Automated Lung Nodule Detection and Classification Using Deep Learning Combined with Multiple Strategies.

Pulmonary cancer is one of the major causes of deaths caused by cancer around the globe. Early stage lung cancer detection can prove to be essential for the patients, for which the computed tomography (CT) images are analyzed by the radiologists to determine the presence of nodules and diagnose the disease. Conventional techniques used by the radiologists for nodule detection in CT images is time-consuming and inefficient; to assist in the diagnosis process and further enhance its efficiency and accuracy, decision support systems have been developed in the past few years. In our paper, we proposed a Multi-Dimension Region-based Fully Convolutional Network based decision support system for detection and classification of lung nodule. The Multi-Dimension RFCN serves as an image classifier backbone for our feature extraction step in addition to the proposed Tri-Level Region Proposal Network (3L-RPN) along with the position-sensitive score maps (PSSM) being explored. A novel median intensity projection method is used to leverage the multi-dimensional information from CT images and introduced an additional deconvolutional layer to adopt the proposed Tri-Level Region Proposal Network in our architecture to automatically identify the potential Region of Interest. We trained and evaluated our proposed decision support system using LIDC-IDRI dataset. The evaluation results demonstrated the high level performance of our proposed model in comparison to the state-of-the-art nodule detection and classification methods by attaining classification accuracy of 97.61% and sensitivity of 97.4%.

Lung Cancer is one of the most dangerous threats for human life since many decades. When compared to other types of cancers, Lung cancer is very prevalent with less survival rate. Computed-Tomography-images are the main source of lung nodule detection as it provides detailed information about the nodules. Though there are many advanced techniques available in the present scenario like CNN, deep Neural Networks etc. For training purposes, all these existing methods use raw chest CT images which contain more complicated information like Blood vessels, muscles, lymph nodes, air ways, bones etc. Hence if the nodule is segmented effectively and that in turn used to train the system to classify cancerous and non-cancerous, then the results will be extraordinarily efficient. This paper presents the simple and efficient segmentation of Lung nodules and automatic classification of Nodules into Benign or Malignant using novel algorithms. The techniques used includes thresholding, morphological operations, pixels closures and mathematical model, for effective detection and classification of Nodules. The proposed algorithm is evaluated on axial plane CT Scans from LIDC-IDRI database, and the comparative study proved the effectiveness of segmentation and classification. The experimental result achieved the specificity of 98.67%, Accuracy of 97.98% and F-score of 1.0859 for β = 0.5.KeywordsComputed TomographyBenignMalignantSegmentationClassificationThresholdingMorphologyImage Processing

Automatic detection of lung nodules in CT datasets based on stable 3D mass–spring models

ObjectiveTo evaluate the relationship between CT dose and the performance of a computer-aided diagnosis (CAD) system, and to determine how best to minimize patient exposure to ionizing radiation while maintaining sufficient image quality for automated lung nodule detection, by the use of lung cancer screening CT.Materials and MethodsTwenty-five asymptomatic volunteers participated in the study. Each volunteer underwent a low-dose CT scan without contrast enhancement (multidetector CT with 16 detector rows, 1.25 mm section thickness, 120 kVp, beam pitch 1.35, 0.6 second rotation time, with 1.25 mm thickness reconstruction at 1.25 mm intervals) using four different amperages 32, 16, 8, and 4 mAs. All series were analyzed using a commercially available CAD system for automatic lung nodule detection and the results were reviewed by a consensus reading by two radiologists. The McNemar test and Kappa analysis were used to compare differences in terms of the abilities to detect pulmonary nodules.ResultsA total of 78 non-calcified true nodules were visualized in the 25 study subjects. The sensitivities for nodule detection were as follows: 72% at 32 mAs, 64% at 16 mAs, 59% at 8 mAs, and 40% at 4 mAs. Although the overall nodule-detecting performance was best at 32 mAs, no significant difference in nodule detectability was observed between scans at 16 mAs or 8 mAs versus 32 mAs. However, scans performed at 4 mAs were significantly inferior to those performed at 32 mAs (p < 0.001).ConclusionReducing the radiation dose (i.e. reducing the amperage) lowers lung nodule detectability by CAD. However, relatively low dose scans were found to be acceptable and to cause no significant reduction in nodule detectability versus usual low-dose CT.

