Free-response Receiver Operating Characteristic Research Articles

Computer‐aided mass detection in mammography: False positive reduction via gray‐scale invariant ranklet texture features

In this work, gray-scale invariant ranklet texture features are proposed for false positive reduction (FPR) in computer-aided detection (CAD) of breast masses. Two main considerations are at the basis of this proposal. First, false positive (FP) marks surviving our previous CAD system seem to be characterized by specific texture properties that can be used to discriminate them from masses. Second, our previous CAD system achieves invariance to linear/nonlinear monotonic gray-scale transformations by encoding regions of interest into ranklet images through the ranklet transform, an image transformation similar to the wavelet transform, yet dealing with pixels' ranks rather than with their gray-scale values. Therefore, the new FPR approach proposed herein defines a set of texture features which are calculated directly from the ranklet images corresponding to the regions of interest surviving our previous CAD system, hence, ranklet texture features; then, a support vector machine (SVM) classifier is used for discrimination. As a result of this approach, texture-based information is used to discriminate FP marks surviving our previous CAD system; at the same time, invariance to linear/nonlinear monotonic gray-scale transformations of the new CAD system is guaranteed, as ranklet texture features are calculated from ranklet images that have this property themselves by construction. To emphasize the gray-scale invariance of both the previous and new CAD systems, training and testing are carried out without any in-between parameters' adjustment on mammograms having different gray-scale dynamics; in particular, training is carried out on analog digitized mammograms taken from a publicly available digital database, whereas testing is performed on full-field digital mammograms taken from an in-house database. Free-response receiver operating characteristic (FROC) curve analysis of the two CAD systems demonstrates that the new approach achieves a higher reduction of FP marks when compared to the previous one. Specifically, at 60%, 65%, and 70% per-mammogram sensitivity, the new CAD system achieves 0.50, 0.68, and 0.92 FP marks per mammogram, whereas at 70%, 75%, and 80% per-case sensitivity it achieves 0.37, 0.48, and 0.71 FP marks per mammogram, respectively. Conversely, at the same sensitivities, the previous CAD system reached 0.71, 0.87, and 1.15 FP marks per mammogram, and 0.57, 0.73, and 0.92 FPs per mammogram. Also, statistical significance of the difference between the two per-mammogram and per-case FROC curves is demonstrated by the p-value < 0.001 returned by jackknife FROC analysis performed on the two CAD systems.

Detection of microcalcifications in mammograms using error of prediction and statistical measures

A two-stage method for detecting microcalcifications in mammograms is presented. In the first stage, the determination of the candidates for microcalcifications is performed. For this purpose, a 2-D linear prediction error filter is applied, and for those pixels where the prediction error is larger than a threshold, a statistical measure is calculated to determine whether they are candidates for microcalcifications or not. In the second stage, a feature vector is derived for each candidate, and after a classification step using a support vector machine, the final detection is performed. The algorithm is tested with 40 mammographic images, from Screen Test: The Alberta Program for the Early Detection of Breast Cancer with 50-µm resolution, and the results are evaluated using a free-response receiver operating characteristics curve. Two different analyses are performed: an individual microcalcification detection analysis and a cluster analysis. In the analysis of individual microcalcifications, detection sensitivity values of 0.75 and 0.81 are obtained at 2.6 and 6.2 false positives per image, on the average, respectively. The best performance is characterized by a sensitivity of 0.89, a specificity of 0.99, and a positive predictive value of 0.79. In cluster analysis, a sensitivity value of 0.97 is obtained at 1.77 false positives per image, and a value of 0.90 is achieved at 0.94 false positive per image.

