Applying a new feature fusion method to classify breast lesions

IntroductionBreast cancer is one of the leading causes of cancer deaths in female patients. Breast lesions can have various morphological diversities, ranging from benign to in situ to malignant. An important histopathological feature that distinguishes benign and malignant lesions is the presence or absence of the myoepithelial cell layer. The tumor protein p63 has been characterized as a reliable immunohistochemical marker and is uniquely displayed in the myoepithelial cells of breast parenchyma. Hence, it helps distinguish benign from malignant lesions.AimTo study the value of the immunohistochemical expression of p63 in the diagnosis of breast lesions, to investigate p63 expression in benign and malignant breast lesions, and to assess if loss of p63 expression is consistently associated with invasive disease.Materials and methodsAbout 98 cases of breast disease were studied, of which 86 were subjected to immunohistochemical staining for p63 and analyzed for its expression, and the staining arrangement was interpreted.ResultsOut of 86 cases, there were 46 benign and 40 malignant breast lesions. Immunohistochemical staining for p63 in the benign lesions showed continuous positivity with a score of three in 30 cases and discontinuous positivity with a score of two in 16 cases. In malignant lesions, p63 was not expressed and scored zero in 35 cases and discontinuous positive with a score of one in five cases. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of p63 in distinguishing the benign from the malignant lesions was 100%, 87.50%, 90.20%, 100%, and 94.19% respectively.ConclusionThere is a statistically significant difference in the expression of p63, a specific nuclear marker for myoepithelial cells, in benign and malignant breast lesions. It was consistently positive in benign breast lesions and negative in the majority of malignant breast lesions. Thus, p63 can act as a helpful immunohistochemical marker in categorizing histopathologically-difficult cases into benign or malignant ones.

Background and objective: To determine the optimum apparent diffusion coefficient (ADC) value in differentiating benign from malignant breast lesions. Materials and Methods: The study is a retrospective review of the patients who underwent breast magnetic resonance imaging (MRI) at King Chulalongkorn Memorial Hospital between January 2017 and May 2020. ADC values were measured by placement of region of interest (ROI) within the breast lesions using Philips DynaCAD breast analysis system and comparing it with histopathological diagnosis. A receiver-operating-characteristics (ROC) analysis was plotted and the area under the curve (AUC) was evaluated to find the ideal ADC value in the differentiation of benign and malignant breast lesions. Results: Two hundred and ten lesions in 163 female patients were included in the present study. One hundred twenty-six lesions (60%) were malignant and eighty-four lesions (40%) were benign. The mean ADC values of malignancy (0.913x10-3 mm2/s) were statistically lower than that of benign lesions (1.080x10-3 mm2/s) (mean difference 0.169x10-3 mm2/s, P < 0.001). According to the ROC analysis, the optimum cut-off ADC value of 0.991x10-3 mm2/s was an excellent predictor for differentiated benign and malignant breast lesions (AUC = 0.835, sensitivity 78.6%, specificity 82.5%, accuracy 81%, PPV 85.3% and NPV 75%). Conclusion: Diffusion-weighted imaging (DWI) was an effective MRI sequence to assess breast cancer by using ADC value as a key parameter in addition to other important imaging findings from MRI. The present study showed the mean ADC value of malignancy was statistically significantly lower than that of benign lesions. The cut-off ADC value of 0.991x10-3 mm2/s had good specificity, accuracy, and PPV to differentiate benign from malignant breast lesions.

