CAD for mammography: the technique, results, current role and further developments

A Statistical Approach to Assess the Robustness of Radiomics Features in the Discrimination of Mammographic Lesions

Despite mammography (MG) being among the most widespread techniques in breast cancer screening, tumour detection and classification remain challenging tasks due to the high morphological variability of the lesions. The extraction of radiomics features has proved to be a promising approach in MG. However, radiomics features can suffer from dependency on factors such as acquisition protocol, segmentation accuracy, feature extraction and engineering methods, which prevent the implementation of robust and clinically reliable radiomics workflow in MG. In this study, the variability and robustness of radiomics features is investigated as a function of lesion segmentation in MG images from a public database. A statistical analysis is carried out to assess feature variability and a radiomics robustness score is introduced based on the significance of the statistical tests performed. The obtained results indicate that variability is observable not only as a function of the abnormality type (calcification and masses), but also among feature categories (first-order and second-order), image view (craniocaudal and medial lateral oblique), and the type of lesions (benign and malignant). Furthermore, through the proposed approach, it is possible to identify those radiomics characteristics with a higher discriminative power between benign and malignant lesions and a lower dependency on segmentation, thus suggesting the most appropriate choice of robust features to be used as inputs to automated classification algorithms.

Physical imaging parameter variation drives domain shift

Statistical learning algorithms strongly rely on an oversimplified assumption for optimal performance, that is, source (training) and target (testing) data are independent and identically distributed. Variation in human tissue, physician labeling and physical imaging parameters (PIPs) in the generative process, yield medical image datasets with statistics that render this central assumption false. When deploying models, new examples are often out of distribution with respect to training data, thus, training robust dependable and predictive models is still a challenge in medical imaging with significant accuracy drops common for deployed models. This statistical variation between training and testing data is referred to as domain shift (DS).To the best of our knowledge we provide the first empirical evidence that variation in PIPs between test and train medical image datasets is a significant driver of DS and model generalization error is correlated with this variance. We show significant covariate shift occurs due to a selection bias in sampling from a small area of PIP space for both inter and intra-hospital regimes. In order to show this, we control for population shift, prevalence shift, data selection biases and annotation biases to investigate the sole effect of the physical generation process on model generalization for a proxy task of age group estimation on a combined 44 k image mammogram dataset collected from five hospitals.We hypothesize that training data should be sampled evenly from PIP space to produce the most robust models and hope this study provides motivation to retain medical image generation metadata that is almost always discarded or redacted in open source datasets. This metadata measured with standard international units can provide a universal regularizing anchor between distributions generated across the world for all current and future imaging modalities.

ABCanDroid: A Cloud Integrated Android App for Noninvasive Early Breast Cancer Detection Using Transfer Learning

Many patients affected by breast cancer die every year because of improper diagnosis and treatment. In recent years, applications of deep learning algorithms in the field of breast cancer detection have proved to be quite efficient. However, the application of such techniques has a lot of scope for improvement. Major works have been done in this field, however it can be made more efficient by the use of transfer learning to get impressive results. In the proposed approach, Convolutional Neural Network (CNN) is complemented with Transfer Learning for increasing the efficiency and accuracy of early detection of breast cancer for better diagnosis. The thought process involved using a pre-trained model, which already had some weights assigned rather than building the complete model from scratch. This paper mainly focuses on ResNet101 based Transfer Learning Model paired with the ImageNet dataset. The proposed framework provided us with an accuracy of 99.58%. Extensive experiments and tuning of hyperparameters have been performed to acquire the best possible results in terms of classification. The proposed frameworks aims to be an efficient tool for all doctors and society as a whole and help the user in early detection of breast cancer.