Agreement of the Order of Overall Performance Levels Under Different Reading Paradigms

To investigate consistency of the orders of performance levels when interpreting mammograms under three different reading paradigms. We performed a retrospective observer study in which nine experienced radiologists rated an enriched set of mammography examinations that they personally had read in the clinic ("individualized") mixed with a set that none of them had read in the clinic ("common set"). Examinations were interpreted under three different reading paradigms: binary using screening Breast Imaging Reporting and Data System (BI-RADS), receiver-operating characteristic (ROC), and free-response ROC (FROC). The performance in discriminating between cancer and noncancer findings under each of the paradigms was summarized using Youden's index/2+0.5 (Binary), nonparameteric area under the ROC curve (AUC), and an overall FROC index (JAFROC-2). Pearson correlation coefficients were then computed to assess consistency in the ordering of observers' performance levels. Statistical significance of the computed correlation coefficients was assessed using bootstrap confidence intervals obtained by resampling sets of examination-specific observations. All but one of the computed pair-wise correlation coefficients were larger than 0.66 and were significantly different from zero. The correlation between the overall performance measures under the Binary and ROC paradigms was the lowest (0.43) and was not significantly different from zero (95% confidence interval -0.078 to 0.733). The use of different evaluation paradigms in the laboratory tends to lead to consistent ordering of the overall performance levels of observers. However, one should recognize that conceptually similar performance indexes resulting from different paradigms often measure different performance characteristics and thus disagreements are not only possible but frequently quite natural.

Performance Assessments of Diagnostic Systems Under the FROC Paradigm: Experimental, Analytical, and Results Interpretation Issues

As use of free response receiver-operating characteristic (FROC) curves gains more acceptance for quantitatively assessing the performance of diagnostic systems, it is important that the experimentalist understands the possible role of this approach as one of the experimental design paradigms that are available to him or her among all other approaches as well as some of the issues associated with FROC type studies. In a number of experimental scenarios, the FROC paradigm and associated analytical tools have theoretical and practical advantages over both the binary and the ROC approaches to performance assessments of diagnostic systems, but it also has some limitations related to experimental design, data analyses, clinical relevance, and complexity in the interpretation of the results. These issues are rarely discussed and are the focus of this work.

Computer‐aided detection of masses in digital tomosynthesis mammography: Comparison of three approaches

The authors are developing a computer-aided detection (CAD) system for masses on digital breast tomosynthesis mammograms (DBT). Three approaches were evaluated in this study. In the first approach, mass candidate identification and feature analysis are performed in the reconstructed three-dimensional (3D) DBT volume. A mass likelihood score is estimated for each mass candidate using a linear discriminant analysis (LDA) classifier. Mass detection is determined by a decision threshold applied to the mass likelihood score. A free response receiver operating characteristic (FROC) curve that describes the detection sensitivity as a function of the number of false positives (FPs) per breast is generated by varying the decision threshold over a range. In the second approach, prescreening of mass candidate and feature analysis are first performed on the individual two-dimensional (2D) projection view (PV) images. A mass likelihood score is estimated for each mass candidate using an LDA classifier trained for the 2D features. The mass likelihood images derived from the PVs are backprojected to the breast volume to estimate the 3D spatial distribution of the mass likelihood scores. The FROC curve for mass detection can again be generated by varying the decision threshold on the 3D mass likelihood scores merged by backprojection. In the third approach, the mass likelihood scores estimated by the 3D and 2D approaches, described above, at the corresponding 3D location are combined and evaluated using FROC analysis. A data set of 100 DBT cases acquired with a GE prototype system at the Breast Imaging Laboratory in the Massachusetts General Hospital was used for comparison of the three approaches. The LDA classifiers with stepwise feature selection were designed with leave-one-case-out resampling. In FROC analysis, the CAD system for detection in the DBT volume alone achieved test sensitivities of 80% and 90% at average FP rates of 1.94 and 3.40 per breast, respectively. With the 2D detection approach, the FP rates were 2.86 and 4.05 per breast, respectively, at the corresponding sensitivities. In comparison, the average FP rates of the system combining the 3D and 2D information were 1.23 and 2.04 per breast, respectively, at 80% and 90% sensitivities. The difference in the detection performances between the 2D and the 3D approach, and that between the 3D and the combined approach were both statistically significant (p = 0.02 and 0.01, respectively) as estimated by alternative FROC analysis. The combined system is a promising approach to improving automated mass detection on DBTs.