Simple SummaryArtificial Intelligence (AI) algorithms can assist clinicians in their daily tasks by automatically detecting and/or classifying nodules in chest CT scans. Bias of such algorithms is one of the reasons why implementation of them in clinical practice is still not widely adopted. There is no published review on the bias that these algorithms may contain. This review aims to present different types of bias in such algorithms and present possible ways to mitigate them. Only then it would be possible to ensure that these algorithms work as intended under many different clinical settings.Artificial Intelligence (AI) algorithms for automatic lung nodule detection and classification can assist radiologists in their daily routine of chest CT evaluation. Even though many AI algorithms for these tasks have already been developed, their implementation in the clinical workflow is still largely lacking. Apart from the significant number of false-positive findings, one of the reasons for that is the bias that these algorithms may contain. In this review, different types of biases that may exist in chest CT AI nodule detection and classification algorithms are listed and discussed. Examples from the literature in which each type of bias occurs are presented, along with ways to mitigate these biases. Different types of biases can occur in chest CT AI algorithms for lung nodule detection and classification. Mitigation of them can be very difficult, if not impossible to achieve completely.

Cancer is known as one of the world’s top reason of mortality in human beings. Lung cancer, notably, has the highest mortality rate. Thus, timely detection of nodule or tumor is a critical and significant job in saving lives. One of the hot topic in current research field is automatic detection of lung nodules. Many methods have been implemented using computer vision-based technologies in the past, but achieving the desired precision still remains a difficult job. In this research, we adopt Convolutional Neural Network (CNN) based UNet image segmentation model and improved its architecture by incorporating convolution mechanisms. Moreover, this scheme uses binary cross entropy as loss function during training process. The proposed mechanism is tested on LIDC-IDRI dataset. The experimental analysis shows the augmented performance of proposed approach when compared with existing segmentation techniques. The qualitative and quantitative comparative analysis shows that the suggested scheme substantially improves the efficiency of segmentation performance.

The classification of benign and malignant lung nodules has great significance for the early detection of lung cancer, since early diagnosis of nodules can greatly increase patient survival. In this paper, we propose a novel classification method for lung nodules based on hybrid features from computed tomography (CT) images. The method fused 3D deep dual path network (DPN) features, local binary pattern (LBP)-based texture features and histogram of oriented gradients (HOG)-based shape features to characterize lung nodules. DPN is a convolutional neural network which integrates the advantages of aggregated residual transformations (ResNeXt) for feature reuse and a densely convolutional network (DenseNet) for exploring new features. LBP is a prominent feature descriptor for texture classification, when combining with the HOG descriptor, it can improve the classification performance considerably. To differentiate malignant nodules from benign ones, a gradient boosting machine (GBM) algorithm is employed. We evaluated the proposed method on the publicly available LUng Nodule Analysis 2016 (LUNA16) dataset with 1004 nodules, achieving an area under the receiver operating characteristic curve (AUC) of 0.9687 and accuracy of 93.78%. The promising results demonstrate that our method has strong robustness on the classification of nodule patterns by virtue of the joint use of texture features, shape features and 3D deep DPN features. The method has the potential to help radiologists to interpret diagnostic data and make decisions in clinical practice.

