To eliminate motion artifacts in dual-shot (DS) contrast-enhanced (CE) breast imaging, a direct-indirect dual-layer flatpanel-detector (DLFPD) has been developed. It acquires spatio-temporally aligned low-energy (LE) and high-energy (HE) images using a single x-ray exposure, with the spectrum shaped by a K-edge filter. In this work, we integrated a DLFPD prototype and a silver K-edge filter into a clinical mammography system. Image acquisition and post-processing pipelines were developed and implemented for CE digital mammography (CEDM) and digital breast tomosynthesis (CEDBT). A 4 cm BR3D phantom with embedded iodine objects was imaged to evaluate lesion detectability relative to a DS-based system. Projective transformation matrices derived from a registration phantom enabled robust LE/HE image alignment. Modulation transfer function (MTF) matching further suppressed inter-layer misalignment artifacts. DL-based lesion detectability exceeded DS performance with motion artifacts and approached ideal motion-free DS results. Detection of the smallest, lowest-contrast target was limited by image blur and high quantum noise in the HE images, which may be improved by task-specific MTF matching or deep learning-based deblurring and denoising. These results demonstrate the feasibility of a DL-based CE breast imaging system and highlight its promise for clinical translation.

Radiographic image segmentation presents unique challenges due to overlapping anatomical structures, projection ambiguity, and the scarcity of high-quality annotations. Recently, segmentation foundation models such as MedSAM have emerged as powerful tools for automated medical image analysis. Trained on large-scale and diverse image-mask pairs, MedSAM has achieved broad generalization across a wide range of medical image segmentation tasks. Despite this, its exposure to X-rays was primarily limited to chest radiographs annotated with lung masks, and the model relied on spatial prompts like bounding boxes, which are labor-intensive to draw precisely during inference and prone to ambiguity. To overcome these limitations, we propose a parameter-efficient adaptation of MedSAM designed for X-ray image segmentation. The approach integrates lightweight low-rank adaptation (LoRA) fine-tuning to enable efficient model updating while incorporating text-based conditioning to guide mask prediction. This design facilitates intuitive, non-expert human interaction without requiring precise geometric prompts. Evaluated on internal chest and lower-limb radiographic datasets, the model achieves a mean Dice (mDice) score of 92.42 and a mean intersection-over-union (mIoU) of 86.46 while unfreezing only a small fraction of parameters. These results demonstrate that parameter-efficient, language-conditioned adaptation offers an effective strategy for enhancing segmentation performance in projection-based medical imaging.

From horizon detection to fibre structures in X-ray imaging, many vision tasks recover lines via peak detection in Hough space H=S^1×ℝ, the domain of orientation-offset pairs (θ,ρ). Differentiable pipelines extract coordinates via soft-argmax, a probability-weighted average that is only meaningful in a globally linear space. However, (θ,ρ) and (θ+π,-ρ) describe the same undirected line, so H double-covers the space of undirected lines H/ℤ_2: a Möbius strip, obtained by identifying each pair under ℤ_2 action. Soft-argmax operates on the cover H, but since H/ℤ_2 admits no linear structure, it tears geometrically adjacent lines apart. Thus we need a ℤ_2-invariant embedding of lines into a linear space, on which soft-argmax is well-defined. We achieve this by parametrising lines via unit-norm homogeneous vectors ℓ=(1+ρ^2)^-1/2(,,-ρ)^⊤∈ℝ^3 and applying the Veronese map v_2(ℓ)=ℓℓ^⊤ that satisfies v_2(ℓ)=v_2(-ℓ). This descends continuously to an embedding of the quotient H/ℤ_2 into the linear space Sym^2(ℝ^3), where the antipodal ambiguity vanishes. Line extraction becomes a barycentre in Sym^2(ℝ^3), projected back via its leading eigenvector. We validate our Veronese soft-argmax in a Hough transform-based network across all resolvable lines, confirming uniform and seam-free recovery. We further derive that the L_2-loss on isometrically weighted Veronese embeddings equals the squared chordal distance between lines in projective space, enabling a geometrically precise training objective.