Objective To discuss the value of conventional ultrasound combined with virtual touch tissue imaging (VTI) in differentiating benign and malignant breast lesions by breast imaging reporting and data system (BI-RADS) diagnostic criteria. Methods From January 2013 to October 2014, 395 breast lesions from 389 patients from the Tenth People's Hospital, Tongji University were included in present study. All the lesions were confirmed by surgery pathology. Conventional ultrasound and VTI examinations were done before surgery. Both benign and malignant lesions were classified into varied categories based on the boundary of lesions under VTI mode. Benign and malignant breast lesions were classified into BI-RADS categories by conventional ultrasound or conventional ultrasound combined with VTI. The χ2 test was used to compare the differences of lesions′ boundaries between VTI categories and VTI mode. The pathology findings were used as golden criteria, to calculate the sensitivity, specificity and accuracy of conventional ultrasound as well as conventional ultrasound combined with VTI. And their diagnostic performances were compared by McNemar′s test. Results Of 395 breast lesions, 282 (71.4%) were benign, 113 (28.6%) were malignant. 75.5% (213/282) benign lesions were classified as Ⅱ-Ⅲ under VTI mode, while 83.2% (94/113) malignant lesions were classified as Ⅳ-Ⅴ under VTI mode. The differences in classification between them was statistically significant (χ2=114.759, P<0.01). 77.7% (219/282) benign lesions had unclear boundaries under VTI mode, while 78.8% (89/113) malignant lesions had clear boundaries under VTI mode. The differences in boundaries between them was statistically significant (χ2=108.480, P<0.01). The sensitivity, specificity and accuracy of conventional ultrasound in differentiating benign and malignant breast lesions were 97.3%, 54.2% and 66.6% respectively; while the sensitivity, specificity and accuracy of conventional ultrasound combined with virtual touch tissue imaging in differentiating benign and malignant breast lesions were 98.2%, 84.4% and 88.4% respectively. The specificity and accuracy were higher when only using conventional ultrasound with statistical significance (χ2=79.425, 67.549, both P<0.01); however the sensitivity had no significantly different (χ2=11.974, P<0.01). Conclusion Conventional ultrasound combined with virtual touch tissue imaging can improve the specificity and the accuracy in differentiating benign and malignant breast lesions. Key words: Elasticity imaging techniques; Ultrasonography; Breast diseases; Diagnosis, differential

Breast cancer is the most common malignant tumor in women and a quantitative contrast-free method is highly desirable for its diagnosis. To investigate the performance of quantitative MRI in differentiating malignant from benign breast lesions and to compare with the Breast Imaging Reporting and Data System (BI-RADS). Retrospective. Eighty patients (56 with malignant lesions and 24 with benign lesions). Diffusion-weighted imaging (DWI) with a single-shot echo planar sequence and synthetic MRI with magnetic resonance image compilation (MAGiC) were performed at 3T. T1 relaxation time (T1 ), T2 relaxation time (T2 ), and proton density (PD) from synthetic MRI and apparent diffusion coefficient (ADC) from DWI were analyzed by two radiologists (Reader A, Reader B). Univariable and multivariable models were developed to optimize differentiation between malignant and benign lesions and their performances compared to BI-RADS. The diagnostic performance was evaluated using multivariate logistic regression analysis and area under the receiver operating characteristic (ROC) curves (AUC). T2 , PD, and ADC values for malignant lesions were significantly lower than those in benign breast lesions for both radiologists (all P < 0.05). The combined T2 , PD, and ADC model had the best performance for differentiating malignant and benign lesions with AUC, sensitivity, specificity, positive predictive value, and negative predictive values of 0.904, 94.6%, 87.5%, 94.6%, and 87.5%, respectively. The corresponding results for BI-RADS were no AUC, 94.6%, 75.0%, 89.8%, and 85.7%, respectively. The approach that combined synthetic MRI and DWI outperformed BI-RADS in the differential diagnosis of malignant and benign breast lesions and was achieved without contrast agents. This approach may serve as an alternative and effective strategy for the improvement of breast lesion differentiation. 3. 3.

BackgroundBreast cancer is a significant global health issue and is a leading cause of female deaths worldwide. Ultrasonography (US) is an adjunct technique for early diagnosis of breast cancer but has low specificity. So, researchers have suggested using US elastography to distinguish between benign and malignant breast lesions, which may lead to a decrease in unnecessary biopsy rates. We aimed to assess the diagnostic performance of strain elastography and if there was an increase in the diagnostic accuracy after adding it to conventional US results in BI-RADS 4 category breast lesions.ResultsThis cross-sectional study was carried out from June 2022 to June 2024 on 240 patients with 240 breast lesions that were diagnosed as BI-RADS 4 category by conventional US. All the cases were examined using conventional US and strain elastography. Out of the 240 lesions, 107 were benign (44.6%) and 133 were malignant (55.4%). There was a significant statistical difference in the mean age of patients with benign breast lesions (37.89 ± 13.42 years) than those with malignant lesions (48.83 ± 13.65 years) (p < 0.001). Lesions’ imaging characteristics detected by conventional US as shape, margin, aspect ratio, vascularity, and calcification all showed significant differences between benign and malignant breast lesions (p < 0.001). The diagnostic accuracy of conventional US alone was 73.3% and increased to 95.3% after combining it with strain elastography. The strain elastography score has a cutoff value of 4 in discriminating malignant from benign lesions (AUC = 0.9).ConclusionsStrain elastography provides an added diagnostic value in discriminating malignant breast lesions from benign lesions. The combination of strain elastography with US BI-RADS increases the diagnostic accuracy for the detection of malignant breast lesions which will decrease the need for unnecessary biopsy.

