No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Automatic classification of pulmonary peri-fissural nodules in computed tomography using an ensemble of 2D views and a convolutional neural network out-of-the-box
Abstract
In this paper, we tackle the problem of automatic classification of pulmonary peri-fissural nodules (PFNs). The classification problem is formulated as a machine learning approach, where detected nodule candidates are classified as PFNs or non-PFNs. Supervised learning is used, where a classifier is trained to label the detected nodule. The classification of the nodule in 3D is formulated as an ensemble of classifiers trained to recognize PFNs based on 2D views of the nodule. In order to describe nodule morphology in 2D views, we use the output of a pre-trained convolutional neural network known as OverFeat. We compare our approach with a recently presented descriptor of pulmonary nodule morphology, namely Bag of Frequencies, and illustrate the advantages offered by the two strategies, achieving performance of AUC = 0.868, which is close to the one of human experts.
... Margin: In general, radiologists use the axial, coronal, and sagittal views of a 3D CT image to comprehensively analyze tumor information within the 3D image ( Fig. 4) [26]. In particular, three 2D images can be extracted from three views and combined as a 2.5D image to approximate the 3D image [26]. ...
... Margin: In general, radiologists use the axial, coronal, and sagittal views of a 3D CT image to comprehensively analyze tumor information within the 3D image ( Fig. 4) [26]. In particular, three 2D images can be extracted from three views and combined as a 2.5D image to approximate the 3D image [26]. Therefore, a 2.5D method is used to extract adequate information of the tumor margin from a 3D image. ...
... On the other hand, AUPRC aims to evaluate the ability of the model in classifying positive samples correctly in an imbalanced data distribution. In particular, the baseline of AUPRC can be calculated by using (26), which is equal to 0.220 in the independent testing set. ...
Macrovascular invasion (MaVI) is a major threat to survival in hepatocellular carcinoma (HCC), which should be treated as early as possible to ensure safety and efficacy. In this aspect, MaVI prediction can be helpful. However, MaVI prediction is difficult because of the inter-class similarity and intra-class variation of HCC in computed tomography (CT) images. Moreover, existing methods fail to include clinical priori knowledge associated with HCC, leading to incomprehensive information extraction. In this paper, we proposed a prior knowledge-aware fusion network (PKAFnet) to accurately achieve MaVI prediction in CT images. First, a perception module was presented to extract features related to tumor marginal heterogeneity in the graph domain, which contributed to rotation invariance and captured intensity variations of tumor margin. Second, a tumor segmentation network was built to obtain global information of a 3D tumor image and information associated with tumor internal heterogeneity in the image domain. Finally, multi-domain features associated with the tumor margin and tumor region were combined by using a multi-domain attentional feature fusion module. Thus, by incorporating MaVI-related prior knowledge, our PKAFnet can alleviate overfitting, which can improve the discriminative ability. The proposed PKAFnet was validated on a multi-center dataset, and remarkable performance was achieved in an independent testing set. Moreover, the interpretabil-ity of perception module and segmentation network were presented in our paper, which illustrated the effectiveness and credibility of PKAFnet. Therefore, the proposed method showed great application potential for MaVI prediction.
... In recent years, deep learning [6,7] specifically Convolution Neural Networks (CNNs) [8] has been demonstrated to perform well and produced promising results in mammogram classification [9,10]. Deep learning CADs have been applied to a variety of medical domains, including Interstitial lung disease, Cervical cancer classification, pulmonary Peri-fissural nodule classification [9], and Thoraco-abdominal lymph node classification [10]. ...
... In recent years, deep learning [6,7] specifically Convolution Neural Networks (CNNs) [8] has been demonstrated to perform well and produced promising results in mammogram classification [9,10]. Deep learning CADs have been applied to a variety of medical domains, including Interstitial lung disease, Cervical cancer classification, pulmonary Peri-fissural nodule classification [9], and Thoraco-abdominal lymph node classification [10]. The works on breast lesions [11,12] piqued our curiosity in particular. ...
Radiologists frequently struggle to define mammography mass lesions, resulting in unneeded breast biopsies to eliminate suspicions, which adds exorbitant costs to an already overburdened patient and medical system..Existing models have limited capability for feature extraction and representation, as well as cancer classification. Therefore, we built deep Convolution neural networks based Computer-aided Diagnosis system to assist radiologists in classifying mammography mass lesions. Here, Two-view LASSO regression feature fusion and fine-tuned transfer learning network model VGG16 were applied for identification of mammogram cancer. First, two independent CNN branches are utilized to extract mammography characteristics from two different perspectives. Feature Extraction is performed by fine-tuning pre-trained deep network models VGG16 which extracts deep convolutional features. Second, the features of the VGG16 models are serially fused using LASSO regression. Lastly, the fused features are entered into the Fully Connected Layer for mammogram classification. The high accuracy of 95.24, senstitivity of 96.11% and AUC score of 97.95% of the proposed approach revealed that it should be used to enhance clinical decision-making.
... -15 [26] Pre-Trained Convolution Neural Networks (CNN) ...
... The compilation of clinical image arrangement techniques is displayed in Table 2 [22][23][24][25][26][27][28][29][30][31][32][33][34][35]. ...
Despite enormous advances in medical technology, the prognosis of Brain Tumour (BT) remains an extremely time-consuming and troublesome assignment for physicians. Early and precise brain tumour identification gives an effective results and leads to increased survival rate. Within this paper, an examination of various techniques in order of priority to classify clinical images is presented to analyse various research gaps and highlights their costs and benefits. Human mortality can be reduced by using an automatic classification scheme. The automatic classification of brain tumours is a difficult task due to the large spatial and structural variability of the brain tumor’s surrounding region. The latest developments have been investigated in image characterization strategies for diagnosing human body disease and addressing the classification of nuclear medical imaging identification techniques like Convolution Neural Network (CNN), Support Vector Machine (SVM), Histogram technique, K-Means Clustering (K-MC) etc., just as the respective parameters like the image modalities employed, the dataset and the trade-offs have been compared for each technique. Among these techniques, CNN model accomplished the highest accuracy of 99% for two sets of data: Brain Tumour Segmentation (BTS) and BD-brain tumour and a high average susceptibility of 0.99 for all datasets. Finally, the review demonstrated that improving image order strategies with regarding accuracy, sensitivity value, and feasibility for Computer-Aided Diagnosis (CAD) is a significant challenge as well as an open research area.
... Secondly, the annotation of medical images (manual labelling) mandates the involvement of senior radiologists, incurring considerable labor and time costs. To mitigate the aforementioned challenges, current strategies primarily involve model complexity reduction, regularization techniques and data augmentation-based enhancement strategies [13][14][15][16]. Nevertheless, such methods exhibit constrained efficacy in alleviating overfitting and are unable to compete with the performance of models trained on large, and high-quality annotation datasets. ...
Deep semi-supervised learning (DSSL) is a machine learning paradigm that blends supervised and unsupervised learning techniques to improve the performance of various models in computer vision tasks. Medical image classification plays a crucial role in disease diagnosis, treatment planning, and patient care. However, obtaining labeled medical image data is often expensive and time-consuming for medical practitioners, leading to limited labeled datasets. DSSL techniques aim to address this challenge, particularly in various medical image tasks, to improve model generalization and performance. DSSL models leverage both the labeled information, which provides explicit supervision, and the unlabeled data, which can provide additional information about the underlying data distribution. That offers a practical solution to resource-intensive demands of data annotation, and enhances the model’s ability to generalize across diverse and previously unseen data landscapes. The present study provides a critical review of various DSSL approaches and their effectiveness and challenges in enhancing medical image classification tasks. The study categorized DSSL techniques into six classes: consistency regularization method, deep adversarial method, pseudo-learning method, graph-based method, multi-label method, and hybrid method. Further, a comparative analysis of performance for six considered methods is conducted using existing studies. The referenced studies have employed metrics such as accuracy, sensitivity, specificity, AUC-ROC, and F1 score to evaluate the performance of DSSL methods on different medical image datasets. Additionally, challenges of the datasets, such as heterogeneity, limited labeled data, and model interpretability, were discussed and highlighted in the context of DSSL for medical image classification. The current review provides future directions and considerations to researchers to further address the challenges and take full advantage of these methods in clinical practices.
... In recent years, the ability to solve complex tasks with high success by processing large data without the need for feature extraction using DL models has made the use of these models quite common in disease classification and recognition of medical images [18][19][20][21][22][23]. However, there is a limited number of studies in the literature that diagnose and classify middle ear disease using DL methods. ...
Otitis media is a medical concept that represents a range of inflammatory middle ear disorders. The high costs of medical devices utilized by field experts to diagnose the disease relevant to otitis media prevent the widespread use of these devices. This makes it difficult for field experts to make an accurate diagnosis and increases subjectivity in diagnosing the disease. To solve these problems, there is a need to develop computer-aided middle ear disease diagnosis systems. In this study, a deep learning-based approach is proposed for the detection of OM disease to meet this emerging need. This approach is the first that addresses the performance of a voting ensemble framework that uses Inception V3, DenseNet 121, VGG16, MobileNet, and EfficientNet B0 pre-trained DL models. All pre-trained CNN models used in the proposed approach were trained using the Public Ear Imagery dataset, which has a total of 880 otoscopy images, including different eardrum cases such as normal, earwax plug, myringosclerosis, and chronic otitis media. The prediction results of these models were evaluated with voting approaches to increase the overall prediction accuracy. In this context, the performances of both soft and hard voting ensembles were examined. Soft voting ensemble framework achieved highest performance in experiments with 98.8% accuracy, 97.5% sensitivity, and 99.1% specificity. Our proposed model achieved the highest classification performance so far in the current dataset. The results reveal that our voting ensemble-based DL approach showed quite high performance for the diagnosis of middle ear disease. In clinical applications, this approach can provide a preliminary diagnosis of the patient's condition just before field experts make a diagnosis on otoscopic images. Thus, our proposed approach can help field experts to diagnose the disease quickly and accurately. In this way, clinicians can make the final diagnosis by integrating automatic diagnostic prediction with their experience.
... A typical example of 2D methods is that proposed in Xie et al (2019) which used a 2D convolutional neural network (CNN) for automated pulmonary nodule detection in CT images. Other 2D examples include the one proposed in Ciompi et al (2015) which utilized a pre-trained CNN to extract features and describe the morphology of nodules in 2D views and the one proposed in Ding et al (2017) which employed the 2D Faster R-CNN model along with the VGG-16 architecture to detect candidate nodules. Besides, in Meraj et al (2021), lung nodules were detected upon performing adaptive thresholding (OTSU) and 2D semantic segmentation located positive samples. ...
