Artificial intelligence evaluation of coronary computed tomography angiography for coronary stenosis classification and diagnosis.

BACKGROUND: Ruling out obstructive coronary artery disease (CAD) using coronary computed tomography angiography (CCTA) is time-consuming and challenging. This study developed a deep learning (DL) model to assist in detecting obstructive CAD on CCTA to streamline workflows. METHODS: In total, 2929 DICOM files and 7945 labels were extracted from curved planar reformatted CCTA images. A modified Inception V3 model was adopted. To validate the artificial intelligence (AI) model, two cardiologists labelled and adjudicated the classification of coronary stenosis on CCTA. The model was trained to differentiate the coronary artery into binary stenosis classifications <50% and ≥50% stenosis. Using the quantitative coronary angiography (QCA) consensus results as a reference standard, the performance of the AI model and CCTA radiology readers was compared by calculating Cohen's kappa coefficients at patient and vessel levels. The net reclassification index was used to evaluate the net benefit of the DL model. RESULTS: The diagnostic accuracy of the AI model was 92.3% and 88.4% at the patient and vessel levels, respectively. Compared with CCTA radiology readers, the AI model had a better agreement for binary stenosis classification at both patient and vessel levels (Cohen kappa coefficient: .79 vs. .39 and .77 vs. .40, p < .0001). The AI model also exhibited significantly improved model discrimination and reclassification (Net reclassification index = .350; Z = 4.194; p < .001). CONCLUSIONS: The developed AI model identified obstructive CAD, and the model results correlated well with QCA results. Incorporating the model into the reporting system of CCTA may improve workflows.

Combined model integrating deep learning, radiomics, and clinical data to classify lung nodules at chest CT.

OBJECTIVES: The study aimed to develop a combined model that integrates deep learning (DL), radiomics, and clinical data to classify lung nodules into benign or malignant categories, and to further classify lung nodules into different pathological subtypes and Lung Imaging Reporting and Data System (Lung-RADS) scores. MATERIALS AND METHODS: The proposed model was trained, validated, and tested using three datasets: one public dataset, the Lung Nodule Analysis 2016 (LUNA16) Grand challenge dataset (n = 1004), and two private datasets, the Lung Nodule Received Operation (LNOP) dataset (n = 1027) and the Lung Nodule in Health Examination (LNHE) dataset (n = 1525). The proposed model used a stacked ensemble model by employing a machine learning (ML) approach with an AutoGluon-Tabular classifier. The input variables were modified 3D convolutional neural network (CNN) features, radiomics features, and clinical features. Three classification tasks were performed: Task 1: Classification of lung nodules into benign or malignant in the LUNA16 dataset; Task 2: Classification of lung nodules into different pathological subtypes; and Task 3: Classification of Lung-RADS score. Classification performance was determined based on accuracy, recall, precision, and F1-score. Ten-fold cross-validation was applied to each task. RESULTS: The proposed model achieved high accuracy in classifying lung nodules into benign or malignant categories in LUNA 16 with an accuracy of 92.8%, as well as in classifying lung nodules into different pathological subtypes with an F1-score of 75.5% and Lung-RADS scores with an F1-score of 80.4%. CONCLUSION: Our proposed model provides an accurate classification of lung nodules based on the benign/malignant, different pathological subtypes, and Lung-RADS system.

A YOLO-based AI system for classifying calcifications on spot magnification mammograms.

OBJECTIVES: Use of an AI system based on deep learning to investigate whether the system can aid in distinguishing malignant from benign calcifications on spot magnification mammograms, thus potentially reducing unnecessary biopsies. METHODS: In this retrospective study, we included public and in-house datasets with annotations for the calcifications on both craniocaudal and mediolateral oblique vies, or both craniocaudal and mediolateral views of each case of mammograms. All the lesions had pathological results for correlation. Our system comprised an algorithm based on You Only Look Once (YOLO) named adaptive multiscale decision fusion module. The algorithm was pre-trained on a public dataset, Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM), then re-trained and tested on the in-house dataset of spot magnification mammograms. The performance of the system was investigated by receiver operating characteristic (ROC) analysis. RESULTS: We included 1872 images from 753 calcification cases (414 benign and 339 malignant) from CBIS-DDSM. From the in-house dataset, 636 cases (432 benign and 204 malignant) with 1269 spot magnification mammograms were included, with all lesions being recommended for biopsy by radiologists. The area under the ROC curve for our system on the in-house testing dataset was 0.888 (95% CI 0.868-0.908), with a sensitivity of 88.4% (95% CI 86.9-8.99%), specificity of 80.8% (95% CI 77.6-84%), and an accuracy of 84.6% (95% CI 81.8-87.4%) at the optimal cutoff value. Using the system with two views of spot magnification mammograms, 80.8% benign biopsies could be avoided. CONCLUSION: The AI system showed good accuracy for classification of calcifications on spot magnification mammograms which were all categorized as suspicious by radiologists, thereby potentially reducing unnecessary biopsies.