Detection of masses in mammograms using a one-stage object detector based on a deep convolutional neural network

Several computer aided diagnosis (CAD) systems have been developed for mammography. They are widely used in certain countries such as the U.S. where mammography studies are conducted more frequently; however, they are not yet globally employed for clinical use due to their inconsistent performance, which can be attributed to their reliance on hand-crafted features. It is difficult to use hand-crafted features for mammogram images that vary due to factors such as the breast density of patients and differences in imaging devices. To address these problems, several studies have leveraged a deep convolutional neural network that does not require hand-crafted features. Among the recent object detectors, RetinaNet is particularly promising as it is a simpler one-stage object detector that is fast and efficient while achieving state-of-the-art performance. RetinaNet has been proven to perform conventional object detection tasks but has not been tested on detecting masses in mammograms. Thus, we propose a mass detection model based on RetinaNet. To validate its performance in diverse use cases, we construct several experimental setups using the public dataset INbreast and the in-house dataset GURO. In addition to training and testing on the same dataset (i.e., training and testing on INbreast), we evaluate our mass detection model in setups using additional training data (i.e., training on INbreast + GURO and testing on INbreast). We also evaluate our model in setups using pre-trained weights (i.e., using weights pre-trained on GURO, training and testing on INbreast). In all the experiments, our mass detection model achieves comparable or better performance than more complex state-of-the-art models including the two-stage object detector. Also, the results show that using the weights pre-trained on datasets achieves similar performance as directly using datasets in the training phase. Therefore, we make our mass detection model's weights pre-trained on both GURO and INbreast publicly available. We expect that researchers who train RetinaNet on their in-house dataset for the mass detection task can use our pre-trained weights to leverage the features extracted from the datasets.

Minimization of annotation work: diagnosis of mammographic masses via active learning

The prerequisite for establishing an effective prediction system for mammographic diagnosis is the annotation of each mammographic image. The manual annotation work is time-consuming and laborious, which becomes a great hindrance for researchers. In this article, we propose a novel active learning algorithm that can adequately address this problem, leading to the minimization of the labeling costs on the premise of guaranteed performance. Our proposed method is different from the existing active learning methods designed for the general problem as it is specifically designed for mammographic images. Through its modified discriminant functions and improved sample query criteria, the proposed method can fully utilize the pairing of mammographic images and select the most valuable images from both the mediolateral and craniocaudal views. Moreover, in order to extend active learning to the ordinal regression problem, which has no precedent in existing studies, but is essential for mammographic diagnosis (mammographic diagnosis is not only a classification task, but also an ordinal regression task for predicting an ordinal variable, viz. the malignancy risk of lesions), multiple sample query criteria need to be taken into consideration simultaneously. We formulate it as a criteria integration problem and further present an algorithm based on self-adaptive weighted rank aggregation to achieve a good solution. The efficacy of the proposed method was demonstrated on thousands of mammographic images from the digital database for screening mammography. The labeling costs of obtaining optimal performance in the classification and ordinal regression task respectively fell to 33.8 and 19.8 percent of their original costs. The proposed method also generated 1228 wins, 369 ties and 47 losses for the classification task, and 1933 wins, 258 ties and 185 losses for the ordinal regression task compared to the other state-of-the-art active learning algorithms. By taking the particularities of mammographic images, the proposed AL method can indeed reduce the manual annotation work to a great extent without sacrificing the performance of the prediction system for mammographic diagnosis.

Classifying symmetrical differences and temporal change for the detection of malignant masses in mammography using deep neural networks

We investigate the addition of symmetry and temporal context information to a deep convolutional neural network (CNN) with the purpose of detecting malignant soft tissue lesions in mammography. We employ a simple linear mapping that takes the location of a mass candidate and maps it to either the contralateral or prior mammogram, and regions of interest (ROIs) are extracted around each location. Two different architectures are subsequently explored: (1) a fusion model employing two datastreams where both ROIs are fed to the network during training and testing and (2) a stagewise approach where a single ROI CNN is trained on the primary image and subsequently used as a feature extractor for both primary and contralateral or prior ROIs. A "shallow" gradient boosted tree classifier is then trained on the concatenation of these features and used to classify the joint representation. The baseline yielded an AUC of 0.87 with confidence interval [0.853, 0.893]. For the analysis of symmetrical differences, the first architecture where both primary and contralateral patches are presented during training obtained an AUC of 0.895 with confidence interval [0.877, 0.913], and the second architecture where a new classifier is retrained on the concatenation an AUC of 0.88 with confidence interval [0.859, 0.9]. We found a significant difference between the first architecture and the baseline at high specificity with [Formula: see text]. When using the same architectures to analyze temporal change, we yielded an AUC of 0.884 with confidence interval [0.865, 0.902] for the first architecture and an AUC of 0.879 with confidence interval [0.858, 0.898] in the second setting. Although improvements for temporal analysis were consistent, they were not found to be significant. The results show our proposed method is promising and we suspect performance can greatly be improved when more temporal data become available.

