Diagnosis of breast tumors with sonographic texture analysis using wavelet transform and neural networks

To increase the ability of ultrasonographic technology for the differential diagnosis of solid breast tumors, we describe a novel computer-aided diagnosis (CADx) system using neural networks for classification of breast tumors. Tumor regions and surrounding tissues are segmented from the physician-located region-of-interest (ROI) images by applying our proposed segmentation algorithm. Cooperating with the segmentation algorithm, three feasible features, including variance contrast, autocorrelation contrast and distribution distortion of wavelet coefficients, were extracted from the ROI images for further classification. A multilayered perceptron (MLP) neural network trained using error back-propagation algorithm with momentum was then used for the differential diagnosis of breast tumors on sonograms. In the experiment, 242 cases (including benign breast tumors from 161 patients and carcinomas from 82 patients) were sampled with k-fold cross-validation (k = 10) to evaluate the performance. The receiver operating characteristic (ROC) area index for the proposed CADx system is 0.9396 +/- 0.0183, the sensitivity is 98.77%, the specificity is 81.37%, the positive predictive value is 72.73% and the negative predictive value is 99.24%. Experimental results showed that our diagnosis model performed very well for breast tumor diagnosis.     

Image retrieval with principal component analysis for breast cancer diagnosis on various ultrasonic systems.

OBJECTIVES: We present a computer-aided diagnostic (CAD) system with textural features and image retrieval strategies for classifying benign and malignant breast tumors on various ultrasonic systems. Effective applications of CAD have used different types of texture analysis. Nevertheless, most approaches performed in a specific ultrasonic machine do not indicate whether the technique functions satisfactorily for other ultrasonic systems. This study evaluated a series of pathologically proven breast tumors using various ultrasonic systems. METHODS: Altogether, 600 ultrasound images of solid breast nodules comprising 230 malignant and 370 benign tumors were investigated. All ultrasound images were acquired from four diverse ultrasonic systems. The suspicious tumor area in the ultrasound image was manually chosen as the region-of-interest (ROI) subimage. Textural features extracted from the ROI subimage are supported in classifying the breast tumor as benign or malignant. However, the textural feature always behaves as a high-dimensional vector. In practice, high-dimensional vectors are unsatisfactory at differentiating breast tumors. This study applied the principal component analysis (PCA) to project the original textural features into a lower dimensional principal vector that summarized the original textural information. The image retrieval techniques were employed to differentiate breast tumors, according to the similarities of the principal vectors. The query ROI subimages were identified as malignant or benign tumors according to characteristics of retrieved images from the ultrasound image database. RESULTS: Using the proposed CAD system, historical cases could be directly added into the database without a retraining program. The area under the receiver-operating characteristics curve for the system was 0.970+/-0.006. CONCLUSION: The CAD system identified solid breast nodules with comparatively high accuracy in the different ultrasound systems investigated.     

Breast cancer diagnosis using three-dimensional ultrasound and pixel relation analysis

Because ultrasound (US) imaging offers benefits compared with other medical imaging techniques, it is used routinely in nearly all hospitals and many clinics. However, the surface features and internal structure of a tumor are not easily demonstrated simultaneously using the traditional 2-D US. The newly developed three-dimensional (3-D) US can capture the morphology of a breast tumor and overcome the limitations of the traditional 2-D US. This study deals with pixel relation analysis techniques for use with 3-D breast US images and compares its performance to 2-D versions of the images. The 3-D US imaging was performed using a Voluson 530 scanner. The rectangular subimages of the volume-of-interest (VOI) were manually selected and the selected VOIs were outlined to include the entire extent of the tumor margin. The databases in this study included 54 malignant and 161 benign tumors. All solid nodules at US belong over C3 (probably benign) according to ACR BI-RADS category. All or some selected 2-D slices were used separately to calculate the diagnosis features for a 3-D US data set. We have proposed and compared several different methods to extract the characteristics of these consecutive 2-D images. As shown in our experiments, the diagnostic results were better than those of the conventional 2-D US. In the experiments, the area index Az under ROC curve of the proposed 3-D US method can achieve 0.9700 +/- 0.0118, but Az of the 2-D US is only 0.8461 +/- 0.0315. The p value of these two Az differences using z test is smaller than 0.01. Furthermore, we can find that the features from only several slices are enough to provide good diagnostic results if the adopted features are modified from the 2-D features.     

