Standardization in Quantitative Imaging: A Multicenter Comparison of Radiomic Features from Different Software Packages on Digital Reference Objects and Patient Data Sets.

Radiomic features are being increasingly studied for clinical applications. We aimed to assess the agreement among radiomic features when computed by several groups by using different software packages under very tightly controlled conditions, which included standardized feature definitions and common image data sets. Ten sites (9 from the NCI's Quantitative Imaging Network] positron emission tomography-computed tomography working group plus one site from outside that group) participated in this project. Nine common quantitative imaging features were selected for comparison including features that describe morphology, intensity, shape, and texture. The common image data sets were: three 3D digital reference objects (DROs) and 10 patient image scans from the Lung Image Database Consortium data set using a specific lesion in each scan. Each object (DRO or lesion) was accompanied by an already-defined volume of interest, from which the features were calculated. Feature values for each object (DRO or lesion) were reported. The coefficient of variation (CV), expressed as a percentage, was calculated across software packages for each feature on each object. Thirteen sets of results were obtained for the DROs and patient data sets. Five of the 9 features showed excellent agreement with CV < 1%; 1 feature had moderate agreement (CV < 10%), and 3 features had larger variations (CV ≥ 10%) even after attempts at harmonization of feature calculations. This work highlights the value of feature definition standardization as well as the need to further clarify definitions for some features.

Aromatase inhibitor-induced modulation of breast density: clinical and genetic effects.

BACKGROUND: Change in breast density may predict outcome of women receiving adjuvant hormone therapy for breast cancer. We performed a prospective clinical trial to evaluate the impact of inherited variants in genes involved in oestrogen metabolism and signalling on change in mammographic percent density (MPD) with aromatase inhibitor (AI) therapy. METHODS: Postmenopausal women with breast cancer who were initiating adjuvant AI therapy were enrolled onto a multicentre, randomised clinical trial of exemestane vs letrozole, designed to identify associations between AI-induced change in MPD and single-nucleotide polymorphisms in candidate genes. Subjects underwent unilateral craniocaudal mammography before and following 24 months of treatment. RESULTS: Of the 503 enrolled subjects, 259 had both paired mammograms at baseline and following 24 months of treatment and evaluable DNA. We observed a statistically significant decrease in mean MPD from 17.1 to 15.1% (P<0.001), more pronounced in women with baseline MPD ≥20%. No AI-specific difference in change in MPD was identified. No significant associations between change in MPD and inherited genetic variants were observed. CONCLUSION: Subjects with higher baseline MPD had a greater average decrease in MPD with AI therapy. There does not appear to be a substantial effect of inherited variants in biologically selected candidate genes.

Treatment response assessment of head and neck cancers on CT using computerized volume analysis.

BACKGROUND AND PURPOSE: Head and neck cancer can cause substantial morbidity and mortality. Our aim was to evaluate the potential usefulness of a computerized system for segmenting lesions in head and neck CT scans and for estimation of volume change of head and neck malignant tumors in response to treatment. MATERIALS AND METHODS: CT scans from a pretreatment examination and a post 1-cycle chemotherapy examination of 34 patients with 34 head and neck primary-site cancers were collected. The computerized system was developed in our laboratory. It performs 3D segmentation on the basis of a level-set model and uses as input an approximate bounding box for the lesion of interest. The 34 tumors included tongue, tonsil, vallecula, supraglottic, epiglottic, and hard palate carcinomas. As a reference standard, 1 radiologist outlined full 3D contours for each of the 34 primary tumors for both the pre- and posttreatment scans and a second radiologist verified the contours. RESULTS: The correlation between the automatic and manual estimates for both the pre- to post-treatment volume change and the percentage volume change for the 34 primary-site tumors was 0.95, with an average error of -2.4 ± 8.5% by automatic segmentation. There was no substantial difference and specific trend in the automatic segmentation accuracy for the different types of primary head and neck tumors, indicating that the computerized segmentation performs relatively robustly for this application. CONCLUSIONS: The tumor size change in response to treatment can be accurately estimated by the computerized segmentation system relative to radiologists' manual estimations for different types of head and neck tumors.

