Performance change of mammographic CAD schemes optimized with most-recent and prior image databases

A Review of Computer-Aided Breast Cancer Diagnosis Using Sequential Mammograms

Radiologists assess the results of mammography, the key screening tool for the detection of breast cancer, to determine the presence of malignancy. They, routinely, compare recent and prior mammographic views to identify changes between the screenings. In case a new lesion appears in a mammogram, or a region is changing rapidly, it is more likely to be suspicious, compared to a lesion that remains unchanged and it is usually benign. However, visual evaluation of mammograms is challenging even for expert radiologists. For this reason, various Computer-Aided Diagnosis (CAD) algorithms are being developed to assist in the diagnosis of abnormal breast findings using mammograms. Most of the current CAD systems do so using only the most recent mammogram. This paper provides a review of the development of methods to emulate the radiological approach and perform automatic segmentation and/or classification of breast abnormalities using sequential mammogram pairs. It begins with demonstrating the importance of utilizing prior views in mammography, through the review of studies where the performance of expert and less-trained radiologists was compared. Following, image registration techniques and their application to mammography are presented. Subsequently, studies that implemented temporal analysis or subtraction of temporally sequential mammograms are summarized. Finally, a description of the open access mammography datasets is provided. This comprehensive review can serve as a thorough introduction to the use of prior information in breast cancer CAD systems but also provides indicative directions to guide future applications.

Applying artificial intelligence technology to assist with breast cancer diagnosis and prognosis prediction

Breast cancer remains the most diagnosed cancer in women. Advances in medical imaging modalities and technologies have greatly aided in the early detection of breast cancer and the decline of patient mortality rates. However, reading and interpreting breast images remains difficult due to the high heterogeneity of breast tumors and fibro-glandular tissue, which results in lower cancer detection sensitivity and specificity and large inter-reader variability. In order to help overcome these clinical challenges, researchers have made great efforts to develop computer-aided detection and/or diagnosis (CAD) schemes of breast images to provide radiologists with decision-making support tools. Recent rapid advances in high throughput data analysis methods and artificial intelligence (AI) technologies, particularly radiomics and deep learning techniques, have led to an exponential increase in the development of new AI-based models of breast images that cover a broad range of application topics. In this review paper, we focus on reviewing recent advances in better understanding the association between radiomics features and tumor microenvironment and the progress in developing new AI-based quantitative image feature analysis models in three realms of breast cancer: predicting breast cancer risk, the likelihood of tumor malignancy, and tumor response to treatment. The outlook and three major challenges of applying new AI-based models of breast images to clinical practice are also discussed. Through this review we conclude that although developing new AI-based models of breast images has achieved significant progress and promising results, several obstacles to applying these new AI-based models to clinical practice remain. Therefore, more research effort is needed in future studies.

Applying a new computer-aided detection scheme generated imaging marker to predict short-term breast cancer risk

This study aims to investigate the feasibility of identifying a new quantitative imaging marker based on false-positives generated by a computer-aided detection (CAD) scheme to help predict short-term breast cancer risk. An image dataset including four view mammograms acquired from 1044 women was retrospectively assembled. All mammograms were originally interpreted as negative by radiologists. In the next subsequent mammography screening, 402 women were diagnosed with breast cancer and 642 remained negative. An existing CAD scheme was applied 'as is' to process each image. From CAD-generated results, four detection features including the total number of (1) initial detection seeds and (2) the final detected false-positive regions, (3) average and (4) sum of detection scores, were computed from each image. Then, by combining the features computed from two bilateral images of left and right breasts from either craniocaudal or mediolateral oblique view, two logistic regression models were trained and tested using a leave-one-case-out cross-validation method to predict the likelihood of each testing case being positive in the next subsequent screening. The new prediction model yielded the maximum prediction accuracy with an area under a ROC curve of AUC  =  0.65  ±  0.017 and the maximum adjusted odds ratio of 4.49 with a 95% confidence interval of (2.95, 6.83). The results also showed an increasing trend in the adjusted odds ratio and risk prediction scores (p  <  0.01). Thus, this study demonstrated that CAD-generated false-positives might include valuable information, which needs to be further explored for identifying and/or developing more effective imaging markers for predicting short-term breast cancer risk.