应用去噪后增强算法辅助乳腺肿块的超声诊断研究

目的探讨超声图像去噪后增强算法对乳腺肿块良恶性检测与分类的价值。方法选用211例603幅乳腺肿块超声图片(良性109例,恶性102例)进行去噪后增强处理,以手术病理结果作为金标准,对乳腺肿块原始图片和处理后图片进行分析,来区分乳腺肿块的良、恶性。利用ROC曲线下面积表现去噪后增强前后的诊断性能,计算超声诊断的准确率。结果通过去噪后增强算法处理后,使腺体和周围组织能分离,突出了腺体和病灶的部位,细节显示更加清晰,超声与病理诊断的各项指标符合率明显提高,准确率提高至92.73%,原片与处理后图片ROC曲线下面积二者之间差异有显著性统计学意义(P0.001)。结论新的超声图像去噪后增强算法可明显地改善了图像质量,提高了乳腺肿块的正确诊断率。     

Ultrasonic Nakagami Imaging: A Strategy to Visualize the Scatterer Properties of Benign and Malignant Breast Tumors

Previous studies have demonstrated the usefulness of the Nakagami parameter in characterizing breast tumors by ultrasound. However, physicians or radiologists may need imaging tools in a clinical setting to visually identify the properties of breast tumors. This study proposed the ultrasonic Nakagami image to visualize the scatterer properties of breast tumors and then explored its clinical performance in classifying benign and malignant tumors. Raw data of ultrasonic backscattered signals were collected from 100 patients (50 benign and 50 malignant cases) using a commercial ultrasound scanner with a 7.5 MHz linear array transducer. The backscattered signals were used to form the B-scan and the Nakagami images of breast tumors. For each tumor, the average Nakagami parameter was calculated from the pixel values in the region-of-interest in the Nakagami image. The receiver operating characteristic (ROC) curve was used to evaluate the clinical performance of the Nakagami image. The results showed that the Nakagami image shadings in benign tumors were different from those in malignant cases. The average Nakagami parameters for benign and malignant tumors were 0.69 ± 0.12 and 0.55 ± 0.12, respectively. This means that the backscattered signals received from malignant tumors tend to be more pre-Rayleigh distributed than those from benign tumors, corresponding to a more complex scatterer arrangement or composition. The ROC analysis showed that the area under the ROC curve was 0.81 ± 0.04 and the diagnostic accuracy was 82%, sensitivity was 92% and specificity was 72%. The results showed that the Nakagami image is useful to distinguishing between benign and malignant breast tumors.     

Improvement in breast tumor discrimination by support vector machines and speckle-emphasis texture analysis

Recent statistics show that breast cancer is a major cause of death among women in developed countries. Hence, finding an accurate and effective diagnostic method is very important. In this paper, we propose a high precision computer-aided diagnosis (CAD) system for sonography. We utilize a support vector machine (SVM) to classify breast tumors according to their texture information surrounding speckle pixels. We test our system with 250 pathologically-proven breast tumors including 140 benign and 110 malignant ones. Also we compare the diagnostic performances of three texture features, i.e., speckle-emphasis texture feature, nonspeckle-emphasis texture feature and conventional all pixels texture feature, applied to breast sonography using SVM. In our experiment, the accuracy of SVM with speckle information for classifying malignancies is 93.2% (233/250), the sensitivity is 95.45% (105/110), the specificity is 91.43% (128/140), the positive predictive value is 89.74% (105/117) and the negative predictive value is 96.24% (128/133). Based on the experimental results, speckle phenomenon is a useful tool to be used in computer-aided diagnosis; its performance is better than those of the other two features. Speckle phenomenon, which is considered as noise in sonography, can intrude into judgments of a physician using naked eyes but it is another story for application in a computer-aided diagnosis algorithm.     

