Computer-aided Detection Methods Research Articles

A new dataset of computed-tomography angiography images for computer-aided detection of pulmonary embolism

The lack of publicly available datasets of computed-tomography angiography (CTA) images for pulmonary embolism (PE) is a problem felt by physicians and researchers. Although a number of computer-aided detection (CAD) systems have been developed for PE diagnosis, their performance is often evaluated using private datasets. In this paper, we introduce a new public dataset called FUMPE (standing for Ferdowsi University of Mashhad's PE dataset) which consists of three-dimensional PE-CTA images of 35 different subjects with 8792 slices in total. For each benchmark image, two expert radiologists provided the ground-truth with the assistance of a semi-automated image processing software tool. FUMPE is a challenging benchmark for CAD methods because of the large number (i.e., 3438) of PE regions and, more especially, because of the location of most of them (i.e., 67%) in lung peripheral arteries. Moreover, due to the reporting of the Qanadli score for each PE-CTA image, FUMPE is the first public dataset which can be used for the analysis of mortality and morbidity risks associated with PE. We also report some complementary prognosis information for each subject.

Predictive features for early cancer detection in Barrett's esophagus using Volumetric Laser Endomicroscopy

The incidence of Barrett cancer is increasing rapidly and current screening protocols often miss the disease at an early, treatable stage. Volumetric Laser Endomicroscopy (VLE) is a promising new tool for finding this type of cancer early, capturing a full circumferential scan of Barrett's Esophagus (BE), up to 3-mm depth. However, the interpretation of these VLE scans can be complicated, due to the large amount of cross-sectional images and the subtle grayscale variations. Therefore, algorithms for automated analysis of VLE data can offer a valuable contribution to its overall interpretation. In this study, we broadly investigate the potential of Computer-Aided Detection (CADe) for the identification of early Barrett's cancer using VLE. We employ a histopathologically validated set of ex-vivo VLE images for evaluating and comparing a considerable set of widely-used image features and machine learning algorithms. In addition, we show that incorporating clinical knowledge in feature design, leads to a superior classification performance and additional benefits, such as low complexity and fast computation time. Furthermore, we identify an optimal tissue depth for classification of 0.5–1.0 mm, and propose an extension to the evaluated features that exploits this phenomenon, improving their predictive properties for cancer detection in VLE data. Finally, we compare the performance of the CADe methods with the classification accuracy of two VLE experts. With a maximum Area Under the Curve (AUC) in the range of 0.90–0.93 for the evaluated features and machine learning methods versus an AUC of 0.81 for the medical experts, our experiments show that computer-aided methods can achieve a considerably better performance than trained human observers in the analysis of VLE data.

A user-guided tool for semi-automated cerebral microbleed detection and volume segmentation: Evaluating vascular injury and data labelling for machine learning.

Background and purposeWith extensive research efforts in place to address the clinical relevance of cerebral microbleeds (CMBs), there remains a need for fast and accurate methods to detect and quantify CMB burden. Although some computer-aided detection algorithms have been proposed in the literature with high sensitivity, their specificity remains consistently poor. More sophisticated machine learning methods appear to be promising in their ability to minimize false positives (FP) through high-level feature extraction and the discrimination of hard-mimics. To achieve superior performance, these methods require sizable amounts of precisely labelled training data. Here we present a user-guided tool for semi-automated CMB detection and volume segmentation, offering high specificity for routine use and FP labelling capabilities to ease and expedite the process of generating labelled training data.Materials and methodsExisting computer-aided detection methods reported by our group were extended to include fully-automated segmentation and user-guided CMB classification with FP labelling. The algorithm's performance was evaluated on a test set of ten patients exhibiting radiotherapy-induced CMBs on MR images.ResultsThe initial algorithm's base sensitivity was maintained at 86.7%. FP's were reduced to inter-rater variations and segmentation results were in 98% agreement with ground truth labelling. There was an approximate 5-fold reduction in the time users spent evaluating CMB burden with the algorithm versus without computer aid. The Intra-class Correlation Coefficient for inter-rater agreement was 0.97 CI[0.92,0.99].ConclusionsThis development serves as a valuable tool for routine evaluation of CMB burden and data labelling to improve CMB classification with machine learning. The algorithm is available to the public on GitHub (https://github.com/LupoLab-UCSF/CMB_labeler).