Selection of an optimal neural network architecture for computer-aided detection of microcalcifications--comparison of automated optimization techniques.

Many computer-aided diagnosis (CAD) systems use neural networks (NNs) for either detection or classification of abnormalities. Currently, most NNs are "optimized" by manual search in a very limited parameter space. In this work, we evaluated the use of automated optimization methods for selecting an optimal convolution neural network (CNN) architecture. Three automated methods, the steepest descent (SD), the simulated annealing (SA), and the genetic algorithm (GA), were compared. We used as an example the CNN that classifies true and false microcalcifications detected on digitized mammograms by a prescreening algorithm. Four parameters of the CNN architecture were considered for optimization, the numbers of node groups and the filter kernel sizes in the first and second hidden layers, resulting in a search space of 432 possible architectures. The area Az under the receiver operating characteristic (ROC) curve was used to design a cost function. The SA experiments were conducted with four different annealing schedules. Three different parent selection methods were compared for the GA experiments. An available data set was split into two groups with approximately equal number of samples. By using the two groups alternately for training and testing, two different cost surfaces were evaluated. For the first cost surface, the SD method was trapped in a local minimum 91% (392/432) of the time. The SA using the Boltzman schedule selected the best architecture after evaluating, on average, 167 architectures. The GA achieved its best performance with linearly scaled roulette-wheel parent selection; however, it evaluated 391 different architectures, on average, to find the best one. The second cost surface contained no local minimum. For this surface, a simple SD algorithm could quickly find the global minimum, but the SA with the very fast reannealing schedule was still the most efficient. The same SA scheme, however, was trapped in a local minimum on the first cost surface. Our CNN study demonstrated that, if optimization is to be performed on a cost surface whose characteristics are not known a priori, it is advisable that a moderately fast algorithm such as a SA using a Boltzman cooling schedule be used to conduct an efficient and thorough search, which may offer a better chance of reaching the global minimum.

Improvement of mammographic mass characterization using spiculation meausures and morphological features.

We are developing new computer vision techniques for characterization of breast masses on mammograms. We had previously developed a characterization method based on texture features. The goal of the present work was to improve our characterization method by making use of morphological features. Toward this goal, we have developed a fully automated, three-stage segmentation method that includes clustering, active contour, and spiculation detection stages. After segmentation, morphological features describing the shape of the mass were extracted. Texture features were also extracted from a band of pixels surrounding the mass. Stepwise feature selection and linear discriminant analysis were employed in the morphological, texture, and combined feature spaces for classifier design. The classification accuracy was evaluated using the area Az under the receiver operating characteristic curve. A data set containing 249 films from 102 patients was used. When the leave-one-case-out method was applied to partition the data set into trainers and testers, the average test Az for the task of classifying the mass on a single mammographic view was 0.83 +/- 0.02, 0.84 +/- 0.02, and 0.87 +/- 0.02 in the morphological, texture, and combined feature spaces, respectively. The improvement obtained by supplementing texture features with morphological features in classification was statistically significant (p = 0.04). For classifying a mass as malignant or benign, we combined the leave-one-case-out discriminant scores from different views of a mass to obtain a summary score. In this task, the test Az value using the combined feature space was 0.91 +/- 0.02. Our results indicate that combining texture features with morphological features extracted from automatically segmented mass boundaries will be an effective approach for computer-aided characterization of mammographic masses.

Computerized image analysis: estimation of breast density on mammograms.