Objective:
Accurate and automatic detection of pulmonary nodules is critical for early lung cancer diagnosis, and promising progress has been achieved in developing effective deep models for nodule detection. However, most existing nodule detection methods merely focus on integrating elaborately designed feature extraction modules into the backbone of the detection network to extract rich nodule features while ignore disadvantages of the structure of detection network itself. This study aims to address these disadvantages and develop a deep learning-based algorithm for pulmonary nodule detection to improve the accuracy of early lung cancer diagnosis.
Approach:
In this paper, an S-shaped network called S-Net is developed with the U-shaped network as backbone, where an information fusion branch is used to propagate lower-level details and positional information critical for nodule detection to higher-level feature maps, head shared scale adaptive detection strategy is utilized to capture information from different scales for better detecting nodules with different shapes and sizes and the feature decoupling detection head is used to allow the classification and regression branches to focus on the information required for their respective tasks. A hybrid loss function is utilized to fully exploit the interplay between the classification and regression branches.
Main results:
The proposed S-Net network with ResSENet and other three U-shaped backbones from SANet, OSAF-YOLOv3 and MSANet (R+SC+ECA) models achieve average CPM scores of 0.914, 0.915, 0.917 and 0.923 on the LUNA16 dataset, which are significantly higher than those achieved with other existing state-of-the-art models.
Significance:
The experimental results demonstrate that our proposed method effectively improves nodule detection performance, which implies potential applications of the proposed method in clinical practice.
... To address this issue, various approaches have been proposed. For instance, some suggest using 2D or 3D image patches as input instead of full-size images [33][34][35][36], while others propose transfer learning by utilizing models trained on a large number of natural images in computer vision [37][38][39]. Due to the limited number of FISH images and the high requirements for segmentation accuracy, we propose the SEAM-Unet++ deep learning model to improve the accuracy of the model for FISH cell image segmentation by fusing the SEAM attention mechanism with Unet++. ...
Fluorescence in situ hybridization (FISH) is a powerful cytogenetic method used to precisely detect and localize nucleic acid sequences. This technique is proving to be an invaluable tool in medical diagnostics and has made significant contributions to biology and the life sciences. However, the number of cells is large and the nucleic acid sequences are disorganized in the FISH images taken using the microscope. Processing and analyzing images is a time-consuming and laborious task for researchers, as it can easily tire the human eyes and lead to errors in judgment. In recent years, deep learning has made significant progress in the field of medical imaging, especially the successful application of introducing the attention mechanism. The attention mechanism, as a key component of deep learning, improves the understanding and interpretation of medical images by giving different weights to different regions of the image, enabling the model to focus more on important features. To address the challenges in FISH image analysis, we combined medical imaging with deep learning to develop the SEAM-Unet++ automated cell contour segmentation algorithm with integrated attention mechanism. The significant advantage of this algorithm is that it improves the accuracy of cell contours in FISH images. Experiments have demonstrated that by introducing the attention mechanism, our method is able to segment cells that are adherent to each other more efficiently.
... CNNs are made up of four main parts: fully-connected layers, convolutional layers, activation functions, and pooling. This paper offers a very concise introduction to CNNs, as described in [123]. The region of the mammalian brain responsible for sight is composed of neurons that isolate image characteristics. ...
Deep learning technologies developed at an exponential rate throughout the years. Starting from Convolutional Neural Networks (CNNs) to Involutional Neural Networks (INNs), there are several neural network (NN) architectures today, including Vision Transformers (ViT), Graph Neural Networks (GNNs), Recurrent Neural Networks (RNNs) etc. However, uncertainty cannot be represented in these architectures, which poses a significant difficulty for decision-making given that capturing the uncertainties of these state-of-the-art NN structures would aid in making specific judgments. Dropout is one method that may be implemented within Deep Learning (DL) networks as a technique to assess uncertainty. Dropout is applied at the inference phase to measure the uncertainty of these neural network models. This approach, commonly known as Monte Carlo Dropout (MCD), works well as a low-complexity estimation to compute uncertainty. MCD is a widely used approach to measure uncertainty in DL models, but majority of the earlier works focus on only a particular application. Furthermore, there are many state-of-the-art (SOTA) NNs that remain unexplored, with regards to that of uncertainty evaluation. Therefore an up-to-date roadmap and benchmark is required in this field of study. Our study revolved around a comprehensive analysis of the MCD approach for assessing model uncertainty in neural network models with a variety of datasets. Besides, we include SOTA NNs to explore the untouched models regarding uncertainty. In addition, we demonstrate how the model may perform better with less uncertainty by modifying NN topologies, which also reveals the causes of a model’s uncertainty. Using the results of our experiments and subsequent enhancements, we also discuss the various advantages and costs of using MCD in these NN designs. While working with reliable and robust models we propose two novel architectures, which provide outstanding performances in medical image diagnosis.
... The former is driven by a global-local refinement structure, and the latter constructs an encoder-decoder to parameterize the variational information bottleneck for providing visual explanations. multifarious 3-D components to learn multidimensional representation from 3-D views [20], [21], [22], [23], [24], 2-D views [25], [26], and a group of patches [27]. Xie et al. [20] designed nine knowledge-based collaborative subnetworks to extract multiview 3-D features and incorporate them for classifying nodules. ...
Computerized tomography (CT) is a clinically primary technique to differentiate benign-malignant pulmonary nodules for lung cancer diagnosis. Early classification of pulmonary nodules is essential to slow down the degenerative process and reduce mortality. The interactive paradigm assisted by neural networks is considered to be an effective means for early lung cancer screening in large populations. However, some inherent characteristics of pulmonary nodules in high-resolution CT images, e.g., diverse shapes and sparse distribution over the lung fields, have been inducing inaccurate results. On the other hand, most existing methods with neural networks are dissatisfactory from a lack of transparency. In order to overcome these obstacles, a united framework is proposed, including the classification and feature visualization stages, to learn distinctive features and provide visual results. Specifically, a bilateral scheme is employed to synchronously extract and aggregate global-local features in the classification stage, where the global branch is constructed to perceive deep-level features and the local branch is built to focus on the refined details. Furthermore, an encoder is built to generate some features, and a decoder is constructed to simulate decision behavior, followed by the information bottleneck viewpoint to optimize the objective. Extensive experiments are performed to evaluate our framework on two publicly available datasets, namely, 1) the Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) and 2) the Lung and Colon Histopathological Image Dataset (LC25000). For instance, our framework achieves 92.98% accuracy and presents additional visualizations on the LIDC. The experiment results show that our framework can obtain outstanding performance and is effective to facilitate explainability. It also demonstrates that this united framework is a serviceable tool and further has the scalability to be introduced into clinical research.
... Deep learning is emerging as the most promising branch of machine learning for complex tasks such as image classification and object detection [8,9]. In medical image analysis, AI and deep learning are being increasingly employed for several tasks, for example, diagnosing developmental anomalies in the brain [10], classifying pulmonary nodules [11][12][13], and in the diagnosis of Alzheimer disease [14]. In spine radiology, deep learning models have been used to detect vertebral fractures [15], to grade the severity of disc degeneration [16,17], and to automatically characterize spinal deformities [18]. ...
Purpose
Radiological degenerative phenotypes provide insight into a patient’s overall extent of disease and can be predictive for future pathological developments as well as surgical outcomes and complications. The objective of this study was to develop a reliable method for automatically classifying sagittal MRI image stacks of cervical spinal segments with respect to these degenerative phenotypes.
Methods
We manually evaluated sagittal image data of the cervical spine of 873 patients (5182 motion segments) with respect to 5 radiological phenotypes. We then used this data set as ground truth for training a range of multi-class multi-label deep learning-based models to classify each motion segment automatically, on which we then performed hyper-parameter optimization.
Results
The ground truth evaluations turned out to be relatively balanced for the labels disc displacement posterior, osteophyte anterior superior, osteophyte posterior superior, and osteophyte posterior inferior. Although we could not identify a single model that worked equally well across all the labels, the 3D-convolutional approach turned out to be preferable for classifying all labels.
Conclusions
Class imbalance in the training data and label noise made it difficult to achieve high predictive power for underrepresented classes. This shortcoming will be mitigated in the future versions by extending the training data set accordingly. Nevertheless, the classification performance rivals and in some cases surpasses that of human raters, while speeding up the evaluation process to only require a few seconds.
... Convolutional neural networks (CNNs), in particular, have revolutionised this field by allowing the automatic extraction of discriminative features straight from raw data [3]. These models have achieved state-of-the-art performance in many medical imaging tasks, such as diagnosing diabetic retinopathy from retinal images [15], detecting lung nodules from CT scans [7], and classifying skin lesions from dermoscopic images [11]. ...
In recent years, deep learning models have revolutionized medical image interpretation, offering substantial improvements in diagnostic accuracy. However, these models often struggle with challenging images where critical features are partially or fully occluded, which is a common scenario in clinical practice. In this paper, we propose a novel curriculum learning-based approach to train deep learning models to handle occluded medical images effectively. Our method progressively introduces occlusion, starting from clear, unobstructed images and gradually moving to images with increasing occlusion levels. This ordered learning process, akin to human learning, allows the model to first grasp simple, discernable patterns and subsequently build upon this knowledge to understand more complicated, occluded scenarios. Furthermore, we present three novel occlusion synthesis methods, namely Wasserstein Curriculum Learning (WCL), Information Adaptive Learning (IAL), and Geodesic Curriculum Learning (GCL). Our extensive experiments on diverse medical image datasets demonstrate substantial improvements in model robustness and diagnostic accuracy over conventional training methodologies.
... The segmentation of ROIs using DL algorithms increases the automation of radiomics [41][42][43][44]. A recently described radiomics method uses a deep transfer network in the feature extraction algorithm, thereby replacing or integrating handcrafted features, which not only reduces the amount of data required, but also provides the advantages of DL [45][46][47]. Zheng et al. [48] demonstrated that classification of kidney ultrasound (US) images based on a combination of TL and handcrafted features was more effective than that based on TL or handcrafted features alone. In the future, AI will bring more innovations and breakthroughs in this type of medical image analysis. ...
