Accuracy of artificial intelligence model for infectious keratitis classification: a systematic review and meta-analysis.

Background: Infectious keratitis (IK) is a sight-threatening condition requiring immediate definite treatment. The need for prompt treatment heavily depends on timely diagnosis. The diagnosis of IK, however, is challenged by the drawbacks of the current "gold standard." The poorly differentiated clinical features, the possibility of low microbial culture yield, and the duration for culture are the culprits of delayed IK treatment. Deep learning (DL) is a recent artificial intelligence (AI) advancement that has been demonstrated to be highly promising in making automated diagnosis in IK with high accuracy. However, its exact accuracy is not yet elucidated. This article is the first systematic review and meta-analysis that aims to assess the accuracy of available DL models to correctly classify IK based on etiology compared to the current gold standards. Methods: A systematic search was carried out in PubMed, Google Scholars, Proquest, ScienceDirect, Cochrane and Scopus. The used keywords are: "Keratitis," "Corneal ulcer," "Corneal diseases," "Corneal lesions," "Artificial intelligence," "Deep learning," and "Machine learning." Studies including slit lamp photography of the cornea and validity study on DL performance were considered. The primary outcomes reviewed were the accuracy and classification capability of the AI machine learning/DL algorithm. We analyzed the extracted data with the MetaXL 5.2 Software. Results: A total of eleven articles from 2002 to 2022 were included with a total dataset of 34,070 images. All studies used convolutional neural networks (CNNs), with ResNet and DenseNet models being the most used models across studies. Most AI models outperform the human counterparts with a pooled area under the curve (AUC) of 0.851 and accuracy of 96.6% in differentiating IK vs. non-IK and pooled AUC 0.895 and accuracy of 64.38% for classifying bacterial keratitis (BK) vs. fungal keratitis (FK). Conclusion: This study demonstrated that DL algorithms have high potential in diagnosing and classifying IK with accuracy that, if not better, is comparable to trained corneal experts. However, various factors, such as the unique architecture of DL model, the problem with overfitting, image quality of the datasets, and the complex nature of IK itself, still hamper the universal applicability of DL in daily clinical practice.

Deep Learning Based Centerline-Aggregated Aortic Hemodynamics: An Efficient Alternative to Numerical Modeling of Hemodynamics.

Image-based patient-specific modelling of hemodynamics are gaining increased popularity as a diagnosis and outcome prediction solution for a variety of cardiovascular diseases. While their potential to improve diagnostic capabilities and thereby clinical outcome is widely recognized, these methods require considerable computational resources since they are mostly based on conventional numerical methods such as computational fluid dynamics (CFD). As an alternative to the numerical methods, we propose a machine learning (ML) based approach to calculate patient-specific hemodynamic parameters. Compared to CFD based methods, our approach holds the benefit of being able to calculate a patient-specific hemodynamic outcome instantly with little need for computational power. In this proof-of-concept study, we present a deep artificial neural network (ANN) capable of computing hemodynamics for patients with aortic coarctation in a centerline aggregated (i.e., locally averaged) form. Considering the complex relation between vessels shape and hemodynamics on the one hand and the limited availability of suitable clinical data on the other, a sufficient accuracy of the ANN may however not be achieved with available data only. Another key aspect of this study is therefore the successful augmentation of available clinical data. Using a statistical shape model, additional training data was generated which substantially increased the ANN's accuracy, showcasing the ability of ML based methods to perform in-silico modelling tasks previously requiring resource intensive CFD simulations.

Deep Learning for Lung Cancer Detection on Screening CT Scans: Results of a Large-Scale Public Competition and an Observer Study with 11 Radiologists.