An automated image analysis tool is being developed for the estimation of mammographic breast density. This tool may be useful for risk estimation or for monitoring breast density change in prevention or intervention programs. In this preliminary study, a data set of 4-view mammograms from 65 patients was used to evaluate our approach. Breast density analysis was performed on the digitized mammograms in three stages. First, the breast region was segmented from the surrounding background by an automated breast boundary-tracking algorithm. Second, an adaptive dynamic range compression technique was applied to the breast image to reduce the range of the gray level distribution in the low frequency background and to enhance the differences in the characteristic features of the gray level histogram for breasts of different densities. Third, rule-based classification was used to classify the breast images into four classes according to the characteristic features of their gray level histogram. For each image, a gray level threshold was automatically determined to segment the dense tissue from the breast region. The area of segmented dense tissue as a percentage of the breast area was then estimated. To evaluate the performance of the algorithm, the computer segmentation results were compared to manual segmentation with interactive thresholding by five radiologists. A "true" percent dense area for each mammogram was obtained by averaging the manually segmented areas of the radiologists. We found that the histograms of 6% (8 CC and 8 MLO views) of the breast regions were misclassified by the computer, resulting in poor segmentation of the dense region. For the images with correct classification, the correlation between the computer-estimated percent dense area and the "truth" was 0.94 and 0.91, respectively, for CC and MLO views, with a mean bias of less than 2%. The mean biases of the five radiologists' visual estimates for the same images ranged from 0.1% to 11%. The results demonstrate the feasibility of estimating mammographic breast density using computer vision techniques and its potential to improve the accuracy and reproducibility of breast density estimation in comparison with the subjective visual assessment by radiologists.

Automated registration of breast lesions in temporal pairs of mammograms for interval change analysis--local affine transformation for improved localization.

Analysis of interval change is important for mammographic interpretation. The aim of this study is to evaluate the use of an automated registration technique for computer-aided interval change analysis in mammography. Previously we developed a regional registration technique for identifying masses on temporal pairs of mammograms. In the current study, we improved lesion registration by including a local alignment step. Initially, the lesion position on the prior mammogram was estimated based on the breast geometry. An initial fan-shaped search region was then defined on the prior mammogram. In the second stage, the location of the fan-shaped region on the prior mammogram was refined by warping, based on an affine transformation and simplex optimization in a local region. In the third stage, a search for the best match between the lesion template from the current mammogram and a structure on the prior mammogram was carried out within the search region. This technique was evaluated on 124 temporal pairs of mammograms containing biopsyproven masses. Eighty-seven percent of the estimated lesion locations resulted in an area overlap of at least 50% with the true lesion locations and an average distance of 2.4 +/- 2.1 mm between their centroids. The average distance between the estimated and the true centroid of the lesions on the prior mammogram over all 124 temporal pairs was 4.2 +/- 5.7 mm. The registration accuracy was improved in comparison with our previous study that used a data set of 74 temporal pairs of mammograms. This improvement in accuracy resulted from the improved geometry estimation and the local affine transformation.

Digital mammography: observer performance study of the effects of pixel size on the characterization of malignant and benign microcalcifications.

RATIONALE AND OBJECTIVES: The authors performed this study to evaluate the effects of pixel size on the characterization of mammographic microcalcifications by radiologists. MATERIALS AND METHODS: Two-view mammograms of 112 microcalcification clusters were digitized with a laser scanner at a pixel size of 35 microm. Images with pixel sizes of 70, 105, and 140 microm were derived from the 35-microm-pixel size images by averaging neighboring pixels. The malignancy or benignity of the microcalcifications had been determined with findings at biopsy or 2-year follow-up. Region-of-interest images containing the microcalcifications were printed with a laser imager. Seven radiologists participated in a receiver operating characteristic (ROC) study to estimate the likelihood of malignancy. The classification accuracy was quantified with the area under the ROC curve (Az). The statistical significance of the differences in the Az values for different pixel sizes was estimated with the Dorfman-Berbaum-Metz method and the Student paired t test. The variance components were analyzed with a bootstrap method. RESULTS: The higher-resolution images did not result in better classification; the average Az with a pixel size of 35 microm was lower than that with pixel sizes of 70 and 105 microm. The differences in Az between different pixel sizes did not achieve statistical significance. CONCLUSION: Pixel sizes in the range studied do not have a strong effect on radiologists' accuracy in the characterization of microcalcifications. The low specificity of the image features of microcalcifications and the large interobserver and intraobserver variabilities may have prevented small advantages in image resolution from being observed.

