Prediction of low-risk breast cancer using perfusion parameters and apparent diffusion coefficient

Hee Jung Shin; Hak Hee Kim; Ki Chang Shin; Yoo Sub Sung; Joo Hee Cha; Jong Won Lee; Byung Ho Son; Sei Hyun Ahn

doi:10.1016/j.mri.2015.10.028

Prediction of low-risk breast cancer using perfusion parameters and apparent diffusion coefficient

Magn Reson Imaging. 2016 Feb;34(2):67-74. doi: 10.1016/j.mri.2015.10.028. Epub 2015 Oct 30.

Authors

Hee Jung Shin¹, Hak Hee Kim², Ki Chang Shin³, Yoo Sub Sung³, Joo Hee Cha², Jong Won Lee⁴, Byung Ho Son⁴, Sei Hyun Ahn⁴

Affiliations

¹ Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan, College of Medicine, 88 Olympic-ro, 43-gil, Songpa-gu, Seoul, 138-736, South Korea. Electronic address: docshin@amc.seoul.kr.
² Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan, College of Medicine, 88 Olympic-ro, 43-gil, Songpa-gu, Seoul, 138-736, South Korea.
³ Department of Radiology and Research Institute of Radiology, Medical Imaging Laboratory, Asan Medical Center, University of Ulsan, College of Medicine, 88 Olympic-ro, 43-gil, Songpa-gu, Seoul, 138-736, South Korea.
⁴ Department of Breast and Endocrine Surgery, Asan Medical Center, University of Ulsan, College of Medicine, 88 Olympic-ro, 43-gil, Songpa-gu, Seoul, 138-736, South Korea.

PMID: 26523654
DOI: 10.1016/j.mri.2015.10.028

Abstract

Purpose: To assess whether perfusion and diffusion parameters were different between low-risk tumors and non-low-risk tumors.

Materials and methods: We prospectively enrolled 87 patients with 91 tumors patients (mean, 49.6 years; range, 29-74 years) who underwent definitive surgery. We defined estrogen receptor (ER)-positive tumors with low histologic grade (HG), low Ki67 (<14%), and negative lymph node metastasis as a low-risk breast cancer. We obtained quantitative and semiquantitative perfusion parameters and apparent diffusion coefficient (ADC) for all tumors. We compared perfusion parameters and ADCs between low-risk tumors (n=33; 36%) and the others (n=58; 64%) using Fisher's exact test, Chi-square test, and student t-test. We developed empirical model to predict low-risk tumor using logistic regression analysis and receiver operating characteristics (ROC) analysis.

Results: On univariate analysis, wash-in and the initial area under the curve on qualitative analysis (iAUCqualitative) were significantly different according to HG, ER, HER-2, Ki67 and lymphovascular invasion (P<.05 for all variables). ADCdiff was significantly different according to HG, HER-2, and Ki67 status (P=.010, .007, and .013). On multivariate analysis, Ktrans, iAUCqualitative, and ADCdiff were the significant variables for the prediction of low-risk tumors, and the area under the ROC curve (AUC) of combined parameters was 0.78, which was higher than those of the individual parameter. ADCdiff was positively correlated with wash-in (r=0.263) and iAUCqualitative (r=0.245), respectively.

Conclusion: The prediction model using Ktrans, wash in, iAUCqualitative, and ADCdiff on DCE-MRI and DWI could be helpful for identifying of low-risk breast cancer and may be used as an imaging biomarker to guide the treatment plan.

Keywords: Breast neoplasm; DCE-MRI; Diffusion-weighted imaging; Low-risk breast cancer.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Adult
Aged
Algorithms*
Computer Simulation
Diffusion Magnetic Resonance Imaging / methods*
Diffusion Magnetic Resonance Imaging / statistics & numerical data
Female
Humans
Image Interpretation, Computer-Assisted / methods*
Magnetic Resonance Angiography / methods*
Magnetic Resonance Angiography / statistics & numerical data
Middle Aged
Models, Statistical*
Prevalence
Prognosis
Proportional Hazards Models*
Reproducibility of Results
Republic of Korea / epidemiology
Sensitivity and Specificity