Chronic kidney disease (CKD) causes irreversible damage to kidney structure and function. Arising from various etiologies, risk factors for CKD include hypertension and diabetes. With a progressively increasing global prevalence, CKD is an important public health problem worldwide. Medical imaging has become an important diagnostic tool for CKD through the non-invasive identification of macroscopic renal structural abnormalities. Artificial intelligence (AI)-assisted medical imaging techniques aid clinicians in the analysis of characteristics that cannot be easily discriminated by the naked eye, providing valuable information for the identification and management of CKD. Recent studies have demonstrated the effectiveness of AI-assisted medical image analysis as a clinical support tool using radiomics- and deep learning-based AI algorithms for improving the early detection, pathological assessment, and prognostic evaluation of various forms of CKD, including autosomal dominant polycystic kidney disease. Herein, we provide an overview of the potential roles of AI-assisted medical image analysis for the diagnosis and management of CKD.
... CNNs have been widely employed in various applications of medical image processing, including classification [2] [3] and segmentation [4] [5]. However, understanding the behavior of CNNs can be challenging, which has significant implications, particularly in sensitive domains like the medical field. ...
Computer aided detection and diagnosis systems based on deep learning have shown promising performance in breast cancer detection. However, there are cases where the obtained results lack justification. In this study, our objective is to highlight the regions of interest used by a convolutional neural network (CNN) for classifying histological images as benign or malignant. We compare these regions with the regions identified by pathologists. To achieve this, we employed the VGG19 architecture and tested three visualization methods: Gradient, LRP Z, and LRP Epsilon. Additionally, we experimented with three pixel selection methods: Bins, K-means, and MeanShift. Based on the results obtained, the Gradient visualization method and the MeanShift selection method yielded satisfactory outcomes for visualizing the images.
... This solution used a deep, fully convolutional neural network to classify detected lung nodules into the stage. This solution was tested on different datasets of different scan conditions and provided an accuracy of 84.58%, higher than the proposed techniques in [1,12]. This solution used a Gabor filter for image quality enhancement, and thresholding was applied to segmentation ROI. ...
The deep learning (DL) classification technique is extensively researched and considered for early lung cancer diagnosis. Despite the encouraging performance reported in the literature, DL models face several challenges to be deployed in real-life systems. These include the DL-Models' stability, the nodule structure's complexity, the lack of proper lung segmentation technique, high false-positive results, and the availability of publically shared medical imaging data. This paper investigates, identifies, and intensively studies DL approaches that yield high performance in the classification of Lung Cancer. We reviewed 338 articles, of which 37 met the inclusion criteria we have set for the proposed framework. In addition, we propose and evaluate a framework to govern the DL model selection and deployment process in real-world systems. The framework consists of four main components; Data, Feature Selection, Classification Technique, and View (DFCV). We discuss the efficiency and the importance of the proposed DFCV framework on 37 state-of-the-art research papers in the field of deep learning-based lung cancer classification systems. The DFCV framework could represent a guide for DL-based systems selection and deployment in medical centers for lung cancer.
... The traditional categorization of Restricted Boltzmann Machine (RBM)-relied approach developed to examining the lung CT scan by integrating both generative and discriminative representation learning (Van et al. 2016). A CNN-related automated classification of peri-fissural nodules (PFN) was introduced in (Ciompi et al. 2015) for performing the lung cancer scanning. Then, 2-stage multi-instance DL model has been projected for classifying diverse body organs (Yan et al. 2016). ...
The advances in biomedical imaging equipment have produced a massive amount of medical images that are generated by the different modalities. Consequently, a huge volume of data has been produced and caused a complex and time-consuming retrieving process of the relevant cases. To resolve this issue, the Content-Based Biomedical Image Retrieval (CBMIR) system is applied to retrieve the related images from the earlier patients’ databases. However, the previous handcrafted features methods that applied the CBMIR model have shown poor performance in many multimodal databases. In this paper, we focus on designing CBMIR technique using Deep Learning (DL) models. We present a new Multimodal Biomedical Image Retrieval and Classification (M-BMIRC) technique for retrieving and classifying the biomedical images from huge databases. The proposed M-BMIRC model involves three dissimilar processes as following: feature extraction, similarity measurement, and classification. It uses an ensemble of handcrafted features from Zernike Moments (ZM) and deep features from Deep Convolutional Neural Networks (DCNN) for feature extraction process. Additionally, the Hausdorff Distance based similarity measure is employed to identify the resemblance between the queried image and the images that exist in the database. Moreover, the classification process gets executed on the retrieval images using the Probabilistic Neural Network (PNN) model, which allocates the class labels of the tested images. Finally, the experimental studies are conducted using two benchmark medical datasets and the results ensure the superior performance of the proposed model in terms of different measures include Average Precision Rate (APR), Average Recall Rate (ARR), F-score, accuracy, and Computation Time (CT).
... As one of the current state-of-the-art machine learning methods, deep learning can achieve higher work efficiency and accuracy in related fields, and has become a hot research topic in clinical applications. The detection of IAC in PR images is important for the diagnosis and treatment of dental diseases and oral surgery [19][20][21][22]. Modern algorithms that can be clinically applied for automatic IAC segmentation should be both fast and accurate. ...
Background:
Manual segmentation of the inferior alveolar canal (IAC) in panoramic images requires considerable time and labor even for dental experts having extensive experience. The objective of this study was to evaluate the performance of automatic segmentation of IAC with ambiguity classification in panoramic images using a deep learning method.
Methods:
Among 1366 panoramic images, 1000 were selected as the training dataset and the remaining 336 were assigned to the testing dataset. The radiologists divided the testing dataset into four groups according to the quality of the visible segments of IAC. The segmentation time, dice similarity coefficient (DSC), precision, and recall rate were calculated to evaluate the efficiency and segmentation performance of deep learning-based automatic segmentation.
Results:
Automatic segmentation achieved a DSC of 85.7% (95% confidence interval [CI] 75.4%-90.3%), precision of 84.1% (95% CI 78.4%-89.3%), and recall of 87.7% (95% CI 77.7%-93.4%). Compared with manual annotation (5.9s per image), automatic segmentation significantly increased the efficiency of IAC segmentation (33 ms per image). The DSC and precision values of group 4 (most visible) were significantly better than those of group 1 (least visible). The recall values of groups 3 and 4 were significantly better than those of group 1.
Conclusions:
The deep learning-based method achieved high performance for IAC segmentation in panoramic images under different visibilities and was positively correlated with IAC image clarity.
... We have used the learned AlexNet as an extractor of biomedical frameworks, utilizing the fully connected layer-6. We use completely connected layer-6 features, since various studies have shown that layer-6 features are more efficient than layer-7 features in biomedical image processing [39][40][41][42][43]. ...
Systems for retrieving and managing content-based medical images are becoming more important, especially as medical imaging technology advances and the medical image database grows. In addition, these systems can also use medical images to better grasp and gain a deeper understanding of the causes and treatments of different diseases, not just for diagnostic purposes. For achieving all these purposes, there is a critical need for an efficient and accurate content-based medical image retrieval (CBMIR) method. This paper proposes an efficient method (RbQE) for the retrieval of computed tomography (CT) and magnetic resonance (MR) images. RbQE is based on expanding the features of querying and exploiting the pre-trained learning models AlexNet and VGG-19 to extract compact, deep, and high-level features from medical images. There are two searching procedures in RbQE: a rapid search and a final search. In the rapid search, the original query is expanded by retrieving the top-ranked images from each class and is used to reformulate the query by calculating the mean values for deep features of the top-ranked images, resulting in a new query for each class. In the final search, the new query that is most similar to the original query will be used for retrieval from the database. The performance of the proposed method has been compared to state-of-the-art methods on four publicly available standard databases, namely, TCIA-CT, EXACT09-CT, NEMA-CT, and OASIS-MRI. Experimental results show that the proposed method exceeds the compared methods by 0.84%, 4.86%, 1.24%, and 14.34% in average retrieval precision (ARP) for the TCIA-CT, EXACT09-CT, NEMA-CT, and OASIS-MRI databases, respectively.
... Machine learning (ML) is an AI subset that presents the "learning" experience which is applicable to human intelligence, while at the same time learning and analyzing it by using computer algorithms [15]. These algorithms are able to detect patterns and improve their learning abilities by analyzing massive amounts of data, so that the system can take autonomous recommendations or decisions. ...
... Because of its strong ability of feature self-learning and expression, deep learning has gradually become the focus of research. With the establishment of various image databases, convolutional neural network (CNN) has gradually become a standard feature extractor to complete computer vision tasks in various fields [6] and has also made many excellent results in the field of medical image analysis [7][8][9][10][11]. Faster R-CNN is a general target detection model based on CNN [12], which replaces the traditional method of extracting candidate targets by Selective Search with a region proposal network (RPN), which greatly improves the detection speed and realizes end-to-end in the true sense. ...
As cancer with the highest morbidity and mortality in the world, lung cancer is characterized by pulmonary nodules in the early stage. The detection of pulmonary nodules is an important method for the early detection of lung cancer, which can greatly improve the survival rate of lung cancer patients. However, the accuracy of conventional detection methods for lung nodules is low. With the development of medical imaging technology, deep learning plays an increasingly important role in medical image detection, and pulmonary nodules can be accurately detected by CT images. Based on the above, a pulmonary nodule detection method based on deep learning is proposed. In the candidate nodule detection stage, the multiscale features and Faster R-CNN, a general-purpose detection framework based on deep learning, were combined together to improve the detection of small-sized lung nodules. In the false-positive nodule filtration stage, a 3D convolutional neural network based on multiscale fusion is designed to reduce false-positive nodules. The experiment results show that the candidate nodule detection model based on Faster R-CNN integrating multiscale features has achieved a sensitivity of 98.6%, 10% higher than that of the other single-scale model, the proposed method achieved a sensitivity of 90.5% at the level of 4 false-positive nodules per scan, and the CPM score reached 0.829. The results are higher than methods in other works of literature. It can be seen that the detection method of pulmonary nodules based on multiscale fusion has a higher detection rate for small nodules and improves the classification performance of true and false-positive pulmonary nodules. This will help doctors when making a lung cancer diagnosis.
... Compared with traditional feature-based CAD systems, the DL-based CAD system can automatically retrieve and extract intrinsic features of a suspicious nodule [38,39], and can model the 3D shape of a nodule ( Figure 2). For example, Ciompi et al. [40] designed a model based on OverFeat [41,42] by extracting three 2D-viewfeature vectors (axial, coronal, and sagittal) of the nodule from CT scans. The recently integrated CNN models facilitate a global and comprehensive inspection of nodules for feature characterization from CT images. ...