Computer-aided characterization of mammographic masses: accuracy of mass segmentation and its effects on characterization.

Mass segmentation is used as the first step in many computer-aided diagnosis (CAD) systems for classification of breast masses as malignant or benign. The goal of this paper was to study the accuracy of an automated mass segmentation method developed in our laboratory, and to investigate the effect of the segmentation stage on the overall classification accuracy. The automated segmentation method was quantitatively compared with manual segmentation by two expert radiologists (R1 and R2) using three similarity or distance measures on a data set of 100 masses. The area overlap measures between R1 and R2, the computer and R1, and the computer and R2 were 0.76 +/- 0.13, 0.74 +/- 0.11, and 0.74 +/- 0.13, respectively. The interobserver difference in these measures between the two radiologists was compared with the corresponding differences between the computer and the radiologists. Using three similarity measures and data from two radiologists, a total of six statistical tests were performed. The difference between the computer and the radiologist segmentation was significantly larger than the interobserver variability in only one test. Two sets of texture, morphological, and spiculation features, one based on the computer segmentation, and the other based on radiologist segmentation, were extracted from a data set of 249 films from 102 patients. A classifier based on stepwise feature selection and linear discriminant analysis was trained and tested using the two feature sets. The leave-one-case-out method was used for data sampling. For case-based classification, the area Az under the receiver operating characteristic (ROC) curve was 0.89 and 0.88 for the feature sets based on the radiologist segmentation and computer segmentation, respectively. The difference between the two ROC curves was not statistically significant.

Analysis of temporal changes of mammographic features: computer-aided classification of malignant and benign breast masses.

A new classification scheme was developed to classify mammographic masses as malignant and benign by using interval change information. The masses on both the current and the prior mammograms were automatically segmented using an active contour method. From each mass, 20 run length statistics (RLS) texture features, 3 speculation features, and 12 morphological features were extracted. Additionally, 20 difference RLS features were obtained by subtracting the prior RLS features from the corresponding current RLS features. The feature space consisted of the current RLS features, the difference RLS features, the current and prior speculation features, and the current and prior mass sizes. Stepwise feature selection and linear discriminant analysis classification were used to select and merge the most useful features. A leave-one-case-out resampling scheme was used to train and test the classifier using 140 temporal image pairs (85 malignant, 55 benign) obtained from 57 biopsy-proven masses (33 malignant, 24 benign) in 56 patients. An average of 10 features were selected from the 56 training subsets: 4 difference RLS features, 4 RLS features, and 1 speculation feature from the current image, and 1 speculation feature from the prior, were most often chosen. The classifier achieved an average training Az of 0.92 and a test Az of 0.88. For comparison, a classifier was trained and tested using features extracted from the 120 current single images. This classifier achieved an average training Az of 0.90 and a test Az of 0.82. The information on the prior image significantly (p = 0.015) improved the accuracy for classification of the masses.

Feature selection and classifier performance in computer-aided diagnosis: the effect of finite sample size.