A Survey of Computer-aided Detection of Breast Cancer with Mammography

Computer-aided detection (CAD) systems can be served as a second view for radiologist. An overview of recent development in CAD methods are presented in this paper. Abnormalities detection, abnormalities classification and content-based image retrieval (CBIR) are briefly reviewed. For the abnormalities detection, micro-calcification detection, mass detection and multi-view based detection are introduced. For the abnormalities classification, micro-calcification classification and mass classification are given.

WE-E-213CD-03: Inverse-Consistent Symmetric Registration of Inner Colon Surfaces Derived from Prone and Supine CT Colonography.

Robust registration of prone and supine colonie surfaces acquired during CT colonography may lead to faster and more accurate detection of colorectal cancer and polyps. Any directional bias when registering one surface to the other could precipitate incorrect anatomical correspondence and engender reader error. Despite this, non-rigid registration methods are often implemented asymmetrically, which could negatively influence the registration. We aimed to reduce directional bias and so increase robustness by adapting a cylindrical registration algorithm to be both symmetric and inverse-consistent. The registration task can be simplified by mapping both prone and supine colonie surfaces onto regular cylinders. Spatial correspondence can then be established in cylindrical space using the original surfaces' local shape indices. We implemented a symmetric formulation of the popular non-rigid B-spline image registration method in cylindrical space. A symmetric similarity measure computes the sum of squared differences between both cylindrical representations of prone-to-supine and supine-to-prone directions simultaneously. Inverse consistency of the transformation is enforced by adding an appropriately weighted penalty term to the optimisation function. We selected 8 CT colonography patient cases with marked variation in luminal distension and surface morphology. We randomly allocated 4 of these for tuning an optimal set of registration parameters and 4 for validation. The mean inverse-consistency error was reduced by 32% from 4.8mm to 3.2mm by the new symmetric formulation. The mean registration error improved from 8.2mm to 7.3mm for 330 manually chosen reference points on the 4 validation sets. A symmetric formulation of prone and supine surface registration improves the quality of registration. Information from both prone-to-supine and supine-to-prone directions helps enforce convergence towards a more accurate solution due to reduced directional bias. A more robust and accurate registration will facilitate interpretation of CT colonography and has the potential to improve existing computer-aided detection methods. The authors gratefully acknowledge financial support for this work from the NIHR program: “Imaging diagnosis of colorectal cancer: Interventions for efficient and acceptable diagnosis in symptomatic and screening populationsâ€.

Accurate computer-aided quantification of left ventricular parameters: experience in 1555 cardiac magnetic resonance studies from the Framingham Heart Study.

Quantitative analysis of short-axis functional cardiac magnetic resonance images can be performed using automatic contour detection methods. The resulting myocardial contours must be reviewed and possibly corrected, which can be time-consuming, particularly when performed across all cardiac phases. We quantified the impact of manual contour corrections on both analysis time and quantitative measurements obtained from left ventricular short-axis cine images acquired from 1555 participants of the Framingham Heart Study Offspring cohort using computer-aided contour detection methods. The total analysis time for a single case was 7.6 ± 1.7 min for an average of 221 ± 36 myocardial contours per participant. This included 4.8 ± 1.6 min for manual contour correction of 2% of all automatically detected endocardial contours and 8% of all automatically detected epicardial contours. However, the impact of these corrections on global left ventricular parameters was limited, introducing differences of 0.4 ± 4.1 mL for end-diastolic volume, -0.3 ± 2.9 mL for end-systolic volume, 0.7 ± 3.1 mL for stroke volume, and 0.3 ± 1.8% for ejection fraction. We conclude that left ventricular functional parameters can be obtained under 5 min from short-axis functional cardiac magnetic resonance images using automatic contour detection methods. Manual correction more than doubles analysis time, with minimal impact on left ventricular volumes and ejection fraction.

