Objective:To explore the feasibility of classification between carcinoma in situ and invasive carcinoma of breast using intratumoral and peritumoral radiomics based on breast dynamic contrast-enhanced (DCE) MRI.Methods:The retrospective study included consecutive invasive breast carcinoma pathological diagnosed by core needle biopsy or surgery from January 2013 to December 2013 and carcinoma in situ of breast diagnosed by surgery from January 2013 to December 2015 in Peking University First Hospital. All patients had pretreatment breast MRI images. A total of 251 cases (251 lesions) were included, with 208 invasive breast carcinoma and 43 carcinoma in situ of breast. They were all females and median age was 53 (23-82) years old. Patients were randomly divided into the training ( n=176) and testing dataset ( n=75) in a 7∶3 ratio. In the training dataset, combined with DCE mask and early enhancement images, intratumoral and peritumoral area were semi-automatic segmentation, and radiomics features were extracted and dimension reduction, finally a prediction model was established. Model performance was tested in the testing dataset. Receiver operating characteristic (ROC) curve and area under curve (AUC) were used to analyze the model prediction performance. Results:The prediction models established by intratumoral, peritumoral and intratumoral combined with peritumoral radiomics had good performance. The AUC of intratumoral, peritumoral and intratumoral combined with peritumoral radiomics prediction models in differentiating breast carcinoma in situ and invasive carcinoma were 0.865, 0.896 and 0.922 in the testing dataset, there was no significant difference in pairwise comparisons ( P>0.05). The sensitivity of intratumoral, peritumoral and intratumoral combined with peritumoral radiomics prediction models were 77.4%, 87.1%, 83.9%, the specificity were 92.3%, 84.6%, 100%, and the accuracy were 80.0%, 85.3%, 86.7%. Conclusion:It is potential feasible for classification between carcinoma in situ and invasive carcinoma of breast using intratumoral and peritumoral radiomics based on breast DCE MRI.

Objective:To compare the imaging features of renal cell carcinoma (RCC) associated with Xp11.2 translocation/TFE3 gene fusion (Xp11.2 RCC) with chromophobe RCC.Methods:From November 2016 to January 2020, 28 patients with Xp11.2 RCC and 28 patients with chromophobe RCC confirmed by pathology were retrospectively analyzed in Peking University First Hospital. All 23 patients underwent preoperative CT examination, and 5 patients underwent routine MRI in each group. The clinical and imaging features were observed and recorded. The CT features including side, location, size, boundary, shape, uniform density, composition (solid, cystic-solid, cystic), hemorrhage, calcification, lymph node metastasis of the lesions and distant metastasis were observed, and the CT value of the solid part of the tumor at each stage was measured. On MRI images, the signal of the lesion in each sequence and enhancement mode were observed. The differences in clinical and imaging characteristics between the 2 groups were compared using independent samples t test or χ 2 test. Results:The Xp11.2 RCC more frequently affected young [(27±10) years] patients, while chromophobe RCC more frequently involved middle-aged [(37±7) years] patients asymptomatically, and the difference was statistically significant ( t=-4.99, P<0.001). The lesion size of Xp11.2 RCC [(5.4±2.2) cm] were significantly smaller than that of chromophobe RCC [(6.9±1.8) cm] ( t=-2.93, P=0.005). There were significant differences in the density and composition of lesions between Xp11.2 RCC and chromophobe RCC (χ 2=4.60, 18.67, P=0.032,<0.001). There were no significant differences in the side, location, boundary, shape, hemorrhage, calcification, fat, lymph node metastasis and distant metastasis between the 2 kind of lesions (all P>0.05). The CT values of solid components in Xp11.2 RCC in cortico-medullary phase and delayed phase were higher than those in chromophobe RCC, and the difference were statistically significant ( t=11.80, 20.15, both P<0.001). Five cases of Xp11.2 RCC showed iso- or slightly hyperintense signal on T 1WI and slightly hypointense signal on T 2WI. Two cases showed delayed enhancement after enhancement, and 3 cases showed a slight decrease in delayed phase enhancement. Conclusion:Compared with chromophobe RCC, Xp11.2 RCC has certain characteristics in imaging manifestations (lesion size, density uniformity, composition, CT value of post-enhanced cortico-medullary phase and delayed phase). Imaging manifestations combining the clinical manifestations (age of onset) are helpful for preoperative diagnosis of Xp11.2 RCC.