A PCB Alignment System Using RST Template Matching with CUDA on Embedded GPU Board.

The fiducial-marks-based alignment process is one of the most critical steps in printed circuit board (PCB) manufacturing. In the alignment process, a machine vision technique is used to detect the fiducial marks and then adjust the position of the vision system in such a way that it is aligned with the PCB. The present study proposed an embedded PCB alignment system, in which a rotation, scale and translation (RST) template-matching algorithm was employed to locate the marks on the PCB surface. The coordinates and angles of the detected marks were then compared with the reference values which were set by users, and the difference between them was used to adjust the position of the vision system accordingly. To improve the positioning accuracy, the angle and location matching process was performed in refinement processes. To overcome the matching time, in the present study we accelerated the rotation matching by eliminating the weak features in the scanning process and converting the normalized cross correlation (NCC) formula to a sum of products. Moreover, the scanning time was reduced by implementing the entire RST process in parallel on threads of a graphics processing unit (GPU) by applying hash functions to find refined positions in the refinement matching process. The experimental results showed that the resulting matching time was around 32× faster than that achieved on a conventional central processing unit (CPU) for a test image size of 1280 × 960 pixels. Furthermore, the precision of the alignment process achieved a considerable result with a tolerance of 36.4µm.

Resistin as a Biomarker for the Prediction of Left Atrial Substrate and Recurrence in Patients with Drug-Refractory Atrial Fibrillation Undergoing Catheter Ablation.

Resistin is an adipocytokine that is abundantly secreted from lipid cells and is related to the inflammatory process and cardiometabolic diseases. This study aimed to examine the role of resistin on inflammation and its effect on the clinical outcome of patients with atrial fibrillation (AF) following catheter ablation.A total of 108 patients (56.9 ± 12.0 years, 76.8% male) with symptomatic and drug-refractory AF undergoing catheter ablation were enrolled. Inflammatory biomarkers and epicardial fat volume by contrast computed tomography (CT) images were assessed in all patients before the procedure. Baseline resistin correlated with epicardial fat volume, tumor necrosis factor-α (TNF-α), and left atrial (LA) scar area. After the index procedure, the univariate analysis revealed that hypertension, persistent AF, LA diameter, and plasma resistin level were related to recurrent atrial arrhythmia. Multivariate regression analysis revealed that persistent AF, LA diameter, and plasma resistin level all independently predicted recurrent atrial arrhythmia after ablation. Plasma resistin with a level higher than 777 (pg/mL) could predict recurrence following catheter ablation of AF.High plasma resistin level is associated with poor left atrial substrate, high epicardial fat volume, and elevated TNF-α level in patients with AF. Plasma resistin may predict the recurrence of atrial arrhythmia after ablation.

Development of a modified 3D region proposal network for lung nodule detection in computed tomography scans: a secondary analysis of lung nodule datasets.

BACKGROUND: Low-dose computed tomography (LDCT) has been shown useful in early lung cancer detection. This study aimed to develop a novel deep learning model for detecting pulmonary nodules on chest LDCT images. METHODS: In this secondary analysis, three lung nodule datasets, including Lung Nodule Analysis 2016 (LUNA16), Lung Nodule Received Operation (LNOP), and Lung Nodule in Health Examination (LNHE), were used to train and test deep learning models. The 3D region proposal network (RPN) was modified via a series of pruning experiments for better predictive performance. The performance of each modified deep leaning model was evaluated based on sensitivity and competition performance metric (CPM). Furthermore, the performance of the modified 3D RPN trained on three datasets was evaluated by 10-fold cross validation. Temporal validation was conducted to assess the reliability of the modified 3D RPN for detecting lung nodules. RESULTS: The results of pruning experiments indicated that the modified 3D RPN composed of the Cross Stage Partial Network (CSPNet) approach to Residual Network (ResNet) Xt (CSP-ResNeXt) module, feature pyramid network (FPN), nearest anchor method, and post-processing masking, had the optimal predictive performance with a CPM of 92.2%. The modified 3D RPN trained on the LUNA16 dataset had the highest CPM (90.1%), followed by the LNOP dataset (CPM: 74.1%) and the LNHE dataset (CPM: 70.2%). When the modified 3D RPN trained and tested on the same datasets, the sensitivities were 94.6%, 84.8%, and 79.7% for LUNA16, LNOP, and LNHE, respectively. The temporal validation analysis revealed that the modified 3D RPN tested on LNOP test set achieved a CPM of 71.6% and a sensitivity of 85.7%, and the modified 3D RPN tested on LNHE test set had a CPM of 71.7% and a sensitivity of 83.5%. CONCLUSION: A modified 3D RPN for detecting lung nodules on LDCT scans was designed and validated, which may serve as a computer-aided diagnosis system to facilitate lung nodule detection and lung cancer diagnosis.