Sonographic Discrimination between Benign and Malignant Breast Lesions with Use of Disparity Processing

The purpose of the current study was to evaluate the performance of a continuous-time random-walk (CTRW) diffusion model for differentiating malignant and benign breast lesions and to consider the potential association between CTRW parameters and the Ki-67 expression. Sixty-four patients (46.2 ± 11.4 years) with breast lesions (29 malignant and 35 benign) were evaluated with the CTRW model, intravoxel incoherent motion model, and diffusion-weighted imaging. Echo planar diffusion-weighted imaging was conducted using 13 b-values (0-3000 s/mm2 ). Three CTRW model parameters, including an anomalous diffusion coefficient Dm , and two parameters related to temporal and spatial diffusion heterogeneity, α and β, respectively, were obtained, and had MRI b-values of 0-3000 s/mm2 . Receiver operating characteristic (ROC) analysis was conducted to determine the sensitivity, specificity, and diagnostic accuracy of CTRW parameters for differentiating malignant from benign breast lesions. In malignant breast lesions, the CTRW parameters Dm , α, and β were significantly lower than the corresponding parameters of benign breast lesions. In the malignant breast lesion group, the CTRW parameter Dm was significantly lower in high Ki-67 expression than in low Ki-67 expression. In ROC analysis, the combination of CTRW parameters (Dm , α, β) demonstrated the highest area under the curve value (0.985) and diagnostic accuracy (94.23%) in differentiating malignant and benign breast lesions. The CTRW model effectively differentiated malignant from benign breast lesions. The CTRW diffusion model offers a new way for noninvasive assessment of breast malignancy and better understanding of the proliferation of malignant lesions.

Axial-Shear Strain Imaging for Differentiating Benign and Malignant Breast Masses

The purpose of this study was to evaluate the diagnostic potential of the sonoelastographic strain index for differentiation of nonpalpable breast masses. Ninety-nine nonpalpable breast masses (79 benign and 20 malignant) in 94 women (mean age, 45 years; range, 21-68 years) who had been scheduled for a sonographically guided core biopsy were examined with B-mode sonography and sonoelastography. Radiologists who had performed the biopsies analyzed the B-mode sonograms and provided American College of Radiology Breast Imaging Reporting and Data System categories. The strain index (fat to lesion strain ratio) was calculated by dividing the strain value of the subcutaneous fat by that of the mass. The histologic result from the sonographically guided core biopsy was used as a reference standard. The diagnostic performance of the strain index and that of B-mode sonography were compared by receiver operating characteristic (ROC) curve analysis. The mean strain index values +/- SD were 6.57 +/- 6.62 (range, 1.29-28.69) in malignant masses and 2.63 +/- 4.57 (range, 0.54-38.76) in benign masses (P = .019). The area under the ROC curve values were 0.835 (95% confidence interval [CI], 0.747-0.902) for B-mode sonography and 0.879 (95% CI, 0.798-0.936) for the strain index (P = .490). The sensitivity, specificity, positive predictive value, and negative predictive value were 95% (19 of 20), 75% (59 of 79), 48% (19 of 39), and 98% (59 of 60), respectively, when a best cutoff point of 2.24 was used. The strain index based on the fat to lesion strain ratio has diagnostic performance comparable with that of B-mode sonography for differentiation of benign and malignant breast masses.

Objectives: This study aimed to evaluate the role of MRI in differentiation between benign and malignant well-circumscribed breast masses.&#x0D; Methods: A prospective cross-sectional study conducted on selected female patients with Breast mass who their age older than 35 years. All patients underwent examination using 1.5 Tesla magnetic resonance unit (MAGNETOM AERA, SIMENS).&#x0D; Results: 30 female patients were selected (mean age of 44.2±9.2 years, range between 35-69). Final findings of MRI revealed that 6 (20%) women had malignant breast mass while 24 (80%) women had benign breast mass. The histopathology revealed that 4 (13.3%) women were having malignancy while 26 (86.6%) were having benign breast tumor. There was a highly significant association between malignancy and hyperdense mammography findings of women (p&lt;0.001), multiple lymph nodes for women (p&lt;0.001), increased size of mass by MRI (p&lt;0.001), circumscribed macrolabular lesion on MRI, rapid early kinetic phase of enhanced MRI (p&lt;0.001), washout late kinetic phase enhanced MRI (p&lt;0.001), and MRI kinetic curve type 3(p&lt;0.001). There was a highly significant association between homogenous enhanced MRI findings of women and benign lesions (p&lt;0.001).&#x0D; Conclusion: The validity results of MRI in differentiation between benign and malignant circumscribed breast masses are high, except for positive predictive value, and application of kinetic curve dynamic contrast enhancement increases magnetic resonance in differentiation between benign and malignant circumscribed breast masses.