Purpose: Scatter reduces the quantitative accuracy of dual-energy (DE) material estimates. However, its impact likely depends on the condition number of the material decomposition problem, which in turn depends on energy channel separation of the DE system. We investigate this tradeoff for various realistic DE radiography configurations. Methods: The study involved high-fidelity Monte Carlo (MC) simulations of multi-material abdomen phantoms presenting a range of vertebral bone mineral (CaHA) concentrations. Polyenergetic DE radiographs were obtained in a standard PA contact-scan setting, with a source-to-detector distance of 180.3 cm, beam collimation of 18 x 43 cm (measured on the detector), and a Pb/Al anti-scatter grid (13:1 ratio, 92 lines/cm). To investigate the impact of spectral separation, four DE configurations at matched doses (similar to 0.1 mGy each) were investigated: (i) a single-exposure (120 kV) protocol with a multi-layer detector consisting of three 0.2 mm CsI layers, with the top providing low-energy (LE) data and bottom providing high-energy (HE) data; (ii)-(iv) three kV-switching protocols with (ii) 100 kV LE and 140 kV (+0.08 mm Ag) HE, resulting in effective energies matching the multi-layer spectral channels, (iii) 60 kV LE and 120 kV HE, and iv) 60 kV LE and HE of 140 kV (+0.08 mm Ag). Moving-average denoising followed by projection-domain decomposition were performed to obtain areal distributions of water and CaHA pathlengths. Scatter correction with a tunable residual error was emulated by subtracting a fraction (ranging 0-200%) of the known simulated scatter from the total measured projections. Accuracy of the vertebral areal Bone Mineral Density (aBMD) estimates were compared for the different DE protocols and scatter estimation errors. Results: Despite the presence of anti-scatter grid, the scatter-to-primary ratio (SPR) was >90% for both LE and HE data in all DE protocols. DE settings with relatively large LE-HE spectral overlap (multi-layer and 100/140 kV) appeared to require more accurate scatter correction than settings with increased spectral separation - e.g., to achieve <0.2 g/cm(2) aBMD bias (compared to aBMD of similar to 1.3 g/cm(2) in healthy bone and similar to 0.5 g/cm(2) in osteoporotic bone), LE and HE scatter errors needed to be <10% for the multi-layer and 100/140 kV protocols, but a 30% error in LE scatter and 15% in HE scatter could be tolerated for the 60/120 and 60/140 kV protocols. Interestingly, accurate aBMD could be obtained for certain ratios of LE-to-HE scatter errors even if the magnitude of the residual uncorrected scatter was high. However, the aBMD bias increased rapidly to >1 g/cm(2) outside of a narrow band of such protocol-specific error ratios. Conclusions: Within the DE configurations tested in this study, quantitative aBMD estimates obtained using protocols with larger LE-HE spectral separation were generally more robust to scatter correction errors.

Collimation in X-ray imaging restricts exposure to the region-of-interest (ROI) and minimizes the radiation dose applied to the patient. The detection of collimator shadows is an essential image-based preprocessing step in digital radiography posing a challenge when edges get obscured by scattered X-ray radiation. Regardless, the prior knowledge that collimation forms polygonal-shaped shadows is evident. For this reason, we introduce a deep learning-based segmentation that is inherently constrained to its geometry. We achieve this by incorporating a differentiable Hough transform-based network to detect the collimation borders and enhance its capability to extract the information about the ROI center. During inference, we combine the information of both tasks to enable the generation of refined, line-constrained segmentation masks. We demonstrate robust reconstruction of collimated regions achieving median Hausdorff distances of 4.3-5.0mm on diverse test sets of real Xray images. While this application involves at most four shadow borders, our method is not fundamentally limited by a specific number of edges.

