Stratifying High-Risk Thyroid Nodules Using a Novel Deep Learning System.

INTRODUCTION: The current ultrasound scan classification system for thyroid nodules is time-consuming, labor-intensive, and subjective. Artificial intelligence (AI) has been shown to increase the accuracy of predicting the malignancy rate of thyroid nodules. This study aims to demonstrate the state-of-the-art Swin Transformer to classify thyroid nodules. MATERIALS AND METHODS: Ultrasound images were collected prospectively from patients who received fine needle aspiration biopsy for thyroid nodules from January 2016 to June 2021. One hundred thirty-nine patients with malignant thyroid nodules were enrolled, while 235 patients with benign nodules served as controls. Images were fed to Swin-T and ResNeSt50 models to classify the thyroid nodules. RESULTS: Patients with malignant nodules were younger and more likely male compared to those with benign nodules. The average sensitivity and specificity of Swin-T were 82.46% and 84.29%, respectively. The average sensitivity and specificity of ResNeSt50 were 72.51% and 77.14%, respectively. Receiver operating characteristics analysis revealed that the area under the curve of Swin-T was higher (AUC=0.91) than that of ResNeSt50 (AUC=0.82). The McNemar test evaluating the performance of these models showed that Swin-T had significantly better performance than ResNeSt50.Swin-T classifier can be a useful tool in helping shared decision-making between physicians and patients with thyroid nodules, particularly in those with high-risk characteristics of sonographic patterns.

Machine Learning: Using Xception, a Deep Convolutional Neural Network Architecture, to Implement Pectus Excavatum Diagnostic Tool from Frontal-View Chest X-rays.

Pectus excavatum (PE), a chest-wall deformity that can compromise cardiopulmonary function, cannot be detected by a radiologist through frontal chest radiography without a lateral view or chest computed tomography. This study aims to train a convolutional neural network (CNN), a deep learning architecture with powerful image processing ability, for PE screening through frontal chest radiography, which is the most common imaging test in current hospital practice. Posteroanterior-view chest images of PE and normal patients were collected from our hospital to build the database. Among them, 80% were used as the training set used to train the established CNN algorithm, Xception, whereas the remaining 20% were a test set for model performance evaluation. The performance of our diagnostic artificial intelligence model ranged between 0.976-1 under the receiver operating characteristic curve. The test accuracy of the model reached 0.989, and the sensitivity and specificity were 96.66 and 96.64, respectively. Our study is the first to prove that a CNN can be trained as a diagnostic tool for PE using frontal chest X-rays, which is not possible by the human eye. It offers a convenient way to screen potential candidates for the surgical repair of PE, primarily using available image examinations.

An improved 3-D attention CNN with hybrid loss and feature fusion for pulmonary nodule classification.

BACKGROUND AND OBJECTIVE: Lung cancer has the highest cancer-related mortality worldwide, and lung nodule usually presents with no symptom. Low-dose computed tomography (LDCT) was an important tool for lung cancer detection and diagnosis. It provided a complete three-dimensional (3-D) chest image with a high resolution.Recently, convolutional neural network (CNN) had flourished and been proven the CNN-based computer-aided diagnosis (CADx) system could extract the features and help radiologists to make a preliminary diagnosis. Therefore, a 3-D ResNeXt-based CADx system was proposed to assist radiologists for diagnosis in this study. METHODS: The proposed CADx system consists of image preprocessing and a 3-D CNN-based classification model for pulmonary nodule classification. First, the image preprocessing was executed to generate the normalized volumn of interest (VOI) only including nodule information and a few surrounding tissues. Then, the extracted VOI was forwarded to the 3-D nodule classification model. In the classification model, the RestNext was employed as the backbone and the attention scheme was embedded to focus on the important features. Moreover, a multi-level feature fusion network incorporating feature information of different scales was used to enhance the prediction accuracy of small malignant nodules. Finally, a hybrid loss based on channel optimization which make the network learn more detailed information was empolyed to replace a binary cross-entropy (BCE) loss. RESULTS: In this research, there were a total of 880 low-dose CT images including 440 benign and 440 malignant nodules from the American National Lung Screening Trial (NLST) for system evaluation. The results showed that our system could achieve the accuracy of 85.3%, the sensitivity of 86.8%, the specificity of 83.9%, and the area-under-curve (AUC) value was 0.9042. It was confirmed that the designed system had a good diagnostic ability. CONCLUSION: In this study, a CADx composed of the image preprocessing and a 3-D nodule classification model with attention scheme, feature fusion, and hybrid loss was proposed for pulmonary nodule classification in LDCT. The results indicated that the proposed CADx system had potential for achieving high performance in classifying lung nodules as benign and malignant.

