Automatic lung nodule detection using profile matching and back-propagation neural network techniques

Deep learning applied to automatic disease detection using chest X-rays

Deep learning (DL) has shown rapid advancement and considerable promise when applied to the automatic detection of diseases using CXRs. This is important given the widespread use of CXRs across the world in diagnosing significant pathologies, and the lack of trained radiologists to report them. This review article introduces the basic concepts of DL as applied to CXR image analysis including basic deep neural network (DNN) structure, the use of transfer learning and the application of data augmentation. It then reviews the current literature on how DNN models have been applied to the detection of common CXR abnormalities (e.g. lung nodules, pneumonia, tuberculosis and pneumothorax) over the last few years. This includes DL approaches employed for the classification of multiple different diseases (multi-class classification). Performance of different techniques and models and their comparison with human observers are presented. Some of the challenges facing DNN models, including their future implementation and relationships to radiologists, are also discussed.

A Modeling Method for Automatic Extraction of Offshore Aquaculture Zones Based on Semantic Segmentation

Monitoring of offshore aquaculture zones is important to marine ecological environment protection and maritime safety and security. Remote sensing technology has the advantages of large-area simultaneous observation and strong timeliness, which provide normalized monitoring of marine aquaculture zones. Aiming at the problems of weak generalization ability and low recognition rate in weak signal environments of traditional target recognition algorithm, this paper proposes a method for automatic extraction of offshore fish cage and floating raft aquaculture zones based on semantic segmentation. This method uses Generative Adversarial Networks to expand the data to compensate for the lack of training samples, and uses ratio of green band to red band (G/R) instead of red band to enhance the characteristics of aquaculture spectral information, combined with atrous convolution and atrous space pyramid pooling to enhance the context semantic information, to extract and identify two types of offshore fish cage zones and floating raft aquaculture zones. The experiment is carried out in the eastern coastal waters of Shandong Province, China, and the overall identification accuracy of the two types of aquaculture zones can reach 94.8%. The results show that the method proposed in this paper can realize high-precision extraction both of offshore fish cage and floating raft aquaculture zones.

Evolving the pulmonary nodules diagnosis from classical approaches to deep learning-aided decision support: three decades' development course and future prospect

PURPOSE

Lung cancer is the commonest cause of cancer deaths worldwide, and its mortality can be reduced significantly by performing early diagnosis and screening. Since the 1960s, driven by the pressing needs to accurately and effectively interpret the massive volume of chest images generated daily, computer-assisted diagnosis of pulmonary nodule has opened up new opportunities to relax the limitation from physicians' subjectivity, experiences and fatigue. And the fair access to the reliable and affordable computer-assisted diagnosis will fight the inequalities in incidence and mortality between populations. It has been witnessed that significant and remarkable advances have been achieved since the 1980s, and consistent endeavors have been exerted to deal with the grand challenges on how to accurately detect the pulmonary nodules with high sensitivity at low false-positive rate as well as on how to precisely differentiate between benign and malignant nodules. There is a lack of comprehensive examination of the techniques' development which is evolving the pulmonary nodules diagnosis from classical approaches to machine learning-assisted decision support. The main goal of this investigation is to provide a comprehensive state-of-the-art review of the computer-assisted nodules detection and benign-malignant classification techniques developed over three decades, which have evolved from the complicated ad hoc analysis pipeline of conventional approaches to the simplified seamlessly integrated deep learning techniques. This review also identifies challenges and highlights opportunities for future work in learning models, learning algorithms and enhancement schemes for bridging current state to future prospect and satisfying future demand.

CONCLUSION

It is the first literature review of the past 30 years' development in computer-assisted diagnosis of lung nodules. The challenges indentified and the research opportunities highlighted in this survey are significant for bridging current state to future prospect and satisfying future demand. The values of multifaceted driving forces and multidisciplinary researches are acknowledged that will make the computer-assisted diagnosis of pulmonary nodules enter into the main stream of clinical medicine and raise the state-of-the-art clinical applications as well as increase both welfares of physicians and patients. We firmly hold the vision that fair access to the reliable, faithful, and affordable computer-assisted diagnosis for early cancer diagnosis would fight the inequalities in incidence and mortality between populations, and save more lives.

An overview of deep learning in medical imaging focusing on MRI

What has happened in machine learning lately, and what does it mean for the future of medical image analysis? Machine learning has witnessed a tremendous amount of attention over the last few years. The current boom started around 2009 when so-called deep artificial neural networks began outperforming other established models on a number of important benchmarks. Deep neural networks are now the state-of-the-art machine learning models across a variety of areas, from image analysis to natural language processing, and widely deployed in academia and industry. These developments have a huge potential for medical imaging technology, medical data analysis, medical diagnostics and healthcare in general, slowly being realized. We provide a short overview of recent advances and some associated challenges in machine learning applied to medical image processing and image analysis. As this has become a very broad and fast expanding field we will not survey the entire landscape of applications, but put particular focus on deep learning in MRI. Our aim is threefold: (i) give a brief introduction to deep learning with pointers to core references; (ii) indicate how deep learning has been applied to the entire MRI processing chain, from acquisition to image retrieval, from segmentation to disease prediction; (iii) provide a starting point for people interested in experimenting and perhaps contributing to the field of deep learning for medical imaging by pointing out good educational resources, state-of-the-art open-source code, and interesting sources of data and problems related medical imaging.