Objective:To explore the feasibility of predicting axillary lymph node metastasis of breast cancer using radiomics analysis based on dynamic contrast-enhanced (DCE) MRI.Methods:The retrospective study enrolled 163 patients (163 lesions) with breast cancer diagnosed by core needle biopsy from January 2013 to December 2013 in Peking University First Hospital. The status of axillary lymph nodes in all patients was pathologically confirmed, and they had complete preoperative breast MRI images. Among the 163 patients, 94 patients were confirmed with axillary lymph node metastasis, and 69 patients without axillary lymph node metastasis. They were randomly divided into the training dataset ( n=115) and testing dataset ( n=48) in a 7∶3 ratio. The radiomics analysis was performed in the training dataset, including image preprocessing and labeling, radiomics feature extraction, radiomics model establishment and model predictive performance inspection. Model performance was tested in the testing dataset. Receiver operating characteristic curve and area under curve (AUC) was used to analyze the model prediction performance. Results:Of the 1 075 features extracted from the training dataset, principal component analyses (PCA) features 8, 41 and 67 were selected by random forest classifier. The radiomics model including 3 PCA features reached an AUC of 0.956 (95%CI 0.907-0.988), with sensitivity of 91.2%, specificity of 100% and accuracy of 94.8%. In the testing dataset, the radiomics model including 3 PCA features reached an AUC of 0.767 (95%CI 0.652-0.890), with sensitivity of 80.8%, specificity of 72.7% and accuracy of 77.1%.Conclusion:It is feasible to predict axillary lymph node metastasis using radiomics features based on DCE-MRI of breast cancer.

Objective To investigate the significance of the objective audit system of mammography diagnostic report in evaluating the diagnostic mammography practice in Peking University First Hospital.Methods According to the regulations issued by the US Food and Drug Administration (FDA 1992) and the mammographic practice parameters recommended by the American College of Radiology (ACR 2013),we retrospectively analyzed the diagnostic mammographic reports in Peking University First Hospital from November 14,2015 to November 14,2016.We calculated the diagnostic sensitivity,specificity,positive predictive value (PPV2),and the third positive predictive value (PPV3).The pathological and follow-up results(the negative patient)were used as the golden standard.All data were compared with the reference data of United States mammography diagnostic report system.Results A total of 1 779 mammography diagnostic reports were analyzed,including 137 malignant cases and 1 642 benign cases.The sensitivity of diagnosis of malignant breast lesions was 93.4％ (128/137),the specificity was 95.2％ (1 563/1 642),PPV2 was 61.8％ (128/207) and PPV3 was 83.1％(128/154).Compared with the recommended target value (sensitivity＞85.0％,specificity＞90.0％,PPV2:25.0％～40.0％),most statistical data were within the recommended range except PPV2 and PPV3,which werehigher.Conclusions The quality of the mammographic report in Peking University First Hospital reached the recommended level in the United States.However,the low proportion of recommendation of biopsy suggested a possibility of missed diagnosis.

