Computerized Assessment of Breast Lesion Malignancy Using DCE-MRI Robustness Study on Two Independent Clinical Datasets from Two Manufacturers

Overview

Journal Acad Radiol

Specialty Radiology

Date 2010 Jun 15

PMID 20540907

Citations 23

Authors

Weijie Chen

Maryellen L Giger

Gillian M Newstead

Ulrich Bick

Sanaz A Jansen

Hui Li

Li Lan

Affiliations

Soon will be listed here.

Abstract

Rationale And Objectives: To conduct a preclinical evaluation of the robustness of our computerized system for breast lesion characterization on two breast magnetic resonance imaging (MRI) databases that were acquired using scanners from two different manufacturers.

Materials And Methods: Two clinical breast MRI databases were acquired from a Siemens scanner and a GE scanner, which shared similar imaging protocols and retrospectively collected under an institutional review board-approved protocol. In our computerized analysis system, after a breast lesion is identified by the radiologist, the computer performs automatic lesion segmentation and feature extraction and outputs an estimated probability of malignancy. We used a Bayesian neural network with automatic relevance determination for joint feature selection and classification. To evaluate the robustness of our classification system, we first used Database 1 for feature selection and classifier training, and Database 2 to test the trained classifier. Then, we exchanged the two datasets and repeated the process. Area under the receiver operating characteristic curve (AUC) was used as a performance figure of merit in the task of distinguishing between malignant and benign lesions.

Results: We obtained an AUC of 0.85 (approximate 95% confidence interval [CI] 0.79-0.91) for (a) feature selection and classifier training using Database 1 and testing on Database 2; and an AUC of 0.90 (approximate 95% CI 0.84-0.96) for (b) feature selection and classifier training using Database 2 and testing on Database 1. We failed to observe statistical significance for the difference AUC of 0.05 between the two database conditions (P = .24; 95% confidence interval -0.03, 0.1).

Conclusion: These results demonstrate the robustness of our computerized classification system in the task of distinguishing between malignant and benign breast lesions on dynamic contrast-enhanced (DCE) MRI images from two manufacturers. Our study showed the feasibility of developing a computerized classification system that is robust across different scanners.

Citing Articles

Vascularity and Dynamic Contrast-Enhanced Breast Magnetic Resonance Imaging.

Frankhouser D, Dietze E, Mahabal A, Seewaldt V Front Radiol. 2023; 1:735567.

PMID: 37492179 PMC: 10364989. DOI: 10.3389/fradi.2021.735567.

Enhancing physicians' radiology diagnostics of COVID-19's effects on lung health by leveraging artificial intelligence.

Gasulla O, Ledesma-Carbayo M, Borrell L, Fortuny-Profitos J, Mazaira-Font F, Allende J Front Bioeng Biotechnol. 2023; 11:1010679.

PMID: 37152658 PMC: 10157246. DOI: 10.3389/fbioe.2023.1010679.

Impact of continuous learning on diagnostic breast MRI AI: evaluation on an independent clinical dataset.

Li H, Whitney H, Ji Y, Edwards A, Papaioannou J, Liu P J Med Imaging (Bellingham). 2022; 9(3):034502.

PMID: 35685120 PMC: 9168763. DOI: 10.1117/1.JMI.9.3.034502.

Clinical Artificial Intelligence Applications: Breast Imaging.

Hu Q, Giger M Radiol Clin North Am. 2021; 59(6):1027-1043.

PMID: 34689871 PMC: 9075017. DOI: 10.1016/j.rcl.2021.07.010.

Radiomics methodology for breast cancer diagnosis using multiparametric magnetic resonance imaging.

Hu Q, Whitney H, Giger M J Med Imaging (Bellingham). 2020; 7(4):044502.

PMID: 32864390 PMC: 7444714. DOI: 10.1117/1.JMI.7.4.044502.

References

Kupinski M, Giger M . Feature selection with limited datasets. Med Phys. 1999; 26(10):2176-82. DOI: 10.1118/1.598821. View

Deurloo E, Muller S, Peterse J, Besnard A, Gilhuijs K . Clinically and mammographically occult breast lesions on MR images: potential effect of computerized assessment on clinical reading. Radiology. 2005; 234(3):693-701. DOI: 10.1148/radiol.2343031580. View

Cook N . Use and misuse of the receiver operating characteristic curve in risk prediction. Circulation. 2007; 115(7):928-35. DOI: 10.1161/CIRCULATIONAHA.106.672402. View