PURPOSE: To determine whether deep learning algorithms developed in a public competition could identify lung cancer on low-dose CT scans with a performance similar to that of radiologists. MATERIALS AND METHODS: In this retrospective study, a dataset consisting of 300 patient scans was used for model assessment; 150 patient scans were from the competition set and 150 were from an independent dataset. Both test datasets contained 50 cancer-positive scans and 100 cancer-negative scans. The reference standard was set by histopathologic examination for cancer-positive scans and imaging follow-up for at least 2 years for cancer-negative scans. The test datasets were applied to the three top-performing algorithms from the Kaggle Data Science Bowl 2017 public competition: grt123, Julian de Wit and Daniel Hammack (JWDH), and Aidence. Model outputs were compared with an observer study of 11 radiologists that assessed the same test datasets. Each scan was scored on a continuous scale by both the deep learning algorithms and the radiologists. Performance was measured using multireader, multicase receiver operating characteristic analysis. RESULTS: The area under the receiver operating characteristic curve (AUC) was 0.877 (95% CI: 0.842, 0.910) for grt123, 0.902 (95% CI: 0.871, 0.932) for JWDH, and 0.900 (95% CI: 0.870, 0.928) for Aidence. The average AUC of the radiologists was 0.917 (95% CI: 0.889, 0.945), which was significantly higher than grt123 (P = .02); however, no significant difference was found between the radiologists and JWDH (P = .29) or Aidence (P = .26). CONCLUSION: Deep learning algorithms developed in a public competition for lung cancer detection in low-dose CT scans reached performance close to that of radiologists.Keywords: Lung, CT, Thorax, Screening, Oncology Supplemental material is available for this article. © RSNA, 2021.

Robust classification from noisy labels: Integrating additional knowledge for chest radiography abnormality assessment.

Chest radiography is the most common radiographic examination performed in daily clinical practice for the detection of various heart and lung abnormalities. The large amount of data to be read and reported, with more than 100 studies per day for a single radiologist, poses a challenge in consistently maintaining high interpretation accuracy. The introduction of large-scale public datasets has led to a series of novel systems for automated abnormality classification. However, the labels of these datasets were obtained using natural language processed medical reports, yielding a large degree of label noise that can impact the performance. In this study, we propose novel training strategies that handle label noise from such suboptimal data. Prior label probabilities were measured on a subset of training data re-read by 4 board-certified radiologists and were used during training to increase the robustness of the training model to the label noise. Furthermore, we exploit the high comorbidity of abnormalities observed in chest radiography and incorporate this information to further reduce the impact of label noise. Additionally, anatomical knowledge is incorporated by training the system to predict lung and heart segmentation, as well as spatial knowledge labels. To deal with multiple datasets and images derived from various scanners that apply different post-processing techniques, we introduce a novel image normalization strategy. Experiments were performed on an extensive collection of 297,541 chest radiographs from 86,876 patients, leading to a state-of-the-art performance level for 17 abnormalities from 2 datasets. With an average AUC score of 0.880 across all abnormalities, our proposed training strategies can be used to significantly improve performance scores.

Deep Learning for Malignancy Risk Estimation of Pulmonary Nodules Detected at Low-Dose Screening CT.

Background Accurate estimation of the malignancy risk of pulmonary nodules at chest CT is crucial for optimizing management in lung cancer screening. Purpose To develop and validate a deep learning (DL) algorithm for malignancy risk estimation of pulmonary nodules detected at screening CT. Materials and Methods In this retrospective study, the DL algorithm was developed with 16 077 nodules (1249 malignant) collected -between 2002 and 2004 from the National Lung Screening Trial. External validation was performed in the following three -cohorts -collected between 2004 and 2010 from the Danish Lung Cancer Screening Trial: a full cohort containing all 883 nodules (65 -malignant) and two cancer-enriched cohorts with size matching (175 nodules, 59 malignant) and without size matching (177 -nodules, 59 malignant) of benign nodules selected at random. Algorithm performance was measured by using the area under the receiver operating characteristic curve (AUC) and compared with that of the Pan-Canadian Early Detection of Lung Cancer (PanCan) model in the full cohort and a group of 11 clinicians composed of four thoracic radiologists, five radiology residents, and two pulmonologists in the cancer-enriched cohorts. Results The DL algorithm significantly outperformed the PanCan model in the full cohort (AUC, 0.93 [95% CI: 0.89, 0.96] vs 0.90 [95% CI: 0.86, 0.93]; P = .046). The algorithm performed comparably to thoracic radiologists in cancer-enriched cohorts with both random benign nodules (AUC, 0.96 [95% CI: 0.93, 0.99] vs 0.90 [95% CI: 0.81, 0.98]; P = .11) and size-matched benign nodules (AUC, 0.86 [95% CI: 0.80, 0.91] vs 0.82 [95% CI: 0.74, 0.89]; P = .26). Conclusion The deep learning algorithm showed excellent performance, comparable to thoracic radiologists, for malignancy risk estimation of pulmonary nodules detected at screening CT. This algorithm has the potential to provide reliable and reproducible malignancy risk scores for clinicians, which may help optimize management in lung cancer screening. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Tammemägi in this issue.