One-stage pulmonary nodule detection using 3-D DCNN with feature fusion and attention mechanism in CT image.

BACKGROUND AND OBJECTIVE: Lung cancer is the most common cause of cancer-related death in the world. Low-dose computed tomography (LDCT) is a widely used modality in lung cancer detection. The nodule is an abnormal tissue and may evolve into lung cancer. Hence, it is crucial to detect nodules in the early detection stage. However, reviewing the LDCT scans to observe suspicious nodules is a time-consuming task. Recently, designing a computer-aided detection (CADe) system with convolutional neural network (CNN) architecture has been proven that it is helpful for radiologists. Hence, in this study, a 3-D YOLO-based CADe system, 3-D OSAF-YOLOv3, is proposed for nodule detection in LDCT images. METHODS: The proposed CADe system consists of data preprocessing, nodule detection, and non-maximum suppression algorithm (NMS). At first, the data preprocessing including the background elimination, the spacing normalization, and the volume of interest (VOI) extraction, are conducted to remove the non-lung region, normalize the image spacing, and divide LDCT image into numerous VOIs. Then, the VOIs are fed into the 3-D OSAF-YOLOv3 model, to detect the suspicious nodules. The proposed model is constructed by integrating the 3-D YOLOv3 with the one-shot aggregation module (OSA), the receptive field block (RFB), and the feature fusion scheme (FFS). Finally, the NMS algorithm is performed to eliminate the duplicated detection generated by the model. RESULTS: In this study, the LUNA-16 dataset composed 1186 nodules from 888 LDCT scans and the competition performance metric (CPM) are used to evaluate our CADe system. In the experiment results, the proposed system can achieve a sensitivities rate of 0.962 with the false positive rate of 8 and complete a CPM value of 0.905. Moreover, according to the ablation study results, the employment of OSA module, RFB, and FFS could improve the detection performance actually. Furthermore, compared to other start-of-the-art (SOTA) models, our detection system could also achieve the higher performance. CONCLUSIONS: In this study, a YOLO-based CADe system for nodule detection in CT image system integrating additional modules and scheme is proposed for nodule detection in LDCT. The result indicates that the proposed the modification can significantly improve detection performance.

3-D Res-CapsNet convolutional neural network on automated breast ultrasound tumor diagnosis.

