep learning-based anomaly detection can help to adjust risk assessment for future lesions. • An appearance anomaly score may be used for adjusting screening interval times.
Identifiants
pubmed: 37280478
doi: 10.1186/s41747-023-00343-y
pii: 10.1186/s41747-023-00343-y
pmc: PMC10244308
doi:
Substances chimiques
Contrast Media
0
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
32Informations de copyright
© 2023. The Author(s).
Références
Radiology. 2012 Jul;264(1):40-50
pubmed: 22523324
Eur Radiol. 2019 Sep;29(9):4825-4832
pubmed: 30993432
Med Image Anal. 2019 May;54:30-44
pubmed: 30831356
Eur J Cancer. 2017 Nov;85:31-38
pubmed: 28886475
Br J Cancer. 2016 Mar 15;114(6):631-7
pubmed: 26908327
J Breast Imaging. 2021 Feb 19;3(2):144-155
pubmed: 33778488
Clin Cancer Res. 2019 Jun 1;25(11):3266-3275
pubmed: 31010833
Biom J. 2005 Aug;47(4):458-72
pubmed: 16161804
Diagn Interv Imaging. 2019 Apr;100(4):219-225
pubmed: 30926444
Sci Rep. 2021 May 13;11(1):10276
pubmed: 33986361
Neuroimage. 2006 Jul 1;31(3):1116-28
pubmed: 16545965
Br J Cancer. 2022 Mar;126(4):533-550
pubmed: 34703006
J Clin Oncol. 2015 Apr 1;33(10):1128-35
pubmed: 25713430
J Med Imaging (Bellingham). 2018 Jan;5(1):014502
pubmed: 29340287
Eur Radiol. 2020 Nov;30(11):6052-6061
pubmed: 32504098
Contrast Media Mol Imaging. 2020 Aug 28;2020:6805710
pubmed: 32934610
Radiology. 2019 Oct;293(1):38-46
pubmed: 31385754
BMJ Open. 2021 Sep 14;11(9):e047755
pubmed: 34521662
Wien Klin Wochenschr. 2012 May;124(9-10):334-9
pubmed: 22644217
Radiology. 2020 Feb;294(2):265-272
pubmed: 31845842
Clin Cancer Res. 2007 Oct 15;13(20):6144-52
pubmed: 17947480
Radiology. 2019 Jul;292(1):60-66
pubmed: 31063083
Diagnostics (Basel). 2020 Jul 04;10(7):
pubmed: 32635547
Breast Cancer Res Treat. 2018 Jun;169(2):323-331
pubmed: 29383629
Med Phys. 2016 Jan;43(1):84
pubmed: 26745902
Am J Surg. 2016 Oct;212(4):660-669
pubmed: 27649974
Cancer J Sci Am. 1999 Sep-Oct;5(5):283-92
pubmed: 10526669
Lancet Oncol. 2006 Mar;7(3):223-9
pubmed: 16510331
JAMA Intern Med. 2015 Nov;175(11):1828-37
pubmed: 26414882
Ann Intern Med. 2020 Nov 3;173(9):704-713
pubmed: 32866413
Int J Cancer. 2019 Apr 15;144(8):1941-1953
pubmed: 30350310
Sci Transl Med. 2021 Jan 27;13(578):
pubmed: 33504648
Br J Cancer. 2012 Jun 26;107(1):24-30
pubmed: 22588560
Acad Radiol. 2018 Nov;25(11):1472-1480
pubmed: 29606338
AJR Am J Roentgenol. 2013 Nov;201(5):1155-63
pubmed: 24147491
AJR Am J Roentgenol. 2020 Feb;214(2):316-323
pubmed: 31714845
Cells. 2019 Nov 26;8(12):
pubmed: 31779139
Eur J Radiol. 2018 Sep;106:150-159
pubmed: 30150038