Computer-aided diagnosis systems for lung cancer: challenges and methodologies

This paper overviews one of the most important, interesting, and challenging problems in oncology, the problem of lung cancer diagnosis. Developing an effective computer-aided diagnosis (CAD) system for lung cancer is of great clinical importance and can increase the patient's chance of survival. For this reason, CAD systems for lung cancer have been investigated in a huge number of research studies. A typical CAD system for lung cancer diagnosis is composed of four main processing steps: segmentation of the lung fields, detection of nodules inside the lung fields, segmentation of the detected nodules, and diagnosis of the nodules as benign or malignant. This paper overviews the current state-of-the-art techniques that have been developed to implement each of these CAD processing steps. For each technique, various aspects of technical issues, implemented methodologies, training and testing databases, and validation methods, as well as achieved performances, are described. In addition, the paper addresses several challenges that researchers face in each implementation step and outlines the strengths and drawbacks of the existing approaches for lung cancer CAD systems.

Development and evaluation of a computer-aided diagnostic scheme for lung nodule detection in chest radiographs by means of two-stage nodule enhancement with support vector classification

PURPOSE

To develop a computer-aided detection (CADe) scheme for nodules in chest radiographs (CXRs) with a high sensitivity and a low false-positive (FP) rate.

METHODS

The authors developed a CADe scheme consisting of five major steps, which were developed for improving the overall performance of CADe schemes. First, to segment the lung fields accurately, the authors developed a multisegment active shape model. Then, a two-stage nodule-enhancement technique was developed for improving the conspicuity of nodules. Initial nodule candidates were detected and segmented by using the clustering watershed algorithm. Thirty-one shape-, gray-level-, surface-, and gradient-based features were extracted from each segmented candidate for determining the feature space, including one of the new features based on the Canny edge detector to eliminate a major FP source caused by rib crossings. Finally, a nonlinear support vector machine (SVM) with a Gaussian kernel was employed for classification of the nodule candidates.

RESULTS

To evaluate and compare the scheme to other published CADe schemes, the authors used a publicly available database containing 140 nodules in 140 CXRs and 93 normal CXRs. The CADe scheme based on the SVM classifier achieved sensitivities of 78.6% (110/140) and 71.4% (100/140) with averages of 5.0 (1165/233) FPs/image and 2.0 (466/233) FPs/image, respectively, in a leave-one-out cross-validation test, whereas the CADe scheme based on a linear discriminant analysis classifier had a sensitivity of 60.7% (85/140) at an FP rate of 5.0 FPs/image. For nodules classified as "very subtle" and "extremely subtle," a sensitivity of 57.1% (24/42) was achieved at an FP rate of 5.0 FPs/image. When the authors used a database developed at the University of Chicago, the sensitivities was 83.3% (40/48) and 77.1% (37/48) at an FP rate of 5.0 (240/48) FPs/image and 2.0 (96/48) FPs/image, respectively.

CONCLUSIONS

These results compare favorably to those described for other commercial and non-commercial CADe nodule detection systems.

Biplane correlation imaging: a feasibility study based on phantom and human data

The objective of this study was to implement and evaluate the performance of a biplane correlation imaging (BCI) technique aimed to reduce the effect of anatomic noise and improve the detection of lung nodules in chest radiographs. Seventy-one low-dose posterior-anterior images were acquired from an anthropomorphic chest phantom with 0.28° angular separations over a range of ±10° along the vertical axis within an 11 s interval. Similar data were acquired from 19 human subjects with institutional review board approval and informed consent. The data were incorporated into a computer-aided detection (CAD) algorithm in which suspect lesions were identified by examining the geometrical correlation of the detected signals that remained relatively constant against variable anatomic backgrounds. The data were analyzed to determine the effect of angular separation, and the overall sensitivity and false-positives for lung nodule detection. The best performance was achieved for angular separations of the projection pairs greater than 5°. Within that range, the technique provided an order of magnitude decrease in the number of false-positive reports when compared with CAD analysis of single-view images. Overall, the technique yielded ~1.1 false-positive per patient with an average sensitivity of 75%. The results indicated that the incorporation of angular information can offer a reduction in the number of false-positives without a notable reduction in sensitivity. The findings suggest that the BCI technique has the potential for clinical implementation as a cost-effective technique to improve the detection of subtle lung nodules with lowered rate of false-positives.