An automated CT based lung nodule detection scheme using geometric analysis of signed distance field

The authors present a new computerized scheme to automatically detect lung nodules depicted on computed tomography (CT) images. The procedure is performed in the signed distance field of the CT images. To obtain an accurate signed distance field, CT images are first interpolated linearly along the axial direction to form an isotropic data set. Then a lung segmentation strategy is applied to smooth the lung border aiming to include as many juxtapleural nodules as possible while minimizing over segmentations of the lung regions. Potential nodule regions are then detected by locating local maximas of signed distances in each subvolume with values and the sizes larger than the smallest nodule of interest in the three-dimensional space. Finally, all detected candidates are scored by computing the similarity distance of their medial axis-like shapes obtained through a progressive clustering strategy combined with a marching cube algorithm from a sphere based shape. A free-response receiver operating characteristics curve is computed to assess the scheme performance. A performance test on 52 low-dose CT screening examinations that depict 184 verified lung nodules showed that during the initial stage the scheme achieved an asymptotic maximum sensitivity of 95.1% (175/184) with an average of 1200 suspicious voxels per CT examination. The nine missed nodules included two small solid nodules (with a diameter < or =3.1 mm) and seven nonsolid nodules. The final performance level after the similarity scoring stage was an absolute sensitivity level, namely, including the nine missed during the initial stage, of 81.5% (150/184) with 6.5 false-positive identifications per CT examination. This preliminary study demonstrates the feasibility of applying a simple and robust geometric model using the signed distance field to identify suspicious lung nodules. In the authors' data set the sensitivity of this scheme is not affected by nodule size. In addition to potentially being a stand alone approach, the signed distance field based method can be easily implemented as an initial filtering step in other computer-aided detection schemes.

Usefulness of Computer-Aided Diagnosis Schemes for Vertebral Fractures and Lung Nodules on Chest Radiographs

We retrospectively evaluated the usefulness of computer-aided diagnosis (CAD) schemes to radiologist performance in the simultaneous detection of vertebral fractures and lung nodules on chest radiographs. We evaluated posteroanterior and lateral chest images of 21 patients with vertebral fractures, 31 patients with lung nodules, and 10 persons acting as controls. The total number of subjects was 60 because both lesions were present in four patients. Eighteen radiologists were asked to detect vertebral fractures and nodules simultaneously on posteroanterior and lateral images. The radiologists indicated their confidence level ratings regarding the presence or absence of lesions and the most likely location of each lesion on either posteroanterior or lateral images, first without and then with CAD output. The observers' performance was evaluated with use of receiver operating characteristic (ROC) and jackknife free-response ROC curves. With the CAD scheme, the average area under the ROC curve for detection of vertebral fractures improved from 0.906 to 0.951 (p = 0.002). That for lung nodules also improved, but the improvement was not statistically significant (0.804-0.816, p = 0.297). The figure-of-merit values obtained with the jackknife free-response ROC program improved from 0.585 to 0.680 (p < 0.001) for vertebral fractures and from 0.622 to 0.650 (p = 0.017) for nodules, both results having statistical significance. Average sensitivity in the detection of lesions improved from 59.8% to 69.3% for vertebral fractures and from 64.9% to 67.6% for nodules. In the detection of vertebral fractures and lung nodules on chest images, diagnostic accuracy among radiologists improves with the use of CAD.

On comparing methods for discriminating between actually negative and actually positive subjects with FROC type data

The task of searching and detecting multiple abnormalities depicted on an image, or a series of images, is a common problem in different areas such as military target detection or diagnostic medical imaging. A free response receiver operating characteristic (FROC) approach for assessing performance in many of these scenarios entails marking the locations of suspected abnormalities and indicating a level of suspicion at each of the marked locations. One of the important characteristics of a system being evaluated under the FROC paradigm is its performance in the conventional ROC domain, namely classifying a subject (or a unit of interest) as "negative" or "positive" in regard to the presence of the abnormality (or any of the abnormalities) of interest. With FROC data we can compare subjects by specifying a function of multiple scores within a subject. This approach allows formulating subject-based ROC type indices that can be estimated using existing ROC concepts. In this article we focus on indices that reflect the ability of the system to discriminate between actually negative and actually positive subjects. We consider a previously proposed index that is based on the comparison of the highest scores on subjects and two new indices that are based on potentially more stable comparison functions, namely comparison of average scores and stochastic dominance. Based on these indices we develop nonparametric procedures for comparing subject-based discriminative ability of diagnostic systems being evaluated under the FROC paradigm. We also investigate the properties of the statistical procedures in a simulation study.