A new and fast image feature selection method for developing an optimal mammographic mass detection scheme

PURPOSE

Selecting optimal features from a large image feature pool remains a major challenge in developing computer-aided detection (CAD) schemes of medical images. The objective of this study is to investigate a new approach to significantly improve efficacy of image feature selection and classifier optimization in developing a CAD scheme of mammographic masses.

METHODS

An image dataset including 1600 regions of interest (ROIs) in which 800 are positive (depicting malignant masses) and 800 are negative (depicting CAD-generated false positive regions) was used in this study. After segmentation of each suspicious lesion by a multilayer topographic region growth algorithm, 271 features were computed in different feature categories including shape, texture, contrast, isodensity, spiculation, local topological features, as well as the features related to the presence and location of fat and calcifications. Besides computing features from the original images, the authors also computed new texture features from the dilated lesion segments. In order to select optimal features from this initial feature pool and build a highly performing classifier, the authors examined and compared four feature selection methods to optimize an artificial neural network (ANN) based classifier, namely: (1) Phased Searching with NEAT in a Time-Scaled Framework, (2) A sequential floating forward selection (SFFS) method, (3) A genetic algorithm (GA), and (4) A sequential forward selection (SFS) method. Performances of the four approaches were assessed using a tenfold cross validation method.

RESULTS

Among these four methods, SFFS has highest efficacy, which takes 3%-5% of computational time as compared to GA approach, and yields the highest performance level with the area under a receiver operating characteristic curve (AUC) = 0.864 ± 0.034. The results also demonstrated that except using GA, including the new texture features computed from the dilated mass segments improved the AUC results of the ANNs optimized using other three feature selection methods. In addition, among 271 features, the shape, local morphological features, fat and calcification based features were the most frequently selected features to build ANNs.

CONCLUSIONS

Although conventional GA is a powerful tool in optimizing classifiers used in CAD schemes of medical images, it is very computationally intensive. This study demonstrated that using a new SFFS based approach enabled to significantly improve efficacy of image feature selection for developing CAD schemes.

Breast mass detection using slice conspicuity in 3D reconstructed digital breast volumes

In digital breast tomosynthesis, the three dimensional (3D) reconstructed volumes only provide quasi-3D structure information with limited resolution along the depth direction due to insufficient sampling in depth direction and the limited angular range. The limitation could seriously hamper the conventional 3D image analysis techniques for detecting masses because the limited number of projection views causes blurring in the out-of-focus planes. In this paper, we propose a novel mass detection approach using slice conspicuity in the 3D reconstructed digital breast volumes to overcome the above limitation. First, to overcome the limited resolution along the depth direction, we detect regions of interest (ROIs) on each reconstructed slice and separately utilize the depth directional information to combine the ROIs effectively. Furthermore, we measure the blurriness of each slice for resolving the degradation of performance caused by the blur in the out-of-focus plane. Finally, mass features are extracted from the selected in focus slices and analyzed by a support vector machine classifier to reduce the false positives. Comparative experiments have been conducted on a clinical data set. Experimental results demonstrate that the proposed approach outperforms the conventional 3D approach by achieving a high sensitivity with a small number of false positives.