A modified 3D RPN for detecting lung nodules on CT images that exhibited greater sensitivity and CPM than did several previously reported CAD detection models was established.

Expression of UCA1 and MALAT1 long-chain non-coding RNAs in esophageal squamous cell carcinoma tissues is predictive of patient prognosis.

INTRODUCTION: The long non-coding RNAs (lncRNAs) urothelial cancer associated 1 (UCA1) and metastasis associated lung adenocarcinoma transcript 1 (MALAT1) are known to impact cancer cell regulation. The aim of the present study was to determine the relationship between the expression of these lncRNAs in esophageal squamous cell carcinoma (ESCC) tissues and disease prognosis. MATERIAL AND METHODS: The expression of UCA1 and MALAT1 lncRNAs was assessed in ESCC and adjacent carcinoma tissues (5 cm away from the tumor) and evaluated in relation to overall survival (OS) and disease-free survival (DFS) of patients. This prospective study included 100 ESCC patients who were admitted to the First Hospital of Yulin City between January 2007 and January 2014. RESULTS: The expression levels of UCA1 and MALAT1 lncRNAs in ESCC tissues were significantly higher than those in adjacent carcinoma tissues, and there were statistically significant differences in TNM staging between the patients with high lncRNA expression and low lncRNA expression. The OS and DFS of patients with high UCA1 and MALAT1 lncRNA expression levels were significantly shorter than those with low expression levels. Furthermore, the OS and DFS of ESCC patients appeared to be correlated with TNM staging. CONCLUSIONS: These results imply that the up-regulation of UCA1 and MALAT1 lncRNAs in ESCC tissues can impact the degree of tumor progression and is predictive of postoperative survival. Therefore, the expression levels of these lncRNAs can be used as measurement indexes to determine the prognosis of ESCC patients.

Electrophysiologic effects of 17ß-estradiol on pacemaker cells in sinoatrial nodes of rabbits.

To investigate the electrophysiological effects of 17ß-estradiol on pacemaker cells in sinoatrial (SA) nodes of rabbits and the underlying mechanism, intracellular microelectrode technique was used to record action potential (AP) in SA node cells of rabbits. The results showed that: (1) 17ß-estradiol (1, 10, 100 µmol/L) not only significantly decreased the amplitude of action potential (APA) and the maximal rate of depolarization (V(max)), but also decreased the velocity of diastolic (phase 4) depolarization (VDD) and rate of pacemaker firing (RPF) in a concentration-dependent manner. The AP duration at 50% repolarization (APD(50)) and at 90% repolarization (APD(90)) were prolonged. But the maximal diastolic potential (MDP) was not affected. (2) Pretreatment with tamoxifen (10 µmol/L), an inhibitor of estrogen receptor, did not block the electrophysiological effects of 17ß-estradiol (10 µmol/L) on SA node cells. (3) Pretreatment with N(G)-nitro-L-arginine methyl ester (L-NAME, 100 µmol/L), a nitric oxide (NO) synthase inhibitor, completely abolished the electrophysiological effects of 17ß-estradiol (10 µmol/L) on SA node cells. The results suggest that 17ß-estradiol inhibits the electrophysiological activity of pacemaker cells in SA nodes of rabbits in a concentration-dependent manner possibly through a non-genomic mechanism related with NO.

Evolutionary Fuzzy Block-Matching-Based Camera Raw Image Denoising.