Operating characteristics predicted by models for diagnostic tasks involving lesion localization

In 1996 Swensson published an observer model that predicted receiver operating characteristic (ROC), localization ROC (LROC), free-response ROC (FROC) and alternative FROC (AFROC) curves, thereby achieving "unification" of different observer performance paradigms. More recently a model termed initial detection and candidate analysis (IDCA) has been proposed for fitting computer aided detection (CAD) generated FROC data, and recently a search model for human observer FROC data has been proposed. The purpose of this study was to derive IDCA and the search model based expressions for operating characteristics, and to compare the predictions to the Swensson model. For three out of four mammography CAD data sets all models yielded good fits in the high-confidence region, i.e., near the lower end of the plots. The search model and IDCA tended to better fit the data in the low-confidence region, i.e., near the upper end of the plots, particularly for FROC curves for which the Swensson model predictions departed markedly from the data. For one data set none of the models yielded satisfactory fits. A unique characteristic of search model and IDCA predicted operating characteristics is that the operating point is not allowed to move continuously to the lowest confidence limit of the corresponding Swensson model curves. This prediction is actually observed in the CAD raw data and it is the primary reason for the poor FROC fits of the Swensson model in the low-confidence region.

Single-exposure dual-energy subtraction chest radiography: Detection of pulmonary nodules and masses in clinical practice

The purpose of this retrospective study was to evaluate the impact of energy subtraction (ES) chest radiography on the detection of pulmonary nodules and masses in daily routine. Seventy-seven patients and 25 healthy subjects were examined with a single exposure digital radiography system. Five blinded readers evaluated first the non-subtracted PA and lateral chest radiographs alone and then together with the subtracted PA soft tissue images. The size, location and number of lung nodules or masses were registered with the confidence level. CT was used as standard of reference. For the 200 total lesions, a sensitivity of 33.5-52.5% was found at non-subtracted and a sensitivity of 43.5-58.5% at energy-subtracted radiography, corresponding to a significant improvement in four of five readers (p < 0.05). However, in three of five readers the rate of false positives was higher with ES. With ES, sensitivity, but not the area under the alternative free-response receiver operating characteristics (AFROC) curve, showed a good correlation with reader experience (R = 0.90, p = 0.026). In four of five readers, the diagnostic confidence improved with ES (p = 0.0036). We conclude that single-exposure digital ES chest radiography improves detection of most pulmonary nodules and masses, but identification of nodules <1 cm and false-positive findings remain a problem.

Multi-scale analysis of lung computed tomography images

A computer-aided detection (CAD) system for the identification of lung internal nodules in low-dose multi-detector helical Computed Tomography (CT) images was developed in the framework of the MAGIC-5 project. The three modules of our lung CAD system, a segmentation algorithm for lung internal region identification, a multi-scale dot-enhancement filter for nodule candidate selection and a multi-scale neural technique for false positive finding reduction, are described. The results obtained on a dataset of low-dose and thin-slice CT scans are shown in terms of free response receiver operating characteristic (FROC) curves and discussed.

Evaluating computer‐aided detection algorithms

Computer-aided detection (CAD) has been attracting extensive research interest during the last two decades. It is recognized that the full potential of CAD can only be realized by improving the performance and robustness of CAD algorithms and this requires good evaluation methodology that would permit CAD designers to optimize their algorithms. Free-response receiver operating characteristic (FROC) curves are widely used to assess CAD performance, however, evaluation rarely proceeds beyond determination of lesion localization fraction (sensitivity) at an arbitrarily selected value of nonlesion localizations (false marks) per image. This work describes a FROC curve fitting procedure that uses a recent model of visual search that serves as a framework for the free-response task. A maximum likelihood procedure for estimating the parameters of the model from free-response data and fitting CAD generated FROC curves was implemented. Procedures were implemented to estimate two figures of merit and associated statistics such as 95% confidence intervals and goodness of fit. One of the figures of merit does not require the arbitrary specification of an operating point at which to evaluate CAD performance. For comparison a related method termed initial detection and candidate analysis was also implemented that is applicable when all suspicious regions are reported. The two methods were tested on seven mammography CAD data sets and both yielded good to excellent fits. The search model approach has the advantage that it can potentially be applied to radiologist generated free-response data where not all suspicious regions are reported, only the ones that are deemed sufficiently suspicious to warrant clinical follow-up. This work represents the first practical application of the search model to an important evaluation problem in diagnostic radiology. Software based on this work is expected to benefit CAD developers working in diverse areas of medical imaging.