Improving the performance of computer-aided detection of subtle breast masses using an adaptive cueing method

Current computer-aided detection (CAD) schemes for detecting mammographic masses have several limitations including high correlation with radiologists' detection and cueing most subtle masses only on one view. To increase CAD sensitivity in cueing more subtle masses that are likely missed and/or overlooked by radiologists without increasing false-positive rates, we investigated a new case-dependent cueing method by combining the original CAD-generated detection scores with a computed bilateral mammographic density asymmetry index. Using the new method, we adaptively raise the CAD-generated scores of the regions detected on 'high-risk' cases to cue more subtle mass regions and reduce the CAD scores of the regions detected on 'low-risk' cases to discard more false-positive regions. A testing dataset involving 78 positive and 338 negative cases was used to test this adaptive cueing method. Each positive case involves two sequential examinations in which the mass was detected in 'current' examination and missed in 'prior' examination but detected in a retrospective review by radiologists. Applying to this dataset, a pre-optimized CAD scheme yielded 75% case-based and 55% region-based sensitivity on 'current' examinations at a false-positive rate of 0.25 per image. CAD sensitivity was reduced to 42% (case based) and 27% (region based) on 'prior' examinations. Using the new cueing method, case-based and region-based sensitivity could maximally increase 9% and 33% on the 'prior' examinations, respectively. The percentages of the masses cued on two views also increased from 27% to 65%. The study demonstrated that using this adaptive cueing method enabled us to help CAD cue more subtle cancers without increasing the false-positive cueing rate.

Mass segmentation using a combined method for cancer detection

Background

Breast cancer is one of the leading causes of cancer death for women all over the world and mammography is thought of as one of the main tools for early detection of breast cancer. In order to detect the breast cancer, computer aided technology has been introduced. In computer aided cancer detection, the detection and segmentation of mass are very important. The shape of mass can be used as one of the factors to determine whether the mass is malignant or benign. However, many of the current methods are semi-automatic. In this paper, we investigate fully automatic segmentation method.

Results

In this paper, a new mass segmentation algorithm is proposed. In the proposed algorithm, a fully automatic marker-controlled watershed transform is proposed to segment the mass region roughly, and then a level set is used to refine the segmentation. For over-segmentation caused by watershed, we also investigated different noise reduction technologies. Images from DDSM were used in the experiments and the results show that the new algorithm can improve the accuracy of mass segmentation.

Conclusions

The new algorithm combines the advantages of both methods. The combination of the watershed based segmentation and level set method can improve the efficiency of the segmentation. Besides, the introduction of noise reduction technologies can reduce over-segmentation.

Detection of architectural distortion in prior mammograms

We present methods for the detection of sites of architectural distortion in prior mammograms of interval-cancer cases. We hypothesize that screening mammograms obtained prior to the detection of cancer could contain subtle signs of early stages of breast cancer, in particular, architectural distortion. The methods are based upon Gabor filters, phase portrait analysis, a novel method for the analysis of the angular spread of power, fractal analysis, Laws' texture energy measures derived from geometrically transformed regions of interest (ROIs), and Haralick's texture features. With Gabor filters and phase portrait analysis, 4224 ROIs were automatically obtained from 106 prior mammograms of 56 interval-cancer cases, including 301 true-positive ROIs related to architectural distortion, and from 52 mammograms of 13 normal cases. For each ROI, the fractal dimension, the entropy of the angular spread of power, 10 Laws' measures, and Haralick's 14 features were computed. The areas under the receiver operating characteristic curves obtained using the features selected by stepwise logistic regression and the leave-one-ROI-out method are 0.76 with the Bayesian classifier, 0.75 with Fisher linear discriminant analysis, and 0.78 with a single-layer feed-forward neural network. Free-response receiver operating characteristics indicated sensitivities of 0.80 and 0.90 at 5.8 and 8.1 false positives per image, respectively, with the Bayesian classifier and the leave-one-image-out method.

Computer-aided detection; the effect of training databases on detection of subtle breast masses

RATIONALE AND OBJECTIVES

Lesion conspicuity is typically highly correlated with visual difficulty for lesion detection, and computer-aided detection (CAD) has been widely used as a "second reader" in mammography. Hence, increasing CAD sensitivity in detecting subtle cancers without increasing false-positive rates is important. The aim of this study was to investigate the effect of training database case selection on CAD performance in detecting low-conspicuity breast masses.