Objective To investigate the early change of MR DWI parameters of breast cancer after neoadjuvant chemotherapy (NAC) and assess the relationship between the changes of the NAC response. Methods Patients with local advanced breast cancer diagnosed by biopsy were enrolled in the prospective study. They received DWI examination before NAC and then received DWI examination at 24 h after the first cycle of NAC. The pathological results of breast cancers were recorded after NAC finished. 14 patients (15 lesions) were enrolled in this study.The minimum ADC values of breast cancer before NAC and after the first cycle of NAC were recorded, and their change rates were calculated. Two weeks after NAC,all cases were divided into clinical complete response (cCR) and non-clinical complete response group according to clinical NAC response evaluation. Then lesions were divided into pathological complete response (pCR) and non-pathological complete response group according to pathological NAC response evaluation. Using the Paired samples t test,we analysed the change of ADC values 24 h after the first cycle of NAC. The ADC values and change rates were compared between cCR and non-cCR group using Mann-Whitney U test. The ADC values and change rates were also compared between pCR and non-pCR group using Mann-Whitney U test.Results The 15 cases were all invasive ductal carcinoma patients. According to clinical NAC response evaluation, 5 cases were cCR and 10 cases were non-cCR. According to pathological NAC response evaluation, 6 cases were pCR and 9 cases were non-pCR. Before NAC, the minimum ADC value of breast cancer was(0.78±0.15)×10-3 mm2/s. After the first cycle of NAC, the minimum ADC value of breast cancer was(0.91 ± 0.16) × 10-3 mm2/s. The minimum ADC value of breast cancer significantly increased 24 h after the first cycle of NAC (t=-2.954,P=0.010). But ADC value and its change rate were not significantly different between cCR and non-cCR groups, also between pCR and non-pCR groups (P>0.05). Conclusion Significant change in the minimum ADC value of breast cancer could be observed 24 h after the first cycle of NAC, but the change do not relate to clinical and pathological NAC response in this study.

BACKGROUNDThis multicenter prospective study aimed to assess the utility of dynamic enhanced magnetic resonance imaging (MRI) prior to breast-conserving surgery for breast cancer.

METHODSThe research subjects were drawn from patients with primary early resectable breast cancer treated in the breast disease centers of six three-level hospitals in Beijing from 1 January 2010 to 31 December 2012. The participants were allocated to a breast-conserving surgery group (breast-conserving group) or a total mastectomy group (total mastectomy group). Enhanced MRI was used to measure breast volume, longest diameter of tumor and tumor volume. The correlations between these measurements and those derived from histopathologic findings were assessed. The relationships between the success rate of breast-conserving surgery and MRI- and pathology-based measurement results were statistically analyzed in the breast-conserving group.

RESULTSThe study included 461 cases in the total mastectomy group and 195 in the breast-conserving group. Allocation to these groups was based on clinical indications and patient preferences. The cut-off for concurrence between MRI- and pathology-based measurements of the longest diameter of tumor was set at 0.3 cm. In the total mastectomy group, the confidence interval for 95% concurrence of these measurements was 35.41%-44.63%. Correlation coefficients for MRI and histopathology-based measurements of breast volume, tumor volume and tumor volume/breast volume ratio were r = 0.861, 0.569, and 0.600, respectively (all P < 0.001). In the breast-conserving group, with 0.30 cm taken as the cut-off for concurrence, the 95% confidence interval for MRI and pathology-based measurements of the longest diameter of tumor was 29.98%-44.01%. The subjective and objective success rates for breast-conserving surgery were 100% and 88.54%, respectively.

CONCLUSIONSThere were significant correlations between dynamic enhanced MRI- and histopathology-based measurements of the longest diameter of breast lesions, breast and tumor volumes, and breast volume/tumor volume ratios. Preoperative MRI examination improves the success rate of breast-conserving surgery.

