| Part1Comparative study of digital breast tomosynthesis and full-field digital mammography for breast benign and malignant lesions [Background and Objective]X-ray mammography is the preferred imaging method for breast examination. However, the traditional X-ray Mammography is a three-dimensional solid breast on two-dimensional projection image, due to the overlap of normal breast tissue, particularly in younger women and those with dense breast patterns, inevitably resulting in a possible miscarriage of justice will be hidden lesion omission false negative or false positive is some overlap of artifacts as a false positive. Thus, with the development of imaging technology, digital mammography tomography came into being, there is a certain influence on the diagnosis accuracy of breast lesions. This part of the study was to investigate the diagnostic value of DBT in benign and malignant breast lesions, comparing with the difference between full-field digital mammography. The comparison between the two diagnostic capabilities, focused on the two imaging methods difference in lesion BI-RADS classification, the sensitivity and the specificity of the different types of breast lesions, and the difference of diagnostic capabilities between the different dense of mammary glands of lesions. [Materials and Methods]1. Study populationAll participant patients provided informed consent. From October2013to February2014, we enrolled256breasts of247diagnostic patients with at least one breast finding found by mammography and/or ultrasound and classified as probably benign, suspicious or highly suspicious for malignancy. All cases were confirmed by biopsy/surgical pathology, or US that showed an anechoic cyst. All patients underwent a DBT and FFDM examination respectively. All patients were female, retrospectively analyzed these X-ray findings and clinical data.2Research Methods2.1Equipment and methodsFull-field digital mammography and tomosynthesis images were acquired by a Selenia DimensionsTM unit (Hologic, American). All patients underwent once FFDM and DBT examination, using combo automatic mode exposure, under the same conditions of compression, one exposure could be obtained FFDM and DBT images simultaneously. The original images reconstructed according to the thickness of lmm, the number of reconstruction images depends on the breast compression thickness. Each FFDM and DBT examination consisted of the MLO and CC views. Studies were displayed on a commercially available digital mammography workstation that included two5-mega-pixel monitors (SecureViewDx; Hologic).2.2Standard for the image evaluationThe FFDM and DBT images were retrospectively analyzed by two senior radiologists with more than5years of experience in breast imaging according to Breast Imaging Reporting and Data System (BI-RADS), recorded the reading results respectively. In order to eliminate the effect of short-term memory effect, all the cases were randomly divided into two groups, the readers read FFDM images of first group and DBT images of second group, and then read the remaining cases two weeks later. Before the formal analysis, every doctor acknowledged the procedure and observation.2.3Reporting SystemThe FFDM and DBT images were described and classified according to the American College of Radiology Breast Imaging-Reporting and Data System (BI-RADS). The analysis included the following aspects:1)The main X-ray findings: mass, calcification, architectural distortion, focal asymmetric density;2)to describe the main X-ray findings, including tumor morphology, edge and accompanied by signs; calcification morphology, distribution and associated signs;focal asymmetric density according to mass standards described;3)The breast density:1, fatty or<25%dense;2, scattered fibro-glandular densities or25-50%dense;3, heterogeneously dense or50-75%dense;4, dense or>75%dense.BI-RADS categories include the following:Category0:need additional imaging evaluation, assessment is incomplete. Category1:negative, no abnormalities detected. Category2:benign findings, some features worthy of description, but are confidently diagnosed as benign. Category3:probably benign, short follow up interval suggested, the lesions have a high probability of being benign, but stability over time is preferably to be established. Category4:suspicious abnormality, biopsy should be considered, they possess a definite probability of being malignant; which is divided into the following three categories, Category4a:malignant lesions may be small; Category4b:malignant lesions may be living in the middle; Category4c:malignant medium but not having the possibility of malignant typical performance. BI-RADS Category5:highly suggestive of malignancy, appropriate action should be taken. Category6:known biopsy proven malignancy, reserved for lesions with biopsy proof of malignancy. 2.4Quality ControlDiagnosis was carried out by two radiologists. If there were any inconformity on the sign observed, the final results were obtained by discussion between the superior doctors.2.5Statistical AnalysisChi-square and receiver operating characteristic (ROC) analysis were performed with the use of statistical software (SPSS, version17.0). Test level of α=0.05, with P <0.05to indicate a significant difference.