This paper introduces a novel approach to quantitative contrast-enhanced spectral mammography (CESM) using triple energy K-edge imaging and a two-pass material decomposition method. The primary objective is to accurately determine local projected iodine concentration by leveraging the iodine K-edge in presence of glandular and adipose breast tissue. Addressing the ill-posed problem of established dual-energy methods, our approach accounts for three base materials with three non-redundant measurements. The proposed method involves two key steps. First, we estimate the local total thickness D via three material decomposition with the use of an iterative beam hardening correction, followed by strong denoising. Second, we use D as a constraint computing the recombined iodine image in classical dual-energy technique. In addition based on noise propagation the dose distribution and tube voltage for each image was optimized with respect to minimize standard deviation. Simulation studies using a numerical CIRS dual-energy phantom demonstrate thickness independent effective background cancellation, accurate reconstruction of relative projected iodine concentrations at moderate image noise levels. Sensitivity to remaining inaccuracies was simulated by adding scatter and deviation of used spectra for reconstruction and beam hardening correction. Furthermore, measurements were performed in laboratory system to test the feasibility. In conclusion, the proposed method provides a robust theoretical framework for accurately reconstructing projected iodine concentration images in CESM. For real measurements the method requires a highly accurate estimation of x-ray spectrum and knowledge of attenuation values.

Purpose The combination of multi-layer flat panel detector (FPDT) X-ray imaging and physics-based material decomposition algorithms allows for the removal of anatomical structures. However, the reliability of these algorithms may be compromised by unaccounted materials or scattered radiation. Approach We investigated the two-material decomposition performance of a multi-layer FPDT in the context of 2D chest radiography without and with a 13:1 anti-scatter grid employed. A matrix-based material decomposition (MBMD) (equivalent to weighted logarithmic subtraction), a matrix-based material decomposition with polynomial beam hardening pre-correction (MBMD-PBC), and a projection domain decomposition were evaluated. The decomposition accuracy of simulated data was evaluated by comparing the bone and soft tissue images to the ground truth using the structural similarity index measure (SSIM). Simulation results were supported by experiments using a commercially available triple-layer FPDT retrofitted to a digital X-ray system. Results Independent of the selected decomposition algorithm, uncorrected scatter leads to negative bone estimates, resulting in small SSIM values and bone structures to remain visible in soft tissue images. Even with a 13:1 anti-scatter grid employed, bone images continue to show negative bone estimates, and bone structures appear in soft tissue images. Adipose tissue on the contrary has an almost negligible effect. Conclusions In a contact scan, scattered radiation leads to negative bone contrast estimates in the bone images and remaining bone contrast in the soft tissue images. Therefore, accurate scatter estimation and correction algorithms are essential when aiming for material decomposition using image data obtained with a multi-layer FPDT.

Radiologists have preferred visual impressions or 'styles' of X-ray images that are manually adjusted to their needs to support their diagnostic performance. In this work, we propose an automatic and interpretable X-ray style transfer by introducing a trainable version of the Local Laplacian Filter (LLF) [1]. From the shape of the LLF's optimized remap function, the characteristics of the style transfer can be inferred and reliability of the algorithm can be ensured. Moreover, we enable the LLF to capture complex X-ray style features by replacing the remap function with a Multi-Layer Perceptron (MLP) and adding a trainable normalization layer. We demonstrate the effectiveness of the proposed method by transforming unprocessed mammographic X-ray images into images that match the style of target mammograms and achieve a Structural Similarity Index (SSIM) of 0.94 compared to 0.82 of the baseline LLF style transfer method from [2].

Deep learning-based image analysis offers great potential in clinical practice. However, it faces mainly two challenges: scarcity of large-scale annotated clinical data for training and susceptibility to adversarial data in inference. As an example, an artificial intelligence (AI) system could check patient positioning, by segmenting and evaluating relative positions of anatomical structures in medical images. Nevertheless, data to train such AI system might be highly imbalanced with mostly well-positioned images being available. Thus, we propose the use of synthetic X-ray images and annotation masks forward projected from 3D photon-counting CT volumes to create realistic non-optimally positioned X-ray images for training. An open-source model (TotalSegmentator) was used to annotate the clavicles in 3D CT volumes. We evaluated model robustness with respect to the internal (simulated) patient rotation α on real-data-trained models and real&synthetic-data-trained models. Our results showed that real&synthetic- data-trained models have Dice score percentage improvements of 3% to 15% across different α groups compared to the real-data-trained model. Therefore, we demonstrated that synthetic data could be supplementary used to train and enrich heavily underrepresented conditions to increase model robustness.