Detection of Hepatocellular Carcinoma on CT in Liver Transplant Candidates: Comparison of PACS Tile and Multisynchronized Stack Modes

The objective of our study was to compare CT image interpretation using PACS tile and multisynchronized stack modes with respect to speed and observer performance for the detection of hepatocellular carcinoma (HCC) in liver transplant candidates. Institutional review board approval was obtained, but informed consent was not required for this retrospective study. Sixty-seven patients underwent dynamic multiphasic CT within 3 months before liver transplantation. Interval reviews using tile and multisynchronized stack modes were performed independently by four reviewers with various levels of experience to determine the presence of HCC using a five-point confidence scale. Observer performance was compared using jackknife free-response receiver operating characteristic (ROC) analysis. The time required to interpret the CT scans using each mode was recorded and compared using the paired Student's t test. Twenty-seven patients had 48 HCC nodules. The mean free-response ROC figures of merit for detecting HCC were significantly higher using the multisynchronized stack mode (0.731) than using the tile mode (0.662) (F-statistic = 6.603, p = 0.012). The 95% CIs for the task were -0.125 - -0.016. The time used for image analysis was also significantly shorter with the stack mode (63 to approximately 75 seconds) than with the tile mode (94 to approximately 191 seconds) for all four reviewers (p < 0.0001). Multisynchronized stack viewing of multiphasic dynamic CT scans significantly increases the detection rate of HCC in liver transplant candidates. It also significantly shortens the interpretation time compared with tile viewing.

Open framework for objective evaluation of crater detection algorithms with first test-field subsystem based on MOLA data

Crater Detection Algorithms (CDAs) applications range from estimation of lunar/planetary surface age to autonomous landing on planets and asteroids and advanced statistical analyses. A large amount of work on CDAs has already been published. However, problems arise when evaluation results of some new CDA have to be compared with already published evaluation results. The problem is that different authors use different test-fields, different Ground-Truth (GT) catalogues, and even different methodologies for evaluation of their CDAs. Re-implementation of already published CDAs or its evaluation environment is a time-consuming and unpractical solution to this problem. In addition, implementation details are often insufficiently described in publications. As a result, there is a need in research community to develop a framework for objective evaluation of CDAs. A scientific question is how CDAs should be evaluated so that the results are easily and reliably comparable. In attempt to solve this issue we first analyzed previously published work on CDAs. In this paper, we propose a framework for solution of the problem of objective CDA evaluation. The framework includes: (1) a definition of the measure for differences between craters; (2) test-field topography based on the 1/64° MOLA data; (3) the GT catalogue wherein each of 17,582 craters is aligned with MOLA data and confirmed with catalogues by N.G. Barlow et al. and J.F. Rodionova et al.; (4) selection of methodology for training and testing; and (5) a Free-response Receiver Operating Characteristics (F-ROC) curves as a way to measure CDA performance. The handling of possible improvements of the framework in the future is additionally addressed as a part of discussion of results. Possible extensions with additional test-field subsystems based on visual images, data sets for other planets, evaluation methodologies for CDAs developed for different purposes than cataloguing of craters, are proposed as well. The goal of the proposed framework is to contribute to the research community by establishing guidelines for objective evaluation of CDAs.