Large scale deep learning for computer aided detection of mammographic lesions

Recent advances in machine learning yielded new techniques to train deep neural networks, which resulted in highly successful applications in many pattern recognition tasks such as object detection and speech recognition. In this paper we provide a head-to-head comparison between a state-of-the art in mammography CAD system, relying on a manually designed feature set and a Convolutional Neural Network (CNN), aiming for a system that can ultimately read mammograms independently. Both systems are trained on a large data set of around 45,000 images and results show the CNN outperforms the traditional CAD system at low sensitivity and performs comparable at high sensitivity. We subsequently investigate to what extent features such as location and patient information and commonly used manual features can still complement the network and see improvements at high specificity over the CNN especially with location and context features, which contain information not available to the CNN. Additionally, a reader study was performed, where the network was compared to certified screening radiologists on a patch level and we found no significant difference between the network and the readers.

Discrimination of Breast Cancer with Microcalcifications on Mammography by Deep Learning

Microcalcification is an effective indicator of early breast cancer. To improve the diagnostic accuracy of microcalcifications, this study evaluates the performance of deep learning-based models on large datasets for its discrimination. A semi-automated segmentation method was used to characterize all microcalcifications. A discrimination classifier model was constructed to assess the accuracies of microcalcifications and breast masses, either in isolation or combination, for classifying breast lesions. Performances were compared to benchmark models. Our deep learning model achieved a discriminative accuracy of 87.3% if microcalcifications were characterized alone, compared to 85.8% with a support vector machine. The accuracies were 61.3% for both methods with masses alone and improved to 89.7% and 85.8% after the combined analysis with microcalcifications. Image segmentation with our deep learning model yielded 15, 26 and 41 features for the three scenarios, respectively. Overall, deep learning based on large datasets was superior to standard methods for the discrimination of microcalcifications. Accuracy was increased by adopting a combinatorial approach to detect microcalcifications and masses simultaneously. This may have clinical value for early detection and treatment of breast cancer.

Role of computer-aided detection in very small screening detected invasive breast cancers

This study aims to assess computer-aided detection (CAD) performance with full-field digital mammography (FFDM) in very small (equal to or less than 1 cm) invasive breast cancers. Sixty-eight invasive breast cancers less than or equal to 1 cm were retrospectively studied. All cases were detected with FFDM in women aged 49-69 years from our breast cancer screening program. Radiological characteristics of lesions following BI-RADS descriptors were recorded and compared with CAD sensitivity. Age, size, BI-RADS classification, breast density type, histological type of the neoplasm, and role of the CAD were also assessed. Per-study specificity and mass false-positive rate were determined by using 100 normal consecutive studies. Thirty-seven (54.4 %) masses, 17 (25 %) calcifications, 6 (8.8 %) masses with calcifications, 7 (10.3 %) architectural distortions, and 1 asymmetry (1.5 %) were found. CAD showed an overall sensitivity of 86.7 % (masses, 86.5 %; calcifications, 100 %; masses with calcifications, 100 %; and architectural distortion, 57.14 %), CAD failed to detect 9 out of 68 cases: 5 of 37 masses, 3 of 7 architectural distortions, and 1 of 1 asymmetry. Fifteen out of 37 masses were hyperdense, and all of them were detected by CAD. No association was seen among mass morphology or margins and detectability. Per-study specificity and CAD false-positive rate was 26 % and 1.76 false marks per study. In conclusion, CAD shows a high sensitivity and a low specificity. Lesion size, histology, and breast density do not influence sensitivity. Mammographic features, mass density, and thickness of the spicules in architectural distortions do influence.