BackgroundOne of the limitations in leveraging the potential of artificial intelligence in X-ray imaging is the limited availability of annotated training data. As X-ray and CT shares similar imaging physics, one could achieve cross-domain data sharing, so to generate labeled synthetic X-ray images from annotated CT volumes as digitally reconstructed radiographs (DRRs). To account for the lower resolution of CT and the CT-generated DRRs as compared to the real X-ray images, we propose the use of super-resolution (SR) techniques to enhance the CT resolution before DRR generation.PurposeAs spatial resolution can be defined by the modulation transfer function kernel in CT physics, we propose to train a SR network using paired low-resolution (LR) and high-resolution (HR) images by varying the kernel's shape and cutoff frequency. This is different to previous deep learning-based SR techniques on RGB and medical images which focused on refining the sampling grid. Instead of generating LR images by bicubic interpolation, we aim to create realistic multi-detector CT (MDCT) like LR images from HR cone-beam CT (CBCT) scans.MethodsWe propose and evaluate the use of a SR U-Net for the mapping between LR and HR CBCT image slices. We reconstructed paired LR and HR training volumes from the same CT scans with small in-plane sampling grid size of '(Res). We used the residual U-Net architecture to train two models. SRUN (k)(Res ): trained with kernel-based LR images, and SRUN'(Res): trained with bicubic downsampled data as baseline. Both models are trained on one CBCT dataset (n = 13 391). The performance of both models was then evaluated on unseen kernel-based and interpolation-based LR CBCT images (n = 10 950), and also on MDCT images (n = 1392).ResultsFive-fold cross validation and ablation study were performed to find the optimal hyperparameters. Both SRUNResk and SRUN'(Res )models show significant improvements (p-value < 0.05) in mean absolute error (MAE), peak signal-to-noise ratio (PSNR) and structural similarity index measures (SSIMs) on unseen CBCT images. Also, the improvement percentages in MAE, PSNR, and SSIM by SRUN (k)(Res) is larger than SRUN '(Res). For SRUN (k)(Res), MAE is reduced by 14%, and PSNR and SSIMs increased by 6 and 8%, respectively. To conclude, SRUN (k)(Res) outperforms SRUN'(Res), which the former generates sharper images when tested with kernel-based LR CBCT images as well as cross-modality LR MDCT data.ConclusionsOur proposed method showed better performance than the baseline interpolation approach on unseen LR CBCT. We showed that the frequency behavior of the used data is important for learning the SR features. Additionally, we showed cross-modality resolution improvements to LR MDCT images. Our approach is, therefore, a first and essential step in enabling realistic high spatial resolution CT-generated DRRs for deep learning training.

