Dan Ruan

Objectives: To evaluate whether synthesized extraoral bitewing views generated from existing cone-beam CT (CBCT) scans can support posterior-tooth assessment when CBCT is the initial imaging modality, without additional radiographic acquisition. Methods: We retrospectively assembled 60 CBCT scans from 3 datasets encompassing diverse posterior jaw morphologies. We developed a computational pipeline to compute a patient-specific dental arch, define a focal trough, and simulate extraoral bitewing acquisition using a panoramic-inspired rotating source trajectory. Synthetic projections were generated using a Beer-Lambert forward model implemented in Tomographic Iterative GPU-based Reconstruction (TIGRE) and reconstructed into right and left synthesized extraoral bitewing views. Eight observers, including dental students and residents in endodontics, periodontics, and oral radiology, scored posterior teeth for caries, periapical lesions, and periodontal bone loss using original CBCT images and synthesized extraoral bitewing views in separate sessions. A consensus of 2 oral radiology experts served as the reference. Performance was evaluated with false-positive and false-negative rates, specificity, F1 score, and Area Under the Curve (AUC). Interobserver agreement was assessed with (weighted) kappa. Differences in AUC (ΔAUC) between CBCT and synthesized extraoral bitewing views were evaluated using bootstrap-derived 95% confidence intervals (CIs). Results: Compared with CBCT alone, synthesized extraoral bitewing views were associated with lower overdiagnosis for dental caries and periapical lesions (95% CI, 5.87-12.00 and 3.87-8.12, respectively) and reduced underdiagnosis for periodontal bone loss (95% CI, 10.12-18.37). Synthesized extraoral bitewing views showed higher ordinal AUC for dental caries and periodontal bone loss (mean ΔAUC,−0.12 to −0.05 and −0.11 to −0.04, respectively). No consistent improvement was observed for periapical lesion AUC, possibly due to small sample size and limited power. Conclusion: When CBCT is the initial imaging modality, synthesized extraoral bitewing views derived from the existing scan may provide a complementary 2D visualization for posterior-tooth assessment. In this reader study, the derived views were associated with shorter interpretation time and task-dependent changes in diagnostic performance.

Background: Existing strategies for free‐breathing cine cardiac MR, including real‐time imaging, struggle with complex arrhythmias such as atrial fibrillation (AF). Dynamic regularized adaptive clustering optimization (DRACO) has been proposed to generate high‐quality and quantifiable breath‐held cine images for patients in AF. Purpose: To tailor DRACO for free‐breathing and evaluate its potential to simultaneously handle cardiac and respiratory motion. Study Type: Prospective. Subjects: 10 sinus rhythm (10 males), 20 AF (18 males). Field Strength/Sequence: 3.0 T, balanced steady‐state free precession (bSSFP) with sorted golden‐step, ECG‐gated segmented cine, and real‐time. Assessment: Patients were imaged using the golden‐step sequence under breath‐held and free‐breathing conditions, and the reference bSSFP cine sequence. Signal‐to‐noise ratio (SNR), contrast‐to‐noise ratio (CNR), edge sharpness, and image quality (4‐point Likert scale) were assessed. Statistical Tests: Paired t ‐test, Friedman test, regression analysis, Fleiss' Kappa. Significance level p < 0.05. Results: The SNR and CNR values of DRACO images showed no significant differences between breath‐held and free‐breathing conditions in both sinus rhythm (SNR p = 0.53; CNR p = 0.87) and in AF (SNR p = 0.56; CNR p = 0.55). Image sharpness was not significantly different between breath‐held and free‐breathing in both sinus rhythm ( p = 0.24) and AF ( p = 0.79). In AF patients, breath‐held DRACO and free‐breathing DRACO had significantly higher image quality scores than real‐time cine. In sinus rhythm patients, there was no significant difference ( p = 0.13) between breath‐held or free‐breathing DRACO and segmented cine. Despite significant within‐patient ejection fraction (EF) differences between breath‐held versus free‐breathing in 80% of sinus rhythm and 70% of AF patients, the average EF between breath‐held versus free‐breathing DRACO showed high correlation (sinus rhythm R ² = 0.71; AF R ² = 0.71), demonstrating DRACO's robustness to motion variation. Data Conclusion: DRACO is robust to respiratory motion and AF‐related irregular cardiac motion. It enables high‐quality cine image generation in both regular and irregular cardiac motion, with or without breath holding. Level of Evidence: 1. Technical Efficacy: Stage 2.