In computer-aided diagnosis (CAD), a frequently used approach for distinguishing normal and abnormal cases is first to extract potentially useful features for the classification task. Effective features are then selected from this entire pool of available features. Finally, a classifier is designed using the selected features. In this study, we investigated the effect of finite sample size on classification accuracy when classifier design involves stepwise feature selection in linear discriminant analysis, which is the most commonly used feature selection algorithm for linear classifiers. The feature selection and the classifier coefficient estimation steps were considered to be cascading stages in the classifier design process. We compared the performance of the classifier when feature selection was performed on the design samples alone and on the entire set of available samples, which consisted of design and test samples. The area Az under the receiver operating characteristic curve was used as our performance measure. After linear classifier coefficient estimation using the design samples, we studied the hold-out and resubstitution performance estimates. The two classes were assumed to have multidimensional Gaussian distributions, with a large number of features available for feature selection. We investigated the dependence of feature selection performance on the covariance matrices and means for the two classes, and examined the effects of sample size, number of available features, and parameters of stepwise feature selection on classifier bias. Our results indicated that the resubstitution estimate was always optimistically biased, except in cases where the parameters of stepwise feature selection were chosen such that too few features were selected by the stepwise procedure. When feature selection was performed using only the design samples, the hold-out estimate was always pessimistically biased. When feature selection was performed using the entire finite sample space, the hold-out estimates could be pessimistically or optimistically biased, depending on the number of features available for selection, the number of available samples, and their statistical distribution. For our simulation conditions, these estimates were always pessimistically (conservatively) biased if the ratio of the total number of available samples per class to the number of available features was greater than five.

Adverse effects of increased body weight on quantitative measures of mammographic image quality.

OBJECTIVE: The purpose of our study was to show that compressed breast thickness on mammograms in overweight and obese women exceeds the thickness in normal-weight women and that increased thickness results in image degradation. SUBJECTS AND METHODS: Three hundred consecutive routine mammograms were reviewed. Patients were categorized according to body mass index. Compression thickness, compressive force, kilovoltage, and milliampere-seconds were recorded. Geometric unsharpness and contrast degradation were calculated for each body mass index category. RESULTS: Body mass index categories were lean (3%), normal (36%), overweight (36%), and obese (25%). Body mass index was directly correlated with compressed thickness. In the mediolateral oblique view, the mean thickness of the obese category exceeded normal thickness by 18 mm (p < 0.01), corresponding to a 32% increase in geometric unsharpness. Mean obese thickness exceeded lean thickness by 33 mm (p < 0.01), corresponding to a 79% increase in unsharpness. Similar trends were observed for the craniocaudal view. In the mediolateral oblique projection, there was an increase of 1.0 kVp (p < 0.01) for obese compared with normal and 1.7 kVp (p < 0.01) between lean and obese, corresponding, respectively, to a 16% and a 25% decrease in image contrast because of scatter and kilovoltage changes. Milliampere-seconds increased by 47% on the mediolateral oblique images in the obese category compared with normal body mass index. CONCLUSION: An increased body mass index was associated with greater compressed breast thickness, resulting in increased geometric unsharpness, decreased image contrast, and greater potential for motion unsharpness.

Stereomammography: evaluation of depth perception using a virtual 3D cursor.

We are evaluating the usefulness of stereomammography in improving breast cancer diagnosis. One area that we are investigating is whether the improved depth perception associated with stereomammography might be significantly enhanced with the use of a virtual 3D cursor. A study was performed to evaluate the accuracy of absolute depth measurements made in stereomammograms with such a cursor. A biopsy unit was used to produce digital stereo images of a phantom containing 50 low contrast fibrils (0.5 mm diam monofilaments) at depths ranging from 1 to 11 mm, with a minimum spacing of 2 mm. Half of the fibrils were oriented perpendicular (vertical) and half parallel (horizontal) to the stereo shift direction. The depth and orientation of each fibril were randomized, and the horizontal and vertical fibrils crossed, simulating overlapping structures in a breast image. Left and right eye images were generated by shifting the x-ray tube from +2.5 degrees to -2.5 degrees relative to the image receptor. Three observers viewed these images on a computer display with stereo glasses and adjusted the position of a cross-shaped virtual cursor to best match the perceived location of each fibril. The x, y, and z positions of the cursor were indicated on the display. The z (depth) coordinate was separately calibrated using known positions of fibrils in the phantom. The observers analyzed images of two configurations of the phantom. Thus, each observer made 50 vertical filament depth measurements and 50 horizontal filament depth measurements. These measurements were compared with the true depths. The correlation coefficients between the measured and true depths of the vertically oriented fibrils for the three observers were 0.99, 0.97, and 0.89 with standard errors of the estimates of 0.39 mm, 0.83 mm, and 1.33 mm, respectively. Corresponding values for the horizontally oriented fibrils were 0.91, 0.28, and 0.08, and 1.87 mm, 4.19 mm, and 3.13 mm. All observers could estimate the absolute depths of vertically oriented objects fairly accurately in digital stereomammograms; however, only one observer was able to accurately estimate the depths of horizontally oriented objects. This may relate to different aptitudes for stereoscopic visualization. The orientations of most objects in actual mammograms are combinations of horizontal and vertical. Further studies are planned to evaluate absolute depth measurements of fibrils oriented at various intermediate angles and of objects of different shapes. The effects of the shape and contrast of the virtual cursor and the stereo shift angle on the accuracy of the depth measurements will also be investigated.