Combining two mammographic projections in a computer aided mass detection method

A method is presented to improve computer aided detection (CAD) results for masses in mammograms by fusing information obtained from two views of the same breast. It is based on a previously developed approach to link potentially suspicious regions in mediolateral oblique (MLO) and craniocaudal (CC) views. Using correspondence between regions, we extended our CAD scheme by building a cascaded multiple-classifier system, in which the last stage computes suspiciousness of an initially detected region conditional on the existence and similarity of a linked candidate region in the other view. We compared the two-view detection system with the single-view detection method using free-response receiver operating characteristic (FROC) analysis and cross validation. The dataset used in the evaluation consisted of 948 four-view mammograms, including 412 cancer cases with a mass, architectural distortion, or asymmetry. A statistically significant improvement was found in the lesion based detection performance. At a false positive (FP) rate of 0.1 FP/image, the lesion sensitivity improved from 56% to 61%. Case based sensitivity did not improve.

Spatial Localization Accuracy of Radiologists in Free-Response Studies: Inferring Perceptual FROC Curves from Mark-Rating Data

Free-response data consist of a set of mark-ratings pairs. Before analysis, the data are classified or "scored" into lesion and non-lesion localizations. The scoring is done by choosing an acceptance-radius and classifying marks within the acceptance-radius of lesion centers as lesion localizations, and all other marks are classified as non-lesion localizations. The scored data are plotted as a free-response receiver operating characteristic (FROC) curve, essentially a plot of appropriately normalized numbers of lesion localizations vs. non-lesion localizations. Scored FROC curves are frequently used to compare imaging systems and computer-aided detection (CAD) algorithms. However, the choice of acceptance-radius is arbitrary. This makes it difficult to compare curves from different studies and to estimate true performance. To resolve this issue the concept of two types of marks is introduced: perceptual hits and perceptual misses. A perceptual hit is a mark made in response to the observer seeing the lesion. A perceptual miss is a mark made in response to the observer seeing a (lesion-like) non-lesion. A method of estimating the most probable numbers of perceptual hits and misses is described. This allows one to plot a perceptual FROC operating point and by extension a perceptual FROC curve. Unlike a scored FROC operating point, a perceptual point is independent of the choice of acceptance-radius. The method does not allow one to identify individual marks as perceptual hits or misses-only the most probable numbers. It is based on a three-parameter statistical model of the spatial distributions of perceptual hits and misses relative to lesion centers. The method has been applied to an observer dataset in which mammographers and residents with different levels of experience were asked to locate lesions in mammograms. The perceptual operating points suggest superior performance for the mammographers and equivalent performance for residents in the first and second mammography rotations. These results and the model validation are preliminary as they are based on a small dataset. The significance of this study is showing that it is possible to probabilistically determine if a mark resulted from seeing a lesion or a non-lesion. Using the method developed in this study one could perform acceptance-radius independent estimation of observer performance.

Task-based comparison of inverse methods in magnetoencephalography

Magnetoencephalography (MEG) provides unique insights into the spatio-temporal dynamics of neural activation in the human brain. Unfortunately, the accuracy with which neural sources can be localized is limited by the highly illposed nature of the inverse problem. A large number of inverse methods have been proposed that deal with this illposedness using a range of different modeling and regularization procedures. Here we describe an objective task-based framework for comparing different inverse methods. Using the free-response receiver operating characteristic (FROC) we compare the performance of matched filters, cortically constrained dipole scanning, and minimum norm imaging methods for the task of detecting focal cortical activation. Our results indicate that the scanning methods outperform matched filters and minimum norm imaging for the case of one and two 2 cm2 patches of cortical activity when the dynamics of the two patches are both strongly and weakly correlated and irrespective of the spacing of the two activated regions.

A completely automated CAD system for mass detection in a large mammographic database