Multiprojection correlation imaging for improved detection of pulmonary nodules

OBJECTIVE

The purpose of this study was the development and preliminary evaluation of multiprojection correlation imaging with 3D computer-aided detection (CAD) on chest radiographs for cost- and dose-effective improvement of early detection of pulmonary nodules.

SUBJECTS AND METHODS

Digital chest radiographs of 10 configurations of a chest phantom and of seven human subjects were acquired in multiple angular projections with an acquisition time of 11 seconds (single breath-hold) and total exposure comparable with that of a posteroanterior chest radiograph. An initial 2D CAD algorithm with two difference-of-gaussians filters and multilevel thresholds was developed with an independent database of 44 single-view chest radiographs with confirmed lesions. This 2D CAD algorithm was used on each projection image to find likely suspect nodules. The CAD outputs were reconstructed in 3D, reinforcing signals associated with true nodules while simultaneously decreasing false-positive findings produced by overlapping anatomic features. The performance of correlation imaging was tested on two to 15 projection images.

RESULTS

Optimum performance of correlation imaging was attained when nine projection images were used. Compared with conventional, single-view CAD, correlation imaging decreased as much as 79% the frequency of false-positive findings in phantom cases at a sensitivity level of 65%. The corresponding reduction in false-positive findings in the cases of human subjects was 78%.

CONCLUSION

Although limited by a relatively simple CAD implementation and a small number of cases, the findings suggest that correlation imaging performs substantially better than single-view CAD and may greatly enhance identification of subtle solitary pulmonary nodules on chest radiographs.

Computer-aided diagnostic scheme for the detection of lung nodules on chest radiographs: localized search method based on anatomical classification

We developed an advanced computer-aided diagnostic (CAD) scheme for the detection of various types of lung nodules on chest radiographs intended for implementation in clinical situations. We used 924 digitized chest images (992 noncalcified nodules) which had a 500 x 500 matrix size with a 1024 gray scale. The images were divided randomly into two sets which were used for training and testing of the computerized scheme. In this scheme, the lung field was first segmented by use of a ribcage detection technique, and then a large search area (448 x 448 matrix size) within the chest image was automatically determined by taking into account the locations of a midline and a top edge of the segmented ribcage. In order to detect lung nodule candidates based on a localized search method, we divided the entire search area into 7 x 7 regions of interest (ROIs: 64 x 64 matrix size). In the next step, each ROI was classified anatomically into apical, peripheral, hilar, and diaphragm/heart regions by use of its image features. Identification of lung nodule candidates and extraction of image features were applied for each localized region (128 x 128 matrix size), each having its central part (64 x 64 matrix size) located at a position corresponding to a ROI that was classified anatomically in the previous step. Initial candidates were identified by use of the nodule-enhanced image obtained with the average radial-gradient filtering technique, in which the filter size was varied adaptively depending on the location and the anatomical classification of the ROI. We extracted 57 image features from the original and nodule-enhanced images based on geometric, gray-level, background structure, and edge-gradient features. In addition, 14 image features were obtained from the corresponding locations in the contralateral subtraction image. A total of 71 image features were employed for three sequential artificial neural networks (ANNs) in order to reduce the number of false-positive candidates. All parameters for ANNs, i.e., the number of iterations, slope of sigmoid functions, learning rate, and threshold values for removing the false positives, were determined automatically by use of a bootstrap technique with training cases. We employed four different combinations of training and test image data sets which was selected randomly from the 924 cases. By use of our localized search method based on anatomical classification, the average sensitivity was increased to 92.5% with 59.3 false positives per image at the level of initial detection for four different sets of test cases, whereas our previous technique achieved an 82.8% of sensitivity with 56.8 false positives per image. The computer performance in the final step obtained from four different data sets indicated that the average sensitivity in detecting lung nodules was 70.1% with 5.0 false positives per image for testing cases and 70.4% sensitivity with 4.2 false positives per image for training cases. The advanced CAD scheme involving the localized search method with anatomical classification provided improved detection of pulmonary nodules on chest radiographs for 924 lung nodule cases.

Computer-aided diagnosis for the detection and classification of lung cancers on chest radiographs ROC analysis of radiologists' performance

RATIONALE AND OBJECTIVES

The aim of the study is to investigate the effect of a computer-aided diagnostic (CAD) scheme on radiologist performance in the detection of lung cancers on chest radiographs.

MATERIALS AND METHODS

We combined two independent CAD schemes for the detection and classification of lung nodules into one new CAD scheme by use of a database of 150 chest images, including 108 cases with solitary pulmonary nodules and 42 cases without nodules. For the observer study, we selected 48 chest images, including 24 lung cancers, 12 benign nodules, and 12 cases without nodules, from the database to investigate radiologist performance in the detection of lung cancers. Nine radiologists participated in a receiver operating characteristic (ROC) study in which cases were interpreted first without and then with computer output, which indicated locations of possible lung nodules, together with a five-color scale illustrating the computer-estimated likelihood of malignancy of the detected nodules.