[Results]1. There are247patients of256breast lesions. The155benign lesions including:30cysts,48fibroadenoma,2intraductal papilloma,2benign phyllodes tumors,7atypical hyperplasia,36breast fibrocystic disease, and30adenosis. The101malignant lesions including:84IDC,10DCIS,3mucinous carcinoma,3cases of invasive ductal carcinoma with micro-invasive papillary carcinoma,1invasive lobular carcinoma,1neuroendocrine carcinoma, and1profusely adenocarcinoma. There was statistically significant in the age of onset between benign and malignant lesions (P=0.000), the highest incidence of breast lesions was40-50years.2. There was not statistically significant in BI-RADS classification of FFDM and DBT for breast lesions. In DBT and FFDM benign lesions mainly distributed in the BI-RADS2-3categoryandBI-RADS category4a, malignant lesions mainly in the BI-RADS4c-5category.3.DBT and FFDM overall diagnostic performance:The ROC showed that the diagnosis ability of breast lesions in DBT was higher than that in FFDM, the difference between the area under the curve was0.046. DBT diagnostic sensitivity, specificity, positive predictive value and negative predictive value were99.0%,97.4%,65.7%, and99.1%; FFDM diagnostic sensitivity, specificity, positive predictive value and negative predictive value were97.0%,92.3%,58.7%and96.6%. The difference of sensitivity between the two methods was not statistically significant (P=0.555), the difference of specificity was statistically significant (P=0.026), the difference of accuracy was statistically significantly (P=0.030).4.DBT and FFDM diagnostic performance in different breast density:DBT diagnostic sensitivity, specificity, PPV and NPV were97.1%,89.5%,94.3%and94.4%on less dense breast, and FFDM diagnostic sensitivity, specificity, PPV and NPV diagnosis were97.1%,73.7%,86.8%and93.3%.The difference of sensitivity, specificity, and accuracy between two methods were not statistically significant (P>0.05). Area under the ROC showed the diagnostic accuracy of DBT was a litter better than that of FFDM, the difference of AUC was0.079. On mostly dense breast, DBT d diagnostic sensitivity, specificity, PPV and NPV were100%,64.7%,58.3%and100%; FFDM diagnostic sensitivity, specificity, PPV and NPV were97.0%,52.9%,50.4%, and97.3%. There was no significant difference in sensitivity (P=0.154), the difference of specificity and accuracy were statistically significant (P=0.049,0.047). The ROC curves showed the diagnostic accuracy of DBT was better than that of FFDM, the difference of AUC was0.077.5. DBT and FFDM diagnostic performance for different lesions:DBT diagnostic sensitivity, specificity, PPV and NPV for non-calcified lesions were respectively98.6%,78.0%,74.0%and99.0%; FFDM diagnostic sensitivity, specificity, PPV and NPV were95.8%,65.0%,61.6%and96.4%.There was no significant difference in sensitivity (P=0.310), the difference of specificity was statistically significant (P=0.010), and the difference of accuracy was statistically significant too (P=0.009). The ROC curves showed the diagnostic accuracy of DBT was better than that of FFDM, the difference of AUC was0.087. For calcified breast lesions, DBT diagnostic sensitivity, specificity, PPV and NPV were100%,21.9%,53.7%and100%; FFDM diagnostic sensitivity, specificity, PPV and NPV were100%,18.8%,52.7%and100%. All the differences of sensitivity, specificity, and accuracy were not statistically significant (P>0.05). The difference of AUC is0.015.[Conclusions]1. The incidence of breast lesions associates with age, the age of onset of malignant lesions is older than benign lesions. Fibroadenoma is the most common benign lesion, followed by the fibrocystic breast disease; IDC is the most common malignant lesion, followed by ductal carcinoma in situ.2. There are some differences of BI-RADS classification between DBT and FFDM diagnosis. DBT will be more malignant lesions attributed to BI-RADS4c category and BI-RADS5category, the more benign lesions attributed to BI-RADS category2and BI-RADS3category.3. In the diagnostic breast examination, the overall diagnostic capability of DBT is better than FFDM.DBT has a higher sensitivity and specificity in breast diagnosis than FFDM, especial in the specificity. The PPV and NPV of DBT are higher than that of FFDM. For the diagnosis of breast lesions in mostly dense breast and non-calcified lesions, DBT also superior to FFDM, the differences of specificity and accuracy are statistically significant. For breast lesions in less dense breast and calcified breast lesions, the diagnostic capability of DBT is not less than FFDM. However, the sensitivity of the two methods is no statistically difference. Therefore, the full potential of DBT is currently unknown and more work is required using robust, comprehensive, and clinically relevant methodologies using the most rigorous analytical methods.Part2Comparative study of DBT and FFDM imaging findings for the mass of breast lesions[Background and Objective] Increasingly higher incidence of breast cancer has been a serious threat to women’s health and life. The X-ray findings of breast cancer are diverse, can divide into mass, calcification, architectural distortion and focal asymmetric density. Mass is the most common X-ray findings of breast cancer. This section focuses on comparing the diagnostic capabilities of DBT and FFDM for mass lesions, compared DBT and FFDM in the characterizing of mass, evaluation of the clinical performance of DBT.