Background: Dual-energy (DE) detection of bone marrow edema (BME) would be a valuable new diagnostic capability for the emerging orthopedic cone-beam computed tomography (CBCT) systems. However, this imaging task is inherently challenging because of the narrow energy separation between water (edematous fluid) and fat (health yellow marrow), requiring precise artifact correction and dedicated material decomposition approaches. Purpose: We investigate the feasibility of BME assessment using kV-switching DE CBCT with a comprehensive CBCT artifact correction framework and a two-stage projection- and image-domain three-material decomposition algorithm. Methods: DE CBCT projections of quantitative BME phantoms (water containers 100-165 mm in size with inserts presenting various degrees of edema) and an animal cadaver model of BME were acquired on a CBCT test bench emulating the standard wrist imaging configuration of a Multitom Rax twin robotic x-ray system. The slow kV-switching scan protocol involved a 60 kV low energy (LE) beam and a 120 kV high energy (HE) beam switched every 0.5 degrees over a 200 degrees angular span. The DE CBCT data preprocessing and artifact correction framework consisted of (i) projection interpolation onto matched LE and HE projections views, (ii) lag and glare deconvolutions, and (iii) efficient Monte Carlo (MC)-based scatter correction. Virtual non-calcium (VNCa) images for BME detection were then generated by projection-domain decomposition into an Aluminium (Al) and polyethylene basis set (to remove beam hardening) followed by three-material image-domain decomposition into water, Ca, and fat. Feasibility of BME detection was quantified in terms of VNCa image contrast and receiver operating characteristic (ROC) curves. Robustness to object size, position in the field of view (FOV) and beam collimation (varied 20-160 mm) was investigated. Results: The MC-based scatter correction delivered > 69% reduction of cupping artifacts for moderate to wide collimations (> 80 mm beam width), which was essential to achieve accurate DE material decomposition. In a forearm-sized object, a 20% increase in water concentration (edema) of a trabecular bone-mimicking mixture presented as similar to 15 HU VNCa contrast using 80-160 mm beam collimations. The variability with respect to object position in the FOV was modest (< 15% coefficient of variation). The areas under the ROC curve were > 0.9. A femur-sized object presented a somewhat more challenging task, resulting in increased sensitivity to object positioning at 160 mm collimation. In animal cadaver specimens, areas of VNCa enhancement consistent with BME were observed in DE CBCT images in regions of MRI-confirmed edema. Conclusion: Our results indicate that the proposed artifact correction and material decomposition pipeline can overcome the challenges of scatter and limited spectral separation to achieve relatively accurate and sensitive BME detection in DE CBCT. This study provides an important baseline for clinical translation of musculoskeletal DE CBCT to quantitative, point-of-care bone health assessment.

Pectoral muscle (PM) segmentation is an important step for improving the accuracy and efficiency of breast cancer screening in digital mammography. In recent years, image-to-image (I2I) deep learning (DL) methods have achieved state-of-the-art performance for automated PM segmentation by representing the PM region as a binary mask. This paper introduces a new curve regression approach by representing the PM boundary as a vector of connected points that lie on the curve separating the PM from surrounding breast tissue. This low-dimensional PM representation is used to introduce a concept of knowledge distillation (KD), which exploits an ensemble of teachers to perform loss weighting based on maximum likelihood (ML) estimation. Experiments with in-house mammography data show that DL based curve regression outperforms a reference I2I DL method (U-net) for PM segmentation. Further, application of the proposed KD concept achieves higher segmentation accuracy with only 16% of parameters and 23% of inference time compared to the U-net.

The existence of metallic implants in projection images for cone-beam computed tomography (CBCT) introduces undesired artifacts which degrade the quality of reconstructed images. In order to reduce metal artifacts, projection in-painting is an essential step in many metal artifact reduction algorithms. In this work, a hybrid network combining the shift window (Swin) vision transformer (ViT) and a convolutional neural network is proposed as a baseline network for the inpainting task. To incorporate metal information for the Swin ViT-based encoder, metal-conscious self-embedding and neighborhood-embedding methods are investigated. Both methods have improved the performance of the baseline network. Furthermore, by choosing appropriate window size, the model with neighborhood-embedding could achieve the lowest mean absolute error of 0.079 in metal regions and the highest peak signal-to-noise ratio of 42.346 in CBCT projections. At the end, the efficiency of metal-conscious embedding on both simulated and real cadaver CBCT data has been demonstrated, where the inpainting capability of the baseline network has been enhanced.

Collimator detection remains a challenging task in X-ray systems with unreliable or non-available information about the detectors position relative to the source. This paper presents a physically motivated image processing pipeline for simulating the characteristics of collimator shadows in X-ray images. By generating randomized labels for collimator shapes and locations, incorporating scattered radiation simulation, and including Poisson noise, the pipeline enables the expansion of limited datasets for training deep neural networks. We validate the proposed pipeline by a qualitative and quantitative comparison against real collimator shadows. Furthermore, it is demonstrated that utilizing simulated data within our deep learning framework not only serves as a suitable substitute for actual collimators but also enhances the generalization performance when applied to real-world data.