The recent development of imaging and sequencing technologies enables systematic advances in the clinical study of lung cancer. Meanwhile, the human mind is limited in effectively handling and fully utilizing the accumulation of such enormous amounts of data. Machine learning-based approaches play a critical role in integrating and analyzing these large and complex datasets, which have extensively characterized lung cancer through the use of different perspectives from these accrued data. In this article, we provide an overview of machine learning-based approaches that strengthen the varying aspects of lung cancer diagnosis and therapy, including early detection, auxiliary diagnosis, prognosis prediction, and immunotherapy practice. Moreover, we highlight the challenges and opportunities for future applications of machine learning in lung cancer.
... By utilizing a nomogram obtained from the United States, a study [39] forecasted morbidity-specific cancer of gastric survival at an institute in Europe. The use of neural networks in the survival evaluation was compared to the Kaplan-Meier and Cox proportional risk images by Ciompi et al. [40]. ...
Cancer is one of the most devastating, fatal, dangerous, and unpredictable ailments. To reduce the risk of fatality in this disease, we need some ways to predict the disease, diagnose it faster and precisely, and predict the prognosis accurately. The incorporation of artificial intelligence (AI), machine learning (ML), and deep learning (DL) algorithms into the healthcare system has already proven to work wonders for patients. Artificial intelligence is a simulation of intelligence that uses data, rules, and information programmed in it to make predictions. The science of machine learning (ML) uses data to enhance performance in a variety of activities and tasks. A bigger family of machine learning techniques built on artificial neural networks and representation learning is deep learning (DL). To clarify, we require AI, ML, and DL to predict cancer risk, survival chances, cancer recurrence, cancer diagnosis, and cancer prognosis. All of these are required to improve patient's quality of life, increase their survival rates, decrease anxiety and fear to some extent, and make a proper personalized treatment plan for the suffering patient. The survival rates of people with diffuse large B-cell lymphoma (DLBCL) can be forecasted. Both solid and non-solid tumors can be diagnosed precisely with the help of AI and ML algorithms. The prognosis of the disease can also be forecasted with AI and its approaches like deep learning. This improvement in cancer care is a turning point in advanced healthcare and will deeply impact patient's life for good.
... Multiple techniques are proposed to overcome the limitation caused by the lack of training data. Some researchers proposed using the transfer learning technique, as they believe that pretraining on other datasets such as ImageNet can help the neural network to better extract meaningful features [27]- [29]. However, the differences between CT images and natural images (e.g. ...
Neural network related methods, due to their unprecedented success in image processing, have emerged as a new set of tools in CT reconstruction with the potential to change the field. However, the lack of high-quality training data and theoretical guarantees, together with increasingly complicated network structures, make its implementation impractical. In this paper, we present a new framework (RBP-DIP) based on Deep Image Prior (DIP) and a special residual back projection (RBP) connection to tackle these challenges. Comparing to other pre-trained neural network related algorithms, the proposed framework is closer to an iterative reconstruction (IR) algorithm as it requires no training data or training process. In that case, the proposed framework can be altered (e.g, different hyperparameters and constraints) on demand, adapting to different conditions (e.g, different imaged objects, imaging instruments, and noise levels) without retraining. Experiments show that the proposed framework has significant improvements over other state-of-the-art conventional methods, as well as pre-trained and untrained models with similar network structures, especially under sparse-view, limited-angle, and low-dose conditions.
... Researchers have registered an accuracy using pre-trained CNN of 76.79% 58 . Ciompi et al. 59 suggested a 2-D CNN architecture for nodule detection called "Over-Feat". 6 convolutional layers were used in the architecture, with lter sizes ranging from 7 *7 to 3 * 3. Finally, a two-stage classi er has been used by researchers for peri ssural nodule detection 59 . ...
Lung cancer is considered as the common cancerous neoplasms across the globe. In 2018, the World Health Organization (WHO) statistics approximated 2.09 million lung cancer cases with 1.76 million deaths globally. Early identification is an important aspect of providing the greatest chance of healing the patients. The objective of this manuscript was to explore how Deep Learning (DL) performs when the method is evaluated on datasets that are not from LUNA 16 for detection of pulmonary nodule and categorization of computed tomography scans. This report covered only peer-reviewed, original research papers using DL technology, and only findings were included from testing on datasets other than LUNA-16 and LIDC-IDRI. Deep learning utilizes Computed-Tomography (CT) to automatically improve the precision of an initial diagnosis of lung cancer. Consequently, this manuscript presents a short yet important review of DL methods to solve the extraordinary challenges of detecting lung cancer. In addition, this paper also traces the various causes, types, and treatment procedures of lung cancer. The fundamental principles of deep learning and CT have been described. A review of the various lung cancer detection methods via deep learning has been presented. Finally, discussions have been provided for further improvisation of the deep learning method. 9 studies investigated pulmonary nodule detection performance, 10 studies investigated the classification of pulmonary nodule performance, and 16 studies documented of pulmonary nodule for both classification and detection. Some of prominent DL methods which have been successful in detection and categorization of lung cancer nodules are Computer Aided Detection (CAD), Wavelet Recurrent Neural Network (WRNN), Optimal Deep Neural Network (ODNN), Massive Artificial Neural Network (MTANN) and Convolutional Neural Network (CNN) Training. Among, these DL methods, in most cases CNN achieved higher accurate results. The reports CNN achieved results between 73%-96.73% for both classification and detection. The CNN achieved results between 76%-99.2% for lung nodules classification and also achieved the results between 74.6%-97.78% for lung nodule detection. In addition to this, it was found that other DL method i.e., MTANN achieved the accurate results between 97%-100% for detection which came out to be superior related to other DL approaches.
... Some datasets were only for the recognition of pneumonia, 6,7 while some datasets were only for the recognition of pulmonary nodules, 8,9 with generally small data volumes. Although many methods [10][11][12] have achieved good accuracy on classification problems, they are not of high practical value because in real life, lung lesions include many types, and a patient may have multiple thorax diseases at the same time. Only identifying a single disease cannot meet the needs of assisting doctors in diagnosis. ...
Deep learning has recently been widely used in medical image analysis due to its ability to learn complex features in images. An increasing number of deep learning methods has been devoted to the classification of chest X‐ray (CXR) images. Most of the existing deep learning methods are based on classical pretraining models trained by global chest X‐ray images. In this paper, we classify 14 different thorax diseases based on the RE‐DSENet model, which applies the image registration algorithm to the multi‐atlas segmentation model. The experimental results on the ChestX‐ray14 dataset show that the RE‐DSENet's classify‐cation precision of 14 thorax diseases is significantly improved compared with three existing classic algorithms, and the average AUC (area under the ROC curve) score is up to 0.857. RE‐DSENet can also reduce the computations in the classification network and improve the training efficiency. In addition, the focus areas diagnosed by the RE‐DSENet model are visualized in this paper, and the results further prove the effectiveness of the model.
... It mentions that their CNN arch itecture is fit for problems of binary classification problem. One such other paper is written by Chung Where he describes the classification of pulmonary nodules in a 2-D v iews [18] and some papers that emphasizes on a similar matter and works on 3-D images are written by Chen H., Jiang. [15] [19] [20]. ...
... Off-the-shelf trained CNNs framework is the one amongst them used for a wide data set of natural images, here the features are extracted from the dataset of medical imaging through executing prescribed layer of a pre-trained model and then training of a separate classification method is performed. References [11,56] amalgamated handcrafted features with the features derived from pre-trained CNN for the recognition of a pulmonary nodule in CT scan. In the second category, CNNs are pre-trained and fine-tuning is applied for the development of medical images database. ...
Understanding and analyzing of Magnetic resonance imaging (MRI) used in detecting the brain anamoly by specialists manually is a time-consuming, cumbersome and susceptible to intra-subject variations. Hence the proposed non-invasive Computer Aided Diagnosis (CAD) based on brain MRI is aimed to aid the radiologists and physicians to detect the presence of Glioma tumors and its variants on pulse sequences of T1, T1C, T2 and Flair. After preprocessing, using segmentation best features that differentiate one class of objects from another are selected by integrating deep learning features with handcrafted features. Later in the Classification phase, the integrated features of deep learning and handcrafted features are optimized by implementing Particle Swarm Optimization (PSO) algorithm. Finally, these integrated features are classified by Classifiers such as MSVM, KNN, ESDN and Softmax. The GoogLeNet is a pre-trained Convolution Neural Network (CNN) model employed for deep features extraction. Two popular datasets BRATS and Figshare is used for Classification of variants of Glioma tumors using a ten-fold cross-validation.The proposed system acheives high classification accuracy with MSVM Classifier when compared with Softmax, KNN and ESDA Classifiers, thus outperforming all state-of-the-art methods. The performance metrics used in this work are the Area Under Curve - Region of Operating Characteristic curve (AUC-ROC), Precision, Recall, and Specificity. Overall outcome clearly reveals that the proposed framework outperforms both the Segmentation and Classification algorithms of Brain tumors mainly in terms of computation time in contrast to the state-of-the-art methods.
Breast cancer’s high mortality rate is often linked to late diagnosis, with mammograms as key but sometimes limited tools in early detection. To enhance diagnostic accuracy and speed, this study introduces a novel computer-aided detection (CAD) ensemble system. This system incorporates advanced deep learning networks—EfficientNet, Xception, MobileNetV2, InceptionV3, and Resnet50—integrated via our innovative consensus-adaptive weighting (CAW) method. This method permits the dynamic adjustment of multiple deep networks, bolstering the system’s detection capabilities. Our approach also addresses a major challenge in pixel-level data annotation of faster R-CNNs, highlighted in a prominent previous study. Evaluations on various datasets, including the cropped DDSM (Digital Database for Screening Mammography), DDSM, and INbreast, demonstrated the system’s superior performance. In particular, our CAD system showed marked improvement on the cropped DDSM dataset, enhancing detection rates by approximately 1.59% and achieving an accuracy of 95.48%. This innovative system represents a significant advancement in early breast cancer detection, offering the potential for more precise and timely diagnosis, ultimately fostering improved patient outcomes.
