Application of an Artificial Neural Network to High-Resolution CT: Usefulness in Differential Diagnosis of Diffuse Lung Disease

Deep learning diagnostic and severity-stratification for interstitial lung diseases and chronic obstructive pulmonary disease in digital lung auscultations and ultrasonography: clinical protocol for an observational case-control study

BACKGROUND

Interstitial lung diseases (ILD), such as idiopathic pulmonary fibrosis (IPF) and non-specific interstitial pneumonia (NSIP), and chronic obstructive pulmonary disease (COPD) are severe, progressive pulmonary disorders with a poor prognosis. Prompt and accurate diagnosis is important to enable patients to receive appropriate care at the earliest possible stage to delay disease progression and prolong survival. Artificial intelligence-assisted lung auscultation and ultrasound (LUS) could constitute an alternative to conventional, subjective, operator-related methods for the accurate and earlier diagnosis of these diseases. This protocol describes the standardised collection of digitally-acquired lung sounds and LUS images of adult outpatients with IPF, NSIP or COPD and a deep learning diagnostic and severity-stratification approach.

METHODS

A total of 120 consecutive patients (≥ 18 years) meeting international criteria for IPF, NSIP or COPD and 40 age-matched controls will be recruited in a Swiss pulmonology outpatient clinic, starting from August 2022. At inclusion, demographic and clinical data will be collected. Lung auscultation will be recorded with a digital stethoscope at 10 thoracic sites in each patient and LUS images using a standard point-of-care device will be acquired at the same sites. A deep learning algorithm (DeepBreath) using convolutional neural networks, long short-term memory models, and transformer architectures will be trained on these audio recordings and LUS images to derive an automated diagnostic tool. The primary outcome is the diagnosis of ILD versus control subjects or COPD. Secondary outcomes are the clinical, functional and radiological characteristics of IPF, NSIP and COPD diagnosis. Quality of life will be measured with dedicated questionnaires. Based on previous work to distinguish normal and pathological lung sounds, we estimate to achieve convergence with an area under the receiver operating characteristic curve of > 80% using 40 patients in each category, yielding a sample size calculation of 80 ILD (40 IPF, 40 NSIP), 40 COPD, and 40 controls.

DISCUSSION

This approach has a broad potential to better guide care management by exploring the synergistic value of several point-of-care-tests for the automated detection and differential diagnosis of ILD and COPD and to estimate severity. Trial registration Registration: August 8, 2022.

CLINICALTRIALS

gov Identifier: NCT05318599.

The Added Effect of Artificial Intelligence on Physicians' Performance in Detecting Thoracic Pathologies on CT and Chest X-ray: A Systematic Review

Our systematic review investigated the additional effect of artificial intelligence-based devices on human observers when diagnosing and/or detecting thoracic pathologies using different diagnostic imaging modalities, such as chest X-ray and CT. Peer-reviewed, original research articles from EMBASE, PubMed, Cochrane library, SCOPUS, and Web of Science were retrieved. Included articles were published within the last 20 years and used a device based on artificial intelligence (AI) technology to detect or diagnose pulmonary findings. The AI-based device had to be used in an observer test where the performance of human observers with and without addition of the device was measured as sensitivity, specificity, accuracy, AUC, or time spent on image reading. A total of 38 studies were included for final assessment. The quality assessment tool for diagnostic accuracy studies (QUADAS-2) was used for bias assessment. The average sensitivity increased from 67.8% to 74.6%; specificity from 82.2% to 85.4%; accuracy from 75.4% to 81.7%; and Area Under the ROC Curve (AUC) from 0.75 to 0.80. Generally, a faster reading time was reported when radiologists were aided by AI-based devices. Our systematic review showed that performance generally improved for the physicians when assisted by AI-based devices compared to unaided interpretation.

Computer-Aided Diagnosis of Pulmonary Fibrosis Using Deep Learning and CT Images

OBJECTIVES

The objective of this study is to assess the performance of a computer-aided diagnosis (CAD) system (INTACT system) for the automatic classification of high-resolution computed tomography images into 4 radiological diagnostic categories and to compare this with the performance of radiologists on the same task.