Purpose:Digital breast tomosynthesis (DBT) has been introduced more than a decade ago. Studies have shown higher breast cancer detection rates and lower recall rates, and it has become an established imaging method in diagnostic settings. However, full-field digital mammography (FFDM) remains the most common imaging modality for screening in many countries, as it delivers high-resolution planar images of the breast. To combine the advantages of DBT with the faster acquisition and the unique in-plane resolution capabilities known from FFDM, a system concept was developed for application in screening and diagnosis. Approach:The concept comprises an X-ray tube with adaptive focal spot position based on the flying focal spot (FFS) technology and optimized X-ray spectra. This is combined with innovative algorithmic concepts for tomosynthesis reconstruction and synthetic mammograms (SMs). Results:An X-ray tube with FFS was incorporated into a DBT system that performs 50-deg wide tomosynthesis scans with 25 projections in 4.85 s. Laboratory evaluations demonstrated significant improvements in the effective modular transfer function (eMTF). The improved eMTF as well as the effectiveness of the algorithmic concepts is shown in images from a clinical evaluation study. Conclusions:The DBT system concept enables high spatial resolution at short acquisition times. This leads to improved microcalcification visibility, reduced risk of motion artifacts, and shorter breast compression times. It shifts the in-plane resolution of DBT into the high-resolution range of FFDM. The presented technology leap might be a key contributor to facilitating the paradigm shift of replacing FFDM with DBT plus SM.

Wide-angle digital breast tomosynthesis (DBT) is well known to offer benefits in mass perceptibility compared to narrow-angle DBT due to reduced anatomical overlap. Regarding the perceptibility of micro-calcifications the situation is somehow inverted. On the one hand this can be related to effects during data acquisition and their impact on the system MTF. On the other hand there is a wider spread of calcifications in depth direction in narrow-angle DBT, which distributes calcifications over more slices. This is equivalent to an inherent thicker slice for high spatial frequencies. In this work we want to assume an equivalent quality of raw data and only focus on the effects of different acquisition angles in the reconstruction. We propose an algorithm which creates so-called hybrid thick DBT slices and optimizes the visualization of calcifications while preserving the high mass perceptibility of thin wide-angle DBT slices. The algorithm is purely based on filtered backprojection (FBP) and can be implemented in an efficient manner. For validation simulation studies using the VICTRE (FDA) pipeline are performed. Our results indicate that hybrid thick-slices in wide-angle DBT enable to successfully solve the contrarian imaging tasks of high mass and high calcification perception within one imaging setup.

In medical imaging, noise is an inherent occurring signal corruption, especially for the X-ray imaging where dose exposure to the patient should be minimal. Besides potential image degeneration, which may hinder accurate diagnoses, the noise can have negative impact on signal processing and evaluation algorithms, especially in deep learning (DL) methods. Furthermore, for the training of DL based noise reduction or to bolster DL methods against degeneration due to unseen types of noise or noise levels, it is inevitable to have a thorough and correct noise simulation available. This paper introduces a comprehensive noise simulation method that integrates the strengths of existing techniques into a more complete solution. Simultaneously, our approach aims to minimize the reliance on device-specific measurements and data, by proposing an automatic detector gain estimation.

The progression of X-ray technology introduces diverse image styles that need to be adapted to the preferences of radiologists. To support this task, we introduce a novel deep learning-based metric that quantifies style differences of non-matching image pairs. At the heart of our metric is an encoder capable of generating X-ray image style representations. This encoder is trained without any explicit knowledge of style distances by exploiting Simple Siamese learning. During inference, the style representations produced by the encoder are used to calculate a distance metric for non-matching image pairs. Our experiments investigate the proposed concept for a disclosed reproducible and a proprietary image processing pipeline along two dimensions: First, we use a t-distributed stochastic neighbor embedding (t-SNE) analysis to illustrate that the encoder outputs provide meaningful and discriminative style representations. Second, the proposed metric calculated from the encoder outputs is shown to quantify style distances for non-matching pairs in good alignment with the human perception. These results confirm that our proposed method is a promising technique to quantify style differences, which can be used for guided style selection as well as automatic optimization of image pipeline parameters.

