Font Size: a A A

Diagnostic Value Study Of MR BI-RADS Combinedwith DWI In Breast Lesions

Posted on:2016-01-10Degree:MasterType:Thesis
Country:ChinaCandidate:F WangFull Text:PDF
GTID:2284330461471971Subject:Medical imaging and nuclear medicine
Abstract/Summary:PDF Full Text Request
Objective:To Analyze the descriptive language of MR Breast Imaging and Reporting Data Systems(MR BI-RADS) and Diffusion-Weighted Imaging(DWI) of breast lesions, and calculate the positive predictive values(PPV) of malignant signs for breast mass lesions and NMLE(non-mass-like lesion), in order to study the diagnostic value of the application combined MR BI-RADS with DWI in breast benign and malignant lesions, which can improve the sensitivity, specificity and accuracy in the diagnosis of breast lesions by MRI.Material and Methods:Collected in 274 cases(290 lesions) of female patients who are from the age of 16 to 81, with the average age of 53, breast MRI was scanned from December 2010 to December 2014 in the hospital; All cases were studied by STIR,T1 WI, DWI and DCE-MRI sequence using Siemens Avanto 1.5T superconductive MRI system and a dedicated four-channel phased array breast surface coil., and confirmed by biopsy or surgical pathology.Two radiologists, with rich experience in reading the breast magnetic resonance imaging read the images in a double-blind method. The descriptions of DCE-MRI analysis was on the basis of ACR MR BI-RADS(the first edition), the MR BI-RADSdata were analyzed by using SPSS 16.0. The morphology analysis of mass lesions included the shape, margin, internal enhancement of the lesions, and the NMLE morphological analysis covered the distribution, internal enhancement of the lesions. The kinetics was analyzed by the time-signal intensity curve(TIC): the early enhanced(1.5min enhancement ratio) characteristics and TIC type. The lesion score was evaluated by Fischer scoring, and matching with MR BI-RADS classification. The results were applied χ2 test to perform statistical analysis, and calculated the positive predictive values of malignant signs and 95% confidence intervals. DWI assessment: measurement of ADC values were measured three times, and taken the lowest ADC value to the last measurement, and applied independent samples t-test to analyze; the mass lesions and NMLE ADC values used by receiver operating curve(ROC) were studied, and to determine the benign and malignant mass lesions and NMLE were the ADC threshold. The diagnosis values of MR BI-RADS, DWI and their combination in sensitivity, specificity, accuracy, positive predictive value and negative predictive value(NPV) of breast lesions.Result:the differences which the shape, margin, distribution, internal enhancement, early enhancement ratio and TIC type of the breast mass and NMLE in benign and malignant lesions showed were statistical significance(P<0.05). Morphological analysis showed that irregular shape(PPV, 0.911), irregular margin(PPV, 0.868) and spiculated margin(PPV, 0.921) were the highest PPV for the malignancy lesions; and the internal heterogeneous enhancement(PPV, 0.877), ring enhancement(PPV, 0.875) were the highest PPV for malignancy. With lobular Segmental distribution(PPV, 0.800) and ductal(PPV, 1.000) of NMLE, the malignant potential was the highest; clumped enhancement(PPV, 0.818) and reticular/dendritic(PPV, 0.909) of internal-enhancedcharacteristics were the highest PPV for the NMLE.Kinetics analysis: early enhancement ratio more than 100% of malignant lesions was the highest positive predictive value(PPV, 0.885), and the early enhancement ratio less than 50% of malignant lesions was lowest predictive value(PPV, 0.080).The washout type in the highest proportion of malignant lesions has highly positive predictive value(PPV, 0.944), and the persistent type is the lowest positive predictive value of(PPV, 0.351) to malignant lesions. The sensitivity, specificity, PPV, NPV and accuracy for breast mass and NMLE of MR BI-RADS are 93.4%, 80.1%, 91.4%, 75.3%, 87.7% and 89.1%, 78.5%, 87.4%, 72.8%, 82.7%, respectively.The ADC value of benign and malignant breast mass lesions is(1.348±0.215)×10-3mm2/s 、(0.897±0.294)×10-3mm2/s. The difference is statistically significant(t=7.596, P=0.000).The optimal diagnostic threshold of ADC value is 1.25×10-3 mm2/s, and the area under the curve of ROC is 0.901.The ADC values of benign and malignant NMLE are(1.582±0.215)×10-3mm2/s、(0.961±0.247)×10-3mm2/s. The difference is statistically significant(t=9.814, P=0.000). The optimal diagnostic threshold of ADC value is 1.35×10-3 mm2/s. The area under the curve of ROC is 0.794. The ADC value of mass and non-mass lesions in the diagnosis sensitivity, specificity, PPV, NPV and accuracy are 86.9%、84.5%、92.6%、76.1%、85.2% and 85.3%、74.8%、86.7%、75.9%、79.3%, respectively.with the combined application, the value of sensitivity, specificity, PPV, NPV and accuracy of mass and NMLE in the diagnosis is 91.8%、88.9%、94.3%、78.9%、92.1% and 90.7%、84.6%、88.5%、78.8%、85.1% respectively.Conclusion:The diagnosis of breast lesions combined MR BI-RADS with DWI can get a high sensitivity, specificity and accuracy. The application of ADC values in diagnosis of breast lesions is an important supplement to standardize of MR BI-RADS, and significantly increased the diagnostic efficiency of MRI.
Keywords/Search Tags:Breast neoplasms, Magnetic resonance imaging, dynamic contrast-enhanced, Diffusion weighted imaging, Apparent diffusion coefficient
PDF Full Text Request
Related items