Combination of computer-aided detection algorithms for automatic lung nodule identification

The aim of this work is to evaluate the potential of combining different computer-aided detection (CADe) methods to increase the actual support for radiologists of automated systems in the identification of pulmonary nodules in CT scans. The outputs of three different CADe systems developed by researchers of the Italian MAGIC-5 collaboration were combined. The systems are: the CAMCADe (based on a Channeler-Ant-Model which segments vessel tree and nodule candidates and a neural classifier), the RGVPCADe (a Region-Growing- Volume-Plateau algorithm detects nodule candidates and a neural network reduces false positives); the VBNACADe (two dedicated procedures, based respectively on a 3D dot-enhancement algorithm and on intersections of pleura surface normals, identifies internal and juxtapleural nodules, and a Voxel-Based-Neural-Approach reduces false positives. A dedicated OsiriX plugin implemented with the Cocoa environments of MacOSX allows annotating nodules and visualizing singles and combined CADe findings. The combined CADe has been tested on thin slice (lower than 2 mm) CTs of the LIDC public research database and the results have been compared with those obtained by the single systems. The FROC (Free Receiver Operating Characteristic) curves show better results than the best of the single approaches. Has been demonstrated that the combination of different approaches offers better results than each single CADe system. A clinical validation of the combined CADe as second reader is being addressed by means of the dedicated OsiriX plugin.

Effect of the Enhancement Threshold on the Computer-Aided Detection of Breast Cancer using MRI

To evaluate the effect that variations in the enhancement threshold have on the diagnostic accuracy of two computer-aided detection (CAD) systems for magnetic resonance based breast cancer screening. Informed consent was obtained from all patients participating in cancer screening and this study was approved by the participating institution's review board. This retrospective study was nested in a prospective, single-institution, high-risk, breast screening study involving dynamic contrast-enhanced magnetic resonance imaging. Only those screening examinations (n = 223) for which a histopathological diagnosis was available were included. Two CAD methods were performed: the signal enhancement ratio (SER) and support vector machines (SVMs). Statistical analysis was performed by tracking changes in each CAD test's diagnostic accuracy (eg, receiver-operating characteristic [ROC] curve area, maximum possible sensitivity) with changes in the enhancement threshold. The enhancement threshold plays a significant role in affecting a CAD test's potential sensitivity, ROC curve area, and number of assumed true and false-positive predictions per cancerous examination. A high threshold can also limit the CAD-based detection of the full size of a lesion. Enhancement thresholds can limit a CAD test's ability to diagnose a lesion's full size and as such should not be raised above 60%. The clinically used SER method exhibits a high rate of false positives at low enhancement thresholds and as such the threshold should not be set lower than 50%. The SVM method yielded better results in our study than the SER method at clinically realistic enhancement thresholds.

The efficiency of the human observer for lesion detection and localization in emission tomography

For the medically relevant task of joint detection and localization of a signal (lesion) in an emission computed tomographic (ECT) image, it is of interest to measure the efficiency, defined as the relative task performance of a human observer versus that of an ideal observer. Efficiency studies can be used in system optimization, improving postprocessing (e.g., reconstruction) algorithms, deriving human-emulating model observers and computer-aided detection methods. Calculation of ideal observer performance for ECT is highly computationally complex. We can, however, compute ideal observer performance exactly using a simplified ‘filtered-noise’ model of ECT. This model results in images whose correlation structure, due to quantum noise, background variability and regularization, is similar to that of real ECT reconstructed images. A two-alternative forced choice test is used to obtain the performance of the human observers. We compare the efficiency of our joint detection–localization task with that of a corresponding signal-known-exactly (SKE) detection task. For the joint task, efficiency is low when the search tolerance is stringent. Efficiency for the joint task rises with signal intensity but is flat for the SKE task. For both tasks, efficiency peaks at a mid-range level of regularization corresponding to a particular noise-resolution tradeoff.

Volume-based feature analysis of mucosa for automatic initial polyp detection in virtual colonoscopy

In this paper, we present a volume-based mucosa-based polyp candidate determination scheme for automatic polyp detection in computed colonography. Different from most of the existing computer-aided detection (CAD) methods where mucosa layer is a one-layer surface, a thick mucosa of 3-5 voxels wide fully reflecting partial volume effect is intentionally extracted, which excludes the direct applications of the traditional geometrical features. In order to address this dilemma, fast marching-based adaptive gradient/curvature and weighted integral curvature along normal directions (WICND) are developed for volume-based mucosa. In doing so, polyp candidates are optimally determined by computing and clustering these fast marching-based adaptive geometrical features. By testing on 52 patients datasets in which 26 patients were found with polyps of size 4-22 mm, both the locations and number of polyp candidates detected by WICND and previously developed linear integral curvature (LIC) were compared. The results were promising that WICND outperformed LIC mainly in two aspects: (1) the number of detected false positives was reduced from 706 to 132 on average, which significantly released our burden of machine learning in the feature space, and (2) both the sensitivity and accuracy of polyp detection have been slightly improved, especially for those polyps smaller than 5mm.