PURPOSE: We propose a 3-D tumor computer-aided diagnosis (CADx) system with U-net and a residual-capsule neural network (Res-CapsNet) for ABUS images and provide a reference for early tumor diagnosis, especially non-mass lesions. METHODS: A total of 396 patients with 444 tumors (226 malignant and 218 benign) were retrospectively enrolled from Sun Yat-sen University Cancer Center. In our CADx, preprocessing was performed first to crop and resize the tumor volumes of interest (VOIs). Then, a 3-D U-net and postprocessing were applied to the VOIs to obtain tumor masks. Finally, a 3-D Res-CapsNet classification model was executed with the VOIs and the corresponding masks to diagnose the tumors. Finally, the diagnostic performance, including accuracy, sensitivity, specificity, and area under the curve (AUC), was compared with other classification models and among three readers with different years of experience in ABUS review. RESULTS: For all tumors, the accuracy, sensitivity, specificity, and AUC of the proposed CADx were 84.9 %, 87.2 %, 82.6 %, and 0.9122, respectively, outperforming other models and junior reader. Next, the tumors were subdivided into mass and non-mass tumors to validate the system performance. For mass tumors, our CADx achieved an accuracy, sensitivity, specificity, and AUC of 85.2 %, 88.2 %, 82.3 %, and 0.9147, respectively, which was higher than that of other models and junior reader. For non-mass tumors, our CADx achieved an accuracy, sensitivity, specificity, and AUC of 81.6 %, 78.3 %, 86.7 %, and 0.8654, respectively, outperforming the two readers. CONCLUSION: The proposed CADx with 3-D U-net and 3-D Res-CapsNet models has the potential to reduce misdiagnosis, especially for non-mass lesions.

Computer-aided tumor detection in automated breast ultrasound using a 3-D convolutional neural network.

BACKGROUND AND OBJECTIVES: Automated breast ultrasound (ABUS) is a widely used screening modality for breast cancer detection and diagnosis. In this study, an effective and fast computer-aided detection (CADe) system based on a 3-D convolutional neural network (CNN) is proposed as the second reader for the physician in order to decrease the reviewing time and misdetection rate. METHODS: Our CADe system uses the sliding window method, a CNN-based determining model, and a candidate aggregation algorithm. First, the sliding window method is performed to split the ABUS volume into volumes of interest (VOIs). Afterward, VOIs are selected as tumor candidates by our determining model. To achieve higher performance, focal loss and ensemble learning are used to solve data imbalance and reduce false positive (FP) and false negative (FN) rates. Because several selected candidates may be part of the same tumor and they may overlap each other, a candidate aggregation method is applied to merge the overlapping candidates into the final detection result. RESULTS: In the experiments, 165 and 81 cases are utilized for training the system and evaluating system performance, respectively. On evaluation with the 81 cases, our system achieves sensitivities of 100% (81/81), 95.3% (77/81), and 90.9% (74/81) with FPs per pass (per case) of 21.6 (126.2), 6.0 (34.8), and 4.6 (27.1) respectively. According to the results, the number of FPs per pass (per case) can be diminished by 56.8% (57.1%) at a sensitivity of 95.3% based on our tumor detection model. CONCLUSIONS: In conclusion, our CADe system using 3-D CNN with the focal loss and ensemble learning may have the capability of being a tumor detection system in ABUS image.

A computer-aided diagnosis system for differentiation and delineation of malignant regions on whole-slide prostate histopathology image using spatial statistics and multidimensional DenseNet.

PURPOSE: Prostate cancer (PCa) is a major health concern in aging males, and proper management of the disease depends on accurately interpreting pathology specimens. However, reading prostatectomy histopathology slides, which is basically for staging, is usually time consuming and differs from reading small biopsy specimens, which is mainly used for diagnosis. Generally, each prostatectomy specimen generates tens of large tissue sections and for each section, the malignant region needs to be delineated to assess the amount of tumor and its burden. With the aim of reducing the workload of pathologists, in this study, we focus on developing a computer-aided diagnosis (CAD) system based on a densely connected convolutional neural network (DenseNet) for whole-slide histopathology images to outline the malignant regions. METHODS: We use an efficient color normalization process based on ranklet transformation to automatically correct the intensity of the images. Additionally, we use spatial probability to segment the tissue structure regions for different tissue recognition patterns. Based on the segmentation, we incorporate a multidimensional structure into DenseNet to determine if a particular prostatic region is benign or malignant. RESULTS: As demonstrated by the experimental results with a test set of 2,663 images from 32 whole-slide prostate histopathology images, our proposed system achieved 0.726, 0.6306, and 0.5209 in the average of the Dice coefficient, Jaccard similarity coefficient, and Boundary F1 score measures, respectively. Then, the accuracy, sensitivity, specificity, and the area under the ROC curve (AUC) of the proposed classification method were observed to be 95.0% (2544/2663), 96.7% (1210/1251), 93.9% (1334/1412), and 0.9831, respectively. DISCUSSIONS: We provide a detailed discussion on how our proposed system demonstrates considerable improvement compared with similar methods considered in previous researches as well as how it can be used for delineating malignant regions.