MATERIALS AND METHODS

For the comparison, a total of 105 cases of pulmonary fibrosis were studied (54 cases of nonspecific interstitial pneumonia and 51 cases of usual interstitial pneumonia). All diagnoses were interstitial lung disease board consensus diagnoses (radiologically or histologically proven cases) and were retrospectively selected from our database. Two subspecialized chest radiologists made a consensual ground truth radiological diagnosis, according to the Fleischner Society recommendations. A comparison analysis was performed between the INTACT system and 2 other radiologists with different years of experience (readers 1 and 2). The INTACT system consists of a sequential pipeline in which first the anatomical structures of the lung are segmented, then the various types of pathological lung tissue are identified and characterized, and this information is then fed to a random forest classifier able to recommend a radiological diagnosis.

RESULTS

Reader 1, reader 2, and INTACT achieved similar accuracy for classifying pulmonary fibrosis into the original 4 categories: 0.6, 0.54, and 0.56, respectively, with P > 0.45. The INTACT system achieved an F-score (harmonic mean for precision and recall) of 0.56, whereas the 2 readers, on average, achieved 0.57 (P = 0.991). For the pooled classification (2 groups, with and without the need for biopsy), reader 1, reader 2, and CAD had similar accuracies of 0.81, 0.70, and 0.81, respectively. The F-score was again similar for the CAD system and the radiologists. The CAD system and the average reader reached F-scores of 0.80 and 0.79 (P = 0.898).

CONCLUSIONS

We found that a computer-aided detection algorithm based on machine learning was able to classify idiopathic pulmonary fibrosis with similar accuracy to a human reader.

Automatic inference model construction for computer-aided diagnosis of lung nodule: Explanation adequacy, inference accuracy, and experts' knowledge

We aimed to describe the development of an inference model for computer-aided diagnosis of lung nodules that could provide valid reasoning for any inferences, thereby improving the interpretability and performance of the system. An automatic construction method was used that considered explanation adequacy and inference accuracy. In addition, we evaluated the usefulness of prior experts’ (radiologists’) knowledge while constructing the models. In total, 179 patients with lung nodules were included and divided into 79 and 100 cases for training and test data, respectively. F-measure and accuracy were used to assess explanation adequacy and inference accuracy, respectively. For F-measure, reasons were defined as proper subsets of Evidence that had a strong influence on the inference result. The inference models were automatically constructed using the Bayesian network and Markov chain Monte Carlo methods, selecting only those models that met the predefined criteria. During model constructions, we examined the effect of including radiologist’s knowledge in the initial Bayesian network models. Performance of the best models in terms of F-measure, accuracy, and evaluation metric were as follows: 0.411, 72.0%, and 0.566, respectively, with prior knowledge, and 0.274, 65.0%, and 0.462, respectively, without prior knowledge. The best models with prior knowledge were then subjectively and independently evaluated by two radiologists using a 5-point scale, with 5, 3, and 1 representing beneficial, appropriate, and detrimental, respectively. The average scores by the two radiologists were 3.97 and 3.76 for the test data, indicating that the proposed computer-aided diagnosis system was acceptable to them. In conclusion, the proposed method incorporating radiologists’ knowledge could help in eliminating radiologists’ distrust of computer-aided diagnosis and improving its performance.

