Breast compression pressure (CP), computed as the force over the paddle contact area paddle, is an important measure of compression quality in mammography evidenced by associations with screening performance, including the odds of interval cancer compared to screen detected cancer. Here we introduce a novel algorithm to determine CP from processed images, the Processed Image Compression Pressure Estimator (PICPE). The aim is for PICPE outputs to align with those from an established method that estimates CP from unprocessed images such that results are comparable between image formats regardless of vendor or modality. Multiple datasets were assembled for testing of PICPE across common digital mammography (DM) and digital breast tomosynthesis (DBT) systems, representing seven different machine models from four vendors. Comparison of CP estimates derived from unprocessed and processed image pairs demonstrated excellent correlations (>0.99), with a relative difference below 5% between results from the different image formats. Uncertainties in CP estimates from variability in calibrated parameters such as the compressed breast thickness readout are expected to be substantially greater than the relative differences in estimates per image format. In future work, further testing of different image types, especially a wider variety of DBT images, should be done to confirm robust general applicability of PICPE. The results suggest that PICPE is a practical alternative algorithm for CP estimation when only processed DM or DBT images are available.
This study investigates the effectiveness of artificial intelligence (AI)-based models in detecting and quantifying Breast Arterial Calcification (BAC) from mammograms, a potential indicator of cardiovascular disease. Using two distinct subsets from the OPTIMAM database, an enriched dataset of 1683 images previously confirmed by expert readers to have lesions with non-BAC calcifications, and a ‘normal’ dataset with 1401 representative screening mammography exams, selected among those that were negative on both the included and prior exams. Manual annotation of the calcification data by four readers established ground truth. Two novel BAC detection and quantification models were tested, a baseline and enhanced model. The models exhibited promising results, particularly in terms of a low false positive rate for the enhanced model at 0.6%, but also highlighted the need for improvements to achieve a balance between sensitivity (51.0%) and specificity (99.4%). Notably, 62% of the findings missed by the enhanced model were classified as single-wall BAC, which is usually scored as minimal based on a lower association with cardiovascular disease. Future work is required to establish the association of the model performance with clinical outcomes. The study also examined the relationship between BAC prevalence and certain patient characteristics such as age and Volpara® Density Grade (VDG) in the ‘normal’ screening dataset. Significant correlations were found between BAC volume and patient age, and between BAC prevalence and VDG, which aligns with existing literature. The findings emphasize the potential of AI in improving the consistency of BAC detection with objective quantitative measures, as well as the developed model’s ability to predict the prevalence of BAC in relation to age.
Purpose: To introduce a novel technique for pretraining deep neural networks on mammographic images, where the network learns to predict multiple metadata attributes and simultaneously to match images from the same patient and study. Further to demonstrate how this network can be used to produce explainable predictions. Methods: We trained a neural network on a dataset of 85,558 raw mammographic images and seven types of metadata, using a combination of supervised and self-supervised learning techniques. We evaluated the performance of our model on a dataset of 4,678 raw mammographic images using classification accuracy and correlation. We also designed an ablation study to demonstrate how the model can produce explainable predictions. Results: The model learned to predict all but one of the seven metadata fields with classification accuracy ranging from 78-99% on the validation dataset. The model was able to predict which images were from the same patient with over 93% accuracy on a balanced dataset. Using a simple X-ray system classifier built on top of the first model, representations learned on the initial X-ray system classification task showed by far the largest effect size on ablation, illustrating a method for producing explainable predictions. Conclusions: It is possible to train a neural network to predict several kinds of mammogram metadata simultaneously. The representations learned by the model for these tasks can be summed to produce an image representation that captures features unique to a patient and study. With such a model, ablation offers a promising method to enhance the explainability of deep learning predictions.