Mass localization plays a crucial role in computer-aided detection (CAD) systems for the classification of suspicious regions in mammograms. In this article we present a completely automated classification system for the detection of masses in digitized mammographic images. The tool system we discuss consists in three processing levels: (a) Image segmentation for the localization of regions of interest (ROIs). This step relies on an iterative dynamical threshold algorithm able to select iso-intensity closed contours around gray level maxima of the mammogram. (b) ROI characterization by means of textural features computed from the gray tone spatial dependence matrix (GTSDM), containing second-order spatial statistics information on the pixel gray level intensity. As the images under study were recorded in different centers and with different machine settings, eight GTSDM features were selected so as to be invariant under monotonic transformation. In this way, the images do not need to be normalized, as the adopted features depend on the texture only, rather than on the gray tone levels, too. (c) ROI classification by means of a neural network, with supervision provided by the radiologist's diagnosis. The CAD system was evaluated on a large database of 3369 mammographic images [2307 negative, 1062 pathological (or positive), containing at least one confirmed mass, as diagnosed by an expert radiologist]. To assess the performance of the system, receiver operating characteristic (ROC) and free-response ROC analysis were employed. The area under the ROC curve was found to be Az = 0.783 +/- 0.008 for the ROI-based classification. When evaluating the accuracy of the CAD against the radiologist-drawn boundaries, 4.23 false positives per image are found at 80% of mass sensitivity.

Effects of JPEG2000 Data Compression on an Automated System for Detecting Clustered Microcalcifications in Digital Mammograms

The functionalities of the JPEG2000 standard have led to its incorporation into digital imaging and communications in medicine (DICOM), which makes this compression method available for medical systems. In this study, we evaluated the compression of mammographic images with JPEG2000 (16 : 1, 20 : 1, 40 : 1, 60.4 : 1, 80: 1, and 106 : 1) for applications with a computer-aided detection (CAD) system for clusters of microcalcifications. Jackknife free-response receiver operating characteristic (JAFROC) analysis indicated that differences in the detection of clusters of microcalcifications were not statistically significant for uncompressed versus 16: 1 (T = -0.7780; p = 0.4370), 20 : 1 (T = 1.0361; p = 0.3007), and 40 : 1 (T = 1.6966; p = 0.0904); and statistically significant for uncompressed versus 60.4 : 1 (T = 5.8883; p < 0.008), 80 : 1 (T = 7.8414; p < 0.008), and 106 : 1 (T = 17.5034; p = < 0.008). Although there is a small difference in peak signal-to-noise ratio (PSNR) between compression ratios, the true-positive (TP) and false-positive (FP) rates, and the free-response receiver operating characteristic (FROC), figure of merit values considerably decreased from a 60 : 1 compression ratio. The performance of the CAD system is significantly reduced when using images compressed at ratios greater than 40 : 1 with JPEG2000 compared to uncompressed images. Mammographic images compressed up to 20 : 1 provide a percentage of correct detections by our CAD system similar to uncompressed images, regardless of the characteristics of the cluster. Further investigation is required to determine how JPEG2000 affects the detectability of clusters of microcalcifications as a function of their characteristics.

Assessment of Two 3D MDCT Colonography Protocols for Observation of Colorectal Polyps

The objective of our study was to assess the value of two-way interpretation (i.e., from rectum to cecum and vice versa) compared with one-way interpretation (i.e., from rectum to cecum only) in terms of polyp detection and interpretation time on MDCT colonography. Fifty consecutive patients underwent both CT colonography and conventional colonoscopy. Three radiologists independently analyzed the CT colonographic examinations of each patient using a primary 3D method. All examinations were analyzed using two techniques: navigation from rectum to cecum only (one-way) and navigation from rectum to cecum and vice versa (two-way). Sensitivity and positive predictive value were calculated on both a per-polyp basis and a per-patient basis. Alternative free-response receiver operating characteristic (ROC) curve analysis was estimated, and image interpretation time was documented. One hundred fifty-five polyps were depicted in 45 patients by colonoscopy. The mean sensitivity of CT colonography for polyp detection with two-way (88.4%) was significantly superior to that with one-way (78.1%) (p < 0.01). The mean positive predictive value of each observer with one-way was 66.7%, whereas that with two-way was 65.8%. The mean area under the alternative free-response ROC curve (A(z) value) with two-way (0.827) was higher than that with one-way (0.816), but there was not a statistically significant difference. The average interpretation time of each observer with two-way (39 min) was statistically significantly longer than that with one-way (25 min) (p < 0.01). When using a primary 3D interpretation technique at CT colonography, complete 3D navigation from rectum to cecum and from cecum to rectum is mandatory to maximize polyp detection. The image interpretation time for two-way interpretation is statistically significantly longer than that with one-way interpretation.