Differentiating early-stage breast cancer from benign breast masses is crucial for radiologists. Additionally, accurately assessing axillary lymph node metastasis (ALNM) plays a significant role in clinical management and prognosis for breast cancer patients. Chest computed tomography (CT) is a commonly used imaging modality in physical and preoperative evaluations. This study aims to develop a deep learning model based on chest CT imaging to improve the preliminary assessment of breast lesions, potentially reducing the need for costly follow-up procedures such as magnetic resonance imaging (MRI) or positron emission tomography-CT and alleviating the financial and emotional burden on patients. We retrospectively collected chest CT images from 482 patients with breast masses, classifying them as benign (n ═ 224) or malignant (n ═ 258) based on pathological findings. The malignant group was further categorized into ALNM-positive (n ═ 91) and ALNM-negative (n ═ 167) subgroups. Patients were randomly divided into training, validation, and test sets in an 8:1:1 ratio, with the test set excluded from model development. All patients underwent non-contrast chest CT before surgery. After preprocessing the images through cropping, scaling, and standardization, we applied ResNet-34, ResNet-50, and ResNet-101 architectures to differentiate between benign and malignant masses and to assess ALNM. Model performance was evaluated using sensitivity, specificity, accuracy, receiver operating characteristic (ROC) curves, and the area under the curve (AUC). The ResNet models effectively distinguished benign from malignant masses, with ResNet-101 achieving the highest performance (AUC: 0.964; 95% CI: 0.948–0.981). It also demonstrated excellent predictive capability for ALNM (AUC: 0.951; 95% CI: 0.926–0.975). In conclusion, these deep learning models show strong diagnostic potential for both breast mass classification and ALNM prediction, offering a valuable tool for improving clinical decision-making.

The incidence of breast cancer ranks first among female malignant tumor. With the increase of the sensitivity of color Doppler ultrasound blood flow, the blood flow distribution in and around the tumor can be clearly displayed, and the analysis of hemodynamic parameters is provided, which provides convenience for the study of tumor blood flow characteristics. Studies have shown that tumor cells can secrete a substance called angiogenesis factor, which makes the tumor site form a rich vascular network to promote tumor growth and metastasis. The tumor has many new blood vessels, abnormal structure, thin wall, lack of muscle layer, and is prone to form arteriovenous rash. These characteristics provide a pathological basis for color Doppler flow imaging (CDFI) for the diagnosis of breast cancer. This article discusses the role of two-dimensional sonographic features in the differential diagnosis of benign and malignant breast masses, CDFI was used to study the blood flow distribution and hemodynamic characteristics in benign and malignant breast masses; explore the value of blood flow characteristics and blood flow parameters in the differential diagnosis of breast masses. The experimental results show that the detection rate of blood flow signals and the classification of blood flow signals in the malignant group are higher than those in the benign group, mainly level II and III blood flow, and the irregular branched blood flow is more common, especially when the tumor appears penetrating blood flow supports the diagnosis of malignancy. PSV, RI and PI have a certain differential meaning in the diagnosis of benign and malignant breast masses. PSV, RI and PI of malignant masses are higher than benign masses. For tumors without obvious necrosis, the larger the tumor diameter, the richer the blood flow and the higher the blood flow grade is. The malignant tumors have more blood flow than the benign ones.