Computer-aided detection of cancer in automated 3-D breast ultrasound

Automated 3-D breast ultrasound (ABUS) has gained a lot of interest and may become widely used in screening of dense breasts, where sensitivity of mammography is poor. However, reading ABUS images is time consuming, and subtle abnormalities may be missed. Therefore, we are developing a computer aided detection (CAD) system to help reduce reading time and prevent errors. In the multi-stage system we propose, segmentations of the breast, the nipple and the chestwall are performed, providing landmarks for the detection algorithm. Subsequently, voxel features characterizing coronal spiculation patterns, blobness, contrast, and depth are extracted. Using an ensemble of neural-network classifiers, a likelihood map indicating potential abnormality is computed. Local maxima in the likelihood map are determined and form a set of candidates in each image. These candidates are further processed in a second detection stage, which includes region segmentation, feature extraction and a final classification. On region level, classification experiments were performed using different classifiers including an ensemble of neural networks, a support vector machine, a k-nearest neighbors, a linear discriminant, and a gentle boost classifier. Performance was determined using a dataset of 238 patients with 348 images (views), including 169 malignant and 154 benign lesions. Using free response receiver operating characteristic (FROC) analysis, the system obtains a view-based sensitivity of 64% at 1 false positives per image using an ensemble of neural-network classifiers.

Breast image analysis for risk assessment, detection, diagnosis, and treatment of cancer

The role of breast image analysis in radiologists' interpretation tasks in cancer risk assessment, detection, diagnosis, and treatment continues to expand. Breast image analysis methods include segmentation, feature extraction techniques, classifier design, biomechanical modeling, image registration, motion correction, and rigorous methods of evaluation. We present a review of the current status of these task-based image analysis methods, which are being developed for the various image acquisition modalities of mammography, tomosynthesis, computed tomography, ultrasound, and magnetic resonance imaging. Depending on the task, image-based biomarkers from such quantitative image analysis may include morphological, textural, and kinetic characteristics and may depend on accurate modeling and registration of the breast images. We conclude with a discussion of future directions.

CAD e mamografia: por que usar esta ferramenta?

A mamografia representa o melhor método de detecção precoce do câncer de mama, porém cerca de 10% a 30% das lesões mamárias são perdidas no rastreamento, devido a limitações próprias dos observadores humanos. A detecção auxiliada por computador (computer-aided detection - CAD) é uma tecnologia relativamente nova que tem sido implementada em alguns serviços de mamografia, com o intuito de prover uma dupla leitura. Estudos clínicos têm demonstrado que o CAD aumenta a sensibilidade de detecção do câncer da mama, por radiologistas, em até 21%. Um sistema CAD é útil em situações em que exista alta variabilidade interobservador, falta de observadores treinados, ou na impossibilidade de se realizar a dupla leitura com dois ou mais radiologistas. O objetivo desta revisão está baseado na necessidade de atualizar a comunidade médica acerca desta ferramenta, como um método auxiliar, quantitativo, não operador-dependente, e que visa a melhorar a qualidade do diagnóstico do câncer de mama.

Recommendations for research design and reporting in computer-assisted diagnosis to facilitate meta-analysis

Computer-assisted diagnosis (CAD) describes a diverse, heterogeneous range of applications rather than a single entity. The aims and functions of CAD systems vary considerably and comparing studies and systems is challenging due to methodological and design differences. In addition, poor study quality and reporting can reduce the value of some publications. Meta-analyses of CAD are therefore difficult and may not provide reliable conclusions. Aiming to determine the major sources of heterogeneity and thereby what CAD researchers could change to allow this sort of assessment, this study reviews a sample of 147 papers concerning CAD used with imaging for cancer diagnosis. It discusses sources of variability, including the goal of the CAD system, learning methodology, study population, design, outcome measures, inclusion of radiologists, and study quality. Based upon this evidence, recommendations are made to help researchers optimize the quality and comparability of their trial design and reporting.

Is computer aided detection (CAD) cost effective in screening mammography? A model based on the CADET II study

BACKGROUND

Single reading with computer aided detection (CAD) is an alternative to double reading for detecting cancer in screening mammograms. The aim of this study is to investigate whether the use of a single reader with CAD is more cost-effective than double reading.