MATERIALS AND METHODS

A full-field digital mammographic image database that included 525 cases depicting malignant masses was randomly partitioned into three subsets. A CAD scheme was applied to detect all initially suspected mass regions and compute region conspicuity. Training samples were iteratively selected from two of the subsets. Four types of training data sets-(1) one including all available true-positive mass regions in the two subsets ("all"), (2) one including 350 randomly selected mass regions ("diverse"), (3) one including 350 high-conspicuity mass regions ("easy"), and (4) one including 350 low-conspicuity mass regions ("difficult")-were assembled. In each training data set, the same number of randomly selected false-positive regions as the true-positives were also included. Two classifiers, an artificial neural network (ANN) and a k-nearest neighbor (KNN) algorithm, were trained using each of the four training data sets and tested on all suspected regions in the remaining data set. Using a threefold cross-validation method, the performance changes of the CAD schemes trained using one of the four training data sets were computed and compared.

RESULTS

CAD initially detected 1025 true-positive mass regions depicted on 507 cases (97% case-based sensitivity) and 9569 false-positive regions (3.5 per image) in the entire database. Using the all training data set, CAD achieved the highest overall performance on the entire testing database. However, CAD detected the highest number of low-conspicuity masses when the difficult training data set was used for training. Results did agree for both ANN-based and KNN-based classifiers in all tests. Compared to the use of the all training data set, the sensitivity of the schemes trained using the difficult data set decreased by 8.6% and 8.4% for the ANN and KNN algorithm on the entire database, respectively, but the detection of low-conspicuity masses increased by 7.1% and 15.1% for the ANN and KNN algorithm at a false-positive rate of 0.3 per image.

CONCLUSIONS

CAD performance depends on the size, diversity, and difficulty level of the training database. To increase CAD sensitivity in detecting subtle cancer, one should increase the fraction of difficult cases in the training database rather than simply increasing the training data set size.

Computer-aided detection of architectural distortion in prior mammograms of interval cancer

Architectural distortion is an important sign of breast cancer, but because of its subtlety, it is a common cause of false-negative findings on screening mammograms. This paper presents methods for the detection of architectural distortion in mammograms of interval cancer cases taken prior to the detection of breast cancer using Gabor filters, phase portrait analysis, fractal analysis, and texture analysis. The methods were used to detect initial candidates for sites of architectural distortion in prior mammograms of interval cancer and also normal control cases. A total of 4,224 regions of interest (ROIs) were automatically obtained from 106 prior mammograms of 56 interval cancer cases, including 301 ROIs related to architectural distortion, and from 52 prior mammograms of 13 normal cases. For each ROI, the fractal dimension and Haralick's texture features were computed. Feature selection was performed separately using stepwise logistic regression and stepwise regression. The best results achieved, in terms of the area under the receiver operating characteristics curve, with the features selected by stepwise logistic regression are 0.76 with the Bayesian classifier, 0.73 with Fisher linear discriminant analysis, 0.77 with an artificial neural network based on radial basis functions, and 0.77 with a support vector machine. Analysis of the performance of the methods with free-response receiver operating characteristics indicated a sensitivity of 0.80 at 7.6 false positives per image. The methods have good potential in detecting architectural distortion in mammograms of interval cancer cases.