Objective To study the value of the semiquantitative-parameter analysis of wash out index of time-intensity curve (Swash-out) in evaluating the therapeutic effect of neoadjuvant chemotherapy for locally advanced breast cancer (LABC).Methods Fifty-nine women with LABC underwent dynamic contrast enhancedt MRI examination before chemotherapy,after the 2nd cycle and the 4th cycle of chemotherapy.All patients were divided into major histological response group (MHR) and non-major histological response group (NMHR) according to the final pathologic response.Swash-out and the variancetrends of Swash-out before NAC,after the 2nd cycle of NAC and after the 4th cycle of NAC were compared in each group and between the two groups.According to the gold standard of Miller & Payne criterion,Receiver operating characteristic curve (ROC) analysis was performed to evaluate the predicting effect of Swash-out for NAC response,and to compare it with Semi-quantitative TIC curve indicators Smax (steepest slope) and PPE (peak percent enhancement).Results Fifty-nine patients of LABC patients were divided into a MHR group of 34 patients and a NMHR group of 25 patients.Swash before NAC of MHR group was-16.99 (-56.72-41.20),Swash-out after the 2nd cycle of NAC was 5.66(-69.45-53.08),Swash-out after 4th cycle of NAC was 15.95 (-7.80-54.23).Swash-out before NAC of NMHR group was-23.08 (-64.24-34.39),Swash-out after the 2nd cycle of NAC of NMHR group was-23.01 (-52.72-28.70),Swash-out after 4th cycle of NAC of NMHR group was-11.45 (-50.49-50.93).Swash-out variance rate of MHR group after the 2nd and the 4th cycle of NAC were-1.18 (-31.32-60.86) and 1.50 (-86.27-3.61),respectively.Swash-out variance rate of NMHR group after the 2nd and the 4th cycle of NAC were-0.28(-3.24-9.46) and 0.27 (-5.34-3.11),respectively.Swash-out was not significantly different between the two groups before NAC (Z =-0.97,P ＞0.05).Swash-out and Swash-out variance rate of MHR group after the 2nd cycle of NAC were significant higher than that of NMHR group (Z =-3.97 and-3.02,P ＜0.01).Swash-out and Swash-out variance rate of MHR group after the 4th cycle of NAC were significant higher than that of NMHR group (Z =-3.96 and-3.16,P ＜ 0.01).Area under curve (Az) after the 2nd and the 4th cycle of NAC were 0.805 and 0.804,respectively,and no significant difference was found between them (Z =0.019,P ＞0.05).Diagnostic cut-off points were-8.670 for the 2nd cycle of NAC and 4.105 for the 4th cycle of NAC.Diagnostic sensitivity was 79.42％,specificity was 76.00％ and Youden index was 0.554,for after the 2nd and the 4th cycle of NAC.Conclusion Swash-out of TIC curve before NAC cannot predict the response of NAC,Swash-out of TIC curve after the 2nd cycle of NAC and after the 4th cycle of NAC are efficient in predicting the response of NAC.