Computerized Classification of Pneumoconiosis on Digital Chest Radiography Artificial Neural Network with Three Stages

It is difficult for radiologists to classify pneumoconiosis from category 0 to category 3 on chest radiographs. Therefore, we have developed a computer-aided diagnosis (CAD) system based on a three-stage artificial neural network (ANN) method for classification based on four texture features. The image database consists of 36 chest radiographs classified as category 0 to category 3. Regions of interest (ROIs) with a matrix size of 32 × 32 were selected from chest radiographs. We obtained a gray-level histogram, histogram of gray-level difference, gray-level run-length matrix (GLRLM) feature image, and gray-level co-occurrence matrix (GLCOM) feature image in each ROI. For ROI-based classification, the first ANN was trained with each texture feature. Next, the second ANN was trained with output patterns obtained from the first ANN. Finally, we obtained a case-based classification for distinguishing among four categories with the third ANN method. We determined the performance of the third ANN by receiver operating characteristic (ROC) analysis. The areas under the ROC curve (AUC) of the highest category (severe pneumoconiosis) case and the lowest category (early pneumoconiosis) case were 0.89 ± 0.09 and 0.84 ± 0.12, respectively. The three-stage ANN with four texture features showed the highest performance for classification among the four categories. Our CAD system would be useful for assisting radiologists in classification of pneumoconiosis from category 0 to category 3.

A study of computer-aided diagnosis for pulmonary nodule: comparison between classification accuracies using calculated image features and imaging findings annotated by radiologists

PURPOSE

In our previous study, we developed a computer-aided diagnosis (CADx) system using imaging findings annotated by radiologists. The system, however, requires radiologists to input many imaging findings. In order to reduce such an interaction of radiologists, we further developed a CADx system using derived imaging findings based on calculated image features, in which the system only requires few user operations. The purpose of this study is to check whether calculated image features (CFT) or derived imaging findings (DFD) can represent information in imaging findings annotated by radiologists (AFD).

METHODS

We calculate 2282 image features and derive 39 imaging findings by using information on a nodule position and its type (solid or ground-glass). These image features are categorized into shape features, texture features and imaging findings-specific features. Each imaging finding is derived based on each corresponding classifier using random forest. To check whether CFT or DFD can represent information in AFD, under an assumption that the accuracies of classifiers are the same if information included in input is the same, we constructed classifiers by using various types of information (CTT, DFD and AFD) and compared accuracies on an inferred diagnosis of a nodule. We employ SVM with RBF kernel as classifier to infer a diagnosis name.

RESULTS

Accuracies of classifiers using DFD, CFT, AFD and CFT [Formula: see text] AFD were 0.613, 0.577, 0.773 and 0.790, respectively. Concordance rates between DFD and AFD of shape findings, texture findings and surrounding findings were 0.644, 0.871 and 0.768, respectively.

CONCLUSIONS

The results suggest that CFT and AFD are similar information and CFT represent only a portion of AFD. Particularly, CFT did not contain shape information in AFD. In order to decrease an interaction of radiologists, a development of a method which overcomes these problems is necessary.

Learning with distribution of optimized features for recognizing common CT imaging signs of lung diseases

Common CT imaging signs of lung diseases (CISLs) are defined as the imaging signs that frequently appear in lung CT images from patients. CISLs play important roles in the diagnosis of lung diseases. This paper proposes a novel learning method, namely learning with distribution of optimized feature (DOF), to effectively recognize the characteristics of CISLs. We improve the classification performance by learning the optimized features under different distributions. Specifically, we adopt the minimum spanning tree algorithm to capture the relationship between features and discriminant ability of features for selecting the most important features. To overcome the problem of various distributions in one CISL, we propose a hierarchical learning method. First, we use an unsupervised learning method to cluster samples into groups based on their distribution. Second, in each group, we use a supervised learning method to train a model based on their categories of CISLs. Finally, we obtain multiple classification decisions from multiple trained models and use majority voting to achieve the final decision. The proposed approach has been implemented on a set of 511 samples captured from human lung CT images and achieves a classification accuracy of 91.96%. The proposed DOF method is effective and can provide a useful tool for computer-aided diagnosis of lung diseases on CT images.