Purpose: To identify mammographic image quality indicators (IQI) predictive of interval breast cancers (IC) as opposed to screen-detected cancers (SDC). Methods: Eligible cases for the study were raw, routine recall, screening exams acquired at two UK sites between 2010- 2018, from the OPTIMAM database. Women were matched 3:1 (SDC, n=965 versus IC, n=326), by age (nearest), screening site, breast density grade, Xray system vendor, and compression paddle. Images of the affected breast for prior (IC only) or incident (SDC only) exams were processed using automated software to obtain volumetric breast density (VBD) and IQI metrics related to compression and breast positioning. Compression pressure (CP) was categorised into tertiles or low/target/high (<7/7-15/<15 kPa) groups. Univariate and logistic regression analyses were used to identify significant predictors of IC versus SDC. Results: Compared to SDC, IC had lower median CP (7.9 versus 8.6 kPa, p<0.05). Multivariate analysis found only CP to be significantly associated with the risk of IC versus SDC, with odds ratios (OR) and 95% confidence intervals of 0.93 (0.89-0.97) per unit CP. Compared to low CP, target CP was significantly associated with a lower IC versus SDC risk at the breast level [OR=0.73 (0.56-0.95)] and for mediolateral oblique views [OR=0.77 (0.59-0.99)]. Comparing the third and first tertile, CP was significantly associated with lower risk of IC versus SDC [0.64 (0.47-0.87)], with very similar results when analysed per view. Conclusions: CP was found to be a significant predictor of IC versus SDC, with higher CP being associated with a lower risk of IC.
The aim of this study is to evaluate the accuracy of one of the commercial breast density quantization software. The estimation is performed by means of mammograms of five different 3D printed breast phantoms obtained from clinical patient images. Mammograms of these phantoms were acquired and analyzed using the density software. In addition, the spectra used by the automatic exposure control for each mammogram was accurately characterized and modeled using a previously published spectral model. The result is that the amount of estimated dense tissue is accurate to within an intra-phantom mean of 10% (std. dev. <4.4%), with negligible bias.
Purpose: To test the association between contrast-to-noise ratio (CNR) measurements made on digital mammograms (DM), human reader performance in a lesion detection task using the same images, and image quality (IQ) as predicted by phantom measurements. Methods: DM from 162 women were evaluated for their CNR using a novel metric for application on clinical images. The original unprocessed images were tested (100% dose), as well as the same images after processing to simulate a 50% and 25% relative dose level. IQ measurements from a CDMAM phantom images, as well as human reader calcification cluster detectability ratings on the clinical image set for the three treatments were used to provide ground truth for human lesion detection performance. Analysis was performed to test for association between DM image CNR at the three dose levels, the CDMAM measurements, and reader performance as quantified by a reader-averaged jack-knifed free response operating characteristic (JAFROC) figure of merit (FoM). Results: The clinical image CNR was strongly correlated with the JAFROC FoM and CDMAM threshold gold thicknesses (r=0.98, and r=0.99 @ 0.25 mm, r=0.94 @ 0.1 mm discs, respectively). On a per-image basis, strong associations between CNR and measures of beam quality and exposure were also found that indicate sensitivity to imaging technique factors while remaining independent of signal variations due to breast parenchyma. Conclusions: Using a clinical image CNR it is possible to objectively predict IQ in mammographic images. As such, this metric could provide a means to perform a practical continuous DM system performance monitoring.
The purpose is to develop and validate an automated method for detecting image unsharpness caused by patient motion blur in digital mammograms. The goal is that such a tool would facilitate immediate re-taking of blurred images, which has the potential to reduce the number of recalled examinations, and to ensure that sharp, high-quality mammograms are presented for reading. To meet this goal, an automated method was developed based on interpretation of the normalized image Wiener Spectrum. A preliminary algorithm was developed using 25 cases acquired using a single vendor system, read by two expert readers identifying the presence of blur, location, and severity. A predictive blur severity score was established using multivariate modeling, which had an adjusted coefficient of determination, R2 =0.63±0.02, for linear regression against the average reader-scored blur severity. A heatmap of the relative blur magnitude showed good correspondence with reader sketches of blur location, with a Spearman rank correlation of 0.70 between the algorithmestimated area fraction with blur and the maximum of the blur area fraction categories of the two readers. Given these promising results, the algorithm-estimated blur severity score and heatmap are proposed to be used to aid observer interpretation. The use of this automated blur analysis approach, ideally with feedback during an exam, could lead to a reduction in repeat appointments for technical reasons, saving time, cost, potential anxiety, and improving image quality for accurate diagnosis.