RESULTS

Performance of the CAD scheme indicated that sensitivity in detecting lung nodules was 80.6%, with 1.2 false-positive results per image, and sensitivity and specificity for classification of nodules by use of the same database for training and testing the CAD scheme were 87.7% and 66.7%, respectively. Average area under the ROC curve value for detection of lung cancers improved significantly (P = .008) from without (0.724) to with CAD (0.778).

CONCLUSION

This type of CAD scheme, which includes two functions, namely detection and classification, can improve radiologist accuracy in the diagnosis of lung cancer.

Quantitative nodule detection in low dose chest CT scans: new template modeling and evaluation for CAD system design

Automatic diagnosis of lung nodules for early detection of lung cancer is the goal of a number of screening studies worldwide. With the improvements in resolution and scanning time of low dose chest CT scanners, nodule detection and identification is continuously improving. In this paper we describe the latest improvements introduced by our group in automatic detection of lung nodules. We introduce a new template for nodule detection using level sets which describes various physical nodules irrespective of shape, size and distribution of gray levels. The template parameters are estimated automatically from the segmented data (after the first two steps of our CAD system for automatic nodule detection) - no a priori learning of the parameters density function is needed. We show quantitatively that this template modeling approach drastically reduces the number of false positives in the nodule detection (the third step of our CAD system for automatic nodule detection), thus improving the overall accuracy of CAD systems. We compare the performance of this approach with other approaches in the literature and with respect to human experts. The impact of the new template model includes: 1) flexibility with respect to nodule topology - thus various nodules can be detected simultaneously by the same technique; 2) automatic parameter estimation of the nodule models using the gray level information of the segmented data; and 3) the ability to provide exhaustive search for all the possible nodules in the scan without excessive processing time - this provides an enhanced accuracy of the CAD system without increase in the overall diagnosis time.

Local contralateral subtraction based on bilateral symmetry of lung for reduction of false positives in computerized detection of pulmonary nodules

A novel method called local contralateral subtraction has been developed for the removal of normal anatomic structures in chest radiographs based on the symmetry between the left and right lung regions. The method was oriented to the reduction of false positives reported by a computer-aided diagnosis (CAD) scheme for detection of lung nodules in chest radiographs. In our method, two regions of interest (ROIs) are extracted, one from the position where a nodule candidate is located, and the other from the anatomically corresponding location in the opposite lung, which contains similar normal structures. A wavelet-based, multiresolution image registration method is employed for matching the two ROIs, and subtraction is performed. If no structure remains in the subtracted ROI, then the original ROI is identified as negative (i.e., it contains only normal structures); otherwise, it is regarded as positive (i.e., it contains a nodule). A measure that quantifies the remaining structures was developed to distinguish between nodules and false positives. Application of the method to clinical chest radiographs showed that it was effective in eliminating normal anatomic structures and reducing the number of false detections in the CAD scheme for detection of lung nodules.

Effect of high sensitivity in a computerized scheme for detecting extremely subtle solitary pulmonary nodules in chest radiographs

RATIONALES AND OBJECTIVES

This study investigated the effect of a high sensitivity in computer-aided diagnosis (CAD) for detecting lung nodules in chest radiographs when extremely subtle cases were presented to radiologists.

MATERIAL AND METHODS

The chest radiographs used in this study consisted of 36 normal images and 54 abnormals containing solitary lung nodules, of which 25 were extremely subtle and 29 were very subtle. Receiver operating characteristic analysis for detecting lung nodules was performed without and with CAD. The levels of CAD output were simulated with a hypothetical ideal performance of 100% sensitivity, but with three or four false positives per image. Six radiologists participated in an observer study in which cases were interpreted first without and then with the use of CAD.

RESULTS

The average A(z) values for radiologists without and with CAD were 0.682 and 0.808, respectively. The performance of radiologists was improved significantly when high sensitivity was used (P = .0003). However, the radiologists were not able to recognize some extremely subtle nodules (5 of 54 nodules by all radiologists), even with the correct CAD output; these nodules were then considered as non-actionable. None of 306 computer-false positives was incorrectly regarded as a nodule by all radiologists, but 63 false positives were incorrectly identified by one or more radiologists.

CONCLUSION

The accuracy of radiologists in the detection of some extremely subtle solitary pulmonary nodules can be improved significantly when the sensitivity of a CAD scheme can be made to be at an extremely high level. However, all of the six radiologists failed to identify some nodules (about 10%), even with the correct output of the CAD.

Patient information extraction in digitized radiography

Digital imagery is gradually replacing the traditional radiograph with the development of digital radiography and film scanner. This report presents a new method to extract the patient information number (PIN) field automatically from the film-scanned image using image analysis technique. To evaluate the PIN field extraction algorithm, 2 formats of label acquired from 2 different hospitals are tested. Given the available films with no constraints on the way the labels are written and positioned, the correct extraction rates are 73% and 84%, respectively. This extracted PIN information can link with Radiology Information System (RIS) or Hospital Information System (HIS), and the image scanned from the film then can be filed into the database automatically. The efficiency this method offers can simplify greatly the image filing process and improve the user friendliness of the overall image digitization system. Moreover, compared with the bar code reader, it solves the automatic information input problem in a very economical way. The authors believe the success of this technique will benefit the development of the PACS (Picture Archiving and Communication System) and teleradiology.

