The Diagnostic Efficiency Analysis Of Multi-parameter Magnetic Resonance Imaging For Breast Cancer

Background and Objective The breast cancer had recently seen a significant growth in China, and had become one of the most common female malignant tumors. Breast MRI was the technique with non-radiation, multi scanning parameters, more imaging sequences, and high resolution of soft tissues, Breast MRI can provide a lot of information of diagnostic values, such as morphology, hemodynamics, water molecules diffusion function and metabolites of tissue cell. Breast MRI had already become an important method of breast lesions examination and attracted more and more attention. The purpose of this research is evaluate the multi-parameter magnetic resonance imaging diagnostic efficiency of routine MRI, dynamic contrast-enhanced MRI(DCE-MRI) and diffusion weighed imaging(DWI) in diagnosis of breast cancer. Materials and methods 1. 95 female patients with 97 breast lesions was collected in our hospital from February,2012 to May,2014, aged from 17 to 69(46.43± 9.34). All patients have been confirmed by biopsy or pathology, among which 43 cases of malignant lesions(including 33 cases of infiltrating ductal carcinoma, 1 case of pleomorphic carcinoma, 2 cases of basaloid carcinoma, 5 cases of intraductal carcinoma, 1 case of invasive cribriform tubular carcinoma, and 1 case of medullary carcinoma), 52 cases of benign lesions with a total of 54 lesions(including 8 cases of adenosis, 36 cases with fibro-adenoma, 2 cases of intraductal papilloma, 2 cases of sclerosing duct disease, 5 cases of atypical ductal hyperplasia, and 1 case of granulomatous lobular mastitis). 2. Equipment and methods: All patients underwent routine MRI(T1WI, T2WI), diffusion weighted imaging(b=0, 500, 800 s/mm2) and dynamic contrast-enhanced MRI. Observed and analyzed routine MRI examination of breast lesions(morphology, boundary, edge and internal enhancement) and TIC curves. Calculated the early enhancement rates(1 minute, 2 minutes and 3 minutes) of the interest areas of the lesions respectively and measured two groups of ADC results(b=500, 800s/mm2). With pathology results as “the gold standard”, attained the receiver operating characteristics(ROC) of the early enhancement rate and the ADC values, Calculated Area under Curve(AUC).According to the largest demarcation point of Youden index, three groups of early enhancement rates and two groups of ADC values optimum diagnostic critical values had been identified and 97 lesions were classified as benign and malignant lesions. Using the software of IBM SPSS Statistics Version 19, the measurement data values were indicated with means and standard deviations(x ±s); t test analysis were used for inter-group measurement data statistical analysis; χ2 test analysis were used for counting data statistical analysis,and P<0.05 indicated statistical significance. With pathology results as “the gold standard”, sensitivity, specificity, accuracy, positive predictive value, negative predictive value and Kappa value have been separately gained. With pathological results as dependent variables and results of multi-MRI parameters being independent variables, Binary Logistic regression analysis was made, and P<0.05 indicated statistical significance. Results 1. The signal distribution of routine MRI T1 WI and T2 WI were largely overlapping in breast benign and malignant lesions. The three observation indexes(shape, boundary and internal enhancement) had no statistical difference(P>0.05), and “burr sign” had remarkable statistical significance in diagnosis of benign and malignant lesions(P<0.01). 2. Type I TIC curve accounted for 83.3% in benign lesions, and type II TIC curve accounted for 53.5%, type III accounting for 44.19% in malignant lesions.1 minute, 2 minutes and 3 minutes early enhancement rates of AUCROC were 0.756, 0.634,0.573, corresponding best diagnostic critical values in benign and malignant lesions were 37.75%, 33.24%, and 34.05% respectively, and 1 minute early enhancement rate had remarkable statistical significance in the diagnosis of benign and malignant lesions(P<0.01). 3. Malignant lesions with high signal DWI accounted for 72.1%, while benign lesions for 50%. AUCROC of two groups ADC values of b=500 s/mm2、800 s/mm2 were 0.907 and 0.919, the best boundary values of two groups of ADC values for diagnosis of malignant and benign lesions were 1.132 x10-3mm2/s and 1.0265 x10-3mm2/s. ADC mean values(x10-3mm2/s) of b=500 s/mm2 and 800 s/mm2 in benign lesions were 1.646 ± 0.440, 1.477 ±0.451, while those in malignant lesions were 1.039± 0.263 and 0.914 ± 0.224. Two groups of b values had remarkable statistical significance in the diagnosis of breast lesions(P<0.01). 4.With pathology results as “the gold standard”, the observation index "burr" sign” had high specificity(96.3%) and Kappa value(0.611) in routine MRI. The sensitivity, specificity and accuracy of the rest routine MRI observation index were less than 80%.Compared with the pathological results, Kappa value of consistency level detection were 0.297~0.410. The sensitivity, specificity, accuracy and Kappa value of TIC curve in the diagnosis of breast lesions were 97.67%, 83.33%, 89.69%, 0.795. Sensitivity of 1minute, 2 minutes and3 minutes early enhancement rates are respectively 88.37%, 90.70%, and 88.37%, while specificity of those were 62.96%, 44.44%, 40.74%, and Kappa values were 0.496, 0.331, 0.274. Sensitivity and specificity of b=500 s/mm2、800 s/mm2 ADC values of were respectively 81.4%, 90.74%, 86.60% and 88.37%, 92.59%, 90.72%. Compared with the pathological results, the ADC index of b=800 s/mm2 had the high level of consistency(Kappa 0.812). 5. Multi-parameter MRI imaging methods were analyzed with Logistic regression analysis, the results showed that TIC curve type had statistical significance in the diagnosis of benign and malignant breast lesions(P<0.01), and “burr sign” and ADC value(b=800s/ mm2) had remarkable statistical significance in the diagnosis of benign and malignant breast lesions(P<0.05). Conclusions: 1. The edge( "burr sign"), 1 minute early enhancement rate and b=500 s/mm2、800 s/mm2 ADC values have notably statistical significance in diagnosis of benign and malignant breast lesions(P<0.01). 2. TIC curve played important role in the diagnosis of benign and malignant breast lesions. 3. The edge( "burr sign"), TIC curve, and ADC value(b=800 s/mm2) have high diagnostic efficiency in the diagnosis of benign and malignant breast lesions.