METHODS

Based on data from the CADET II study, the cost-effectiveness of single reading with CAD versus double reading was measured in terms of cost per cancer detected. Cost (Pound (£), year 2007/08) of single reading with CAD versus double reading was estimated assuming a health and social service perspective and a 7 year time horizon. As the equipment cost varies according to the unit size a separate analysis was conducted for high, average and low volume screening units. One-way sensitivity analyses were performed by varying the reading time, equipment and assessment cost, recall rate and reader qualification.

RESULTS

CAD is cost increasing for all sizes of screening unit. The introduction of CAD is cost-increasing compared to double reading because the cost of CAD equipment, staff training and the higher assessment cost associated with CAD are greater than the saving in reading costs. The introduction of single reading with CAD, in place of double reading, would produce an additional cost of £227 and £253 per 1,000 women screened in high and average volume units respectively. In low volume screening units, the high cost of purchasing the equipment will results in an additional cost of £590 per 1,000 women screened.One-way sensitivity analysis showed that the factors having the greatest effect on the cost-effectiveness of CAD with single reading compared with double reading were the reading time and the reader's professional qualification (radiologist versus advanced practitioner).

CONCLUSIONS

Without improvements in CAD effectiveness (e.g. a decrease in the recall rate) CAD is unlikely to be a cost effective alternative to double reading for mammography screening in UK. This study provides updated estimates of CAD costs in a full-field digital system and assessment cost for women who are re-called after initial screening. However, the model is highly sensitive to various parameters e.g. reading time, reader qualification, and equipment cost.

Downgrading BIRADS 3 to BIRADS 2 category using a computer-aided microcalcification analysis and risk assessment system for early breast cancer

This paper explores the potential of a computer-aided diagnosis system to discriminate the real benign microcalcifications among a specific subset of 109 patients with BIRADS 3 mammograms who had undergone biopsy, thus making it possible to downgrade them to BIRADS 2 category. The system detected and quantified critical features of microcalcifications and classified them on a risk percentage scale for malignancy. The system successfully detected all cancers. Nevertheless, it suggested biopsy for 11/15 atypical lesions. Finally, the system characterized as definitely benign (BIRADS 2) 29/88 benign lesions, previously assigned to BIRADS 3, and thus achieved a reduction of 33% in unnecessary biopsies.

Invasive breast cancers detected by screening mammography: a detailed comparison of computer-aided detection-assisted single reading and double reading

To compare double reading plus arbitration for discordance, (currently best practice, (BP)) with computer-aided-detection (CAD)-assisted single reading (CAD-R) for detection of invasive cancers detected within BreastScreen Australia. Secondarily, to examine characteristics of cancers detected/rejected using each method. Mammograms of 157 randomly selected double-read invasive cancers were mixed 1:9 with normal cancers (total 1569), all detected in a BreastScreen service. Cancers were detected by two readers or one reader (C2 and C1 cancers, ratio 70:30%) in the program. The 1569 film-screen mammograms were read by two radiologists (reader A (RA) and reader B(RB)), with findings recorded before and after CAD. Discordant findings with BP were resolved by arbitration. We compared CAD-assisted reading (CAD-RA, CAD-RB) with BP, and CAD and arbitration contribution to findings. We correlated cancer size, sensitivity and mammographic density with detection methods. BP sensitivity 90.4% compared with CAD-RA sensitivity 86.6% (P = 0.12) and CAD-RB 94.3% (P = 0.14). CAD-RB specificity was less than BP (P = 0.01). CAD sensitivity was 93%, but readers rejected most positive CAD prompts. After CAD, reader's sensitivity increased 1.9% and specificity dropped 0.2% and 0.8%. Arbitration decreased specificity 4.7%. Receiving operator curves analysis demonstrated BP accuracy better than CAD-RA, borderline significance (P = 0.07), but not CAD-RB. Secondarily, cancer size was similar for BP and CAD-R. Cancers recalled after arbitration (P = 0.01) and CAD-R (P = 0.10) were smaller. No difference in cancer size or sensitivity between reading methods was found with increasing breast density. CAD-R and BP sensitivity and cancer detection size were not significantly different. CAD-R specificity was significantly lower for one reader.