Computerized classification of suspicious regions in chest radiographs using subregion Hotelling observers

We propose to investigate the use of subregion Hotelling observers (SRHOs) in conjunction with perceptrons for the computerized classification of suspicious regions in chest radiographs for being nodules requiring follow up. Previously, 239 regions of interest (ROIs), each containing a suspicious lesion with proven classification, were collected. We chose to investigate the use of SRHOs as part of a multilayer classifier to determine the presence of a nodule. Each SRHO incorporates information about signal, background, and noise correlation for classification. For this study, 225 separate Hotelling observers were set up in a grid across each ROI. Each separate observer discriminates an 8 by 8 pixel area. A round robin sampling scheme was used to generate the 225 features, where each feature is the output of the individual observers. These features were then rank ordered by the magnitude of the weights of a perceptron. Once rank ordered, subsets of increasing number of features were selected to be used in another perceptron. This perceptron was trained to minimize mean squared error and the output was a continuous variable representing the likelihood of the region being a nodule. Performance was evaluated by receiver operating characteristic (ROC) analysis and reported as the area under the curve (Az). The classifier was optimized by adding additional features until the Az declined. The optimized subset of observers then were combined using a third perceptron. A subset of 80 features was selected which gave an Az of 0.972. Additionally, at 98.6% sensitivity, the classifier had a specificity of 71.3% and increased the positive predictive value from 60.7% to 84.1 %. Preliminary results suggest that using SRHOs in combination with perceptrons can provide a successful classification scheme for pulmonary nodules. This approach could be incorporated into a larger computer aided detection system for decreasing false positives.

Computer-aided diagnosis in chest radiography: a survey

The traditional chest radiograph is still ubiquitous in clinical practice, and will likely remain so for quite some time. Yet, its interpretation is notoriously difficult. This explains the continued interest in computer-aided diagnosis for chest radiography. The purpose of this survey is to categorize and briefly review the literature on computer analysis of chest images, which comprises over 150 papers published in the last 30 years. Remaining challenges are indicated and some directions for future research are given.

Computerized analysis of the likelihood of malignancy in solitary pulmonary nodules with use of artificial neural networks

PURPOSE

To develop a computer-aided diagnostic scheme by using an artificial neural network (ANN) to assist radiologists in the distinction of benign and malignant pulmonary nodules.

MATERIALS AND METHODS

Fifty-six chest radiographs of 34 primary lung cancers and 22 benign nodules were digitized with a 0.175-mm pixel size and a 10-bit gray scale. Eight subjective image features were evaluated and recorded by radiologists in each case. A computerized method was developed to extract objective features that could be correlated with the subjective features. An ANN was used to distinguish benign from malignant nodules on the basis of subjective or objective features. The performance of the ANN was compared with that of the radiologists by means of receiver operating characteristic (ROC) analysis.

RESULTS

Performance of the ANN was considerably greater with objective features (area under the ROC curve, Az = 0.854) than with subjective features (Az = 0.761). Performance of the ANN was also greater than that of the radiologists (Az = 0.752).

CONCLUSION

The computerized scheme has the potential to improve the diagnostic accuracy of radiologists in the distinction of benign and malignant solitary pulmonary nodules.

Automated localisation of the optic disc, fovea, and retinal blood vessels from digital colour fundus images

AIM

To recognise automatically the main components of the fundus on digital colour images.

METHODS

The main features of a fundus retinal image were defined as the optic disc, fovea, and blood vessels. Methods are described for their automatic recognition and location. 112 retinal images were preprocessed via adaptive, local, contrast enhancement. The optic discs were located by identifying the area with the highest variation in intensity of adjacent pixels. Blood vessels were identified by means of a multilayer perceptron neural net, for which the inputs were derived from a principal component analysis (PCA) of the image and edge detection of the first component of PCA. The foveas were identified using matching correlation together with characteristics typical of a fovea-for example, darkest area in the neighbourhood of the optic disc. The main components of the image were identified by an experienced ophthalmologist for comparison with computerised methods.

RESULTS

The sensitivity and specificity of the recognition of each retinal main component was as follows: 99.1% and 99.1% for the optic disc; 83.3% and 91.0% for blood vessels; 80.4% and 99.1% for the fovea.

CONCLUSIONS

In this study the optic disc, blood vessels, and fovea were accurately detected. The identification of the normal components of the retinal image will aid the future detection of diseases in these regions. In diabetic retinopathy, for example, an image could be analysed for retinopathy with reference to sight threatening complications such as disc neovascularisation, vascular changes, or foveal exudation.