Machine learning has significantly impacted the analysis of biological images and is now an important part of many biological data analysis pipelines. A variety of biological and biomedical domain-related tasks is gaining benefit from image analysis and pattern recognition tools developed currently. Applications include diagnostic histopathology, environmental monitoring, synthetic biology, genomics, and proteomics. Particularly in the last decade, several deep learning and advanced computer vision methods such as convolutional neural networks (CNNs), typically trained in a supervised fashion, have started to be largely employed in biological image classification. Moreover, the advancement of automatic acquisition systems has been generating a massive amount of biological data, which requires to be analyzed by domain experts. However, the cost of manual annotation of such data has become a bottleneck, impairing the application of supervised machine learning algorithms. Biological images generally have an intrinsic high variability, whose identity is sometimes hard to assign and strongly dependent on the annotator’s expertise. In this context, a limited number of annotation-free (i.e., unsupervised) learning solutions have been proposed, typically based on hand-crafted features, specifically tailored for a certain biological domain. Nonetheless, a successful unsupervised learning approach must be accurate, and sufficiently robust to deal with different biological domains. This paper aims at providing a viable solution to these issues, proposing an unsupervised learning algorithm based on compressed deep features for image classification. We exploit features extracted from ImageNet pre-trained transformers and CNNs, further compressed with a customized 𝛽-Variational AutoEncoder (𝛽-VAE), that we call reconstruction VAE (R-VAE). We test our algorithm on biological images coming from diverse domains characterized by high variability in shape and texture information and acquired with widely differing imaging platforms. Considered image datasets range from multi-cellular organisms (plankton, coral) to sub-cellular organelles (budding yeast vacuoles, human cells’ nuclei, etc.). Our results show that the compressed deep features extracted from different pre-trained vision models establish new unsupervised learning state-of-the-art performances for the investigated datasets.
The use of machine learning (ML) is continuously growing in healthcare to analyze millions of different data points and predict results, show risk scores on time, resource allocation precisely, and other applications. It has already been proved by the different applications that ML has tremendous potential in health care, including various parts of nutrition cum diet recommendations. Many research papers also showed the role and prospects of ML in food, nutrition, and diet management,
such as food recognition and nutrition estimation. Machine learning techniques can also identify leading dietary determinants for metabolic risks associated with a healthy diet score. However, nutritionists and researchers have yet to apply these techniques considering better prospects for the well-being of patients (following the diet), and all communities need the proper nutrition. Hence there is a need for awareness and basic level learning of ML for nutritionists and all stakeholders.
Because of the great responsiveness of aspiratory knob location, computerized tomography (CT) is generally used to analyze cellular breakdown in the lungs without performing biopsy, which could make actual harm nerves and vessels. Notwithstanding, recognizing threatening and harmless aspiratory knobs stays troublesome. Since CT checks are regularly of low goal, it is challenging for radiologists to peruse the output picture’s subtleties. The proceeded with quick development of CT examine examination frameworks lately has made a squeezing need for cutting edge computational apparatuses to remove helpful highlights to help the radiologist in understanding advancement. PC-supported discovery (CAD) frameworks have been created to diminish notable mistakes by distinguishing the dubious highlights a radiologist searches for in a case survey. Our project aims to compare performance of various low memories, lightweight deep neural net (DNN) architectures for biomedical image analysis. It will involve networks like vanilla 2D CNN, U-Net, 2D SqueezeNet, and 2D MobileNet for two case classifications to discover the existence of lung cancer in patient CT scans of lungs with and without primary phase lung cancer.KeywordPulmonary noduleComputer tomographyAI modelCAD system
Minimally invasive repair of pectus excavatum (MIRPE) is an effective method for correcting pectus excavatum (PE), a congenital chest wall deformity characterized by concave depression of the sternum. In MIRPE, a long, thin, curved stainless plate (implant) is placed across the thoracic cage to correct the deformity. However, the implant curvature is difficult to accurately determine during the procedure. This implant depends on the surgeon's expert knowledge and experience and lacks objective criteria. Moreover, tedious manual input by surgeons is required to estimate the implant shape. In this study, a novel three-step end-to-end automatic framework is proposed to determine the implant shape during preoperative planning: (1) The deepest depression point (DDP) in the sagittal plane of the patient's CT volume is automatically determined using Sparse R-CNN-R101, and the axial slice containing the point is extracted. (2) Cascade Mask R-CNN-X101 segments the anterior intercostal gristle of the pectus, sternum and rib in the axial slice, and the contour is extracted to generate the PE point set. (3) Robust shape registration is performed to match the PE shape with a healthy thoracic cage, which is then utilized to generate the implant shape. The framework was evaluated on a CT dataset of 90 PE patients and 30 healthy children. The experimental results show that the average error of the DDP extraction was 5.83 mm. The end-to-end output of our framework was compared with surgical outcomes of professional surgeons to clinically validate the effectiveness of our method. The results indicate that the root mean square error (RMSE) between the midline of the real implant and our framework output was less than 2 mm.
Lung cancer is the primary reason of cancer deaths worldwide, and the percentage of death rate is increasing step by step. There are chances of recovering from lung cancer by detecting it early. In any case, because the number of radiologists is limited and they have been working overtime, the increase in image data makes it hard for them to evaluate the images accurately. As a result, many researchers have come up with automated ways to predict the growth of cancer cells using medical imaging methods in a quick and accurate way. Previously, a lot of work was done on computer-aided detection (CADe) and computer-aided diagnosis (CADx) in computed tomography (CT) scan, magnetic resonance imaging (MRI), and X-ray with the goal of effective detection and segmentation of pulmonary nodule, as well as classifying nodules as malignant or benign. But still, no complete comprehensive review that includes all aspects of lung cancer has been done. In this paper, every aspect of lung cancer is discussed in detail, including datasets, image preprocessing, segmentation methods, optimal feature extraction and selection methods, evaluation measurement matrices, and classifiers. Finally, the study looks into several lung cancer-related issues with possible solutions.
Lung cancer, the deadliest disease worldwide, poses a massive threat to humankind. Various researchers have designed Computer-Aided-Diagnosis systems for the early-stage detection of lung cancer. However, patients are primarily diagnosed in advanced stages when treatment becomes complicated and dependent on multiple factors like size, nature, location of the tumor, and proper cancer staging. TNM (Tumor, Node, and Metastasis) staging provides all this information. This study aims to develop a novel and efficient approach to classify lung cancer stages based on TNM standards. We propose a multi-level 3D deep convolutional neural network, LCSCNet (Lung Cancer Stage Classification Network). The proposed network architecture consists of three similar classifier networks to classify three labels, T, N, and M-labels. First, we pre-process the data, in which the CT images are augmented, and the label files are processed to get the corresponding TNM labels. For the classification network, we implement a dense convolutional neural network with a concurrent squeeze & excitation module and asymmetric convolutions for classifying each label separately. The overall stage is determined by combining all three labels. The concurrent squeeze & excitation module helps the network focus on the essential information of the image, due to which the classification performance is enhanced. The asymmetric convolutions are introduced to reduce the computation complexity of the network. Two publicly available datasets are used for this study. We achieved average accuracies of 96.23% for T-Stage, 97.63% for N-Stage, and 96.92% for M-Stage classification. Furthermore, an overall stage classification accuracy of 97% is achieved.
Data mining of medical imaging approaches makes it difficult to determine their value in the disease's insight, analysis, and diagnosis. Image classification presents a significant difficulty in image analysis and plays a vital part in computer‐aided diagnosis. This task concerned the use of optimization techniques for the utilization of image processing, pattern recognition, and classification techniques, as well as the validation of image classification results in medical expert reports. The primary intention of this study is to analyze the performance of optimization techniques explored in the area of medical image analysis. For this motive, the optimization techniques employed in existing literature from 2012 to 2021 are reviewed in this study. The contribution of optimization‐based medical image classification and segmentation utilized image modalities, data sets, and tradeoffs for each technique are also discussed in this review study. Finally, this review study provides the gap analysis of optimization techniques used in medical image analysis with the possible future research direction.
Automatic pulmonary nodule detection is the most crucial technology in early diagnosis of lung cancer and early treatment. However, the various nodule type, shape and size, especially the diameter of lung nodules (ranging from 3 mm to 30 mm), make the lung nodule detection results with high false positives. It significantly affects the detection performance of lung nodules. In this paper, an adaptive and attentive 3D Convolutional Neural Network (CNN) is proposed for automatic pulmonary nodule detection, which contains two parts: the candidate nodule detection and false positive reduction. In the first stage, the globule, spital and fine-grained information of focal nodules are scratched by the high-resolution fused attention module in the proposed method. In the second stage, an adaptive 3D CNN structure is designed to further reduce the false positives, which extracts the multilevel contextual information via an adaptive 3D convolution kernel. Extensive experiments are conducted on publicly available LUNA16. The results demonstrate that the proposed method can increase the sensitivity and decrease the false positives rate for automated pulmonary nodule detection effectively.
Background and Objective
Recent studies have shown that colorectal cancer (CRC) patients with microsatellite instability high (MSI-H) are more likely to benefit from immunotherapy. However, current MSI testing methods are not available for all patients due to the lack of available equipment and trained personnel, as well as the high cost of the assay. Here, we developed an improved deep learning model to predict MSI-H in CRC from whole slide images (WSIs).
Methods
We established the MSI-H prediction model based on two stages: tumor detection and MSI classification. Previous works applied fine-tuning strategy directly for tumor detection, but ignoring the challenge of vanishing gradient due to the large number of convolutional layers. We added auxiliary classifiers to intermediate layers of pre-trained models to help propagate gradients back through in an effective manner. To predict MSI status, we constructed a pair-wise learning model with a synergic network, named parameter partial sharing network (PPsNet), where partial parameters are shared among two deep convolutional neural networks (DCNNs). The proposed PPsNet contained fewer parameters and reduced the problem of intra-class variation and inter-class similarity. We validated the proposed model on a holdout test set and two external test sets.
Results
144 H&E-stained WSIs from 144 CRC patients (81 cases with MSI-H and 63 cases with MSI-L/MSS) were collected retrospectively from three hospitals. The experimental results indicate that deep supervision based fine-tuning almost outperforms training from scratch and utilizing fine-tuning directly. The proposed PPsNet always achieves better accuracy and area under the receiver operating characteristic curve (AUC) than other solutions with four different neural network architectures on validation. The proposed method finally achieves obvious improvements than other state-of-the-art methods on the validation dataset with an accuracy of 87.28% and AUC of 94.29%.
Conclusions
The proposed method can obviously increase model performance and our model yields better performance than other methods. Additionally, this work also demonstrates the feasibility of MSI-H prediction using digital pathology images based on deep learning in the Asian population. It is hoped that this model could serve as an auxiliary tool to identify CRC patients with MSI-H more time-saving and efficiently.
