Intravoxel Incoherent Motion Mr Imaging In Breast Cancer: Quantitative Analysis For Characterizing Lesions

Naier Lin,Junyan Chen,Jia Hua,Jinhua Zhao,Jun Zhao,Jinsong Lu
2017-01-01
Abstract:Objectives: We compared the monoexponential model diffusion weight imaging (DWI) and intravoxel incoherent motion DWI (IVIM-DWI) parameters for characterizing different subtypes and grades of breast cancer lesions. Methods: We analyzed 93 patients (51 with breast cancers and 47 with benign lesions) using monoexponential (b=0.800 s/mm(2)) and biexponential analysis (b=0-800 s/mm(2)) and 3.0T MRI. We compared the apparent diffusion coefficient (ADC), true diffusion coefficient (D), perfusion fraction (f) and pseudo-diffusion coefficient (D*) in different subtypes and grades of cancers. We used receiver operating characteristics (ROC) to identify the best cutoff for differentiation between lesions. Results: All data showed good fit (R-2>0.90). The D value was significantly lower than ADC in all the lesions except cysts (P<0.05). ADC, D and D* in malignant tumors were significantly lower than those in benign ones (P<0.001 for ADC and D; 0.002 for D*). The f value was higher than that of benign lesions (P<0.001). D showed higher accuracy than ADC in differentiating malignancy from benign lesions (AUC=0.945 for D; 0.931 for ADC), or ductal carcinoma in situ (DCIS) from invasive ductal carcinoma (IDC) (AUC=0.791 for D; 0.675 for ADC). Furthermore, the D value was significantly lower in high grade IDC (TNM III) than in low grade IDC (TNM I+II) (P=0.006), while no other parameters showed a difference between varied grades of IDC (P>0.05). Conclusions: IVIM provides quantitative measurement of cellularity and vascularity for characterizing breast lesions. In particular, D shows good potential for grading breast cancer.
What problem does this paper attempt to address?