Computer-aided diagnosis of endobronchial ultrasound images using convolutional neural network.

BACKGROUND AND OBJECTIVE: In the United States, lung cancer is the leading cause of cancer death. The survival rate could increase by early detection. In recent years, the endobronchial ultrasonography (EBUS) images have been utilized to differentiate between benign and malignant lesions and guide transbronchial needle aspiration because it is real-time, radiation-free and has better performance. However, the diagnosis depends on the subjective judgment from doctors. In some previous studies, which using the grayscale image textures of the EBUS images to classify the lung lesions but it belonged to semi-automated system which still need the experts to select a part of the lesion first. Therefore, the main purpose of this study was to achieve full automation assistance by using convolution neural network. METHODS: First of all, the EBUS images resized to the input size of convolution neural network (CNN). And then, the training data were rotated and flipped. The parameters of the model trained with ImageNet previously were transferred to the CaffeNet used to classify the lung lesions. And then, the parameter of the CaffeNet was optimized by the EBUS training data. The features with 4096 dimension were extracted from the 7th fully connected layer and the support vector machine (SVM) was utilized to differentiate benign and malignant. This study was validated with 164 cases including 56 benign and 108 malignant. RESULTS: According to the experiment results, applying the classification by the features from the CNN with transfer learning had better performance than the conventional method with gray level co-occurrence matrix (GLCM) features. The accuracy, sensitivity, specificity, and the area under ROC achieved 85.4% (140/164), 87.0% (94/108), 82.1% (46/56), and 0.8705, respectively. CONCLUSIONS: From the experiment results, it has potential ability to diagnose EBUS images with CNN.

Tumor Detection in Automated Breast Ultrasound Using 3-D CNN and Prioritized Candidate Aggregation.

Automated whole breast ultrasound (ABUS) has been widely used as a screening modality for examination of breast abnormalities. Reviewing hundreds of slices produced by ABUS, however, is time consuming. Therefore, in this paper, a fast and effective computer-aided detection system based on 3-D convolutional neural networks (CNNs) and prioritized candidate aggregation is proposed to accelerate this reviewing. First, an efficient sliding window method is used to extract volumes of interest (VOIs). Then, each VOI is estimated the tumor probability with a 3-D CNN, and VOIs with higher estimated probability are selected as tumor candidates. Since the candidates may overlap each other, a novel scheme is designed to aggregate the overlapped candidates. During the aggregation, candidates are prioritized based on estimated tumor probability to alleviate over-aggregation issue. The relationship between the sizes of VOI and target tumor is optimally exploited to effectively perform each stage of our detection algorithm. On evaluation with a test set of 171 tumors, our method achieved sensitivities of 95% (162/171), 90% (154/171), 85% (145/171), and 80% (137/171) with 14.03, 6.92, 4.91, and 3.62 false positives per patient (with six passes), respectively. In summary, our method is more general and much faster than preliminary works and demonstrates promising results.

Computer-Aided tumor diagnosis in 3-D breast elastography.