Introduction: In this study, the performance of computed tomography lung image segmentation using image processing and Markov Random Field was investigated. Before cancer segmentation and analysis, lung segmentation is an important initial process. Thus, the aim of this study is to find the optimal Markov Random Field setting for lung segmentation. Methods: The Centre for Diagnostic Nuclear Imaging at UPM provided 11 anonymous sets of cancerous lung CT images for this study. The thresholding technique is an effective method for medical image segmentation when the priori information for the region of interest is known, such as the Hounsfield Unit value of lung. Due to the large differences in grey levels in the image, the thresholding approach is difficult to apply in segmentation, especially for lung. Thus, for the segmentation process, this study used multilevel thresholding with Markov Random Field with three settings; Iterated Condition Mode, Metropolis algorithm, and Gibbs sampler. The images then went through image processing procedures which were binarization, small object removal, lung region extraction and lung segmentation. The output from the experiments were analyzed and compared to determine the ideal lung segmentation setting. Results: The Jaccard index average values; Markov Random Field -Metropolis = 0.9464, Markov Random Field -ICM = 0.9499 and Markov Random Field -Gibbs = 0.9512. The Dice index average values; Markov Random Field - Metropolis = 0.9743, Markov Random Field - ICM = 0.9724 and Markov Random Field - Gibbs = 0.9749. Conclusion: Markov Random Field using Gibbs sampler delivered the best results for lung segmentation.
Rupture of intracranial aneurysm is the first cause of subarachnoid hemorrhage, second only to cerebral thrombosis and hypertensive cerebral hemorrhage, and the mortality rate is very high. MRI technology plays an irreplaceable role in the early detection and diagnosis of intracranial aneurysms and supports evaluating the size and structure of aneurysms. The increase in many aneurysm images, may be a massive workload for the doctors, which is likely to produce a wrong diagnosis. Therefore, we proposed a simple and effective comprehensive residual attention network (CRANet) to improve the accuracy of aneurysm detection, using a residual network to extract the features of an aneurysm. Many experiments have shown that the proposed CRANet model could detect aneurysms effectively. In addition, on the test set, the accuracy and recall rates reached 97.81% and 94%, which significantly improved the detection rate of aneurysms.
In this study, the performance of computed tomography lung classification using image processing and Markov Random Field were investigated. For lung classification, the process must first be going thru lung segmentation process. Lung segmentation is important as an initial process before lung cancer segmentation and analysis. Image processing was employed to the input image. We propose multilevel thresholding and Markov Random Field to improve the segmentation process. Three setting for Markov Random Field was used for segmentation process that is Iterated Condition Mode, Metropolis algorithm and Gibbs sampler. The output from the experiments were analysed and compared to get the best performance. The results revealed that for CT image lung classification, Markov Random Field using Metropolis algorithm gives the best results. In view of the result obtained, the average accuracy is 94.75% while the average sensitivity and specificity are 76.34% and 99.80%. The output from this study can be implemented in lung cancer analysis research and computer aided diagnosis development.
Unets have become the standard method for semantic segmentation of medical images, along with fully convolutional networks (FCN). Unet++ was introduced as a variant of Unet, in order to solve some of the problems facing Unet and FCNs. Unet++ provided networks with an ensemble of variable depth Unets, hence eliminating the need for professionals estimating the best suitable depth for a task. While Unet and all its variants, including Unet++ aimed at providing networks that were able to train well without requiring large quantities of annotated data, none of them attempted to eliminate the need for pixel-wise annotated data altogether. Obtaining such data for each disease to be diagnosed comes at a high cost. Hence such data is scarce. In this paper we use contrastive learning to train Unet++ for semantic segmentation of medical images using medical images from various sources including magnetic resonance imaging (MRI) and computed tomography (CT), without the need for pixel-wise annotations. Here we describe the architecture of the proposed model and the training method used. This is still a work in progress and so we abstain from including results in this paper. The results and the trained model would be made available upon publication or in subsequent versions of this paper on arxiv.
To check out the health of the patient, digital images are generated every single day and are used by the radiologist for extracting out the details and anomalies. The complicated part is to figure out the disease in those images. By the manual diagnosis of the images through the radiologists, the doctors can get to know exact scenario of the abnormalities in images, but is considerably more difficult with Content Based Image Retrieval (CBIR) to get those finer details from MR images. These CBIR approaches are now frequently employed in the automatic diagnosis of disease from MR images, mammograms, and other sources. Bridging this gap can be done with deep learning feature extraction algorithm and the canny edge detection technique we propose, and accuracy closer to the manual results of a human evaluator can be achieved to a significant extent as part of the goal of sustainable development through innovation.
Deep learning has received extensive research interest in developing new medical image processing algorithms, and deep learning based models have been remarkably successful in a variety of medical imaging tasks to support disease detection and diagnosis. Despite the success, the further improvement of deep learning models in medical image analysis is majorly bottlenecked by the lack of large-sized and well-annotated datasets. In the past five years, many studies have focused on addressing this challenge. In this paper, we reviewed and summarized these recent studies to provide a comprehensive overview of applying deep learning methods in various medical image analysis tasks. Especially, we emphasize the latest progress and contributions of state-of-the-art unsupervised and semi-supervised deep learning in medical image analysis, which are summarized based on different application scenarios, including classification, segmentation, detection, and image registration. We also discuss the major technical challenges and suggest the possible solutions in future research efforts.
Recent results indicate that the generic descriptors extracted from the
convolutional neural networks are very powerful. This paper adds to the
mounting evidence that this is indeed the case. We report on a series of
experiments conducted for different recognition tasks using the publicly
available code and model of the OverFeat network which was trained to perform
object classification on ILSVRC13. We use features extracted from the OverFeat
network as a generic image representation to tackle the diverse range of
recognition tasks of object image classification, scene recognition, fine
grained recognition, attribute detection and image retrieval applied to a
diverse set of datasets. We selected these tasks and datasets as they gradually
move further away from the original task and data the OverFeat network was
trained to solve. Remarkably we report better or competitive results compared
to the state-of-the-art in all the tasks on various datasets. The results are
achieved using a linear SVM classifier applied to a feature representation of
size 4096 extracted from a layer in the net. The results strongly suggest that
features obtained from deep learning with convolutional nets should be the
primary candidate in most visual classification tasks.
Major issues in the implementation of screening for lung cancer by means of low-dose computed tomography (CT) are the definition of a positive result and the management of lung nodules detected on the scans. We conducted a population-based prospective study to determine factors predicting the probability that lung nodules detected on the first screening low-dose CT scans are malignant or will be found to be malignant on follow-up.
We analyzed data from two cohorts of participants undergoing low-dose CT screening. The development data set included participants in the Pan-Canadian Early Detection of Lung Cancer Study (PanCan). The validation data set included participants involved in chemoprevention trials at the British Columbia Cancer Agency (BCCA), sponsored by the U.S. National Cancer Institute. The final outcomes of all nodules of any size that were detected on baseline low-dose CT scans were tracked. Parsimonious and fuller multivariable logistic-regression models were prepared to estimate the probability of lung cancer.
In the PanCan data set, 1871 persons had 7008 nodules, of which 102 were malignant, and in the BCCA data set, 1090 persons had 5021 nodules, of which 42 were malignant. Among persons with nodules, the rates of cancer in the two data sets were 5.5% and 3.7%, respectively. Predictors of cancer in the model included older age, female sex, family history of lung cancer, emphysema, larger nodule size, location of the nodule in the upper lobe, part-solid nodule type, lower nodule count, and spiculation. Our final parsimonious and full models showed excellent discrimination and calibration, with areas under the receiver-operating-characteristic curve of more than 0.90, even for nodules that were 10 mm or smaller in the validation set.
Predictive tools based on patient and nodule characteristics can be used to accurately estimate the probability that lung nodules detected on baseline screening low-dose CT scans are malignant. (Funded by the Terry Fox Research Institute and others; ClinicalTrials.gov number, NCT00751660.).
Purpose: The development of computer-aided diagnostic (CAD) methods for lung nodule detection, classification, and quantitative assessment can be facilitated through a well-characterized repository of computed tomography (CT) scans. The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI) completed such a database, establishing a publicly available reference for the medical imaging research community. Initiated by the National Cancer Institute (NCI), further advanced by the Foundation for the National Institutes of Health (FNIH), and accompanied by the Food and Drug Administration (FDA) through active participation, this public-private partnership demonstrates the success of a consortium founded on a consensus-based process.
Methods: Seven academic centers and eight medical imaging companies collaborated to identify, address, and resolve challenging organizational, technical, and clinical issues to provide a solid foundation for a robust database. The LIDC/IDRI Database contains 1018 cases, each of which includes images from a clinical thoracic CT scan and an associated XML file that records the results of a two-phase image annotation process performed by four experienced thoracic radiologists. In the initial blinded-read phase, each radiologist independently reviewed each CT scan and marked lesions belonging to one of three categories (“nodule ≥ 3 mm,” “nodule<3 mm,” and “non-nodule ≥ 3 mm”). In the subsequent unblinded-read phase, each radiologist independently reviewed their own marks along with the anonymized marks of the three other radiologists to render a final opinion. The goal of this process was to identify as completely as possible all lung nodules in each CT scan without requiring forced consensus.
Results: The Database contains 7371 lesions marked “nodule” by at least one radiologist. 2669 of these lesions were marked “nodule ≥ 3 mm” by at least one radiologist, of which 928 (34.7%) received such marks from all four radiologists. These 2669 lesions include nodule outlines and subjective nodule characteristic ratings.
Conclusions: The LIDC/IDRI Database is expected to provide an essential medical imaging research resource to spur CAD development, validation, and dissemination in clinical practice.
In this work, we investigate the visual appearance of real-world surfaces and the dependence of appearance on the geometry of imaging conditions. We discuss a new texture representation called the BTF (bidirectional texture function) which captures the variation in texture with illumination and viewing direction. We present a BTF database with image textures from over 60 different samples, each observed with over 200 different combinations of viewing and illumination directions. We describe the methods involved in collecting the database as well as the importqance and uniqueness of this database for computer graphics. A related quantity to the BTF is the familiar BRDF (bidirectional reflectance distribution function). The measurement methods involved in the BTF database are conducive to simultaneous measurement of the BRDF. Accordingly, we also present a BRDF database with reflectance measurements for over 60 different samples, each observed with over 200 different combinations of viewing and illumination directions. Both of these unique databases are publicly available and have important implications for computer graphics.
Purpose:
Existing computer-aided detection schemes for lung nodule detection require a large number of calculations and tens of minutes per case; there is a large gap between image acquisition time and nodule detection time. In this study, we propose a fast detection scheme of lung nodule in chest CT images using cylindrical nodule-enhancement filter with the aim of improving the workflow for diagnosis in CT examinations.