Investigation of reading mode and relative sensitivity as factors that influence reader performance when using computer-aided detection software

RATIONALE AND OBJECTIVES

The aim of this study was to investigate the effects of relative sensitivity (reader without computer-aided detection [CAD] vs stand-alone CAD) and reading mode on reader performance when using CAD software.

MATERIALS AND METHODS

Two sets of 100 images (low-contrast and high-contrast sets) were created by adding low-contrast or high-contrast simulated masses to random locations in 100 normal mammograms. This produced a relative sensitivity, substantially less for the low-contrast set and similar for the high-contrast set. Seven readers reviewed every image in each set and specified location and probability scores using three reading modes (without CAD, second read with CAD, and concurrent read with CAD). Reader detection accuracy was analyzed using areas under free-response receiver operating characteristic curves, sensitivity, and the number of false-positive findings per image.

RESULTS

For the low-contrast set, average differences in areas under free-response receiver operating characteristic curves, sensitivity, and false-positive findings per image without CAD were 0.02, 0.12, and 0.11, respectively, compared to second read and 0.05, 0.17, and 0.09 (not statistically significant), respectively, compared to concurrent read. For the high-contrast set, average differences were 0.002 (not statistically significant), 0.04, and 0.05, respectively, compared to second read and -0.004 (not statistically significant), 0.04, and 0.08 (not statistically significant), respectively, compared to concurrent read (all differences were statistically significant except as noted). Differences were greater in the low-contrast set than the high-contrast set. Differences between second read and concurrent read were not significant.

CONCLUSIONS

Relative sensitivity is a critical factor that determines incremental improvement in reader performance when using CAD and appears to be more important than reading mode. Relative sensitivity may determine the clinical usefulness of CAD in different clinical applications and for different types of users.

Mammographic features and histopathological findings of interval breast cancers

BACKGROUND

Interval cancers are considered a shortcoming in screening mammography due to less favorable prognostic tumor characteristics compared to screening-detected cancers and consequently a lower chance of survival from the disease.

PURPOSE

To describe the mammographic features and prognostic histopathological tumor characteristics of interval breast cancers.

MATERIAL AND METHODS

A total of 231 interval breast cancer cases diagnosed in prevalently screened women aged 50-69 years old were examined. Thirty-five percent of the cases were retrospectively classified as missed cancers, 23% as minimal sign, and 42% as true negative (including occult cancers) in a definitive classification performed by six experienced breast radiologists. The retrospective classification described the mammographic features of the baseline screening mammograms in missed and minimal-sign interval cancers, while histopathological reports were used to describe the tumor characteristics in all the subgroups of interval cancers.

RESULTS

Fifty percent of the missed and minimal-sign interval cancers combined presented poorly defined mass or asymmetric density, and 26% calcifications with or without associated density or mass at baseline screening. Twenty-seven percent of invasive tumors were <15 mm for missed and 47% for true interval cancers (P<0.001). Lymph node involvement was more common in missed (49%) compared with the true cases (33%, P<0.05).

CONCLUSION

Missed interval cancers have less prognostically favorable histopathological tumor characteristics compared with true interval cancers. Improving the radiologist's perception and interpretation by establishing systematic collection of features and implementation of organized reviews may decrease the number of interval cancers in a screening program.

Single reading with computer-aided detection for screening mammography

BACKGROUND

The sensitivity of screening mammography for the detection of small breast cancers is higher when the mammogram is read by two readers rather than by a single reader. We conducted a trial to determine whether the performance of a single reader using a computer-aided detection system would match the performance achieved by two readers.

METHODS

The trial was designed as an equivalence trial, with matched-pair comparisons between the cancer-detection rates achieved by single reading with computer-aided detection and those achieved by double reading. We randomly assigned 31,057 women undergoing routine screening by film mammography at three centers in England to double reading, single reading with computer-aided detection, or both double reading and single reading with computer-aided detection, at a ratio of 1:1:28. The primary outcome measures were the proportion of cancers detected according to regimen and the recall rates within the group receiving both reading regimens.