Accurate pulmonary nodule detection is a crucial step in lung cancer screening. Computer-aided detection (CAD) systems are not routinely used by radiologists for pulmonary nodule detection in clinical practice despite their potential benefits. Maximum intensity projection (MIP) images improve the detection of pulmonary nodules in radiological evaluation with computed tomography (CT) scans. Inspired by the clinical methodology of radiologists, we aim to explore the feasibility of applying MIP images to improve the effectiveness of automatic lung nodule detection using convolutional neural networks (CNNs). We propose a CNN-based approach that takes MIP images of different slab thicknesses (5 mm, 10 mm, 15 mm) and 1 mm axial section slices as input. Such an approach augments the two-dimensional (2-D) CT slice images with more representative spatial information that helps discriminate nodules from vessels through their morphologies. Our proposed method achieves sensitivity of 92.7% with 1 false positive per scan and sensitivity of 94.2% with 2 false positives per scan for lung nodule detection on 888 scans in the LIDC-IDRI dataset. The use of thick MIP images helps the detection of small pulmonary nodules (3 mm-10 mm) and results in fewer false positives. Experimental results show that utilizing MIP images can increase the sensitivity and lower the number of false positives, which demonstrates the effectiveness and significance of the proposed MIP-based CNNs framework for automatic pulmonary nodule detection in CT scans. The proposed method also shows the potential that CNNs could gain benefits for nodule detection by combining the clinical procedure.

NAS FD Lung: A novel lung assist diagnostic system based on neural architecture search

Early detection of lung cancer is an effective way to improve the survival rate of patients. It is a critical step to have accurate detection of lung nodules in computed tomography (CT) images for the diagnosis of lung cancer. However, due to the heterogeneity of the lung nodules and the complexity of the surrounding environment, it is a challenge to develop a robust nodule detection method. In this study, we propose a two-stage convolutional neural networks (TSCNN) for lung nodule detection. The first stage based on the improved U-Net segmentation network is to establish an initial detection of lung nodules. During this stage, in order to obtain a high recall rate without introducing excessive false positive nodules, we propose a new sampling strategy for training. Simultaneously, a two-phase prediction method is also proposed in this stage. The second stage in the TSCNN architecture based on the proposed dual pooling structure is built into three 3D-CNN classification networks for false positive reduction. Since the network training requires a significant amount of training data, we designed a random mask as the data augmentation method in this study. Furthermore, we have improved the generalization ability of the false positive reduction model by means of ensemble learning. We verified the proposed architecture on the LUNA dataset in our experiments, which showed that the proposed TSCNN architecture did obtain competitive detection performance.

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.

A survey and taxonomy of 2.5D approaches for lung segmentation and nodule detection in CT images

Adaptive morphology aided 2-pathway convolutional neural network for lung nodule classification

The main objective of the proposed work is to develop an automated CAD system for classification of lung nodules using various classifiers from CT images. The classification of nodule and non-nodule patterns in CT is one of the most significant processes during the detection of lung nodule. This helps in detecting the disease at early stage thereby decrease the mortality rate. The developed CAD systems consist of segmentation, feature extraction and classification. For Segmentation we used Fuzzy C Means (FCM) for effective extraction infected region. Later, we extracted features through First order statistics (FOS) and Second order statistics (SOS) and fed into classifiers like DS, RF and BPNN. The experimentation is conducted on LIDC-IDRI dataset and results outperforms well with BPNN when compare to other classifiers.KeywordsSegmentationClassificationCADFeature extraction

Deep learning-based methods have achieved promising performance in early\ndetection and classification of lung nodules, most of which discard unsure\nnodules and simply deal with a binary classification -- malignant vs benign.\nRecently, an unsure data model (UDM) was proposed to incorporate those unsure\nnodules by formulating this problem as an ordinal regression, showing better\nperformance over traditional binary classification. To further explore the\nordinal relationship for lung nodule classification, this paper proposes a meta\nordinal regression forest (MORF), which improves upon the state-of-the-art\nordinal regression method, deep ordinal regression forest (DORF), in three\nmajor ways. First, MORF can alleviate the biases of the predictions by making\nfull use of deep features while DORF needs to fix the composition of decision\ntrees before training. Second, MORF has a novel grouped feature selection (GFS)\nmodule to re-sample the split nodes of decision trees. Last, combined with GFS,\nMORF is equipped with a meta learning-based weighting scheme to map the\nfeatures selected by GFS to tree-wise weights while DORF assigns equal weights\nfor all trees. Experimental results on the LIDC-IDRI dataset demonstrate\nsuperior performance over existing methods, including the state-of-the-art\nDORF.\n