Potential clinical impact of advanced imaging and computer-aided diagnosis in chest radiology: importance of radiologist's role and successful observer study

This review paper is based on our research experience in the past 30 years. The importance of radiologists' role is discussed in the development or evaluation of new medical images and of computer-aided detection (CAD) schemes in chest radiology. The four main topics include (1) introducing what diseases can be included in a research database for different imaging techniques or CAD systems and what imaging database can be built by radiologists, (2) understanding how radiologists' subjective judgment can be combined with technical objective features to improve CAD performance, (3) sharing our experience in the design of successful observer performance studies, and (4) finally, discussing whether the new images and CAD systems can improve radiologists' diagnostic ability in chest radiology. In conclusion, advanced imaging techniques and detection/classification of CAD systems have a potential clinical impact on improvement of radiologists' diagnostic ability, for both the detection and the differential diagnosis of various lung diseases, in chest radiology.

Development of CAD based on ANN analysis of power spectra for pneumoconiosis in chest radiographs: effect of three new enhancement methods

We have been developing a computer-aided detection (CAD) scheme for pneumoconiosis based on a rule-based plus artificial neural network (ANN) analysis of power spectra. In this study, we have developed three enhancement methods for the abnormal patterns to reduce false-positive and false-negative values. The image database consisted of 2 normal and 15 abnormal chest radiographs. The International Labour Organization standard chest radiographs with pneumoconiosis were categorized as subcategory, size, and shape of pneumoconiosis. Regions of interest (ROIs) with a matrix size of 32 × 32 were selected from normal and abnormal lungs. Three new enhanced methods were obtained by window function, top-hat transformation, and gray-level co-occurrence matrix analysis. We calculated the power spectrum (PS) of all ROIs by Fourier transform. For the classification between normal and abnormal ROIs, we applied a combined analysis using the ruled-based plus the ANN method. To evaluate the overall performance of this CAD scheme, we employed ROC analysis for distinguishing between normal and abnormal ROIs. On the chest radiographs of the highest categories (severe pneumoconiosis) and the lowest categories (early pneumoconiosis), this CAD scheme achieved area under the curve (AUC) values of 0.93 ± 0.02 and 0.72 ± 0.03. The combined rule-based plus ANN method with the three new enhanced methods obtained the highest classification performance for distinguishing between abnormal and normal ROIs. Our CAD system based on the three new enhanced methods would be useful in assisting radiologists in the classification of pneumoconiosis.

[CT imaging of chronic interstitial lung diseases: from diagnosis to automated quantification]

Computed tomography is important for the diagnosis and follow-up of chronic diffuse interstitial lung diseases. Image quality has improved from each generation of scanner to the next and this continues to allow a better characterization of extent of pathology, or even the nature of the pathological process (potentially reversible inflammatory lesions compared to fibrotic lesions). The diagnostic imaging approach has evolved at the same time as technological developments. We initially thought in terms of the predominant lesions (nodular, alveolar consolidation, ground-glass opacity), and then moved to reasoning based on patterns, which are a combination of several elementary lesions (typically for the diagnosis of idiopathic pulmonary fibrosis). Nowadays, studies are focused on building models characterizing a specific disease and which combine several distinct patterns (typically for ground-glass opacity analysis). CT also allows a quantification of the extent of lung disease, which is linked to the prognosis of the disease and helps to monitor its progression. This quantification is usually based on visual criteria, the principles of which are summarized here. The development of automated quantification software could in the near future, be a support for the radiologist.

Computer-assisted detection of infectious lung diseases: a review

Respiratory tract infections are a leading cause of death and disability worldwide. Although radiology serves as a primary diagnostic method for assessing respiratory tract infections, visual analysis of chest radiographs and computed tomography (CT) scans is restricted by low specificity for causal infectious organisms and a limited capacity to assess severity and predict patient outcomes. These limitations suggest that computer-assisted detection (CAD) could make a valuable contribution to the management of respiratory tract infections by assisting in the early recognition of pulmonary parenchymal lesions, providing quantitative measures of disease severity and assessing the response to therapy. In this paper, we review the most common radiographic and CT features of respiratory tract infections, discuss the challenges of defining and measuring these disorders with CAD, and propose some strategies to address these challenges.