RESULTS

The proportion of cancers detected was 199 of 227 (87.7%) for double reading and 198 of 227 (87.2%) for single reading with computer-aided detection (P=0.89). The overall recall rates were 3.4% for double reading and 3.9% for single reading with computer-aided detection; the difference between the rates was small but significant (P<0.001). The estimated sensitivity, specificity, and positive predictive value for single reading with computer-aided detection were 87.2%, 96.9%, and 18.0%, respectively. The corresponding values for double reading were 87.7%, 97.4%, and 21.1%. There were no significant differences between the pathological attributes of tumors detected by single reading with computer-aided detection alone and those of tumors detected by double reading alone.

CONCLUSIONS

Single reading with computer-aided detection could be an alternative to double reading and could improve the rate of detection of cancer from screening mammograms read by a single reader. (ClinicalTrials.gov number, NCT00450359.)

Variable size computer-aided detection prompts and mammography film reader decisions

Introduction

The purpose of the present study was to investigate the effect of computer-aided detection (CAD) prompts on reader behaviour in a large sample of breast screening mammograms by analysing the relationship of the presence and size of prompts to the recall decision.

Methods

Local research ethics committee approval was obtained; informed consent was not required. Mammograms were obtained from women attending routine mammography at two breast screening centres in 1996. Films, previously double read, were re-read by a different reader using CAD. The study material included 315 cancer cases comprising all screen-detected cancer cases, all subsequent interval cancers and 861 normal cases randomly selected from 10,267 cases. Ground truth data were used to assess the efficacy of CAD prompting. Associations between prompt attributes and tumour features or reader recall decisions were assessed by chi-squared tests.

Results

There was a highly significant relationship between prompting and a decision to recall for cancer cases and for a random sample of normal cases (P < 0.001). Sixty-four per cent of all cases contained at least one CAD prompt. In cancer cases, larger prompts were more likely to be recalled (P = 0.02) for masses but there was no such association for calcifications (P = 0.9). In a random sample of 861 normal cases, larger prompts were more likely to be recalled (P = 0.02) for both mass and calcification prompts. Significant associations were observed with prompting and breast density (p = 0.009) for cancer cases but not for normal cases (P = 0.05).

Conclusions

For both normal cases and cancer cases, prompted mammograms were more likely to be recalled and the prompt size was also associated with a recall decision.

Digital mammography: what do we and what don't we know?

High-quality full-field digital mammography has been available now for several years and is increasingly used for both diagnostic and screening mammography. A number of different detector technologies exist, which all have their specific advantages and disadvantages. Diagnostic accuracy of digital mammography has been shown to be at least equivalent to film-screen mammography in a general screening population. Digital mammography is superior to screen-film mammography in younger women with dense breasts due to its ability to selectively optimize contrast in areas of dense parenchyma. This advantage is especially important in women with a genetic predisposition for breast cancer, where intensified early detection programs may have to start from 25 to 30 years of age. Tailored image processing and computer-aided diagnosis hold the potential to further improve the early detection of breast cancer. However, at present no consensus exists among radiologists on which processing is optimal for digital mammograms. Image processing may also vary significantly among vendors with so far limited interoperability. This review aims to summarize the available information regarding the impact of digital mammography on workflow and breast cancer diagnosis.

CAD in clinical trials: Current role and architectural requirements

Computer-aided diagnosis (CAD) technology is becoming an important tool to assess treatment response in clinical trials. However, CAD software alone is not sufficient to conduct an imaging-based clinical trial. There are a number of architectural requirements such as image receive (from multiple field sites), a database for storing quantitative measures, and data mining and reporting capabilities. In this paper we describe the architectural requirements to incorporate CAD into clinical trials and illustrate their functionality in therapeutic trials for emphysema.

Current status and future directions of computer-aided diagnosis in mammography

The concept of computer-aided detection (CADe) was introduced more than 50 years ago; however, only in the last 20 years there have been serious and successful attempts at developing CADe for mammography. CADe schemes have high sensitivity, but poor specificity compared to radiologists. CADe has been shown to help radiologists find more cancers both in observer studies and in clinical evaluations. Clinically, CADe increases the number of cancers detected by approximately 10%, which is comparable to double reading by two radiologists.