Artificial neural networks in chest radiography: application to the differential diagnosis of interstitial lung disease

RATIONALE AND OBJECTIVES

The authors evaluated the usefulness of artificial neural networks (ANNs) in the differential diagnosis of interstitial lung disease.

MATERIALS AND METHODS

The authors used three-layer, feed-forward ANNs with a back-propagation algorithm. The ANNs were designed to distinguish between 11 interstitial lung diseases on the basis of 10 clinical parameters and 16 radiologic findings extracted by chest radiologists. Thus, the ANNs consisted of 26 input units and 11 output units. One hundred fifty actual clinical cases, 110 cases from previously published articles, and 110 hypothetical cases were used for training and testing the ANNs by using a round-robin (or leave-one-out) technique. ANN performance was evaluated with receiver operating characteristic (ROC) analysis.

RESULTS

The Az (area under the ROC curve) obtained with actual clinical cases was 0.947, and both the sensitivity and specificity of the ANNs were approximately 90% in terms of indicating the correct diagnosis with the two largest output values among the 11 diseases.

CONCLUSION

ANNs using clinical parameters and radiologic findings may be useful for making the differential diagnosis of interstitial lung disease on chest radiographs.

Characteristics of regions suspicious for pulmonary nodules at chest radiography

RATIONALE AND OBJECTIVES

This study was performed to determine physical characteristics of areas on chest radiographs that are suspicious but not definitive for the presence of a pulmonary nodule and the characteristics of areas that contain an obvious nodule.

MATERIALS AND METHODS

Two groups of patients were identified: those who had an area at plain radiography that was suspicious for a pulmonary nodule and underwent fluoroscopy for further evaluation (138 patients, 142 areas) and those who had an obvious nodule at plain radiography who underwent computed tomography for further evaluation (72 patients, 97 areas). The measured characteristics of the region of interest included size, circularity, compactness, contrast, and location.

RESULTS

A comparison of the data show that while there was some difference between these groups of patients with regard to location of the nodules, there were essentially no differences with regard to size, circularity, compactness, and contrast of the regions of interest.

CONCLUSION

Size, circularity, compactness, contrast, and location are not sufficient to distinguish pulmonary nodules from other suspicious regions on the chest radiograph.

Understanding artificial neural networks and exploring their potential applications for the practicing urologist

Artificial neural networks (ANNs) are complex mathematical models that are distantly based on the human neuronal structure. They are capable of modeling elaborate biologic systems without making assumptions based on statistical distributions. Preliminary work has been reported on their application in urology. The initial results have been promising, particularly as an additional tool in the detection of early prostate cancer using the ProstAsure Index, which has been the most extensively studied urologic ANN to date. We review the basic concepts behind ANNs and examine currently existing and potential future applications of this new dynamic technology both in urology and in general clinical medicine.

Automated segmentation of anatomic regions in chest radiographs using an adaptive-sized hybrid neural network

The purposes of this research are to investigate the effectiveness of our novel image features for segmentation of anatomic regions such as the lungs and the mediastinum in chest radiographs and to develop an automatic computerized method for image processing. A total of 85 screening chest radiographs from Johns Hopkins University Hospital were digitized to 2 K by 2.5 K pixels with 12 bit gray scale. To reduce the amount of information, the images were smoothed and subsampled to 256 by 310 pixels with 8 bit. The determination approach consists of classifying each pixel into two anatomic classes (lung and others) on the basis of several image features: (1) relative pixel address (Rx, Ry) based on lung edges extracted through image processing using profile, (2) density normalized from lungs and mediastinum density, and (3) histogram equalized entropy. The combinations of image features were evaluated using an adaptive-sized hybrid neural network consisting of an input, a hidden, and an output layer. Fourteen images were used for the training of the neural network and the remaining 71 images for testing. Using four features of relative address (Rx, Ry), normalized density, and histogram equalized entropy, the neural networks classified lungs at 92% accuracy against test images following the same rules as for the training images.

Predicting outcome of anterior temporal lobectomy using simulated neural networks

PURPOSE

Anterior temporal lobectomy (ATL) is an important option for treatment of medically refractory seizures. Patient selection is not always clear-cut, and there is inherent morbidity and mortality associated with the invasive and expensive surgical protocols. To determine whether patient selection might be facilitated by application of artificial intelligence, we developed a model that predicted seizure outcome after ATL, using a simulated neural network (SNN).

METHODS

Predictions of the model were compared with predictions derived from conventional discriminant function analysis. Neural networks and discriminant functions were devised that would predict the occurrence of both Class 1 outcomes (totally seizure-free), and Class 1 or Class 2 outcomes (nearly or totally seizure-free), using data from 87 patients from three surgical centers. The SNNs and discriminant functions were developed using data from a randomly selected subsample of 65 patients, and both models were cross-validated, using the remaining 22 patients.

RESULTS

The discriminant functions showed overall predictive accuracy of 78.5% and 72.7%, while the neural networks demonstrated overall accuracy of 81.8% and 95.4%.