Neural network evaluation of PET scans of the liver: a potentially useful adjunct in clinical interpretation

PURPOSE

To assess the performance of an artificial neural network in the evaluation of fluorine 18 fluorodeoxyglucose (FDG) uptake in the liver, compared with the results of expert interpretation of abdominal liver magnetic resonance (MR) images.

MATERIALS AND METHODS

The study was approved by the institutional human research committee and was HIPAA compliant, with waiver of informed consent. Digital data from positron emission tomographic (PET)/computed tomographic (CT) examinations, along with patient demographics, were obtained from 98 consecutive patients who underwent both whole-body PET/CT examinations and liver MR imaging examinations within 2 months. Interpretations of the scans from PET/CT examinations by trained neural networks were cross-classified with expert interpretations of the findings on images from MR examinations for intrahepatic benignity or malignancy. Receiver operating characteristic (ROC) curves were obtained for the designed networks. The significance of the difference between neural network ROC curves and the ROC curves detailing the performance of two expert blinded observers in the interpretation of liver FDG uptake was determined.

RESULTS

A neural network incorporating lesion data demonstrated an ROC curve with an area under the curve (AUC) of 0.905 (standard error, 0.0370). A network independent of lesion data demonstrated an ROC curve with an AUC of 0.896 (standard error, 0.0386). These results compare favorably with results of expert blinded observers 1 and 2 who demonstrated ROCs with AUCs of 0.786 (standard error, 0.0522) and 0.796 (standard error, 0.0514), respectively. Following unblinding to network data, the AUCs for readers 1 and 2 improved to 0.924 (standard error, 0.0335) and 0.881 (standard error, 0.0409), respectively.

CONCLUSION

Computers running artificial neural networks employing PET/CT scan data are sensitive and specific in the designation of the presence of intrahepatic malignancy, with comparison with interpretation by expert observers. When used in conjunction with human expertise, network data improve accuracy of the human interpreter.

Esophageal stenosis associated with tumor regression in radiotherapy for esophageal cancer: frequency and prediction

PURPOSE

To determine clinical factors for predicting the frequency and severity of esophageal stenosis associated with tumor regression in radiotherapy for esophageal cancer.

METHODS AND MATERIALS

The study group consisted of 109 patients with esophageal cancer of T1-4 and Stage I-III who were treated with definitive radiotherapy and achieved a complete response of their primary lesion at Kyushu University Hospital between January 1998 and December 2007. Esophageal stenosis was evaluated using esophagographic images within 3 months after completion of radiotherapy. We investigated the correlation between esophageal stenosis after radiotherapy and each of the clinical factors with regard to tumors and therapy. For validation of the correlative factors for esophageal stenosis, an artificial neural network was used to predict the esophageal stenotic ratio.

RESULTS

Esophageal stenosis tended to be more severe and more frequent in T3-4 cases than in T1-2 cases. Esophageal stenosis in cases with full circumference involvement tended to be more severe and more frequent than that in cases without full circumference involvement. Increases in wall thickness tended to be associated with increases in esophageal stenosis severity and frequency. In the multivariate analysis, T stage, extent of involved circumference, and wall thickness of the tumor region were significantly correlated to esophageal stenosis (p = 0.031, p < 0.0001, and p = 0.0011, respectively). The esophageal stenotic ratio predicted by the artificial neural network, which learned these three factors, was significantly correlated to the actual observed stenotic ratio, with a correlation coefficient of 0.864 (p < 0.001).

CONCLUSION

Our study suggested that T stage, extent of involved circumference, and esophageal wall thickness of the tumor region were useful to predict the frequency and severity of esophageal stenosis associated with tumor regression in radiotherapy for esophageal cancer.