KEYWORDS: Breast imaging, Digital mammography, Digital x-ray imaging, Digital breast tomosynthesis, Image analysis, Image quality, Mammography, Spatial frequencies
Mammographic image quality is important to monitor to maximize diagnostic performance while minimizing patient exposure to ionizing radiation. Phantom imaging for quality control permits practical monitoring of signal and noise, and to optimize use of dose via the contrast-to-noise ratio (CNR). However, it remains a challenge to directly and objectively evaluate CNR in clinical images due to subject variability. A novel clinical image CNR metric has been developed that derives an estimate of system-dependent image noise and references contrast to tissue composition. The present work uses phantom images to validate the noise estimates and to demonstrate sensitivity to imaging conditions. Images of 1 cm adipose-equivalent blocks with 2, 3, and 5 cm 50/50 swirl phantoms and uniform 50/50 blocks were acquired using AEC-selected parameters, and at 0.33 and 0.5 of the AEC-selected mAs at 6 cm. Digital mammograms (DM) were acquired on a GE Essential with and without FineView processing, and in conventional and digital breast tomosynthesis (DBT) views on a Hologic Selenia Dimensions. The CNR was computed using contrast between a 0.4 mm CaCO3 speck in a target slab and adjacent background signal, and noise derived from paired raw and subtracted swirl phantom images. Swirl phantom CNR was estimated to within ±10% of uniform image CNR for GE and Hologic DM, and ±3% for Hologic DBT, and showed good sensitivity to acquisition technique. These results demonstrate promise for objective and efficient image quality evaluation from patient images, using noise estimates that effectively avoid signal related to tissue structure.
Whole-mount pathology imaging has the potential to revolutionize clinical practice by preserving context lost when tissue is cut to fit onto conventional slides. Whole-mount digital images are very large, ranging from 4GB to greater than 50GB, making concurrent processing infeasible. Block-processing is a method commonly used to divide the image into smaller blocks and process them individually. This approach is useful for certain tasks, but leads to over-counting objects located on the seams between blocks. This issue is exaggerated as the block size decreases. In this work we apply a novel technique to enumerate vessels, a clinical task that would benefit from automation in whole-mount images. Whole-mount sections of rabbit VX2 tumors were digitized. Color thresholding was used to segment the brown CD31- DAB stained vessels. This vessel enumeration was applied to the entire whole-mount image in two distinct phases of block-processing. The first (whole-processing) phase used a basic grid and only counted objects that did not intersect the block’s borders. The second (seam-processing) phase used a shifted grid to ensure all blocks captured the block-seam regions from the original grid. Only objects touching this seam-intersection were counted. For validation, segmented vessels were randomly embedded into a whole-mount image. The technique was tested on the image using 24 different block-widths. Results indicated that the error reaches a minimum at a block-width equal to the maximum vessel length, with no improvement as the block-width increases further. Object-density maps showed very good correlation between the vessel-dense regions and the pathologist outlined tumor regions.
Dual-energy contrast-enhanced digital breast tomosynthesis (DE
CE-DBT) image quality is affected by a large parameter
space including the tomosynthesis acquisition geometry, imaging technique factors, the choice of reconstruction
algorithm, and the subject breast characteristics. The influence of most of these factors on reconstructed image quality is
well understood for DBT. However, due to the contrast agent uptake kinetics in CE imaging, the subject breast
characteristics change over time, presenting a challenge for optimization . In this work we experimentally evaluate the
sensitivity of the reconstructed image quality to timing of the
low-energy and high-energy images and changes in iodine
concentration during image acquisition. For four contrast uptake patterns, a variety of acquisition protocols were tested
with different timing and geometry. The influence of the choice of reconstruction algorithm (SART or FBP) was also
assessed. Image quality was evaluated in terms of the lesion
signal-difference-to-noise ratio (LSDNR) in the central slice
of DE CE-DBT reconstructions. Results suggest that for maximum image quality, the low- and high-energy image
acquisitions should be made within one x-ray tube sweep, as separate low- and high-energy tube sweeps can degrade
LSDNR. In terms of LSDNR per square-root dose, the image quality is nearly equal between SART reconstructions with
9 and 15 angular views, but using fewer angular views can result in a significant improvement in the quantitative
accuracy of the reconstructions due to the shorter imaging time interval.