No Surprises: Training Robust Lung Nodule Detection for Low-Dose CT Scans by Augmenting With Adversarial Attacks.

Detecting malignant pulmonary nodules at an early stage can allow medical interventions which may increase the survival rate of lung cancer patients. Using computer vision techniques to detect nodules can improve the sensitivity and the speed of interpreting chest CT for lung cancer screening. Many studies have used CNNs to detect nodule candidates. Though such approaches have been shown to outperform the conventional image processing based methods regarding the detection accuracy, CNNs are also known to be limited to generalize on under-represented samples in the training set and prone to imperceptible noise perturbations. Such limitations can not be easily addressed by scaling up the dataset or the models. In this work, we propose to add adversarial synthetic nodules and adversarial attack samples to the training data to improve the generalization and the robustness of the lung nodule detection systems. To generate hard examples of nodules from a differentiable nodule synthesizer, we use projected gradient descent (PGD) to search the latent code within a bounded neighbourhood that would generate nodules to decrease the detector response. To make the network more robust to unanticipated noise perturbations, we use PGD to search for noise patterns that can trigger the network to give over-confident mistakes. By evaluating on two different benchmark datasets containing consensus annotations from three radiologists, we show that the proposed techniques can improve the detection performance on real CT data. To understand the limitations of both the conventional networks and the proposed augmented networks, we also perform stress-tests on the false positive reduction networks by feeding different types of artificially produced patches. We show that the augmented networks are more robust to both under-represented nodules as well as resistant to noise perturbations.

Efficient organ localization using multi-label convolutional neural networks in thorax-abdomen CT scans.

Automatic localization of organs and other structures in medical images is an important preprocessing step that can improve and speed up other algorithms such as organ segmentation, lesion detection, and registration. This work presents an efficient method for simultaneous localization of multiple structures in 3D thorax-abdomen CT scans. Our approach predicts the location of multiple structures using a single multi-label convolutional neural network for each orthogonal view. Each network takes extra slices around the current slice as input to provide extra context. A sigmoid layer is used to perform multi-label classification. The output of the three networks is subsequently combined to compute a 3D bounding box for each structure. We used our approach to locate 11 structures of interest. The neural network was trained and evaluated on a large set of 1884 thorax-abdomen CT scans from patients undergoing oncological workup. Reference bounding boxes were annotated by human observers. The performance of our method was evaluated by computing the wall distance to the reference bounding boxes. The bounding boxes annotated by the first human observer were used as the reference standard for the test set. Using the best configuration, we obtained an average wall distance of [Formula: see text] mm in the test set. The second human observer achieved [Formula: see text] mm. For all structures, the results were better than those reported in previously published studies. In conclusion, we proposed an efficient method for the accurate localization of multiple organs. Our method uses multiple slices as input to provide more context around the slice under analysis, and we have shown that this improves performance. This method can easily be adapted to handle more organs.

A survey on deep learning in medical image analysis.

Deep learning algorithms, in particular convolutional networks, have rapidly become a methodology of choice for analyzing medical images. This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year. We survey the use of deep learning for image classification, object detection, segmentation, registration, and other tasks. Concise overviews are provided of studies per application area: neuro, retinal, pulmonary, digital pathology, breast, cardiac, abdominal, musculoskeletal. We end with a summary of the current state-of-the-art, a critical discussion of open challenges and directions for future research.

Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: The LUNA16 challenge.