Computerized detection of diffuse lung disease in MDCT: the usefulness of statistical texture features

Accurate detection of diffuse lung disease is an important step for computerized diagnosis and quantification of this disease. It is also a difficult clinical task for radiologists. We developed a computerized scheme to assist radiologists in the detection of diffuse lung disease in multi-detector computed tomography (CT). Two radiologists selected 31 normal and 37 abnormal CT scans with ground glass opacity, reticular, honeycombing and nodular disease patterns based on clinical reports. The abnormal cases in our database must contain at least an abnormal area with a severity of moderate or severe level that was subjectively rated by the radiologists. Because statistical texture features may lack the power to distinguish a nodular pattern from a normal pattern, the abnormal cases that contain only a nodular pattern were excluded. The areas that included specific abnormal patterns in the selected CT images were then delineated as reference standards by an expert chest radiologist. The lungs were first segmented in each slice by use of a thresholding technique, and then divided into contiguous volumes of interest (VOIs) with a 64 x 64 x 64 matrix size. For each VOI, we determined and employed statistical texture features, such as run-length and co-occurrence matrix features, to distinguish abnormal from normal lung parenchyma. In particular, we developed new run-length texture features with clear physical meanings to considerably improve the accuracy of our detection scheme. A quadratic classifier was employed for distinguishing between normal and abnormal VOIs by the use of a leave-one-case-out validation scheme. A rule-based criterion was employed to further determine whether a case was normal or abnormal. We investigated the impact of new and conventional texture features, VOI size and the dimensionality for regions of interest on detecting diffuse lung disease. When we employed new texture features for 3D VOIs of 64 x 64 x 64 voxels, our system achieved the highest performance level: a sensitivity of 86% and a specificity of 90% for the detection of abnormal VOIs, and a sensitivity of 89% and a specificity of 90% for the detection of abnormal cases. Our computerized scheme would be useful for assisting radiologists in the diagnosis of diffuse lung disease.

Subjective similarity of patterns of diffuse interstitial lung disease on thin-section CT: an observer performance study

RATIONALE AND OBJECTIVES

The aim of this study was to investigate the subjective similarity for pairs of images with various abnormal patterns of diffuse interstitial lung disease on thin-section computed tomography by experienced radiologists to explore a basis for selecting similar images to assist radiologists' interpretation.

MATERIALS AND METHODS

Four major patterns (ground-glass opacity, nodular opacity, reticular opacity, and honeycombing) on thin-section computed tomographic images were identified by at least two of three radiologists. One radiologist manually selected 104 image pairs, in which the images in each pair had the same pattern and were similar in appearance. An additional 208 image pairs were randomly selected and evenly divided among the four patterns. These pairs were then rated for subjective similarity (on a continuous scale ranging from 0 = not similar at all to 1.0 = almost identical) by 12 radiologists.

RESULTS

For radiologist-selected pairs, the mean similarity rated by the 12 radiologists was 0.72. For randomly selected pairs, the mean similarity was higher for the same pattern (0.47) than for the varying patterns (0.27) (P < .001), and among the same pattern, the mean similarity was 0.63 for ground-glass opacity, 0.58 for honeycombing, 0.45 for nodular opacity, and 0.32 for reticular opacity. The mean standard deviation for similarity ratings on all pairs given by the 12 radiologists was 0.05 (rang, 0.01-0.09).

CONCLUSION

Subjective similarity ratings for pairs of abnormal images can be measured reliably and reproducibly by radiologists and will provide a basis for the selection of similar images to assist radiologists' interpretation.

Differentiation of common large sellar-suprasellar masses effect of artificial neural network on radiologists' diagnosis performance

RATIONALE AND OBJECTIVES

When pituitary adenoma, craniopharyngioma, and Rathke's cleft cyst grow in the sellar and suprasellar region, it is often difficult to differentiate among these three lesions on magnetic resonance (MR) images. The purpose of this study was to apply an artificial neural network (ANN) for differential diagnosis among these three lesions with MR images and retrospectively evaluate the effect of ANN output on radiologists' performance.

MATERIALS AND METHODS

Forty-three patients with sellar-suprasellar masses were studied. The ANN was designed to differentiate among pituitary adenoma, craniopharyngioma, and Rathke's cleft cyst by using patients' ages and nine MR image findings obtained by three neuroradiologists using a subjective rating scale. In the observer performance test, MR images were viewed by nine radiologists, including four neuroradiologists and five general radiologists, first without and then with ANN output. The radiologists' performance was evaluated using receiver-operating characteristic analysis with a continuous rating scale.