In contrast-enhanced digital mammography (CEDM) an iodinated contrast agent is employed to increase lesion contrast
and to provide tissue functional information. Here, we present the details of a software phantom that can be used as a
tool for the simulation of CEDM images, and compare the degree of anatomic noise present in images simulated using
the phantom to that associated with breast parenchyma in clinical CEDM images. Such a phantom could be useful for
multiparametric investigations including characterization of CEDM imaging performance and system optimization. The
phantom has a realistic mammographic appearance based on a clustered lumpy background and models contrast agent
uptake according to breast tissue physiology. Fifty unique phantoms were generated and used to simulate regions of
interest (ROI) of pre-contrast images and logarithmically subtracted CEDM images using monoenergetic ray tracing.
Power law exponents, β, were used as a measure of anatomic noise and were determined using a linear least-squares fit
to log-log plots of the square of the modulus of radially averaged image power spectra versus spatial frequency. The
power spectra for ROI selected from regions of normal parenchyma in 10 pairs of clinical CEDM pre-contrast and
subtracted images were also measured for comparison with the simulated images. There was good agreement between
the measured β in the simulated CEDM images and the clinical images. The values of β were consistently lower for the
logarithmically subtracted CEDM images compared to the pre-contrast images, indicating that the subtraction process
reduced anatomical noise.
Computed tomography (CT) enables high resolution, whole-body imaging with excellent depth penetration. The
development of new targeted radiopaque CT contrast agents can provide the required sensitivity and localization for the
successful detection and diagnosis of smaller lesions representing earlier disease. Nanoscale, perfluorooctylbromide
(C8F17Br, PFOB) droplets have previously been used as untargeted contrast agents in X-ray imaging, and form the basis
of a promising new group of agents that can be developed for targeted CT imaging. For successful targeting to disease
sites, new PFOB droplet formulations tailored for ideal in vivo performance (e.g., biodistribution, toxicity, and
pharmacokinetics) must be developed. However, the direct assessment of PFOB agents in biological environments early
in their development is difficult using CT, as its sensitivity is not adequate for identification of single probes in vitro or
in vivo. In order to allow single droplet interactions with cells to be directly assessed using standard cellular imaging
tools, we integrate an optical marker within the PFOB agent. In this work, a new method to label a PFOB agent with
fluorescent quantum dot (QD) nanoparticles is presented. These composite PFOB-QD droplets loaded into macrophage
cells result in fluorescence on a cellular level that correlates well to the strong CT contrast exhibited in corresponding
tissue-mimicking cell pellets. QD loading within the PFOB droplet core allows optical labeling without influencing the
surface-dependent properties of the PFOB droplets in vivo, and may be used to follow PFOB localization from in vitro
cell studies to histopathology.
The use of contrast agents can help to overcome a lack of intrinsic radiographic contrast between malignant and benign breast tissue by taking advantage of the properties of tumour angiogenesis. Studies of contrast-enhanced mammography have demonstrated increased lesion conspicuity and have shown that this technique provides information on contrast uptake kinetics. It has been suggested that malignant and benign lesions can be differentiated in part by their uptake kinetics, so this additional data may lead to more accurate diagnoses. Tomosynthesis is a 3D x-ray imaging technique that permits lesion depth localization and increased conspicuity in comparison with 2D x-ray projection techniques. This modality, used in combination with contrast agents, promises to be a sensitive method of breast cancer detection. To develop the technique of contrast-enhanced breast tomosynthesis, a dynamic flow phantom has been constructed to provide the same types of imaging challenges anticipated in the clinical setting. These challenges include a low-contrast tumour space, relevant temporal contrast agent uptake and washout profiles, and a need for quantitative analysis of enhancement levels. The design of a flow phantom will be presented that includes a dynamic tumour space, a background that masks the tumour space in images without contrast enhancement, and flow characteristics that simulate tumour contrast agent uptake and washout kinetics. The system is calibrated to relate signal to concentration of the contrast agent using a well plate filled with iodinated water. Iodine detectability in the flow phantom is evaluated in terms of the signal-difference-to-noise ratio for various tomosynthesis image acquisition parameters including number of acquired angular views, angular extent, and reconstruction voxel size.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.