Evaluate microcalcification detectability in digital breast tomosynthesis (DBT) and synthetic 2D mammography (SM) for different acquisition setups using a virtual imaging trial (VIT) approach. Medio-lateral oblique (MLO) DBT acquisitions on eight patients were performed at twice the automatic exposure controlled (AEC) dose. The noise was added to the projections to simulate a given dose trajectory. Virtual microcalcification models were added to a given projection set using an in-house VIT framework. Three setups were evaluated: (1) standard acquisition with 25 projections at AEC dose, (2) 25 projections with a convex dose distribution, and (3) sparse setup with 13 projections, every second one over the angular range. The total scan dose and angular range remained constant. DBT volume reconstruction and synthetic mammography image generation were performed using a Siemens prototype algorithm. Lesion detectability was assessed through a Jackknife-alternative free-response receiver operating characteristic (JAFROC) study with six observers. For DBT, the area under the curve (AUC) was 0.97 ± 0.01 for the standard, 0.95 ± 0.02 for the convex, and 0.89 ± 0.03 for the sparse setup. There was no significant difference between standard and convex dose distributions (p = 0.309). Sparse projections significantly reduced detectability (p = 0.001). Synthetic images had a higher AUC with the convex setup, though not significantly (p = 0.435). DBT required four times more reading time than synthetic mammography. A convex setup did not significantly improve detectability in DBT compared to the standard setup. Synthetic images exhibited a non-significant increase in detectability with the convex setup. Sparse setup significantly reduced detectability in both DBT and synthetic mammography. This virtual imaging trial study allowed the design and efficient testing of different dose distribution trajectories with real mammography images, using a dose-neutral protocol. • In DBT, a convex dose distribution did not increase the detectability of microcalcifications compared to the current standard setup but increased detectability for the SM images. • A sparse setup decreased microcalcification detectability in both DBT and SM images compared to the convex and current clinical setups. • Optimal microcalcification cluster detection in the system studied was achieved using either the standard or convex dose setting, with the default number of projections.

Contrast-enhanced mammography (CEM) offers a promising alternative to address the limitations of digital mammography, particularly in cases of dense breast tissue, which compromises the performance of non-contrast x-ray imaging modalities. CEM uses iodinated contrast material to enhance cancer detection in denser breasts and provides critical functional information about suspicious findings. However, the process of combining images acquired with different x-ray energy spectra in CEM can introduce artifacts, challenging interpretation and confidence in CEM images. This study presents novel approaches to improve CEM image quality. First, deep learning (DL)-based algorithms for scatter correction in both low-energy and high-energy images are proposed to enhance contrast-enhancement patterns and iodine quantification. Additionally, a unique deep learning network is introduced to predict pixel-by-pixel the compressed breast thickness, enabling the use of local thickness-based image subtraction-weighting maps throughout the breast area. Results in phantom cases demonstrate the effectiveness of the scatter correction models in predicting the scatter signal, even in cases with the anti-scatter grid present. The thickness map model accurately estimates the local thickness, particularly in the constant thickness area of the breast. Comparison with clinical data revealed good agreement between estimated thickness maps and ground truth, with minor discrepancies attributed to alignment issues. Furthermore, the study explored the combined use of scatter correction and thickness-based weighting maps in creating recombined CEM images. This approach showed a marginal positive impact due to scatter correction, with larger improvements observed in the signal intensity homogeneity at the border of the breast. These advancements aim to enhance the CEM diagnostic accuracy, making it a valuable tool for breast cancer detection and evaluation, especially in cases with dense breast tissue.