Methods:
Proposed detection scheme involves segmentation of the lung region, preprocessing, nodule enhancement, further segmentation, and false-positive (FP) reduction. As a nodule enhancement, our method employs a cylindrical shape filter to reduce the number of calculations. False positives (FPs) in nodule candidates are reduced using support vector machine and seven types of characteristic parameters.
Results:
The detection performance and speed were evaluated experimentally using Lung Image Database Consortium publicly available image database. A 5-fold cross-validation result demonstrates that our method correctly detects 80 % of nodules with 4.2 FPs per case, and detection speed of proposed method is also 4-36 times faster than existing methods.
Conclusion:
Detection performance and speed indicate that our method may be useful for fast detection of lung nodules in CT images.
The aggressive and heterogeneous nature of lung cancer has thwarted efforts to reduce mortality from this cancer through the use of screening. The advent of low-dose helical computed tomography (CT) altered the landscape of lung-cancer screening, with studies indicating that low-dose CT detects many tumors at early stages. The National Lung Screening Trial (NLST) was conducted to determine whether screening with low-dose CT could reduce mortality from lung cancer.
From August 2002 through April 2004, we enrolled 53,454 persons at high risk for lung cancer at 33 U.S. medical centers. Participants were randomly assigned to undergo three annual screenings with either low-dose CT (26,722 participants) or single-view posteroanterior chest radiography (26,732). Data were collected on cases of lung cancer and deaths from lung cancer that occurred through December 31, 2009.
The rate of adherence to screening was more than 90%. The rate of positive screening tests was 24.2% with low-dose CT and 6.9% with radiography over all three rounds. A total of 96.4% of the positive screening results in the low-dose CT group and 94.5% in the radiography group were false positive results. The incidence of lung cancer was 645 cases per 100,000 person-years (1060 cancers) in the low-dose CT group, as compared with 572 cases per 100,000 person-years (941 cancers) in the radiography group (rate ratio, 1.13; 95% confidence interval [CI], 1.03 to 1.23). There were 247 deaths from lung cancer per 100,000 person-years in the low-dose CT group and 309 deaths per 100,000 person-years in the radiography group, representing a relative reduction in mortality from lung cancer with low-dose CT screening of 20.0% (95% CI, 6.8 to 26.7; P=0.004). The rate of death from any cause was reduced in the low-dose CT group, as compared with the radiography group, by 6.7% (95% CI, 1.2 to 13.6; P=0.02).
Screening with the use of low-dose CT reduces mortality from lung cancer. (Funded by the National Cancer Institute; National Lung Screening Trial ClinicalTrials.gov number, NCT00047385.).
The National Lung Screening Trial (NLST) is a randomized multicenter study comparing low-dose helical computed tomography (CT) with chest radiography in the screening of older current and former heavy smokers for early detection of lung cancer, which is the leading cause of cancer-related death in the United States. Five-year survival rates approach 70% with surgical resection of stage IA disease; however, more than 75% of individuals have incurable locally advanced or metastatic disease, the latter having a 5-year survival of less than 5%. It is plausible that treatment should be more effective and the likelihood of death decreased if asymptomatic lung cancer is detected through screening early enough in its preclinical phase. For these reasons, there is intense interest and intuitive appeal in lung cancer screening with low-dose CT. The use of survival as the determinant of screening effectiveness is, however, confounded by the well-described biases of lead time, length, and overdiagnosis. Despite previous attempts, no test has been shown to reduce lung cancer mortality, an endpoint that circumvents screening biases and provides a definitive measure of benefit when assessed in a randomized controlled trial that enables comparison of mortality rates between screened individuals and a control group that does not undergo the screening intervention of interest. The NLST is such a trial. The rationale for and design of the NLST are presented.
Lung cancer screening with low dose computed tomography (CT) has not yet been evaluated in randomized clinical trials, although several are underway.
In The Danish Lung Cancer Screening Trial, 4104 smokers and previous smokers from 2004 to 2006 were randomized to either screening with annual low dose CT scans for 5 years or no screening. A history of cigarette smoking of at least 20 pack years was required. All participants have annual lung function tests, and questionnaires regarding health status, psychosocial consequences of screening, smoking habits, and smoking cessation. Baseline CT scans were performed in 2052 participants. Pulmonary nodules were classified according to size and morphology: (1) Nodules smaller than 5 mm and calcified (benign) nodules were tabulated, (2) Noncalcified nodules between 5 and 15 mm were rescanned after 3 months. If the nodule increased in size or was larger than 15 mm the participant was referred for diagnostic workup.
At baseline 179 persons showed noncalcified nodules larger than 5 mm, and most were rescanned after 3 months: The rate of false-positive diagnoses was 7.9%, and 17 individuals (0.8%) turned out to have lung cancer. Ten of these had stage I disease. Eleven of 17 lung cancers at baseline were treated surgically, eight of these by video assisted thoracic surgery resection.
Screening may facilitate minimal invasive treatment and can be performed with a relatively low rate of false-positive screen results compared with previous studies on lung cancer screening.
Convolutional neural networks (CNNs) have emerged as the most powerful technique for a range of different tasks in computer vision. Recent work suggested that CNN features are generic and can be used for classification tasks outside the exact domain for which the networks were trained. In this work we use the features from one such network, OverFeat, trained for object detection in natural images, for nodule detection in computed tomography scans. We use 865 scans from the publicly available LIDC data set, read by four thoracic radiologists. Nodule candidates are generated by a state-of-the-art nodule detection system. We extract 2D sagittal, coronal and axial patches for each nodule candidate and extract 4096 features from the penultimate layer of OverFeat and classify these with linear support vector machines. We show for various configurations that the off-the-shelf CNN features perform surprisingly well, but not as good as the dedicated detection system. When both approaches are combined, significantly better results are obtained than either approach alone. We conclude that CNN features have great potential to be used for detection tasks in volumetric medical data.
In this work, we examine the strength of deep learning approaches for pathology detection in chest radiographs. Convolutional neural networks (CNN) deep architecture classification approaches have gained popularity due to their ability to learn mid and high level image representations. We explore the ability of CNN learned from a non-medical dataset to identify different types of pathologies in chest x-rays. We tested our algorithm on a 433 image dataset. The best performance was achieved using CNN and GIST features. We obtained an area under curve (AUC) of 0.87-0.94 for the different pathologies. The results demonstrate the feasibility of detecting pathology in chest x-rays using deep learning approaches based on non-medical learning. This is a first-of-its-kind experiment that shows that Deep learning with ImageNet, a large scale non-medical image database may be a good substitute to domain specific representations, which are yet to be available, for general medical image recognition tasks.
We present a novel descriptor for the characterization of pulmonary nodules in computed tomography (CT) images. The descriptor encodes information on nodule morphology and has scale-invariant and rotation-invariant properties. Information on nodule morphology is captured by sampling intensity profiles along circular patterns on spherical surfaces centered on the nodule, in a multi-scale fashion. Each intensity profile is interpreted as a periodic signal, where the Fourier transform is applied, obtaining a spectrum. A library of spectra is created and labeled via unsupervised clustering, obtaining a Bag-of- Frequencies, which is used to assign each spectra a label. The descriptor is obtained as the histogram of labels along all the spheres. Additional contributions are a technique to estimate the nodule size, based on the sampling strategy, as well as a technique to choose the most informative plane to cut a 2-D view of the nodule in the 3-D image. We evaluate the descriptor on several nodule morphology classification problems, namely discrimination of nodules versus vascular structures and characterization of spiculation. We validate the descriptor on data from European screening trials NELSON and DLCST and we compare it with state-of-the-art approaches for 3-D shape description in medical imaging and computer vision, namely SPHARM and 3-D SIFT, outperforming them in all the considered experiments.
We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 dif-ferent classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make train-ing faster, we used non-saturating neurons and a very efficient GPU implemen-tation of the convolution operation. To reduce overfitting in the fully-connected layers we employed a recently-developed regularization method called "dropout" that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.
Subsolid pulmonary nodules occur less often than solid pulmonary nodules, but show a much higher malignancy rate. Therefore, accurate detection of this type of pulmonary nodules is crucial. In this work, a computer-aided detection (CAD) system for subsolid nodules in computed tomography images is presented and evaluated on a large data set from a multi-center lung cancer screening trial. The paper describes the different components of the CAD system and presents experiments to optimize the performance of the proposed CAD system. A rich set of 128 features is defined for subsolid nodule candidates. In addition to previously used intensity, shape and texture features, a novel set of context features is introduced. Experiments show that these features significantly improve the classification performance. Optimization and training of the CAD system is performed on a large training set from one site of a lung cancer screening trial. Performance analysis on an independent test from another site of the trial shows that the proposed system reaches a sensitivity of 80% at an average of only 1.0 false positive detections per scan. A retrospective analysis of the output of the CAD system by an experienced thoracic radiologist shows that the CAD system is able to find subsolid nodules which were not contained in the screening database.
Random forests are a combination of tree predictors such that each tree depends on the values of a random vector sampled independently and with the same distribution for all trees in the forest. The generalization error for forests converges a.s. to a limit as the number of trees in the forest becomes large. The generalization error of a forest of tree classifiers depends on the strength of the individual trees in the forest and the correlation between them. Using a random selection of features to split each node yields error rates that compare favorably to Adaboost (Y. Freund & R. Schapire, Machine Learning: Proceedings of the Thirteenth International conference, ***, 148–156), but are more robust with respect to noise. Internal estimates monitor error, strength, and correlation and these are used to show the response to increasing the number of features used in the splitting. Internal estimates are also used to measure variable importance. These ideas are also applicable to regression.
Purpose:
To assess the prevalence, natural course, and malignancy rate of perifissural nodules (PFNs) in smokers participating in a lung cancer screening trial.
Materials and methods:
As part of the ethics-committee approved Dutch-Belgian Randomised Lung Cancer Multi-Slice Screening Trial (NELSON), computed tomography (CT) was used to screen 2994 current or former heavy smokers, aged 50-74 years, for lung cancer. CT was repeated after 1 and 3 years, with additional follow-up CT scans if necessary. All baseline CT scans were screened for nodules. Nodule volume was determined with automated volumetric analysis. Homogeneous solid nodules, attached to a fissure with a lentiform or triangular shape, were classified as PFNs. Nodules were considered benign if they did not grow during the total follow-up period or were proved to be benign in a follow-up by a pulmonologist. Prevalence, growth, and malignancy rate of PFNs were assessed.