An evolutionary fuzzy block-matching-based image denoising algorithm is proposed to remove noise from a camera raw image. Recently, a variance stabilization transform is widely used to stabilize the noise variance, so that a Gaussian denoising algorithm can be used to remove the signal-dependent noise in camera sensors. However, in the stabilized domain, the existed denoising algorithm may blur too much detail. To provide a better estimate of the noise-free signal, a new block-matching approach is proposed to find similar blocks by the use of a type-2 fuzzy logic system (FLS). Then, these similar blocks are averaged with the weightings which are determined by the FLS. Finally, an efficient differential evolution is used to further improve the performance of the proposed denoising algorithm. The experimental results show that the proposed denoising algorithm effectively improves the performance of image denoising. Furthermore, the average performance of the proposed method is better than those of two state-of-the-art image denoising algorithms in subjective and objective measures.

3D localization of clustered microcalcifications using cranio-caudal and medio-lateral oblique views.

This paper presents a 3D localization method to register clustered microcalcifications on mammograms from cranio-caudal (CC) and medio-lateral oblique (MLO) views. The method consists of three major components: registration of clustered microcalcifications in CC and MLO views, 3D localization of clustered microcalcifications and 3D visualization of clustered microcalcifications. The registration is performed based on three features, gradient, energy and local entropy codes that are independent of spatial locations of microcalcifications in two different views and are prioritized by discriminability in a binary decision tree. The 3D localization is determined by a sequence of coordinate corrections of calcified pixels using the breast nipple as a controlling point. Finally, the 3D visualization implements a virtual reality modeling language viewer (VRMLV) to view the exact location of the lesion as a guide for needle biopsy. In order to validate our proposed 3D localization system, a set of breast lesions, which appear both in mammograms and in MR Images is used for experiments where the depth of clustered microcalcifications can be verified by the MR images.

Genetic-based fuzzy image filter and its application to image processing.

In this paper, we propose a Genetic-based Fuzzy Image Filter (GFIF) to remove additive identical independent distribution (i.i.d.) impulse noise from highly corrupted images. The proposed filter consists of a fuzzy number construction process, a fuzz filtering process, a genetic learning process, and an image knowledge base. First, the fuzzy number construction process receives sample images or the noise-free image and then constructs an image knowledge base for the fuzzy filtering process. Second, the fuzzy filtering process contains a parallel fuzzy inference mechanism, a fuzzy mean process, and a fuzzy decision process to perform the task of noise removal. Finally, based on the genetic algorithm, the genetic learning process adjusts the parameters of the image knowledge base. By the experimental results, GFIF achieves a better performance than the state-of-the-art filters based on the criteria of Peak-Signal-to-Noise-Ratio (PSNR), Mean-Square-Error (MSE), and Mean-Absolute-Error (MAE). On the subjective evaluation of those filtered images, GFIF also results in a higher quality of global restoration.

Resveratrol inhibits neuronal discharges in rat hippocampal CA1 area.

The effects of resveratrol on the discharges of neurons in CA1 area of rat hippocampal slices were examined by using extracellular recording technique. The results are as follows: (1) In response to the application of resveratrol (0.05, 0.5, 5.0 micromol/L, n=52) into the superfusate for 2 min, the spontaneous discharge rate of 46/52 (88.5%) neurons was significantly decreased in a dose-dependent manner; (2) Application of L-glutamate (0.2 mmol/L) into the superfusate led to a marked increase in discharge rate of all 8 (100%) slices in an epileptiform pattern. The increased discharges were suppressed by application of resveratrol (5.0 micromol/L); (3) In 7 slices, perfusion of the selective L-type calcium channel agonist, Bay K8644 (0.1 micromol/L), induced a significant increase in the discharge rate of 6/7 (85.7%) slices. The increased discharges were suppressed by application of resveratrol (5.0 micromol/L); (4) In 9 slices, perfusion of nitric oxide synthase (NOS) inhibitor N(G)-nitro-L-arginine methyl ester (L-NAME, 50 micromol/L) into the superfusate significantly augmented the discharge rate of 7/9 (77.8%) slices. Resveratrol (5.0 micromol/L) applied into the superfusate reduced the increased discharges of all 7/7 (100%) neurons; (5) In 10 units, the large-conductance Ca(2+)-activated K(+) channel blocker (tetraethylammonium chloride, TEA, 1 mmol/L) significantly increased the discharge rate of 9/10 (90%) slices. Resveratrol (5.0 micromol/L) applied into the superfusate inhibited the discharges of 8/9 (88.9%) slices. These results suggest that resveratrol inhibits the electrical activity of CA1 neurons. This effect may be related to the blockade of L-type calcium channel and a subsequent reduction of calcium influx, and probably has no association with large-conductance Ca(2+)-activated K(+) channel.