Colon polyp detection using smoothed shape operators: Preliminary results

Computer-aided detection (CAD) algorithms identify locations in computed tomographic (CT) images of the colon that are most likely to contain polyps. Existing CAD methods treat the CT data as a voxelized, volume image. They estimate a curvature-based feature at the mucosal surface voxels. However, curvature is a smooth notion, while our data are discrete and noisy. As a second order differential quantity, curvature amplifies noise. In this paper, we present the smoothed shape operators method (SSO), which uses a geometry processing approach. We extract a triangle mesh representation of the colon surface, and estimate curvature on this surface using the shape operator. We then smooth the shape operators on the surface iteratively. Throughout, we use techniques explicitly designed for discrete geometry. All our computation occurs on the surface, rather than in the voxel grid. We evaluate our algorithm on patient data and provide free-response receiver-operating characteristic performance analysis over all size ranges of polyps. We also provide confidence intervals for our performance estimates. We compare our performance with the surface normal overlap (SNO) method for the same data. A preliminary evaluation of our method on 35 patients yielded the following results (polyp diameter range; sensitivity; false positives/case): (10mm; 100%; 17.5), (5-10 mm; 89.7%, 21.23), (<5 mm; 59.1%; 23.9) and (overall; 80.3%; 23.9). The evaluation of the SNO method yielded: (10 mm; 75%; 17.5), (5-10 mm; 43.1%; 21.23), (<5 mm; 15.9%; 23.9) and (overall; 38.5%; 23.9).

Computer-Aided Mass Detection Based on Ipsilateral Multiview Mammograms

Recent reports on advances in computer-aided detection (CAD) indicate that current schemes miss early-stage breast cancers and result in a relatively large false-positive detection rate in order to achieve a high sensitivity rate for mass detection. This paper is inspired by the interpretation procedure from mammographers. The abnormal diagnosis can be derived from multiple views but is not available through single-view image analysis. A new multiview CAD system for early-stage breast cancer detection, which is based on modifying the optimized CAD algorithms from our prior single-view CAD system for constructing an adaptive ipsilateral multiview concurrent CAD system, is presented in this paper. The selection and design for the training and testing ipsilateral multiview mammogram databases are described here. The performance evaluation of the developed ipsilateral multiview CAD system using free-response receiver operating characteristic analysis and computerized receiver operating characteristic experiments are presented. The results indicated that the proposed multiview CAD system is significantly superior to the single-view CAD systems based on statistically standard P-values. This paper addresses a very important and timely project. It is related to two main problems regarding the development of breast cancer detection and diagnosis: early-stage detection and diagnosis of breast cancer with digital mammogram, and overall improvement of CAD system performance for clinical implementation. In order to improve the efficacy, accuracy, and efficiency of the current CAD scheme, an entirely new class of CAD method is required. This paper is unique in that a comprehensive and state-of-the-art approach is proposed for the CAD scheme of digital mammography. From the design aspect of the CAD scheme, the proposed ipsilateral multiview CAD method is innovative and quite different from current single-view CAD methods.

Full-field digital mammography compared with screen-film mammography in the detection of breast cancer: rays of light through DMIST or more fog?