RESULTS

The ANN showed high performance in differentiation among the three lesions (area under the receiver-operating characteristic curve, 0.990). The average area under the curve for all radiologists for differentiation among the three lesions increased significantly from 0.910 to 0.985 (P = .0024) when they used the computer output. Areas under the curves for the general radiologists and neuroradiologists increased from 0.876 to 0.983 (P = .0083) and from 0.952 to 0.989 (P = .038), respectively.

CONCLUSION

In diagnostic performance for differentiation among pituitary macroadenoma, craniopharyngioma, and Rathke's cleft cyst with MR imaging, the ANN resulted in parity between neuroradiologists and general radiologists.

Computer-aided diagnosis for the differentiation of malignant from benign thyroid nodules on ultrasonography

RATIONALE AND OBJECTIVES

We sought to evaluate the diagnostic performance of an artificial neural network (ANN) and binary logistic regression (BLR) in differentiating malignant from benign thyroid nodules on ultrasonography.

MATERIALS AND METHODS

Two experienced radiologists, who were unaware of the histopathological diagnosis, analyzed ultrasonographic (US) features of 109 pathologically proven thyroid lesions (49 malignant and 60 benign) in 96 patients. Each radiologist was asked to evaluate US findings and categorize nodules into one of the two groups (malignant vs. benign) in each case. The following 8 US parameters were assessed for each nodule: size, shape, margin, echogenicity, cystic change, microcalcification, macrocalcification, and halo sign. Statistically significant US findings were obtained with backward stepwise logistic regression and were used for training and testing of the ANN and the BLR. The performance of the ANN and BLR was compared to that of the radiologists using receiver-operating characteristic (ROC) analysis.

RESULTS

Statistically significant US findings were size, margin, echogenicity, cystic change, and macrocalcification of the nodules. The area under the ROC curve (Az) values of ANN and BLR were 0.9492 +/- 0.0195 and 0.9046 +/- 0.0289, respectively. The Az value was 0.8300 +/- 0.0359 for reader 1 and 0.7600 +/- 0.0409 for reader 2. The Az values for ANN and BLR were significantly higher than those for both radiologists (all p < .05).

CONCLUSION

The performance of the ANN and the BLR was better than that of the radiologists in the distinction of benign and malignant thyroid nodules.

Performance evaluation of radiologists with artificial neural network for differential diagnosis of intra-axial cerebral tumors on MR images

BACKGROUND AND PURPOSE

Previous studies have suggested that use of an artificial neural network (ANN) system is beneficial for radiological diagnosis. Our purposes in this study were to construct an ANN for the differential diagnosis of intra-axial cerebral tumors on MR images and to evaluate the effect of ANN outputs on radiologists' diagnostic performance.

MATERIALS AND METHODS

We collected MR images of 126 patients with intra-axial cerebral tumors (58 high-grade gliomas, 37 low-grade gliomas, 19 metastatic tumors, and 12 malignant lymphomas). We constructed a single 3-layer feed-forward ANN with a Levenberg-Marquardt algorithm. The ANN was designed to differentiate among 4 categories of tumors (high-grade gliomas, low-grade gliomas, metastases, and malignant lymphomas) with use of 2 clinical parameters and 13 radiologic findings in MR images. Subjective ratings for the 13 radiologic findings were provided independently by 2 attending radiologists. All 126 cases were used for training and testing of the ANN based on a leave-one-out-by-case method. In the observer test, MR images were viewed by 9 radiologists, first without and then with ANN outputs. Each radiologist's performance was evaluated through a receiver operating characteristic (ROC) analysis on a continuous rating scale.

RESULTS

The averaged area under the ROC curve for ANN alone was 0.949. The diagnostic performance of the 9 radiologists increased from 0.899 to 0.946 (P < .001) when they used ANN outputs.