Automatic detection of pulmonary nodules in thoracic computed tomography (CT) scans has been an active area of research for the last two decades. However, there have only been few studies that provide a comparative performance evaluation of different systems on a common database. We have therefore set up the LUNA16 challenge, an objective evaluation framework for automatic nodule detection algorithms using the largest publicly available reference database of chest CT scans, the LIDC-IDRI data set. In LUNA16, participants develop their algorithm and upload their predictions on 888 CT scans in one of the two tracks: 1) the complete nodule detection track where a complete CAD system should be developed, or 2) the false positive reduction track where a provided set of nodule candidates should be classified. This paper describes the setup of LUNA16 and presents the results of the challenge so far. Moreover, the impact of combining individual systems on the detection performance was also investigated. It was observed that the leading solutions employed convolutional networks and used the provided set of nodule candidates. The combination of these solutions achieved an excellent sensitivity of over 95% at fewer than 1.0 false positives per scan. This highlights the potential of combining algorithms to improve the detection performance. Our observer study with four expert readers has shown that the best system detects nodules that were missed by expert readers who originally annotated the LIDC-IDRI data. We released this set of additional nodules for further development of CAD systems.

Towards automatic pulmonary nodule management in lung cancer screening with deep learning.

The introduction of lung cancer screening programs will produce an unprecedented amount of chest CT scans in the near future, which radiologists will have to read in order to decide on a patient follow-up strategy. According to the current guidelines, the workup of screen-detected nodules strongly relies on nodule size and nodule type. In this paper, we present a deep learning system based on multi-stream multi-scale convolutional networks, which automatically classifies all nodule types relevant for nodule workup. The system processes raw CT data containing a nodule without the need for any additional information such as nodule segmentation or nodule size and learns a representation of 3D data by analyzing an arbitrary number of 2D views of a given nodule. The deep learning system was trained with data from the Italian MILD screening trial and validated on an independent set of data from the Danish DLCST screening trial. We analyze the advantage of processing nodules at multiple scales with a multi-stream convolutional network architecture, and we show that the proposed deep learning system achieves performance at classifying nodule type that surpasses the one of classical machine learning approaches and is within the inter-observer variability among four experienced human observers.

Using deep learning to segment breast and fibroglandular tissue in MRI volumes.

PURPOSE: Automated segmentation of breast and fibroglandular tissue (FGT) is required for various computer-aided applications of breast MRI. Traditional image analysis and computer vision techniques, such atlas, template matching, or, edge and surface detection, have been applied to solve this task. However, applicability of these methods is usually limited by the characteristics of the images used in the study datasets, while breast MRI varies with respect to the different MRI protocols used, in addition to the variability in breast shapes. All this variability, in addition to various MRI artifacts, makes it a challenging task to develop a robust breast and FGT segmentation method using traditional approaches. Therefore, in this study, we investigated the use of a deep-learning approach known as "U-net." MATERIALS AND METHODS: We used a dataset of 66 breast MRI's randomly selected from our scientific archive, which includes five different MRI acquisition protocols and breasts from four breast density categories in a balanced distribution. To prepare reference segmentations, we manually segmented breast and FGT for all images using an in-house developed workstation. We experimented with the application of U-net in two different ways for breast and FGT segmentation. In the first method, following the same pipeline used in traditional approaches, we trained two consecutive (2C) U-nets: first for segmenting the breast in the whole MRI volume and the second for segmenting FGT inside the segmented breast. In the second method, we used a single 3-class (3C) U-net, which performs both tasks simultaneously by segmenting the volume into three regions: nonbreast, fat inside the breast, and FGT inside the breast. For comparison, we applied two existing and published methods to our dataset: an atlas-based method and a sheetness-based method. We used Dice Similarity Coefficient (DSC) to measure the performances of the automated methods, with respect to the manual segmentations. Additionally, we computed Pearson's correlation between the breast density values computed based on manual and automated segmentations. RESULTS: The average DSC values for breast segmentation were 0.933, 0.944, 0.863, and 0.848 obtained from 3C U-net, 2C U-nets, atlas-based method, and sheetness-based method, respectively. The average DSC values for FGT segmentation obtained from 3C U-net, 2C U-nets, and atlas-based methods were 0.850, 0.811, and 0.671, respectively. The correlation between breast density values based on 3C U-net and manual segmentations was 0.974. This value was significantly higher than 0.957 as obtained from 2C U-nets (P < 0.0001, Steiger's Z-test with Bonferoni correction) and 0.938 as obtained from atlas-based method (P = 0.0016). CONCLUSIONS: In conclusion, we applied a deep-learning method, U-net, for segmenting breast and FGT in MRI in a dataset that includes a variety of MRI protocols and breast densities. Our results showed that U-net-based methods significantly outperformed the existing algorithms and resulted in significantly more accurate breast density computation.