[Materials and Methods]1. Study populationAll participant patients provided informed consent. From October2013to February2014, we enrolled174mass breast lesions of170patiens with at least one breast finding found by mammography and/or ultrasound and classified as probably benign, suspicious or highly suspicious for malignancy. All cases were confirmed by biopsy/surgical pathology, or US that showed an anechoic cyst. All patients underwent a DBT and FFDM examination respectively. All patients were female, retrospectively analyzed these X-ray findings and clinical data.2Research Methods2.1Equipment and MethodsIt is the same as the first part.2.2Standard of the image evaluationThe FFDM and DBT images were retrospectively analyzed by two senior radiologists with more than5years of experience in breast imaging according to Breast Imaging Reporting and Data System (BI-RADS), analysis of breast lesions characteristic. Observation contents included:tumor size, shape, edge, structural distortion, increase blood flow, vascular penetration, and other accompany signs. In order to eliminate the effect of short-term memory effect, all the cases were randomly divided into two groups, the readers read FFDM images of first group and DBT images of second group, and then read the remaining cases two weeks later. Before the formal analysis, every doctor acknowledged the procedure and observation.2.3Reporting SystemIt is the same as the first part.2.4Quality ControlDiagnosis was carried out by two radiologists. If there were any inconformity on the sign observed, the final results were obtained by discussion between the superior doctors.2.5Statistical AnalysisChi-square and T-test ware performed with the use of statistical software (SPSS, versionl7.0). Test level of α=0.05, with P<0.05to indicate a significant difference.[Results]1. There were102benign breast lesions and72malignant breast lesions. DBT diagnostic sensitivity, specificity, and accuracy were98.6%,91.2%, and92.7%; FFDM diagnostic sensitivity, specificity, and accuracy were95.8%,73.5%and82.8%.The was no significant difference in the sensitivity (P=0.310), there was statistically significant in specificity (P=0.002), and the difference of accuracy was statistically significant (P=0.002).2. There was statistically significant in the BI-RADS classification between DBT and FFDM for benign breast lesions.92cases of benign mass in the DBT were classified as BI-RADS category2and category3,74cases of benign mass in FFDM were classified as Class2and Class3.The difference of BI-RADS classification between DBT and FFDM for malignant mass was not statistically significant. In DBT58malignant masses were classified as BI-RADS5category, and in FFDM46malignant masses were classified as BI-RADS5category.3.DBT and FFDM for the margin of mass:In DBT the margin of benign mass was characterized clear, fuzzy, shadow and lobulated were87.3%(89/102),5.9%(6/102),4.9%(5/102) and1.9%(2/102) respectively; and in FFDM the margin of benign mass was characterized clear, fuzzy, shadow and lobulated were10.2%(41/02),18.6%(19/102),41.1%(42/102) and0%(0/102) respectively. There was statistically difference between the two groups (P=0.000). The difference between DBT and FFDM for the margin of malignant masses also was statistically significant(P=0.001).The speculated margin of mass were62.5%(45/72) in DBT, and the fuzzy margin of mass were33.3%(24/72); In FFDM the fuzzy margin of mass were65.3%(47/70),and speculated margin of mass were only27.8%(20/72).4. In72cases of malignant masses, there were45cases margin of mass display speculated in DBT,47cases were observed structural distortion,40cases of mass were displayed increased blood flow,16cases of mass could be observed the vessels penetration; In FFDM the margins of20masses were displayed speculated,35masses could be visible structural distortion,38masses were showed increased blood flow, vascular penetration were observed in4masses. There was statistically significant of speculated structural distortions and vascular penetration between DBT and FFDM (P<0.05).5. In DBT the average longest diameter of masses was2.30±0.96cm, and in FFDM the average longest diameter of masses was2.41±0.99cm, the difference was statistically significant (T=9.520, P=0.000).[Conclusion]1. DBT diagnostic sensitivity and specificity for masses is higher than FFDM, particularly to improve the specificity. For the BI-RADS classification of benign and malignant masses, DBT is also more accurate than FFDM, especially for benign masses, DBT can reduced some lesions which are classified as BI-RADS category4in FFDM to BI-RADS category2or category3. 2. DBT can reduce the overlap of tissue that the margin of lesion can be showed more clearly, especially for speculated, structural distortions and vascular penetration. DBT can improve the diagnostic accuracy of benign and malignant breast masses. Further, because of the advantages of the margins of masses display, DBT in measurement of masses size is also superior FFDM. |
PDF Full Text Request