BACKGROUND AND OBJECTIVE: Breast cancer is the major cause of cancer-related mortality in women. However, the death rate can be effectively decreased if the breast cancer can be detected early and treated appropriately. In recent years, many studies have indicated that the elastography has the better diagnosis performance than conventional ultrasound (US). METHOD: In this study, the 3-D tumor contour is obtained by using the proposed segmentation methods and then the features containing texture information, shape information, ellipsoid fitting information are extracted respectively by using the segmented 3-D tumor contour and B-mode images, and the features containing elasticity information are calculated using the same contour and elastographic images. RESULTS: In this experiment, totally 40 biopsy-proved lesions containing 20 benign tumors and 20 malignant tumors are used to evaluate the proposed computer-aided diagnosis (CAD) system. From the experimental results, the combination of shape, ellipsoid fitting and elastographic features has the best performance with accuracy 90.50% (36/40), sensitivity 85.00% (17/20), specificity 95.00% (19/20), and the area under the ROC curve Az 0.987. CONCLUSION: The result shows that tumors can be diagnosed more precisely by using the elastography images.

Computer-aided prediction of axillary lymph node status in breast cancer using tumor surrounding tissue features in ultrasound images.

BACKGROUND AND OBJECTIVE: The presence or absence of axillary lymph node (ALN) metastasis is the most important prognostic factor for patients with early-stage breast cancer. In this study, a computer-aided prediction (CAP) system using the tumor surrounding tissue features in ultrasound (US) images was proposed to determine the ALN status in breast cancer. METHODS: The US imaging database used in this study contained 114 cases of invasive breast cancer and 49 of them were ALN metastasis. After the tumor region segmentation by the level set method, image matting method was used to extract surrounding abnormal tissue of tumor from the acquired images. Then, 21 features composed of 2 intensity, 3 morphology, and 16 textural features are extracted from the surrounding tissue and processed by a logistic regression model. Finally, the prediction model is trained and tested from the selected features. RESULTS: In the experiments, the textural feature set extracted from surrounding tissue showed higher performance than intensity and morphology feature sets (Az, 0.7756 vs 0.7071 and 0.6431). The accuracy, sensitivity, specificity and the area index Az under the receiver operating characteristic (ROC) curve for the CAP system were 81.58% (93/114), 81.63% (40/49), 81.54% (53/65), and 0.8269 for using combined feature set. CONCLUSIONS: These results indicated that the proposed CAP system can be helpful to determine the ALN status in patients with breast cancer.

Computer-aided tumor diagnosis using shear wave breast elastography.

The shear wave elastography (SWE) uses the acoustic radiation force to measure the stiffness of tissues and is less operator dependent in data acquisition compared to strain elastography. However, the reproducibility of the result is still interpreter dependent. The purpose of this study is to develop a computer-aided diagnosis (CAD) method to differentiate benign from malignant breast tumors using SWE images. After applying the level set method to automatically segment the tumor contour and hue-saturation-value color transformation, SWE features including average tissue elasticity, sectional stiffness ratio, and normalized minimum distance for grouped stiffer pixels are calculated. Finally, the performance of CAD based on SWE features are compared with those based on B-mode ultrasound (morphologic and textural) features, and a combination of both feature sets to differentiate benign from malignant tumors. In this study, we use 109 biopsy-proved breast tumors composed of 57 benign and 52 malignant cases. The experimental results show that the sensitivity, specificity, accuracy and the area under the receiver operating characteristic ROC curve (Az value) of CAD are 86.5%, 93.0%, 89.9%, and 0.905 for SWE features whereas they are 86.5%, 80.7%, 83.5% and 0.893 for B-mode features and 90.4%, 94.7%, 92.3% and 0.961 for the combined features. The Az value of combined feature set is significantly higher compared to the B-mode and SWE feature sets (p=0.0296 and p=0.0204, respectively). Our results suggest that the CAD based on SWE features has the potential to improve the performance of classifying breast tumors with US.

Computer-aided diagnosis for distinguishing between triple-negative breast cancer and fibroadenomas based on ultrasound texture features.