Cone Beam Computed Tomography (CBCT) image quality is degraded by the increased scatter from the broad beam geometry, while conventional anti-scatter grid (ASG) only provides partial mitigation at the cost of elevated imaging dose. Objective: In this study, we exploit the smooth behavior of the scatter signal with a novel sparse anti-scatter grid to improve the image quality of CBCT. Methods: We achieve sparse sampling of the primary beam signal by sparsely inserting individual collimators focusing on the X-ray source into a template in front of the detector. The novel sparse primary sampling (SPS) grid is evaluated via Monte Carlo simulations with a synthetic CT phantom, patient head phantom, and patient pelvis phantom. Image reconstruction based on SPS was formulated as a constrained optimization problem with fidelity terms on the total signal and the sparsely sampled primary signals. Image quality improvement was benchmarked using ideal primary signal reconstructed images, worst-case scatter degraded reconstructed images, and the previously studied 3-D Richardson-Lucy fitting scatter correction method using low count Monte Carlo. Results: The novel SPS grid and reconstruction method demonstrated recovery of HU values and image resolution with sampling densities under 0.1%. Conclusion: The hybrid hardware-software method supports flexible sampling density and pattern with minimal primary signal loss and image dose increase. Significance: A novel SPS grid was introduced, and a successful demonstration in the Monte Carlo study shows the feasibility of significantly improving CBCT image quality for interventional and radiotherapy procedures via the SPS strategy.

Background: Wearable devices enable continuous acquisition of physiological and behavioral signals with broad utility in precision and population health. However, real-world datasets often have privacy constraints, irregular sampling, and activity-dependent recording, complicating data sharing and modeling. We present a synthesis framework that preserves statistical realism by matching marginal distributions and inter-record distances. Method: We propose a hierarchical heated Markov model (HHMM) that captures conditional dependencies and time-varying behavioral patterns. Tier 1 generates minute-level activity types using hour-specific priors and a heating mechanism to achieve a target ergodic distribution. Tier 2 samples heart rate (via Metropolis–Hastings conditioned on activity) and activity duration from empirical conditional distributions. Tier 3 synthesizes additional variables (e.g., sleep duration, calories) conditioned on Tiers 1–2, with deterministic rules for activity codes and floors climbed. Activity-conditioned Poisson subsampling emulates device-driven irregular timestamps. We benchmark HHMM against CTGAN and TVAE using a synthetic IEEE BHI dataset and validate on a real Fitbit dataset. Fidelity is assessed by comparing distributions of pairwise inter-record distances—within synthetic vs. within reference cohorts—via Wasserstein distances for Kolmogorov–Smirnov (KS), Jensen–Shannon (JSD), and distance-correlation (DC) metrics. Results: On the BHI dataset, mean WD(KS)/WD(JSD)/WD(DC) were 0.125/0.242/0.327 for HHMM, 0.130/0.245/0.257 for CTGAN, and 0.129/0.246/0.249 for TVAE. On the Fitbit dataset, values were 0.295/0.489/0.233 (HHMM), 0.293/0.488/0.209 (CTGAN), and 0.293/0.588/0.160 (TVAE). Discussion: HHMM offers task-specific gains on synthetic benchmarks. Real-world results highlight a need for domain adaptation. The method is computationally efficient and privacy-preserving, supporting scalable synthetic data generation for wearable health research.

The segmentation of pelvic fracture fragments in CT and X-ray images is crucial for trauma diagnosis, surgical planning, and intraoperative guidance. However, accurately and efficiently delineating the bone fragments remains a significant challenge due to complex anatomy and imaging limitations. The PENGWIN challenge, organized as a MICCAI 2024 satellite event, aimed to advance automated fracture segmentation by benchmarking state-of-the-art algorithms on these complex tasks. A diverse dataset of 150 CT scans was collected from multiple clinical centers, and a large set of simulated X-ray images was generated using the DeepDRR method. Final submissions from 16 teams worldwide were evaluated under a rigorous multi-metric testing scheme. The top-performing CT algorithm achieved an average fragment-wise intersection over union (IoU) of 0.930, demonstrating satisfactory accuracy. However, in the X-ray task, the best algorithm achieved an IoU of 0.774, which is promising but not yet sufficient for intra-operative decision-making, reflecting the inherent challenges of fragment overlap in projection imaging. Beyond the quantitative evaluation, the challenge revealed methodological diversity in algorithm design. Variations in instance representation, such as primary-secondary classification versus boundary-core separation, led to differing segmentation strategies. Despite promising results, the challenge also exposed inherent uncertainties in fragment definition, particularly in cases of incomplete fractures. These findings suggest that interactive segmentation approaches, integrating human decision-making with task-relevant information, may be essential for improving model reliability and clinical applicability.