Breast cancer is the most common malignancy and second most common cause of cancer death in women in the US [1]. To date, screen-film mammography (SFM) has been considered the gold standard for breast cancer screening and detection [1]. Over 70% of women in the US over the age of 40 have had a mammogram within the past 2 years [2]. Prior research has investigated the efficacy of SFM at reducing mortality from breast cancer [3–17]. While the interpretation of these results has been hotly debated [18–25], the consensus in recent systematic reviews and practice guidelines is that SFM reduces breast cancer mortality among women who are 50–75 years of age, and may reduce mortality among women aged 40 years and older [26–30]. A series of complex models supports the hypothesis that SFM has contributed to a reduction in breast cancer mortality in the US [31]. Although film mammography is the established diagnostic tool, it has a number of limitations. Breast cancer screening is less effective in younger women than in older women, most likely because younger women have a higher proportion of mammographically dense breast tissue [32– 35]. Up to a third of breast cancers are missed, even though as many as 20 women are called back due to false-positive results for every one cancer identified; this increases patient anxiety and leads to additional breast imaging, radiation exposure, and biopsies. Furthermore, with SFM, image acquisition, storage, and comparison can be cumbersome. As a result, much attention has been devoted to developing improved radiographic techniques for breast cancer screening and evaluation. Full-field digital mammography (FFDM) addresses some of the limitations seen with SFM. It has a wider range of contrast resolution, which holds particular promise for improving the detection of low-contrast lesions in radiographically dense breasts. FFDM allows for the separation of image acquisition, presentation, and storage, thus enabling these tasks to be independently optimized. The implementation of computer-aided detection methods is potentially simplified with FFDM, as there is no longer the separate step of digitizing film mammograms [36]. Computer-aided enhancement of images at computer workstations may also improve the accuracy of mammographic interpretation [37]. FFDM also has the potential to improve workflow by allowing for electronic transmission, storage, and retrieval of the images. Use of this modality may facilitate the interpretation of mammograms, as the images can be obtained locally, but sent to a central location for interpretation by experts at centers that specialize in mammographic interpretation. Digital storage and retrieval may increase the likelihood that comparison images from prior mammograms would be available to aid the radiologist in the interpretation of a new mammogram. Digital images can either be printed on film for review (hard copy) or read on computer monitors (soft copy). Digital image acquisition may improve the signal-to-noise ratio of X-ray detection over a wider range of intensities, as compared to that of film acquisition over the same range [38–40]. There is also the possibility of lower radiation exposure than that required for SFM. In spite of these potential advantages, adoption of a digital approach to mammography has been slow, in part due to the high spatial resolution required by J. A. Tice (&) M. D. Feldman Division of General Internal Medicine, Department of Medicine, University of California, 1701 Divisadero Street, Suite 554, San Francisco, CA 94143-1732, USA e-mail: jtice@medicine.ucsf.edu

Colonic polyp characterization and detection based on both morphological and texture features

In this paper, a method is presented for detection of colonic polyps using both morphological and texture features of the colon wall from the abdominal computed tomography (CT) images. This method consists of two steps. In the first step, suspicious patches of the colon wall are quickly identified by utilizing special local and global geometrical information of the segmented inner-wall mucosa layer. In the second step, a special mapping strategy is employed to identify the growing region of each suspected polyp. Then its internal CT density textures are extracted and quantitatively analyzed based on an assumed ellipsoid polyp model. Finally, both the extracted texture and morphological information are applied to eliminate the false-positives from the identified suspicious patches. With all the extracted geometrical, morphological and texture features, this presented computer-aided detection (CAD) method has demonstrated a significant improvement in detection of the colonic polyps over the previously reported geometry-based CAD methods for CT-based virtual colonoscopy.

Computer-aided lesion detection with statistical model-based features in PET images

Positron emission tomography (PET) with the glucose analog [/sup 18/F] fluorodeoxyglucose is proving to be useful in cancer diagnosis and treatment. However, as in all nuclear medicine imaging technologies, lesion detection with PET is often hindered by limited spatial resolution and low signal-to-noise ratios. Under such conditions, conventional diagnosis by visual inspection usually becomes difficult and potentially inaccurate. In this paper, we propose use of computer-aided lesion detection methods for PET imaging by applying a maximum likelihood ratio test and a composite hypothesis test, assuming that the mean positron emission rate is deterministic or random, respectively. In our approach, different statistical models characterizing the mean positron emission rate, the raw sinogram data and the filtered backprojection (FBP) reconstructed image are used to derive the test criteria. Three methods to estimate the unknown parameters of the test functions from observations are presented. The performance of one of the proposed methods is evaluated and compared with both simulated and experimental phantom data. In the preliminary trials, the methods detect correctly (with a high probability >0.9) lesions of diameter/spl ges/15 mm with lesion-to-background contrast 1.1:1. Under the same conditions, the test lesion could not be detected by visual inspection alone in the images reconstructed by either the FBP or the maximum likelihood iterative algorithms. The methods may also be used for the objective assessment of the quality of images reconstructed from different algorithms.