应用计算机人工神经网络系统结合CT与PET鉴别良、恶性肺结节

目的 利用肺结节CT、PET特征,开发计算机人工神经网络(ANN)辅助诊断系统,评价其对肺结节良恶性的鉴别能力。方法 连续收集112例肺内单发小结节(<3.0 cm)患者,均接受PET/CT及胸部CT检查,二者间隔小于1个月。112例患者中恶性肺结节52例,良性60例,均经组织学或临床随诊证实。利用结节的CT特征及PET特征开发计算机ANN辅助诊断系统。计算机ANN的训练及测试采用Round-Robin方法。采用ROC方法评价计算机ANN输出结果并进行统计学分析。结果 CT计算机ANN程序采用20个输入单元,包括4个临床特征及16个CT特征,ROC曲线下面积(Az)为0.83;PET计算机ANN程序采用4个临床特征及1个PET特征作为5个输入单元,Az值为0.91;CT+PET计算机ANN程序采用临床特征CT及PET所有21个输入单元,Az值为0.95。与CT计算机ANN程序、PET计算机ANN程序相比,CT+PET计算机ANN程序输出结果明显提高(P＝0.015、0.037)。 结论 CT+PET ANN计算机辅助诊断程序输出结果优于单纯PET或CT计算机ANN结果。当PET对肺结节诊断有困难时,结节的CT特征有助于鉴别诊断。     

Computer-Aided Diagnosis for Breast Tumors by Using Vascularization of 3-D Power Doppler Ultrasound

This study assessed the accuracy of three-dimensional (3-D) power Doppler ultrasound in differentiating between benign and malignant breast tumors by using a support vector machine (SVM). A 3-D power Doppler ultrasonography was performed on 164 patients with 86 benign and 78 malignant breast tumors. The volume-of-interest (VOI) in 3-D ultrasound images was automatically generated from three rectangular regions-of-interest (ROI). The vascularization index (VI), flow index (FI) and vascularization-flow index (VFI) on 3-D power-Doppler ultrasound images were evaluated for the entire volume area, computer extracted VOI area and the area outside the VOI. Furthermore, patient's age and VOI volume were also applied for breast tumor classifications. Each ultrasonography in this study was classified as benign or malignant based on the features using the SVM model. All the tumors were sampled using k-fold cross-validation (k = 10) to evaluate the diagnostic performance with receiver operating characteristic (ROC) curves. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of SVM for classifying malignancies were 94%, 69%, 73%, 92% and 81%, respectively. The classification performance in terms of Az value for the ROC curve of the features derived from 3-D power Doppler is 0.91. This study indicates that combining 3-D power Doppler vascularity with patient's age and tumor size offers a good method for differentiating benign andmalignant breast tumors. (E-mail: ylhuang@thu.edu.tw (Y.-L.H.); darren_chen@cch.org.tw (D.-R.C.))     

CAD Algorithms for Solid Breast Masses Discrimination: Evaluation of the Accuracy and Interobserver Variability

For a successful computer-aided diagnosis (CAD) approach, investigating the benefit of the output for radiologist diagnosis is as important as developing the computer algorithm itself. To evaluate the accuracy and the interobserver variability of two newly developed CAD algorithms for breast mass discrimination, eight radiologists with varied experience in breast ultrasonography (US) independently reviewed the lesions according to Breast Imaging Reporting and Data System (BI-RADS)-US. They interpreted the original ultrasound images, provided a final assessment category to indicate the probability of malignancy and then made a further diagnosis using the images processed by the proposed CAD algorithms. The receiver operating characteristic (ROC) curve and Cohen's κ statistics were employed to evaluate the effect of the CAD algorithms on radiologist diagnoses. By using the proposed CAD approach, the quality of the images was improved and more information was provided to the observers. With the processed images, the areas under the ROC (Az) of each reader (0.86∼0.89) were greater than those with the original ultrasound images (0.81∼0.86) and all the radiologists improved their performance significantly (p < 0.05) except two senior radiologists (p > 0.05). The Az values of the junior radiologists with CAD were comparable to those of the senior radiologists. Cohen's κ statistics showed that better interobserver agreement was obtained by using the processed images. We conclude that the proposed CAD method is more helpful for the junior radiologists than for the senior ones and it also showed the advantage of decreasing interobserver variability. (E-mail: jwtian2004@yahoo.com.cn)     