Resveratrol inhibits the electrical activity of subfornical organ neurons in rat.

The effects of resveratrol on the discharges of neurons in rat subfornical organ (SFO) slices were examined by using extracellular recording technique. The results are as follows: (1) In response to the application of resveratrol (1, 5, 10 mumol/L, n=65) into the superfusate for 2 min, the spontaneous discharge rate of 60/65 (92.3%) neurons was significantly decreased in a dose-dependent manner;(2) Application of L-glutamate (0.3 mmol/L) into the superfusate led to a marked increase in discharge rate of all 12 (100%) neurons in an epileptiform pattern. The increased discharges of 10/12 (83.3%) neurons were suppressed by application of resveratrol (5 mumol/L);(3) In 8 neurons, the selective L-type calcium channel agonist, Bay K8644 (0.1 mumol/L), induced a significant increase in discharge rate of all 8 (100%) neurons. The increased discharges of all 8 (100%) neurons were suppressed by resveratrol (5 mumol/L);(4) In 14 neurons, nitric oxide synthase (NOS) inhibitor N(G)-nitro-L-arginine methyl ester (L-NAME) 50 mumol/L significantly increased the discharge rate of 11/14 (78.6%) neurons. Resveratrol (5 ?mol/L) applied into the superfusate reduced the increased discharges of 9/11 (81.8%) neurons;(5) In 12 neurons, the large-conductance Ca(2+)-activated K(+) channel blocker tetraethylammonium chloride (TEA) 1 mmol/L significantly increased the discharge rate of 10/12 (83.3%) neurons. Resveratrol (5 mumol/L) inhibited the increased discharges of 9/10 (90%) neurons. These results suggest that resveratrol inhibits the electrical activity of SFO neurons. This effect may be related to its properties of blockade of L-type voltage-gated calcium channel and nitric oxide (NO) promoting, and probably has no association with large-conductance Ca(2+)-activated K(+) channel.

High-frequency electrical stimulation of femoral nerve reduces infarct size following myocardial ischemia-reperfusion in rats.

The effects of femoral nerve electrostimulation (FNES) on ischemia-reperfused myocardium were examined in the urethane- anesthetized rats to determine whether FNES may provide cardioprotection and to observe the possible mechanism. The area at risk (AR) and infarct area (IA) were determined using Evans blue and nitro-blue tetrazolium staining, respectively. Infarct size (IS) was defined as 100xIA/AR (%). The results are as follows: (1) During 30 min myocardial ischemia and subsequent 120 min reperfusion, the myocardial infarct size occupied (54.96+/-0.82)% of the area at risk. (2) FNES of high frequency (10 V, 100 Hz, 1 ms) significantly reduced myocardial infarct size to (36.94+/-1.34)% (P<0.01), indicating the cardioprotective effect FNES of high frequency on myocardial ischemia-reperfusion, while FNES of low frequency (10 V, 10 Hz, 1 ms) had no effect on myocardial infarct size. (3) Pretreatment with either naloxone (5 mg /kg, i.v), a nonselective opioid receptor antagonist, or glibenclamide (5 mg /kg, i.v), a K(ATP) channel antagonist, completely abolished the cardioprotection of FNES (100 Hz) from myocardial ischemia-reperfusion. It is suggested that FNES of high frequency can protect myocardium from ischemia-reperfusion injury. The possible mechanism is that FNES of high frequency may induce the release of opioids from the central nervous system, and the activation of opioid receptors in the heart results in an opening of myocardial K(ATP) channels which can protect myocardium.

Digital redesign of analog smith predictor for systems with input time delays.

This paper presents a new methodology for digitally redesigning an existing analog Smith predictor control system, such that the cascaded analog controller with input delay can be implemented with a digital controller. A traditional analog Smith predictor system is reformulated into an augmented system, which is then digitally redesigned using the predicted intersampling states. The paper extends the prediction-based digital redesign method from a delay free feedback system to an input time-delay cascaded system. A tuning parameter v is optimally determined online such that in any sampling period, the output response error between the original analogously controlled time-delay system and the digitally controlled sampled-data time-delay system is significantly reduced. The proposed method gives very good performance in dealing with systems with delays in excess of several integer sampling periods and shows good robustness to sampling period selection.