Improving airway segmentation in computed tomography using leak detection with convolutional networks.

We propose a novel method to improve airway segmentation in thoracic computed tomography (CT) by detecting and removing leaks. Leak detection is formulated as a classification problem, in which a convolutional network (ConvNet) is trained in a supervised fashion to perform the classification task. In order to increase the segmented airway tree length, we take advantage of the fact that multiple segmentations can be extracted from a given airway segmentation algorithm by varying the parameters that influence the tree length and the amount of leaks. We propose a strategy in which the combination of these segmentations after removing leaks can increase the airway tree length while limiting the amount of leaks. This strategy therefore largely circumvents the need for parameter fine-tuning of a given airway segmentation algorithm. The ConvNet was trained and evaluated using a subset of inspiratory thoracic CT scans taken from the COPDGene study. Our method was validated on a separate independent set of the EXACT'09 challenge. We show that our method significantly improves the quality of a given leaky airway segmentation, achieving a higher sensitivity at a low false-positive rate compared to all the state-of-the-art methods that entered in EXACT09, and approaching the performance of the combination of all of them.

Pulmonary Nodule Detection in CT Images: False Positive Reduction Using Multi-View Convolutional Networks.

We propose a novel Computer-Aided Detection (CAD) system for pulmonary nodules using multi-view convolutional networks (ConvNets), for which discriminative features are automatically learnt from the training data. The network is fed with nodule candidates obtained by combining three candidate detectors specifically designed for solid, subsolid, and large nodules. For each candidate, a set of 2-D patches from differently oriented planes is extracted. The proposed architecture comprises multiple streams of 2-D ConvNets, for which the outputs are combined using a dedicated fusion method to get the final classification. Data augmentation and dropout are applied to avoid overfitting. On 888 scans of the publicly available LIDC-IDRI dataset, our method reaches high detection sensitivities of 85.4% and 90.1% at 1 and 4 false positives per scan, respectively. An additional evaluation on independent datasets from the ANODE09 challenge and DLCST is performed. We showed that the proposed multi-view ConvNets is highly suited to be used for false positive reduction of a CAD system.

Automatic detection of large pulmonary solid nodules in thoracic CT images.

PURPOSE: Current computer-aided detection (CAD) systems for pulmonary nodules in computed tomography (CT) scans have a good performance for relatively small nodules, but often fail to detect the much rarer larger nodules, which are more likely to be cancerous. We present a novel CAD system specifically designed to detect solid nodules larger than 10 mm. METHODS: The proposed detection pipeline is initiated by a three-dimensional lung segmentation algorithm optimized to include large nodules attached to the pleural wall via morphological processing. An additional preprocessing is used to mask out structures outside the pleural space to ensure that pleural and parenchymal nodules have a similar appearance. Next, nodule candidates are obtained via a multistage process of thresholding and morphological operations, to detect both larger and smaller candidates. After segmenting each candidate, a set of 24 features based on intensity, shape, blobness, and spatial context are computed. A radial basis support vector machine (SVM) classifier was used to classify nodule candidates, and performance was evaluated using ten-fold cross-validation on the full publicly available lung image database consortium database. RESULTS: The proposed CAD system reaches a sensitivity of 98.3% (234/238) and 94.1% (224/238) large nodules at an average of 4.0 and 1.0 false positives/scan, respectively. CONCLUSIONS: The authors conclude that the proposed dedicated CAD system for large pulmonary nodules can identify the vast majority of highly suspicious lesions in thoracic CT scans with a small number of false positives.