Comparative Study Of Digital Mammography, Ultrasonography And MR Imaging For The Diagnosis Of Breast Carcinoma

Purpose:To (a) determine the performance of digital mammography (DM) and ultrasonography (US), (b) analyze the influence of age, hormonal status, breast density, pathological type, T stage, and calcification or not; (c) assess the value of BI-RADS categories.Materials and Methods:Between December 2007 and December 2008, the suspected and risk patients with breast carcinoma received DM and US examination. Until December 2009 follow-up ended,1432 female were rolled in this study who had complete medical history and definite diagnosis. Both examinations were randomly performed without knowing results of the other. All imagines were assessed according to BI-RADS categories. With pathological diagnosis as golden standard, the performance of each modality was evaluated and associated factors were analyzed. X2 test and the ROC curve were used to compare the diagnosis power of each modality.Results:In 1432 patients,1500 lesions including 737 benign and 763 malignant were found.1. Sensitivity, specificity, and accuracy of mammography were 90.0%,83.0%, and 86.6%, respectively; and those of US,95.4%,84.8%, and 90.2%, respectively. There was significant difference between DM and US performance. DM and US together had significantly higher sensitivity (98.4%) than a single modality (P<0.05).2. Mammographic sensitivity declined significantly with increasing breast density (P <0.05), but not for US. There was no sensitivity difference for both modalities in patients older than 50 years or with fat and scattered fibroglandular density, or with the same pathological T stage. The sensitivity of both modalities was higher for invasive ductal carcinoma (IDC) than ductal carcinoma in situ (DCIS). The sensitivity of DM for DCIS and IDC were 74.6% and 92.2%, those of US were 79.1% and 98.4%. US had higher sensitivity for IDC or palpable lesions than DM (P=0.000). DM was more sensitive for lesions with calcification (97.0% VS 95.1%, P=0.001) or pure calcification (91.5% VS 85.1%,P=0.016).3. The positive predictive value of DM BI-RADS category 1-5 were 12.9%,3.1%, 8.9%,70.9% and 98.3%, respectively, those of US were 3.9%,3.9%,6.3%,62.8% and 95.7%, respectively, those of combined modalities were 1.6%,2.7%,2.2%,45.0% and 95.5%. The area under ROC value of DM and US BI-RADS was 0.901 and 0.945.Conclusion:US has shown better performance than DM for clinic patients. Combined exams improve the sensitivity. BI-RADS categories are useful for predicting the presence of malignancy. Purpose:To assess the value of the signal intensity time course and associated quantitative parameters for the differential diagnosis of benign and malignant breast disease in the dynamic contrast-enhanced MRI (DCE-MRI).Materials and Methods:188 patients underwent breast DCE-MRI which included mask and nine-phase-enhancing scanning with vibrant technique. The type of DCE-MRI kinetic curve was assessed qualitatively by one radiologist. For quantitative assessment, the initial enhancement rate, peak enhancement rate, phase to peak instead of time to peak were calculated. X2 test and ROC curve were used to compare quantitative versus qualitative methods for establishing or excluding malignancy.Results:There were 119 malignant (57.5%) and 88 benign lesions (42.5%).1. The distribution of curve types for breast cancers was type 1,1.7%; type 2,26.5%; and type 3,71.8%. The distribution of curve types for benign diseases was type 1,60.2%; type 2,23.9%; and type 3,15.9%. The distribution proved significantly different (P<0.05).The positive predictive value of curve type 1,2, and 3 was 3.6%,59.6%, and 85.7%. The diagnostic indices for curve type were sensitivity,98.3%; specificity,62.5% and accuracy,82.9%.2. The initial enhancement rate, peak enhancement rate, phase to peak showed significant difference between breast cancer and benign disease (p<0.05), with an area under the ROC curve (AUC) of 0.759,0.702, and 0.802. Using threshold initial enhancement rate of 100% alone, the sensitivity, specificity and accuracy were 99.1%, 35.5%, and 76.7%.Conclusion:The type of contrast enhancement kinetic curve on breast DCE-MRI results in significantly higher diagnostic performance for differential diagnosis compared with assessment based on the quantitative parameters. The curve type is not an independent diagnostic tool, and architectural features should be combined. Purpose:to assess the accuracy of mammography, US and MR for the tumor extent of breast cancer.Materials and Methods:DM, US and MR imaging were performed in 188 patients. The MM and MR images were independently interpreted by two imaging doctors together without knowing the result of the other. The US was performed with knowing MM and MR results. The correlation of the results with histological findings was examined.Results:1. There were 119 malignant lesions.101(84.9%) were detected on DM,105(88.2%) were detected on US, and 117 (98.3%) were detected on MR. There was no significant difference of sensitivity between each other.115 lesions (96.6%) were detected when combined DM and US,118 (99.2%) were detected when combined US and MR, and 119 (100%) were detected when combined DM and MR.2. The pathologic size of 92 lesions was acquired with mean of 1.96cm. MR measurements correlated better to pathologic size (r=0.56) than DM (r=0.39) or US (r=0.37). Underestimation of tumor size was more common on US (14.1%). Overestimation occurred in MR and DM with frequency of 15.2%, and 11.0%. There was no significant difference of accuracy among 3 modalities.Conclusion:Although there was no significant difference of accuracy among 3 modalities, MR had better performance in the detection and extent measurement than MM and US. MR had a substantial risk of overestimation and US had a risk of underestimation.