CONCLUSIONS

The ANN can provide useful output as a second opinion to improve radiologists' diagnostic performance in the differential diagnosis of intra-axial cerebral tumors seen on MR imaging.

Usefulness of artificial neural network for differential diagnosis of hepatic masses on CT images

RATIONALE AND OBJECTIVE

Our purpose in this study is to apply an artificial neural network (ANN) for differential diagnosis of certain hepatic masses on computed tomographic (CT) images and evaluate the effect of ANN output on radiologist diagnostic performance.

MATERIALS AND METHODS

We collected 120 cases of hepatic disease. We used a single three-layer feed-forward ANN with a back-propagation algorithm. The ANN is designed to differentiate four hepatic masses (hepatocellular carcinoma, intrahepatic peripheral cholangiocarcinoma, hemangioma, and metastasis) by using nine clinical parameters and 24 radiological findings in dual-phase contrast-enhanced CT images. Thus, the ANN consisted of 33 input units and four output units. Subjective ratings for the 24 radiological findings were provided independently by two attending radiologists. All clinical cases were used for training and testing of the ANN by implementation of a round-robin technique. In the observer test, CT images of all 120 cases (30 cases for each disease) were used. CT images were viewed by seven radiologists first without and then with ANN output. Radiologist performance was evaluated by using receiver operating characteristic (ROC) analysis on a continuous rating scale.

RESULTS

Averaged area under the ROC curve for ANN alone was 0.961. The diagnostic performance of seven radiologists increased from 0.888 to 0.934 (P < .02) when they used ANN output.

CONCLUSION

The ANN can provide useful output as a second opinion to improve radiologist diagnostic performance in the differential diagnosis of hepatic masses seen on contrast-enhanced CT.

Computer analysis of computed tomography scans of the lung: a survey

Current computed tomography (CT) technology allows for near isotropic, submillimeter resolution acquisition of the complete chest in a single breath hold. These thin-slice chest scans have become indispensable in thoracic radiology, but have also substantially increased the data load for radiologists. Automating the analysis of such data is, therefore, a necessity and this has created a rapidly developing research area in medical imaging. This paper presents a review of the literature on computer analysis of the lungs in CT scans and addresses segmentation of various pulmonary structures, registration of chest scans, and applications aimed at detection, classification and quantification of chest abnormalities. In addition, research trends and challenges are identified and directions for future research are discussed.

High-Resolution CT in Diagnosis of Diffuse Infiltrative Lung Disease

The chest radiograph remains the first imaging modality for the approach to diffuse infiltrative lung disease (DILD), but, 23 years after its introduction, high-resolution CT (HRCT) is still considered the best imaging tool for the evaluation of the pulmonary interstitium and to diagnose and assess DILD. The introduction of multidetector computed tomography (MDCT) has provided the thoracic radiologist with a powerful tool with which to image the lung. Moreover MDCT has enabled radiologists to understand better the functional information contained within CT images of DILD. By focusing on the HRCT signs, patterns, and distributions of abnormalities, and mentioning the clinical aspects and the new recent advances in pulmonary imaging, in this article we provide an overview of a practical approach to the interpretation of the DILD.

Computer-Aided Diagnosis in Thoracic CT

Computer-aided diagnosis (CAD) provides a computerized diagnostic result as a "second opinion" to assist radiologists in the diagnosis of various diseases by use of medical images. CAD has become a practical clinical approach in diagnostic radiology, although, at present, primarily in the area of detection of breast cancer in mammograms. Currently, a large research effort has been devoted to the detection and classification of various lung diseases in thoracic computed tomography (CT) images. We describe in this article the current status of the development of CAD schemes in thoracic CT, including nodule detection, distinction between benign and malignant nodules, and detection, characterization, and differential diagnosis of diffuse lung disease. Observer performance studies indicate that these CAD schemes would be useful in clinical practice by providing radiologists with computer output as a "second opinion."