A modified NARMAX model-based self-tuner with fault tolerance for unknown nonlinear stochastic hybrid systems with an input-output direct feed-through term.

A modified nonlinear autoregressive moving average with exogenous inputs (NARMAX) model-based state-space self-tuner with fault tolerance is proposed in this paper for the unknown nonlinear stochastic hybrid system with a direct transmission matrix from input to output. Through the off-line observer/Kalman filter identification method, one has a good initial guess of modified NARMAX model to reduce the on-line system identification process time. Then, based on the modified NARMAX-based system identification, a corresponding adaptive digital control scheme is presented for the unknown continuous-time nonlinear system, with an input-output direct transmission term, which also has measurement and system noises and inaccessible system states. Besides, an effective state space self-turner with fault tolerance scheme is presented for the unknown multivariable stochastic system. A quantitative criterion is suggested by comparing the innovation process error estimated by the Kalman filter estimation algorithm, so that a weighting matrix resetting technique by adjusting and resetting the covariance matrices of parameter estimate obtained by the Kalman filter estimation algorithm is utilized to achieve the parameter estimation for faulty system recovery. Consequently, the proposed method can effectively cope with partially abrupt and/or gradual system faults and input failures by the fault detection.

Efficient decentralized iterative learning tracker for unknown sampled-data interconnected large-scale state-delay system with closed-loop decoupling property.

In this paper, an efficient decentralized iterative learning tracker is proposed to improve the dynamic performance of the unknown controllable and observable sampled-data interconnected large-scale state-delay system, which consists of N multi-input multi-output (MIMO) subsystems, with the closed-loop decoupling property. The off-line observer/Kalman filter identification (OKID) method is used to obtain the decentralized linear models for subsystems in the interconnected large-scale system. In order to get over the effect of modeling error on the identified linear model of each subsystem, an improved observer with the high-gain property based on the digital redesign approach is developed to replace the observer identified by OKID. Then, the iterative learning control (ILC) scheme is integrated with the high-gain tracker design for the decentralized models. To significantly reduce the iterative learning epochs, a digital-redesign linear quadratic digital tracker with the high-gain property is proposed as the initial control input of ILC. The high-gain property controllers can suppress uncertain errors such as modeling errors, nonlinear perturbations, and external disturbances (Guo et al., 2000) [18]. Thus, the system output can quickly and accurately track the desired reference in one short time interval after all drastically-changing points of the specified reference input with the closed-loop decoupling property.

Iterative learning-based decentralized adaptive tracker for large-scale systems: a digital redesign approach.

In this paper, a digital redesign methodology of the iterative learning-based decentralized adaptive tracker is proposed to improve the dynamic performance of sampled-data linear large-scale control systems consisting of N interconnected multi-input multi-output subsystems, so that the system output will follow any trajectory which may not be presented by the analytic reference model initially. To overcome the interference of each sub-system and simplify the controller design, the proposed model reference decentralized adaptive control scheme constructs a decoupled well-designed reference model first. Then, according to the well-designed model, this paper develops a digital decentralized adaptive tracker based on the optimal analog control and prediction-based digital redesign technique for the sampled-data large-scale coupling system. In order to enhance the tracking performance of the digital tracker at specified sampling instants, we apply the iterative learning control (ILC) to train the control input via continual learning. As a result, the proposed iterative learning-based decentralized adaptive tracker not only has robust closed-loop decoupled property but also possesses good tracking performance at both transient and steady state. Besides, evolutionary programming is applied to search for a good learning gain to speed up the learning process of ILC.

3D ROC Analysis for Medical Imaging Diagnosis.

Receiver operating characteristics (ROC) has been widely used as a performance evaluation tool to measure effectiveness of medical modalities. It is derived from a standard detection theory with false alarm and detection power interpreted as false positive (FP) and true positive (TP) respectively in terms of medical diagnosis. The ROC curve is plotted based on TP versus FP via hard decisions. This paper presents a three dimensional (3D) ROC analysis which extends the traditional two-dimensional (2D) ROC analysis by including a threshold parameter in a third dimension resulting from soft decisions, (SD). As a result, a 3D ROC curve can be plotted based on three parameters, TP, FP and SD. By virtue of such a 3D ROC curve three two-dimensional (2D) ROC curves can be derived, one of which is the traditional 2D ROC curve of TP versus FP with SD reduced to hard decision. In order to illustrate its utility in medical diagnosis, its application to magnetic resonance (MR) image classification is demonstrated.