CONCLUSIONS

Simulated neural networks show promise as adjuncts to decision-making in the selection of epilepsy surgery patients.

Neural networks analysis of astrocytic gliomas from MRI appearances

A three-layered backpropagation neural network was developed to differentiate malignant from benign brain tumors in a group of patients with astrocytic gliomas. The MRI findings of 43 patients were reviewed before biopsy by three neuroradiologists independently. This provided a database made up of 129 patients' records each of which comprised 13 parameters derived from pre- and post-contrast MR images. The network's generalizing ability was then tested to predict the outcome of biopsy in 36 new cases and its performance compared to that of radiologist using ROC analysis. The output of the network with and without radiologists' impression yielded a better diagnostic performance with relative ROC areas of 0.94 and 0.91, respectively; compared to 0.84 obtained by radiologist. These results demonstrate that the neural network can effectively differentiate malignant from benign brain tumors.

Neural network based on adaptive resonance theory as compared to experts in suggesting treatment for schizophrenic and unipolar depressed in-patients

A modified neural network based on adaptive resonance theory (ART) was trained with the records of 211 psychiatric inpatients (74 schizophrenic, 50 unipolar depressed, 34 bipolar depressed, 20 bipolar manic, 33 other) who improved by at least 40 points on the GAFS during 8 weeks of treatment. Thereafter, a comparison was made between the clinical response of another 26 schizophrenic patients and 28 unipolar depressed inpatients, to treatment suggested by the trained ART (N = 21) and by the consensus of two senior psychiatrists (N = 33). The patients were allocated blindly and randomly to the two treatment groups. The BPRS (for the schizophrenic patients) or the HDRS (for the unipolar depressed patients) was completed weekly for 5 weeks. Results showed no difference between decisions regarding treatment by the ART network and by the experts. Length of hospital stay was also similar. All ART suggestions included supportive psychotherapy. High potency antipsychotics were suggested for 7 schizophrenic inpatients, clozapine for one and the addition of community therapy for another. Depressed patients got a variety of treatment suggestions. No contraindicated treatment was suggested by ART; however, two incomplete treatment suggestions were dropped from the study. In conclusion, in a prospective study ART was successful in learning treatment strategies and performed under supervision similar to experts.

Automatic detection of diabetic retinopathy using an artificial neural network: a screening tool

AIMS

To determine if neural networks can detect diabetic features in fundus images and compare the network against an ophthalmologist screening a set of fundus images.

METHODS

147 diabetic and 32 normal images were captured from a fundus camera, stored on computer, and analysed using a back propagation neural network. The network was trained to recognise features in the retinal image. The effects of digital filtering techniques and different network variables were assessed. 200 diabetic and 101 normal images were then randomised and used to evaluate the network's performance for the detection of diabetic retinopathy against an ophthalmologist.

RESULTS

Detection rates for the recognition of vessels, exudates, and haemorrhages were 91.7%, 93.1%, and 73.8% respectively. When compared with the results of the ophthalmologist, the network achieved a sensitivity of 88.4% and a specificity of 83.5% for the detection of diabetic retinopathy.

CONCLUSIONS

Detection of vessels, exudates, and haemorrhages was possible, with success rates dependent upon preprocessing and the number of images used in training. When compared with the ophthalmologist, the network achieved good accuracy for the detection of diabetic retinopathy. The system could be used as an aid to the screening of diabetic patients for retinopathy.

Reduction of false positives in lung nodule detection using a two-level neural classification

The authors have developed a neural-digital computer-aided diagnosis system, based on a parameterized two-level convolution neural network (CNN) architecture and on a special multilabel output encoding procedure. The developed architecture was trained, tested, and evaluated specifically on the problem of diagnosis of lung cancer nodules found on digitized chest radiographs. The system performs automatic "suspect" localization, feature extraction, and diagnosis of a particular pattern-class aimed at a high degree of "true-positive fraction" detection and low "false-positive fraction" detection. In this paper, the authors aim at the presentation of the two-level neural classification method in reducing false-positives in their system. They employed receiver operating characteristics (ROC) method with the area under the ROC curve (A(z)) as the performance index to evaluate all the simulation results. The two-level CNN showed superior performance (A(z)=0.93) to the single-level CNN (A(z)=0.85). The proposed two-level CNN architecture is proven to be promising and to be extensible, problem-independent, and therefore, applicable to other medical or difficult diagnostic tasks in two-dimensional (2-D) image environments.