Tomography recovers internal volume from projection measurements. Formulated as inverse problems, classic computed tomography generally reconstructs attenuation property in a preset cartesian grid coordinate. While this is intuitive and convenient for digital display, such discretization leads to forward-backward projection inconsistency, and discrepancy between digital and effective resolution. We take a different perspective by considering the image volume as continuous and modelling forward projection as a hybrid continuous-to-discrete mapping from volume to detector elements, which we call "ray bundles". The ray bundle can be regarded as an unconventional heterogenous coordinate. Projections are modeled as line integrations along ray bundles in the continuous volume space and approximated by numerical integration using customized sample points. This modeling approach is conveniently supported with an implicit neural representation approach. By representing the volume as a function mapping spatial coordinates to attenuation properties and leveraging ray bundle projection, this approach reflects transmission physics and eliminates the need for explicit interpolation, intersection calculations, or matrix inversions. A novel sampling strategy is further developed to adaptively distribute points along the ray bundles, emphasizing high gradient regions to allocate computational resources to heterogenous structures and details. We call this system T-ReX to indicate Transmission Ray bundles for X-ray geometry. We validate T-ReX through comprehensive experiments across three scenarios: simulated full-fan projections with primary signal only, half-fan setups with simulated scatter and noise, and an in-house dataset with realistic acquisition conditions. These results highlight the effectiveness of T-ReX in sparse view X-ray tomography.

In this study, we propose a novel perturbed resonance frequency analysis method where the modified, THz frequency, Fabry–P´ erot resonances of a tissue loaded quartz window are used to quantify superficial skin layer thickness and water content. The superficial layer is treated as an unknown matching layer between the quartz window and deep epidermis and resolution of the longitudinal modes provides a unique mapping from resonance frequency and width to layer water content and thickness. The approach is experimentally validated with a telecentric beam scanning system coupled to VNA extender operating in the 330– 500 GHz band. The system is used to observe the dynamic effects of stinging nettle exposure on in vivo human skin over a 52 x 52 mm2 field-of-view (FOV). Spectroscopic imaging revealed an immediate decrease in resonance frequency and width of up to 15 GHz and 20 GHz, respectively, compared to unperturbed skin, which maps to a 30μm increase in thickness and 25% decrease in water content. This immediate response was followed by a gradual recovery to pre-exposed values. These findings demonstrate additional dielectric window utility in terahertz biomedical spectroscopy in imaging beyond the standard field flattening and system calibration roles.

While 4 π radiotherapy has shown dosimetric advantages over coplanar delivery, prior studies have assumed a fixed source-to-isocenter distance (SID), as varying SID is cumbersome to implement on conventional C-arm linacs. Radiotherapy platforms based on articulated robotic arms eliminate this constraint, enabling non-coplanar, non-isocentric delivery with variable SIDs without moving the patient. Larger SIDs offer increased per-field tumor coverage with a larger field-of-view (FOV), while shorter SIDs improve modulation resolution. This study combines multiple SIDs to simultaneously achieve high target coverage and modulation resolution in robotic radiotherapy. A Multi-SID optimization framework was developed for simultaneous beam orientation optimization, isocenter selection, SID selection, and fluence map optimization. The planning cost includes a least-squares dose fidelity objective, a total variation term for fluence smoothness, and a group sparsity term for beam orientation, SID, and isocenter selection. The collision-free candidate beam pool includes over 1000 beam orientations, 10 SID levels (50–95 cm), and 5–9 isocenters. Multi-SID 4 π intensity-modulated radiation therapy (IMRT) plans were compared against fixed-SID (SID-50 cm and SID-100 cm) 4 π IMRT plans on 10 head-and-neck (H&N) cancer cases. Multi-SID consistently achieved superior plan quality across all 10 patients using varying beams (10–30). For 15-beam plans, the planning target volume homogeneity (defined as D95/D5) is (0.898, 0.893, 0.889), dose conformity (using the Paddick Conformity Index) is (0.867, 0.853, 0.837), the mean Organs-at-Risk (OAR) dose is (12.67, 15.05, 14.12), and the R50 is (2.429, 2.575, 2.485) for Multi-SID, SID-100, and SID-50 respectively. Wilcoxon signed-rank testing showed that the improvements in homogeneity, conformity, and OAR sparing with Multi-SID were statistically significant. Our Multi-SID method leverages both high target coverage and modulation resolution to achieve better dosimetry and/or more efficient treatment plans than fixed SID methods.

Next-generation sequencing (NGS) technologies have become integral for high-throughput transcriptomic studies. Among these, single-cell RNA sequencing (scRNA-seq) is especially valuable for quantifying gene expression at the individual cell level, enabling the identification of rare cell populations and cellular differentiation pathways. However, the high cost of scRNA-seq often limits its broader application. Bulk RNA sequencing (RNA-seq) provides a more affordable alternative but lacks the single-cell resolution needed to elucidate cellular heterogeneity. Here, we present a cycle-consistent generative adversarial network (cycleGAN) approach to generate synthetic single-cell-like transcriptomes from bulk RNA-seq data. By adversarially training two sets of generators and discriminators, our framework attempts to learn the relationship between bulk and single-cell transcriptome distributions. Although this approach does not replace real scRNA-seq experiments, it can be a usefult tool to generate synthetic single-cell-like data for preliminary exploratory investigations and other machine learning applications. We further discuss the performance, limitations, and ethical considerations of our method.