Results:
At baseline screening, 4026 nodules were detected in 1729 participants, and 19.7% (794 of 4026) of the nodules were classified as PFNs. The mean size of the PFNs was 4.4 mm (range: 2.8-10.6 mm) and the mean volume was 43 mm3 (range: 13-405 mm3). None of the PFNs were found to be malignant during follow-up. Between baseline and the first follow-up CT scan, 15.5% (123 of 794) were found to have grown, and 8.3% (66 of 794) had a volume doubling time of less than 400 days. One PFN was resected and proved to be a lymph node.
Conclusion:
PFNs are frequently found at CT scans for lung cancer. They can show growth rates in the range of malignant nodules, but none of the PFNs in the present study turned out to be malignant. Recognition of PFNs can reduce the number of follow-up examinations required for the workup of suspicious nodules.
The efficacy and cost-effectiveness of low-dose spiral computed tomography (LDCT) screening in heavy smokers is currently under evaluation worldwide. Our screening program started with a pilot study on 1035 volunteers in Milan in 2000 and was followed up in 2005 by a randomized trial comparing annual or biennial LDCT with observation, named Multicentric Italian Lung Detection. This included 4099 participants, 1723 randomized to the control group, 1186 to biennial LDCT screening, and 1190 to annual LDCT screening. Follow-up was stopped in November 2011, with 9901 person-years for the pilot study and 17 621 person-years for Multicentric Italian Lung Detection. Forty-nine lung cancers were detected by LDCT (20 in biennial and 29 in the annual arm), of which 17 were identified at baseline examination; 63% were of stage I and 84% were surgically resectable. Stage distribution and resection rates were similar in the two LDCT arms. The cumulative 5-year lung cancer incidence rate was 311/100 000 in the control group, 457 in the biennial, and 620 in the annual LDCT group (P=0.036); lung cancer mortality rates were 109, 109, and 216/100 000 (P=0.21), and total mortality rates were 310, 363, and 558/100 000, respectively (P=0.13). Total mortality in the pilot study was similar to that observed in the annual LDCT arm at 5 years. There was no evidence of a protective effect of annual or biennial LDCT screening. Furthermore, a meta-analysis of the four published randomized trials showed similar overall mortality in the LDCT arms compared with the control arm.
A new computer algorithm is presented to distinguish a special, most probably benign, subclass of lung nodules called perifissural opacities (PFO's), from potentially malignant nodules. The method focuses on the quantification of two characteristic properties of PFO's, namely the typical flattened surface of the nodule and its attachment to plate-like structures in the direct neighborhood of the nodule (the lung fissures). For the detection of fissures in the proximity of the nodule, an analysis based on the eigenvalues of the Hessian matrix has been developed. Further processing with a voxel grouping algorithm is shown to substantially improve the results of the fissure detection. Through a comparison of Hough transforms of the nodule boundary and the detected fissure voxels, features are constructed that enable a reliable separation of benign PFO from other lesions
Computer-aided detection (CAD) is increasingly used in clinical practice and for many applications a multitude of CAD systems have been developed. In practice, CAD systems have different strengths and weaknesses and it is therefore interesting to consider their combination. In this paper, we present generic methods to combine multiple CAD systems and investigate what kind of performance increase can be expected. Experimental results are presented using data from the ANODE09 and ROC09 online CAD challenges for the detection of pulmonary nodules in computed tomography scans and red lesions in retinal images, respectively. For both applications, combination results in a large and significant increase in performance when compared to the best individual CAD system.
Early detection of lung nodules is extremely important for the diagnosis and clinical management of lung cancer. In this paper, a novel computer aided detection (CAD) system for the detection of pulmonary nodules in thoracic computed tomography (CT) imagery is presented. The paper describes the architecture of the CAD system and assesses its performance on a publicly available database to serve as a benchmark for future research efforts. Training and tuning of all modules in our CAD system is done using a separate and independent dataset provided courtesy of the University of Texas Medical Branch (UTMB). The publicly available testing dataset is that created by the Lung Image Database Consortium (LIDC). The LIDC data used here is comprised of 84 CT scans containing 143 nodules ranging from 3 to 30mm in effective size that are manually segmented at least by one of the four radiologists. The CAD system uses a fully automated lung segmentation algorithm to define the boundaries of the lung regions. It combines intensity thresholding with morphological processing to detect and segment nodule candidates simultaneously. A set of 245 features is computed for each segmented nodule candidate. A sequential forward selection process is used to determine the optimum subset of features for two distinct classifiers, a Fisher Linear Discriminant (FLD) classifier and a quadratic classifier. A performance comparison between the two classifiers is presented, and based on this, the FLD classifier is selected for the CAD system. With an average of 517.5 nodule candidates per case/scan (517.5+/-72.9), the proposed front-end detector/segmentor is able to detect 92.8% of all the nodules in the LIDC/testing dataset (based on merged ground truth). The mean overlap between the nodule regions delineated by three or more radiologists and the ones segmented by the proposed segmentation algorithm is approximately 63%. Overall, with a specificity of 3 false positives (FPs) per case/patient on average, the CAD system is able to correctly identify 80.4% of the nodules (115/143) using 40 selected features. A 7-fold cross-validation performance analysis using the LIDC database only shows CAD sensitivity of 82.66% with an average of 3 FPs per CT scan/case.
Purpose:
To describe and characterize the potential for malignancy of noncalcified lung nodules adjacent to fissures that are often found in current or former heavy smokers who undergo computed tomography (CT) for lung cancer screening.
Materials and methods:
Institutional review board approval and informed consent were obtained. Baseline and follow-up thin-section multidetector CT scans obtained in 146 consecutive subjects at high risk for lung cancer (age range, 50-75 years; > 30 pack-year smoking history) were retrospectively reviewed. Noncalcified nodules (NCNs) were categorized according to location (parenchymal, perifissural), shape, septal connection, manually measured diameter, diameter change, and lung cancer outcome at 7(1/2) years.
Results:
Retrospective review of images from 146 baseline and 311 follow-up CT examinations revealed 837 NCNs in 128 subjects. Of those 837 nodules, 234 (28%), in 98 subjects, were adjacent to a fissure and thus classified as perifissural nodules (PFNs). Multiple (range, 2-14) PFNs were seen in 47 subjects. Most PFNs were triangular (102/234, 44%) or oval (98/234, 42%), were located inferior to the carina (196/234, 84%), and had a septal connection (171/234, 73%). The mean maximal length was 3.2 mm (range, 1-13 mm). During 2-year follow-up in 71 subjects, seven of 159 PFNs increased in size on one scan but were then stable. The authors searched a lung cancer registry 7(1/2) years after study entry and found 10 lung cancers in 139 of 146 study subjects who underwent complete follow-up; none of these cancers had originated from a PFN.
Conclusion:
PFNs are frequently seen on screening CT scans obtained in high-risk subjects. Although PFNs may show increased size at follow-up CT, the authors in this study found none that had developed into lung cancer; this suggests that the malignancy potential of PFNs is low. (c) RSNA, 2010.
The purpose of this study was to describe a simple bootstrap approach for estimating sensitivity, specificity, and the area under the receiver operating characteristic curve for multisite test outcome data.
The performance of bootstrap estimates was evaluated and compared with that of analytic estimates by using a simulation study. Bootstrapping was demonstrated by using data from a previous study comparing two angiographic methods.
Analytic and bootstrap estimates had similar coverage rates for 95% confidence intervals. With many sites per patient, bootstrap estimates had slightly better coverage than analytic estimates. Bootstrap percentile intervals had better coverage than asymptotic normal bootstrap intervals.
Bootstrapping is a useful method of estimating confidence intervals for the area under the receiver operating characteristic curve, sensitivity, and specificity when data are correlated.
Volumetric growth assessment of pulmonary lesions is crucial to both lung cancer screening and oncological therapy monitoring. While several methods for small pulmonary nodules have previously been presented, the segmentation of larger tumors that appear frequently in oncological patients and are more likely to be complexly interconnected with lung morphology has not yet received much attention. We present a fast, automated segmentation method that is based on morphological processing and is suitable for both small and large lesions. In addition, the proposed approach addresses clinical challenges to volume assessment such as variations in imaging protocol or inspiration state by introducing a method of segmentation-based partial volume analysis (SPVA) that follows on the segmentation procedure. Accuracy and reproducibility studies were performed to evaluate the new algorithms. In vivo interobserver and interscan studies on low-dose data from eight clinical metastasis patients revealed that clinically significant volume change can be detected reliably and with negligible computation time by the presented methods. In addition, phantom studies were conducted. Based on the segmentation performed with the proposed method, the performance of the SPVA volumetry method was compared with the conventional technique on a phantom that was scanned with different dosages and reconstructed with varying parameters. Both systematic and absolute errors were shown to be reduced substantially by the SPVA method. The method was especially successful in accounting for slice thickness and reconstruction kernel variations, where the median error was more than halved in comparison to the conventional approach.
Computed tomography (CT) screening is an important new tool for the early detection of lung cancer. In the current study, the authors assessed the discomfort associated with CT scanning and the subsequent wait for results and health-related quality of life (HRQoL) over time.
A total of 351 participants in the Dutch-Belgian randomized controlled trial for lung cancer screening in high-risk subjects (the NELSON trial) who had an appointment for a baseline CT scan were asked to complete questionnaires regarding their experienced discomfort and HRQoL before, 1 day after, and approximately 6 months after the CT scan. HRQoL was measured as generic HRQoL (12-item Short Form [SF-12] and EuroQol questionnaire [EQ-5D]), generic anxiety (State-Trait Anxiety Inventory [STAI-6]), and lung cancer-specific distress (Impact of Event Scale [IES]). Approximately 76.9% of the participants completed all 3 questionnaires.
Approximately 87% to 99% of participants reported experiencing no discomfort related to the CT scan. The median SF-12, EQ-5D, STAI-6, and IES scores did not appear to change relevantly over time. Approximately 46.0% and 51.3%, respectively, of the participants reported discomfort in connection with having to wait for the results of the CT scan and dreading those results. These patients had relevantly higher STAI-6 and IES scores (P < .01) (unfavorable) at all 3 assessments.
The current evaluation of the potential adverse effects of CT screening for lung cancer on HRQoL demonstrated no negative effects. However, waiting for the CT scan results was reported to be discomforting by approximately half of the participants. Minimizing the waiting time for the test results is therefore recommended.