High false-positive recall rate is an important clinical issue that reduces efficacy of screening mammography. Aiming to help improve accuracy of classification between the benign and malignant breast masses and then reduce false-positive recalls, we developed and tested a new computer-aided diagnosis (CAD) scheme for mass classification using a database including 600 verified mass regions. The mass regions were segmented from regions of interest (ROIs) with a fixed size of 512×512 pixels. The mass regions were first segmented by an automated scheme, with manual corrections to the mass boundary performed if there was noticeable segmentation error. We randomly divided the 600 ROIs into 400 ROIs (200 malignant and 200 benign) for training, and 200 ROIs (100 malignant and 100 benign) for testing. We computed and analyzed 124 shape, texture, contrast, and spiculation based features in this study. Combining with previously computed 27 regional and shape based features for each of the ROIs in our database, we built an initial image feature pool. From this pool of 151 features, we extracted 13 features by applying the Sequential Forward Floating Selection algorithm on the ROIs in the training dataset. We then trained a multilayer perceptron model using these 13 features, and applied the trained model to the ROIs in the testing dataset. Receiver operating characteristic (ROC) analysis was used to evaluate classification accuracy. The area under the ROC curve was 0.8814±0.025 for the testing dataset. The results show a higher CAD mass classification performance, which needs to be validated further in a more comprehensive study.

Objective To investigate the application value of three-dimensional shear wave elastography(3D-SWE) in diagnosis of malignant and benign breast masses. Methods Sixty-seven patients with total 96 masses diagnosed by ultrasound and verified with surgical pathological analysis were selected. The masses were divided into benign group (54 masses) and malignant group (42 masses). The quantitative diagnosis was performed using two-dimensional shear wave elastography(2D-SWE) and 3D-SWE to evaluate the hardness of the masses. Receiver operating characteristic (ROC) curves of 2D-SWE and 3D-SWE were drawn to calculate the sensibility and specificity in diagnosis of breast masses. The qualitative diagnosis was performed according to the elastic color image classification to classify the elastic images of the lesions into type Ⅰ-Ⅳ. Results The quantitative diagnosis: ① There were significant differences in Emean, Emax and SD on planes of 2D-SWE and three orthogonal planes of 3D-SWE between the two groups (P 0.05); ③ Emean, Emax and SD on planes of 2D-SWE and three orthogonal planes of 3D- SWE both exhibited high diagnostic performance.However, there was no significant difference in Emean, Emax and SD between 2D-SWE and 3D-SWE (P>0.05). The qualitative diagnosis of elastic color mode: type Ⅰ and type Ⅱ account for most planes of 2D-SWE and three orthogonal planes of 3D-SWE in benign masses, while type Ⅲ and type Ⅳ account for most in malignant masses.Among the malignant group, type Ⅲ on the 3D-SWE coronal planes was significantly more than other three types (P<0.05). Conclusions For quantitative diagnosis, 2D-SWE and 3D-SWE are both beneficial to differentiate benign from malignant masses, but there is no difference in diagnostic efficacy; for qualitative diagnosis, 3D-SWE can provide more diagnostic information than 2D-SWE, especially on the coronal planes of 3D-SWE. Key words: Ultrasonography, mammary; Breast neoplasms; Shear wave elastography

Evaluating the value of screening breast masses by separate or combined use of multimodal Sound Touch Elastography. Women with 159 masses (mean size, 14.86 ± 6.57 mm; range, 5.30-30.00 mm) were enrolled in the study. The pathology results were adopted as diagnostic standards. The abilities of Young's modulus (E), shear modulus (G), and shear wave (C) to differentiate malignant and benign breast masses based on receiver operating characteristic curves were evaluated, and the optimal cutoff values were obtained. Sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio were calculated. Then, the values were combined to perform an overall analysis of Sound Touch Elastography using evidence-based medicine, construct forest plots, and calculate areas under the summary receiver operating characteristic curves, pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic scores. A total of 159 masses with a mean size of 14.86 ± 6.57 mm (range, 5.30-30.00 mm) were included. For the various parameters, the diagnostic values were as follows: Gmax > Emax > Cmax > Csd > Esd > Gsd > Emean > Gmean > Cmean . There were no significant differences in Emin , Gmin , or Cmin . When the 9 parameters were combined, the pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, diagnostic scores and areas under the summary receiver operating characteristic curves were 84% (95% confidence interval [CI], 79%-88%), 82% (95% CI, 80%-84%), 4.75 (95% CI, 4.15-5.43), 0.20 (95% CI, 0.15-0.25), 3.19 (95% CI, 2.84-3.54) and 90.2% (95% CI, 87%-92%), respectively. Sound Touch Elastography can be recognized as a new ultrasound-based diagnostic method for differentiation between benign and malignant breast masses.