Dual system approach to computer-aided detection of breast masses on mammograms

In this study, our purpose was to improve the performance of our mass detection system by using a new dual system approach which combines a computer-added detection (CAD) system optimized with "average" masses with another CAD system optimized with "subtle" masses. The two single CAD systems have similar image processing steps, which include prescreening, object segmentation, morphological and texture feature extraction, and false positive (FP) reduction by rule-based and linear discriminant analysis (LDA) classifiers. A feed-forward backpropagation artificial neural network was trained to merge the scores from the LDA classifiers in the two single CAD systems and differentiate true masses from normal tissue. For an unknown test mammogram, the two single CAD systems are applied to the image in parallel to detect suspicious objects. A total of three data sets were used for training and testing the systems. The first data set of 230 current mammograms, referred to as the average mass set, was collected from 115 patients. We also collected 264 mammograms, referred to as the subtle mass set, which were one to two years prior to the current exam from these patients. Both the average and the subtle mass sets were partitioned into two independent data sets in a cross validation training and testing scheme. A third data set containing 65 cases with 260 normal mammograms was used to estimate the FP marker rates during testing. When the single CAD system trained on the average mass set was applied to the test set with average masses, the FP marker rates were 2.2, 1.8, and 1.5 per image at the case-based sensitivities of 90%, 85%, and 80%, respectively. With the dual CAD system, the FP marker rates were reduced to 1.2, 0.9, and 0.7 per image, respectively, at the same case-based sensitivities. Statistically significant (p < 0.05) improvements on the free response receiver operating characteristic curves were observed when the dual system and the single system were compared using the test sets with either average masses or subtle masses.

Comparison of similarity measures for the task of template matching of masses on serial mammograms

We conducted a study to evaluate the effectiveness of twelve different similarity measures in matching the corresponding masses on temporal pairs of current and prior mammograms. To perform this comparison we implemented each of the twelve similarity measures in the final stage of our multistage registration technique for automated registration of breast lesions in serial mammograms. The multistage technique consists of three stages. In the first stage an initial fan-shape search region was estimated on the prior mammogram based on the geometrical position of the mass on the current mammogram. In the second stage, the location of the fan-shape region was refined by warping, based on an affine transformation and simplex optimization. A new refined search region was defined on the prior mammogram. 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. Our data set consisted of 318 temporal pairs. We performed three experiments, using a different subset of the 318 temporal pairs for each experiment. In each experiment we further tested how the performance of the similarity measures varied as the size of the search region increased or decreased. We evaluated the twelve similarity measures based on four criteria. The first criterion was the mean Euclidean distance, which was the average distance of the true location of the mass to the location detected by the similarity measure. The second criterion was the percentage of temporal pairs that were aligned so that 50% or more of the lesion area overlapped. The third criterion was the percentage of pairs that were aligned so that 75% or more of the lesion area overlapped. The fourth and final criterion was the robustness of the similarity measure. Our results showed that three of the similarity measures, Pearson's correlation, the cosine coefficient, and Goodman and Kruskal's Gamma coefficient, provide significantly higher accuracy (p < 0.05) in the task of matching the corresponding masses on serial mammograms than the other nine similarity measures.

Detection and classification performance levels of mammographic masses under different computer-aided detection cueing environments1

RATIONALE AND OBJECTIVES

The authors evaluated the impact of different computer-aided detection (CAD) cueing conditions on radiologists' performance levels in detecting and classifying masses depicted on mammograms.

MATERIALS AND METHODS

In an observer performance study, eight radiologists interpreted 110 subtle cases six times under different display conditions to detect depicted masses and classify them as benign or malignant. Forty-five cases depicted biopsy-proven masses and 65 were negative. One mass-based cueing sensitivity of 80% and two false-positive cueing rates of 1.2 and 0.5 per image were used in this study. In one mode, radiologists first interpreted images without CAD results, followed by the display of cues and reinterpretation. In another mode, radiologists viewed CAD cues as images were presented and then interpreted images. Free-response receiver operating characteristic method was used to analyze and compare detection performance. The receiver operating characteristic method was used to evaluate classification performance.

RESULTS

At these performance levels, providing cues after initial interpretation had little effect on the overall performance in detecting masses. However, in the mode with the highest false-positive cueing rate, viewing CAD cues immediately upon display of images significantly reduced average performance for both detection and classification tasks (P < .05). Viewing CAD cues during the initial display consistently resulted in fewer abnormalities being identified in noncued regions.

CONCLUSION

CAD systems with low sensitivity (< or = 80% on mass-based detection) and high false-positive rate (> or = 0.5 per image) in a dataset with subtle abnormalities had little effect on radiologists' performance in the detection and classification of mammographic masses.