Kupinski M, EDWARDS D, Giger M, Metz C . Ideal observer approximation using Bayesian classification neural networks. IEEE Trans Med Imaging. 2001; 20(9):886-99. DOI: 10.1109/42.952727. View

Chen W, Giger M, Lan L, Bick U . Computerized interpretation of breast MRI: investigation of enhancement-variance dynamics. Med Phys. 2004; 31(5):1076-82. DOI: 10.1118/1.1695652. View

Jiang Y, Nishikawa R, Schmidt R, Metz C, Giger M, Doi K . Improving breast cancer diagnosis with computer-aided diagnosis. Acad Radiol. 1999; 6(1):22-33. DOI: 10.1016/s1076-6332(99)80058-0. View

Lucht R, Knopp M, Brix G . Classification of signal-time curves from dynamic MR mammography by neural networks. Magn Reson Imaging. 2001; 19(1):51-7. DOI: 10.1016/s0730-725x(01)00222-3. View

McLaren C, Chen W, Nie K, Su M . Prediction of malignant breast lesions from MRI features: a comparison of artificial neural network and logistic regression techniques. Acad Radiol. 2009; 16(7):842-51. PMC: 2832583. DOI: 10.1016/j.acra.2009.01.029. View

Metz C . Quantification of failure to demonstrate statistical significance. The usefulness of confidence intervals. Invest Radiol. 1993; 28(1):59-63. DOI: 10.1097/00004424-199301000-00017. View

10.

Jansen S, Shimauchi A, Zak L, Fan X, Wood A, Karczmar G . Kinetic curves of malignant lesions are not consistent across MRI systems: need for improved standardization of breast dynamic contrast-enhanced MRI acquisition. AJR Am J Roentgenol. 2009; 193(3):832-9. PMC: 2938789. DOI: 10.2214/AJR.08.2025. View

11.

Schnall M, Blume J, Bluemke D, DeAngelis G, DeBruhl N, Harms S . Diagnostic architectural and dynamic features at breast MR imaging: multicenter study. Radiology. 2005; 238(1):42-53. DOI: 10.1148/radiol.2381042117. View

12.

Chen W, Giger M, Bick U, Newstead G . Automatic identification and classification of characteristic kinetic curves of breast lesions on DCE-MRI. Med Phys. 2006; 33(8):2878-87. DOI: 10.1118/1.2210568. View

13.

Jiang Y, Nishikawa R, Schmidt R, Toledano A, Doi K . Potential of computer-aided diagnosis to reduce variability in radiologists' interpretations of mammograms depicting microcalcifications. Radiology. 2001; 220(3):787-94. DOI: 10.1148/radiol.220001257. View

14.

Penn A, Bolinger L, SCHNALL M, Loew M . Discrimination of MR images of breast masses with fractal-interpolation function models. Acad Radiol. 2000; 6(3):156-63. DOI: 10.1016/S1076-6332(99)80401-2. View

15.

Twellmann T, Lichte O, Nattkemper T . An adaptive tissue characterization network for model-free visualization of dynamic contrast-enhanced magnetic resonance image data. IEEE Trans Med Imaging. 2005; 24(10):1256-66. DOI: 10.1109/TMI.2005.854517. View

16.

Woods B, Clymer B, Kurc T, Heverhagen J, Stevens R, Orsdemir A . Malignant-lesion segmentation using 4D co-occurrence texture analysis applied to dynamic contrast-enhanced magnetic resonance breast image data. J Magn Reson Imaging. 2007; 25(3):495-501. DOI: 10.1002/jmri.20837. View

17.

Kuhl C . The current status of breast MR imaging. Part I. Choice of technique, image interpretation, diagnostic accuracy, and transfer to clinical practice. Radiology. 2007; 244(2):356-78. DOI: 10.1148/radiol.2442051620. View

18.

Kuhl C . Current status of breast MR imaging. Part 2. Clinical applications. Radiology. 2007; 244(3):672-91. DOI: 10.1148/radiol.2443051661. View

19.

Nie K, Chen J, Yu H, Chu Y, Nalcioglu O, Su M . Quantitative analysis of lesion morphology and texture features for diagnostic prediction in breast MRI. Acad Radiol. 2008; 15(12):1513-25. PMC: 2791407. DOI: 10.1016/j.acra.2008.06.005. View

20.

Szabo B, Kristoffersen Wiberg M, Bone B, Aspelin P . Application of artificial neural networks to the analysis of dynamic MR imaging features of the breast. Eur Radiol. 2004; 14(7):1217-25. DOI: 10.1007/s00330-004-2280-x. View