PURPOSE: Triple-negative breast cancer (TNBC), an aggressive subtype, is frequently misclassified as fibroadenoma due to benign morphologic features on breast ultrasound (US). This study aims to develop a computer-aided diagnosis (CAD) system based on texture features for distinguishing between TNBC and benign fibroadenomas in US images. METHODS: US images of 169 pathology-proven tumors (mean size, 1.65 cm; range, 0.7-3.0 cm) composed of 84 benign fibroadenomas and 85 TNBC tumors are used in this study. After a tumor is segmented out using the level-set method, morphological, conventional texture, and multiresolution gray-scale invariant texture feature sets are computed using a best-fitting ellipse, gray-level co-occurrence matrices, and the ranklet transform, respectively. The linear support vector machine with leave-one-out cross-validation schema is used as a classifier, and the diagnostic performance is assessed with receiver operating characteristic curve analysis. RESULTS: The Az values of the morphology, conventional texture, and multiresolution gray-scale invariant texture feature sets are 0.8470 [95% confidence intervals (CIs), 0.7826-0.8973], 0.8542 (95% CI, 0.7911-0.9030), and 0.9695 (95% CI, 0.9376-0.9865), respectively. The Az of the CAD system based on the combined feature sets is 0.9702 (95% CI, 0.9334-0.9882). CONCLUSIONS: The CAD system based on texture features extracted via the ranklet transform may be useful for improving the ability to discriminate between TNBC and benign fibroadenomas.

Intrinsic subtypes and tumor grades in breast cancer are associated with distinct 3-D power Doppler sonographic vascular features.

PURPOSE: This study aimed to investigate the three-dimensional (3-D) power Doppler ultrasonographic (PDUS) vascular features of breast carcinoma according to intrinsic subtypes, nodal stage, and tumor grade. MATERIALS AND METHODS: Total 115 receiving mastectomy breast carcinomas (mean size, 2.5 cm; range, 0.7-6.5 cm), including 102 invasive ductal carcinomas (IDC), 10 ductal carcinomas in situ (DCIS), and 3 invasive lobular carcinomas (ILC) diagnosed after mastectomy, were used in this retrospective study. Sixty IDC had nodal status and histopathologic tumor grades available for analysis. Vascular features, including number of vascular trees (NV), longest path length (LPL), total vessel length (TVL), number of bifurcations (NB), distance metric (DM), inflection count metric (ICM), vessel diameter (VD), and vessel-to-volume ratio (VVR) were extracted using 3-D thinning method. The Mann-Whitney U test, Student's t-test, one-way ANOVA, and Kruskal-Wallis test were performed as appropriate. RESULTS: There was no significant difference of vascular features among IDC, DCIS and ILC. Except VD, vascular features in luminal type were significantly lower compared to HER2-enriched or triple negative types (p<0.05). Compared to ER+ (estrogen receptor positive) tumors, all features in ER- (estrogen receptor negative) tumors were significantly higher (p<0.01). Despite some significantly higher vascular features in high grade IDC compared to low and intermediate grade, there was no significant correlation between vascular features and nodal stages. CONCLUSION: Differences in 3-D PDUS vascular features among intrinsic types of IDC are attributed to their ER status. Vascular features extracted by 3-D PDUS correlate with tumor grades but not nodal stage in IDC.

Computer-aided diagnosis for 3-d power Doppler breast ultrasound.

In recent studies, both tumor morphology and vascularity played an important role in differentiating breast tumors. In this article, a computer-aided diagnosis (CAD) system was proposed to quantify the tumor morphology of vascularity on three-dimensional (3-D) power Doppler breast ultrasound (PDUS) images. We segmented the tumor margin by the level set method and skeletonized vessels by the 3-D thinning algorithm from 3-D PDUS data to capture the B-mode and vascularity features. The B-mode features including texture, shape and ellipsoid fitting and the vascularity features containing volume, complexity, length, radius and tortuosity were used to differentiate breast tumors. In the experiment, 82 biopsy-verified lesions including 41 benign and 41 malignant lesions were used to test the performance of the proposed system. The proposed method performed well, achieving accuracy, sensitivity, specificity and Az values of 85.37% (70/82), 85.37% (35/41), 85.37% (35/41) and 0.9104, respectively.