ObjectiveStudy the ADC value and the maximal diameter and their changes of breast cancer before and after neoadjuvant chemotherapy,to determine the relationship with different expression level of Ki-67.Methods Forty eight patients with breast cancer confirmed by biopsy underwent MR DWI and enhanced scan before and after 4 cyclesneoadjuvant chemotherapy.ReviewtheMRimages retrospectively.The ADC value and the maximum diameter( D)of the cancer foci were measured before and after chemotherapy,and the rate of their changes △ADC％ and △D％ were calculated.Using different Ki-67 index level,all the foci were divided into three groups:group A with Ki-67 ＜ 20％,group B with Ki-67 between 20％ and 60％,and group C with Ki-67 ＞ 60％.Using nonparameter test to compare the ADC values,△ADC％,D and △D％ of the three groups before and after chemotherapy,determine whether there were differences.ResultsBefore chemotherapy,the ADC value of group A ( n = 15 ) was 1.1 ×10-3 mm2/s[ (0.9 × 10-3—1.2 × 10-3) mm2/s],which was higher than that of group B[n = 8,0.9 ×10-3 mm2/s(0.9 × 10-3-1.0 × 10-3) mm2/s] and C [n =25,0.9 × 10-3 mm2/s(0.7 × 10-3—1.2 ×10-3) mm2/s],and the difference was statistically significant (P ＜0.05 ) ; while the ADC value of group C after chemotherapy was 1.3 × 10 -3 mm2/s[ (0.2 × 10 -3—1.4 × 10 -3 ) mm2/s],which was higher than that of group A [1.1 × 10-3 mm2/s,(1.0 × 10-3—1.2 × 10-3) mm2/s] and B[1.1 × 10-3 mm2/s,( 1.0 × 10-3-1.1 × 10 -3 ) mm2/s],and the differences were statistically significance ( P ＜ 0.01 ) ; the ADC change rate( △ADC％ ) of group C was 45.5％ ( - 12.0％ —78.6％ ),which was greater than group A [45.5％ ( - 12.0％—78.6％ ) ] and B [ 45.5％ ( - 12.0％—78.6％ ) ],the difference was significant (P ＜ 0.01 ).The maximum diameters of group A were 2.2 cm (2.0—2.4 cm)and 1.0 cm(0.0—1.4 cm)before and after chemotherapy,lower than those of group B [ 3.7 cm ( 3.6—3.9 cm ) before NAC,2.9 cm (0.0-3.1 em) after NAC] and group C[3.4 cm(2.7—4.2 cm) before NAC,1.9 cm(0.0—2.2 cm) after NAC ],and the differences were statistically significant ( P ＜ 0.05 ) ; the change rate of the maximum diameter in group B was 21.6％ ( - 15.2％—27.5％ ),which was less than group A [52.7％ ( -23.6％—72.1％)] (P＜0.01) and C [51.2％ ( -10.3％—92.6％)] (P ＜0.05),and the difference was statistically significant.Conclusion The ADC values and the maximal diameter of breast cancer differed with different expression levels of Ki-67 index before and after neoadjuvant chemotherapy,and the response to neoadjuvant chemotherapy of which varied as well.

Objective To Compare different methods of quantitative breast density measurement.Methods The study included sixty patients who underwent both mammography and breast MRI. The breast density was computed automatically on digital mammograms with R2 workstation. Two experienced radiologists read the mammograms and assessed the breast density with Wolfe and ACR classification respectively. Fuzzy C-means clustering algorithm (FCM) was used to assess breast density on MRI. Each assessment method was repeated after 2 weeks. Spearman and Pearson correlations of inter- and intrareader and intermodality were computed for density estimates. Results Inter- and intrareader correlation of Wolfe classification were 0. 74 and 0. 65, and they were 0. 74 and 0. 82 for ACR classification respectively.Correlation between Wolfe and ACR classification was 0. 77. High interreader correlation of 0. 98 and intrareader correlation of 0. 96 was observed with MR FCM measurement. And the correlation between digital mammograms and MRI was high in the assessment of breast density (r = 0. 81, P ＜ 0. 01). Conclusion High correlation of breast density estimates on digital mammograms and MRI FCM suggested the former could be used as a simple and accurate method.

Objective To assess the value of initial MRI measurements of breast tumor diameter and pathological responses for predicting response to neoadjuvant chemotherapy (NAC). Methods Fortynine patients who underwent neoadjuvant chemotherapy were included in this study. The longest diameter of each tumor was measured on MRI before and after two cycles of NAC and just before operation. Pathological responses were evaluated by Miller and Payne grading system criteria by comparing post-operation breast tissue with large core biopsy tissue. Changes of diameter after two cycles of NAC and before operation were compared and pathological responses were evaluated. ROC and spearman correlation analysis was used.Results The sensitivity and specificity of initial diameter for predicting response to NAC were 96. 7% (29/30) and 84.2% (16/19). There was a moderate correlation between pathological responses and diameter measured pre-operation ( r = 0. 613, P = 0. 000 ). Conclusion The initial diameter changes of breast tumor could predict response to neoadjuvant chemotherapy, pathological response have moderate correlation with changes of tumor diameter.