Breast tumor classification using fuzzy clustering for breast elastography

Elastography is a new ultrasound imaging technique to provide the information about relative tissue stiffness. The elasticity information provided by this dynamic imaging method has proven to be helpful in distinguishing benign and malignant breast tumors. In previous studies for computer-aided diagnosis (CAD), the tumor contour was manually segmented and each pixel in the elastogram was classified into hard or soft tissue using the simple thresholding technique. In this paper, the tumor contour was automatically segmented by the level set method to provide more objective and reliable tumor contour for CAD. Moreover, the elasticity of each pixel inside each tumor was classified by the fuzzy c-means clustering technique to obtain a more precise diagnostic result. The test elastography database included 66 benign and 31 malignant biopsy-proven tumors. In the experiments, the accuracy, sensitivity, specificity and the area index Az under the receiver operating characteristic curve for the classification of solid breast masses were 83.5% (81/97), 83.9% (26/31), 83.3% (55/66) and 0.902 for the fuzzy c-means clustering method, respectively, and 59.8% (58/97), 96.8% (30/31), 42.4% (28/66) and 0.818 for the conventional thresholding method, respectively. The differences of accuracy, specificity and Az value were statistically significant (p < 0.05). We conclude that the proposed method has the potential to provide a CAD tool to help physicians to more reliably and objectively diagnose breast tumors using elastography.(E-mail: rfchang@csie.ntu.edu.tw)     

Automated Segmentation of Ultrasonic Breast Lesions Using Statistical Texture Classification and Active Contour Based on Probability Distance

Because of its complicated structure, low signal/noise ratio, low contrast and blurry boundaries, fully automated segmentation of a breast ultrasound (BUS) image is a difficult task. In this paper, a novel segmentation method for BUS images without human intervention is proposed. Unlike most published approaches, the proposed method handles the segmentation problem by using a two-step strategy: ROI generation and ROI segmentation. First, a well-trained texture classifier categorizes the tissues into different classes, and the background knowledge rules are used for selecting the regions of interest (ROIs) from them. Second, a novel probability distance-based active contour model is applied for segmenting the ROIs and finding the accurate positions of the breast tumors. The active contour model combines both global statistical information and local edge information, using a level set approach. The proposed segmentation method was performed on 103 BUS images (48 benign and 55 malignant). To validate the performance, the results were compared with the corresponding tumor regions marked by an experienced radiologist. Three error metrics, true-positive ratio (TP), false-negative ratio (FN) and false-positive ratio (FP) were used for measuring the performance of the proposed method. The final results (TP = 91.31%, FN = 8.69% and FP = 7.26%) demonstrate that the proposed method can segment BUS images efficiently, quickly and automatically. (E-mail: hengda.cheng@usu.edu)     

Computer-aided diagnosis for the classification of breast masses in automated whole breast ultrasound images

New automated whole breast ultrasound (ABUS) machines have recently been developed and the ultrasound (US) volume dataset of the whole breast can be acquired in a standard manner. The purpose of this study was to develop a novel computer-aided diagnosis system for classification of breast masses in ABUS images. One hundred forty-seven cases (76 benign and 71 malignant breast masses) were obtained by a commercially available ABUS system. Because the distance of neighboring slices in ABUS images is fixed and small, these continuous slices were used for reconstruction as three-dimensional (3-D) US images. The 3-D tumor contour was segmented using the level-set segmentation method. Then, the 3-D features, including the texture, shape and ellipsoid fitting were extracted based on the segmented 3-D tumor contour to classify benign and malignant tumors based on the logistic regression model. The Student’s t test, Mann-Whitney U test and receiver operating characteristic (ROC) curve analysis were used for statistical analysis. From the Az values of ROC curves, the shape features (0.9138) are better than the texture features (0.8603) and the ellipsoid fitting features (0.8496) for classification. The difference was significant between shape and ellipsoid fitting features (p = 0.0382). However, combination of ellipsoid fitting features and shape features can achieve a best performance with accuracy of 85.0% (125/147), sensitivity of 84.5% (60/71), specificity of 85.5% (65/76) and the area under the ROC curve Az of 0.9466. The results showed that ABUS images could be used for computer-aided feature extraction and classification of breast tumors. (E-mail: rfchang@csie.ntu.edu.tw)     

Tissue classification with generalized spectrum parameters.