Application of artificial neural networks to modeling and prediction of ambient ozone concentrations.

The deterministic modeling of ambient O3 concentrations is difficult because of the complexity of the atmospheric system in terms of the number of chemical species; the availability of accurate, time-resolved emissions data; and the required rate constants. However, other complex systems have been successfully approximated using artificial neural networks (ANNs). In this paper, ANNs are used to model and predict ambient O3 concentrations based on a limited number of measured hydrocarbon species, NOx compounds, temperature, and radiant energy. In order to examine the utility of these approaches, data from the Coastal Oxidant Assessment for Southeast Texas (COAST) program in Houston, TX, have been used. In this study, 53 hydrocarbon compounds, along with O3, nitrogen oxides, and meteorological data were continuously measured during summer 1993. Steady-state ANN models were developed to examine the ability of these models to predict current O3 concentrations from measured VOC and NOx concentrations. To predict the future concentrations of O3, dynamic models were also explored and were used for extraction of chemical information such as reactivity estimations for the VOC species. The steady-state model produced an approximation of O3 data and demonstrated the functional relationship between O3 and VOC-NOx concentrations. The dynamic models were able to the adequately predict the O3 concentration and behavior of VOC-NOx-O3 system a number of hourly intervals into the future. For 3 hr into the future, O3 concentration could be predicted with a root-mean squared error (RMSE) of 8.21 ppb. Extending the models further in time led to an RMSE of 11.46 ppb for 5-hr-ahead values. This prediction capability could be useful in determining when control actions are needed to maintain measured concentrations within acceptable value ranges.

Classification of malignant and benign masses based on hybrid ART2LDA approach.

A new type of classifier combining an unsupervised and a supervised model was designed and applied to classification of malignant and benign masses on mammograms. The unsupervised model was based on an adaptive resonance theory (ART2) network which clustered the masses into a number of separate classes. The classes were divided into two types: one containing only malignant masses and the other containing a mix of malignant and benign masses. The masses from the malignant classes were classified by ART2. The masses from the mixed classes were input to a supervised linear discriminant classifier (LDA). In this way, some malignant masses were separated and classified by ART2 and the less distinguishable benign and malignant masses were classified by LDA. For the evaluation of classifier performance, 348 regions of interest (ROI's) containing biopsy proven masses (169 benign and 179 malignant) were used. Ten different partitions of training and test groups were randomly generated using an average of 73% of ROI's for training and 27% for testing. Classifier design, including feature selection and weight optimization, was performed with the training group. The test group was kept independent of the training group. The performance of the hybrid classifier was compared to that of an LDA classifier alone and a backpropagation neural network (BPN). Receiver operating characteristics (ROC) analysis was used to evaluate the accuracy of the classifiers. The average area under the ROC curve (A(z)) for the hybrid classifier was 0.81 as compared to 0.78 for the LDA and 0.80 for the BPN. The partial areas above a true positive fraction of 0.9 were 0.34, 0.27 and 0.31 for the hybrid, the LDA and the BPN classifier, respectively. These results indicate that the hybrid classifier is a promising approach for improving the accuracy of classification in CAD applications.