Classification of mass and normal breast tissue: a convolution neural network classifier with spatial domain and texture images

The authors investigated the classification of regions of interest (ROI's) on mammograms as either mass or normal tissue using a convolution neural network (CNN). A CNN is a backpropagation neural network with two-dimensional (2-D) weight kernels that operate on images. A generalized, fast and stable implementation of the CNN was developed. The input images to the CNN were obtained from the ROI's using two techniques. The first technique employed averaging and subsampling. The second technique employed texture feature extraction methods applied to small subregions inside the ROI. Features computed over different subregions were arranged as texture images, which were subsequently used as CNN inputs. The effects of CNN architecture and texture feature parameters on classification accuracy were studied. Receiver operating characteristic (ROC) methodology was used to evaluate the classification accuracy. A data set consisting of 168 ROIs containing biopsy-proven masses and 504 ROI's containing normal breast tissue was extracted from 168 mammograms by radiologists experienced in mammography. This data set was used for training and testing the CNN. With the best combination of CNN architecture and texture feature parameters, the area under the test ROC curve reached 0.87, which corresponded to a true-positive fraction of 90% at a false positive fraction of 31%. The authors' results demonstrate the feasibility of using a CNN for classification of masses and normal tissue on mammograms.

Detection of lung nodules in digital chest radiographs using artificial neural networks: a pilot study

Radiologists can fail to detect up to 30% of pulmonary nodules in chest radiographs. A back-propagation neural network was used to detect lung nodules in digital chest radiographs to assist radiologists in the diagnosis of lung cancer. Regions of interest (ROIs) that contained nodules and normal tissues in the lung were selected from digitized chest radiographs by a previously developed computer-aided diagnosis (CAD) scheme. Different preprocessing techniques were used to produce input data to the neural network. The performance of the neural network was evaluated by receiver operating characteristic (ROC) analysis. We found that subsampling of original 64- x 64-pixel ROIs to smaller 8- x 8-pixel ROIs provides the optimal preprocessing for the neural network to distinguish ROIs containing nodules from false-positive ROIs containing normal regions. The neural network was able to detect obvious nodules very well with an Az value (area under ROC curve) of 0.93, but was unable to detect subtle nodules. However, with a training method that uses different orientations of the original ROIs, we were able to improve the performance of the neural network to detect subtle nodules. Artificial neural networks have the potential to serve as a useful classifier to help to eliminate the false-positive detections of the CAD scheme.

Teleradiology/telepathology requirements and implementation

Teleradiology and telepathology form an integral part of the telemedicine concept. Teleradiology is becoming a mature technology because of advances in imaging technology, database design and communications infrastructure and capabilities. Telepathology has also made significant progress but more development is needed in the definition of required images, database design and standards. While the requirements of most clinical applications of teleradiology are well established, telemammography still presents some impediments. Technical difficulties in telemammography are presented in terms of the lack of a clinically accepted digital imaging system and large data volume required per image. Another important aspect in tele-imaging is the database question. Workstations constitute a window into database. Comprehensive database development is the most difficult and expensive technology for tele-imaging and operational features of such systems are discussed. Finally, we explore current examples of the use of telepathology and teleradiology in the global telemedicine context.

Reduction of false positives in computerized detection of lung nodules in chest radiographs using artificial neural networks, discriminant analysis, and a rule-based scheme

A computer-aided diagnosis (CAD) scheme is being developed to identify image regions considered suspicious for lung nodules in chest radiographs to assist radiologists in making correct diagnoses. Automated classifiers--an artificial neural network, discriminant analysis, and a rule-based scheme--are used to reduce the number of false-positive detections of the CAD scheme. The CAD scheme first detects nodule candidates from chest radiographs based on a difference image technique. Nine image features characterizing nodules are extracted automatically for each of the nodule candidates. The extracted image features are then used as input data to the classifiers for distinguishing actual nodules from the false-positive detections. The performances of the classifiers are evaluated by receiver-operating characteristic analysis. On the basis of the database of 30 normal and 30 abnormal chest images, the neural network achieves an AZ value (area under the receiver-operating-characteristic curve) of 0.79 in detecting lung nodules, as tested by the round-robin method. The neural network, after being trained with a training database, is able to eliminate more than 83% of the false-positive detections reported by the CAD scheme. Moreover, the combination of the trained neural network and a rule-based scheme eliminates 96% of the false-positive detections of the CAD scheme.

Simulated neural networks to predict outcomes, costs, and length of stay among orthopedic rehabilitation patients

Our purpose was to develop a set of simulated neural networks that would predict functional outcomes, length of stay, and costs among orthopedic patients admitted to an inpatient rehabilitation hospital. We used retrospective data for a sample of 387 patients between the ages of 60 and 89 who had been admitted to a single rehabilitation facility over a period of 12 months. Using age and data on functional capacity at admission from the Functional Independence Measure, we were successful in constructing networks that were 86%, 87%, and 91% accurate in predicting functional outcome, length of stay, and costs to within +/- 15% of the actual value. In each case the accuracy of the network exceeded that of a multiple regression equation using the same variables. Our results show the feasibility of using simulated neural networks to predict rehabilitation outcomes, and the advantages of neural networks over conventional linear models. Networks of this kind may be of significant value to administrators and clinicians in predicting outcomes and resource usage as rehabilitation hospitals are faced with capitation and prospective payment schemes.