This paper presents performance comparisons between breast tumor classifiers based on parameters from a conventional texture analysis (CTA) and the generalized spectrum (GS). The computations of GS-based parameters from radiofrequency (RF) ultrasonic scans and their relationship to underlying scatterer properties are described. Clinical experiments demonstrate classifier performances using 22 benign and 24 malignant breast mass regions taken from 40 patients. Linear classifiers based on parameters from the front edge, back edge and interior tumor regions are examined. Results show significantly better performances for GS-based classifiers, with improvements in empirical receiver operating characteristic (ROC) areas of greater than 10%. The ROC curves show GS-based classifiers achieving a 90% sensitivity level at 50% specificity when applied to the back-edge tumor regions, an 80% sensitivity level at 65% specificity when applied to the front-edge tumor regions, and a 100% sensitivity level at 45% specificity when applied to the interior tumor regions.     

Breast tissue characterization using FARMA modeling of ultrasonic RF echo

A number of empirical and analytical studies demonstrated that the ultrasound RF echo reflected from tissue exhibits 1/f characteristics. In this paper, we propose to model 1/f characteristics of the ultrasonic RF echo by a novel parsimonious model, namely the fractional differencing auto regressive moving average (FARMA) process, and evaluated diagnostic value of model parameters for breast cancer malignancy differentiation. FARMA model captures the fractal and long term correlated nature of the backscattered speckle texture and facilitates robust efficient estimation of fractal parameters. In our study, in addition to the computer generated FARMA model parameters, we included patient age and radiologist's prebiopsy level of suspicion (LOS) as potential indicators of malignant and benign masses. We evaluated the performance of the proposed set of features using various classifiers and training methods using 120 in vivo breast images. Our study shows that the area under the receiver operating characteristics (ROC) curve of FARMA model parameters alone is superior to the area under the ROC curve of the radiologist's prebiopsy LOS. The area under the ROC curve of the three sets of features yields a value of 0.87, with a confidence interval of [0.85, 0.89], at a significance level of 0.05. Our results suggest that the proposed method of ultrasound RF echo model leads to parameters that can differentiate breast tumors with a relatively high precision. This set of RF echo features can be incorporated into a comprehensive computer-aided diagnostic system to aid physicians in breast cancer diagnosis.     

Semi-supervised segmentation of echocardiography videos via noise-resilient spatiotemporal semantic calibration and fusion

We present a novel model for left ventricle endocardium segmentation from echocardiography video, which is of great significance in clinical practice and yet a challenging task due to (1) the severe speckle noise in echocardiography videos, (2) the irregular motion of pathological heart, and (3) the limited training data caused by high annotation cost. The proposed model has three compelling characteristics. First, we propose a novel adaptive spatiotemporal semantic calibration method to align the feature maps of consecutive frames, where the spatiotemporal correspondences are figured out based on feature maps instead of pixels, thereby mitigating the adverse effects of speckle noise in the calibration. Second, we further learn the importance of each feature map of neighbouring frames to the current frame from the temporal perspective so as to distinctively rather than uniformly harness the temporal information to tackle the irregular and anisotropic motions. Third, we integrate these techniques into the mean teacher semi-supervised architecture to leverage a large amount of unlabeled data to improve the segmentation accuracy. We extensively evaluate the proposed method on two public echocardiography video datasets (EchoNet-Dynamic and CAMUS), where the average dice coefficient on the left ventricular endocardium segmentation achieves 92.87% and 93.79%, respectively. Comparisons with state-of-the-art methods also demonstrate the effectiveness of the proposed method by achieving a better segmentation performance with a faster speed.     

Texture analysis of breast tumors on sonograms.

We performed a feasibility study to determine if the texture features extracted from sonograms can be used to predict malignant or benign breast pathology by the proposed artificial neural network and to compare the diagnostic results with the radiologists' results. A total of 1,020 images (4 different rectangular regions from the 2 orthogonal imaging planes of each tumor) from 255 patients were used as samples. When a sonogram was performed, 1 physician identified the region of interest in the sonogram; then, a neural network model, using 24 autocorrelation texture features, classified the tumor as benign or malignant. Three radiologists who were unfamiliar with the samples also classified these images. The receiver operating characteristic (ROC) area index for the proposed neural network system is 0.9840 +/- 0.0072. The neural network identified 35 of 36 malignancies and 211 of 219 benign tumors using all 4 regions of interest. The radiologists, on average, identified 19 of 36 malignancies, with 12 tumors called indeterminate and 4 tumors called benign. We conclude that benign and malignant breast tumors can be distinguished using interpixel correlation in digital ultrasonic images.     

数字胸片多频域后处理对计算机辅助检测系统肺结节检测的影响

目的评价数字胸片多频域后处理对计算机辅助检测(CAD)系统肺结节检出的影响。方法对经CT证实的54例肺结节病例及54例正常者的数字胸片选择三种不同的structure preference参数进行多频域后处理,在1次曝光的条件下各得到标准图像、高通过图像及低通过图像三组图像,然后,采用计算机辅助检测系统(IQQATM-Chest V1.2)进行阅片,首先由2名观察者共同根据CT结果客观分析并记录CAD检出的结节数及假阳性数。其次,由另外4名观察者(高年资及低年资放射医师各2名)应用CAD系统所提供的肺结节的智能质量和数量分析功能,对CAD输出图像独立进行分析,并记录检出结节数及假阳性数,采用受试者操作特征(ROC)曲线分析观察结果。结果三组图像中,低通过组图像的平均ROC曲线下面积最大,人-机交互前、高年资及低年资人-机交互后的平均ROC曲线下面积分别为0.77、0.81、0.80。高通过组图像的平均ROC曲线下面积为最小,人-机交互前、高年资及低年资人-机交互后的平均ROC曲线下面积分别为0.57、0.67、0.71。三组图像的平均ROC曲线下面积存在差异(P<0.01)。结论数字胸片多频域后处理影响计算机辅助检测系统肺结节的检出。     

基于非下采样双树复轮廓波变换的小波纹理特征识别肺良恶性结节CT图像

目的 观察基于非下采样双树复轮廓波变换（NSDTCT）的小波纹理特征在识别肺良恶性结节CT图像中的应用价值。方法 从肺结节患者的CT图像中分别提取基于NSDTCT和基于Contourlet变换的小波纹理参数,对高维纹理参数采用单因素分析、Lasso回归等方法进行降维。对降维后的纹理参数分别构建诊断良恶性肺结节的支持向量机分类诊断模型,绘制ROC曲线,比较2种方法的诊断效能。结果 采用NSDTCT方法,基于经Lasso降维且自变量数目较少的纹理参数构建的诊断模型分类效果最好,判断良恶性肺结节的准确率为98.37%,AUC为1.00;采用Contourlet变换方法,基于全部提取纹理参数构建的模型分类效果最好,诊断准确率为56.05%,AUC为0.73;2个模型的ROC曲线的AUC差异有统计学意义（Z=6.430,P<0.001）。结论 基于NSDTCT的纹理分析方法对判断良恶性肺结节的准确性较高。     

Stepwise logistic regression analysis of tumor contour features for breast ultrasound diagnosis.

To assist the ultrasound (US) differential diagnosis of solid breast tumors by using stepwise logistic regression (SLR) analysis of tumor contour features, we retrospectively reviewed 111 medical records of digitized US images of breast pathologies. They were pathologically proved benign breast tumors from 40 patients (i.e., 40 fibroadenomas) and malignant breast tumors from 71 patients (i.e., 71 infiltrative ductal carcinomas). Radiologists, before analysis by the computer-aided diagnosis (CAD) system, segmented the tumors manually. The contour features were calculated by measuring the radial length of tumor boundaries. The features selection process was accomplished using a stepwise analysis procedure. Then, an SLR model with contour features was used to classify tumors as benign or malignant. In this experiment, cases were sampled with "leave-one-out" test methods to evaluate the SLR performance using a receiver operating characteristic (ROC) curve. The accuracy of our SLR model with contour features for classifying malignancies was 91.0% (101 of 111 tumors), the sensitivity was 97.2% (69 of 71), the specificity was 80.0% (32 of 40), the positive predictive value was 89.6% (69 of 77), and the negative predictive value was 94.1% (32 of 34). The CAD system using SLR can differentiate solid breast nodules with relatively high accuracy and its high negative predictive value could potentially help inexperienced operators to avoid misdiagnoses. Because the SLR model is trainable, it could be optimized if a larger set of tumor images were supplied.