3D Modalities Research Articles

Precise Embodied Data Selection with Haptic Feedback while Retaining Room-Scale Visualisation Context.

Room-scale immersive data visualisations provide viewers a wide-scale overview of a large dataset, but to interact precisely with individual data points they typically have to navigate to change their point of view. In traditional screen-based visualisations, focus-and-context techniques allow visualisation users to keep a full dataset in view while making detailed selections. Such techniques have been studied extensively on desktop to allow precise selection within large data sets, but they have not been explored in immersive 3D modalities. In this paper we develop a novel immersive focus-and-context technique based on a "magic portal" metaphor adapted specifically for data visualisation scenarios. An extendable-hand interaction technique is used to place a portal close to the region of interest. The other end of the portal then opens comfortably within the user's physical reach such that they can reach through to precisely select individual data points. Through a controlled study with 12 participants, we find strong evidence that portals reduce overshoots in selection and overall hand trajectory length, reducing arm and shoulder fatigue compared to ranged interaction without the portal. The portals also enable us to use a robot arm to provide haptic feedback for data within the limited volume of the portal region. In a second study with another 12 participants we found that haptics provided a positive experience (qualitative feedback) but did not significantly reduce fatigue. We demonstrate applications for portal-based selection through two use-case scenarios.

Recent trends in the additive manufacturing of polyurethanes

AbstractPolyurethanes are remarkably versatile materials that offer exceptional control over structure–property relationships, making them the subject of extensive research and exploration across diverse applications. These materials have garnered significant attention due to their inherent chemical, mechanical, thermomechanical, biological and physical properties, further fueling interest in their potential uses. However, conventional processing methods, involving molds, high temperatures or solvents, impose limitations on geometric complexity, hindering their potential applications. Additive manufacturing, or 3D printing, has emerged as a transformative solution, enabling the fabrication of intricate geometries, unparalleled design flexibility, dematerialization and enhanced material properties. This mini‐review explores recent advancements in additive manufacturing techniques applied to polyurethanes, focusing on three prominent 3D printing modalities: vat photopolymerization, direct ink write and fused filament fabrication. Examining the successful integration of polyurethanes with these cutting‐edge 3D printing methods illuminates the remarkable progress achieved in tailoring part design, expanding the range of applications and unlocking novel material−object functionalities. This mini‐review aims to provide valuable insight into the latest trends and development in 3D printing polyurethanes, paving the way for their future utilization in diverse industries. © 2023 Society of Chemical Industry.

Machine-learning-based performance comparison of two-dimensional (2D) and three-dimensional (3D) CT radiomics features for intracerebral haemorrhage expansion

To investigate the value of non-contrast CT (NCCT)-based two-dimensional (2D) radiomics features in predicting haematoma expansion (HE) after spontaneous intracerebral haemorrhage (ICH) and compare its predictive ability with the three-dimensional (3D) signature. Three hundred and seven ICH patients who received baseline NCCT within 6 h of ictus from two stroke centres were analysed retrospectively. 2D and 3D radiomics features were extracted in the manner of one-to-one correspondence. The 2D and 3D models were generated by four different machine-learning algorithms (regularised L1 logistic regression, decision tree, support vector machine and AdaBoost), and the receiver operating characteristic (ROC) curve was used to compare their predictive performance. A robustness analysis was performed according to baseline haematoma volume. Each feature type of 2D and 3D modalities used for subsequent analyses had excellent consistency (mean ICC >0.9). Among the different machine-learning algorithms, pairwise comparison showed no significant difference in both the training (mean area under the ROC curve [AUC] 0.858 versus 0.802, all p>0.05) and validation datasets (mean AUC 0.725 versus 0.678, all p>0.05), and the 10-fold cross-validation evaluation yielded similar results. The AUCs of the 2D and 3D models were comparable either in the binary or tertile volume analysis (all p>0.5). NCCT-derived 2D radiomics features exhibited acceptable and similar performance to the 3D features in predicting HE, and this comparability seemed unaffected by initial haematoma volume. The 2D signature may be preferred in future HE-related radiomic works given its compatibility with emergency condition of ICH.

DILF: Differentiable rendering-based multi-view Image–Language Fusion for zero-shot 3D shape understanding

Zero-shot 3D shape understanding aims to recognize “unseen” 3D categories that are not present in training data. Recently, Contrastive Language–Image Pre-training (CLIP) has shown promising open-world performance in zero-shot 3D shape understanding tasks by information fusion among language and 3D modality. It first renders 3D objects into multiple 2D image views and then learns to understand the semantic relationships between the textual descriptions and images, enabling the model to generalize to new and unseen categories. However, existing studies in zero-shot 3D shape understanding rely on predefined rendering parameters, resulting in repetitive, redundant, and low-quality views. This limitation hinders the model’s ability to fully comprehend 3D shapes and adversely impacts the text–image fusion in a shared latent space. To this end, we propose a novel approach called Differentiable rendering-based multi-view Image–Language Fusion (DILF) for zero-shot 3D shape understanding. Specifically, DILF leverages large-scale language models (LLMs) to generate textual prompts enriched with 3D semantics and designs a differentiable renderer with learnable rendering parameters to produce representative multi-view images. These rendering parameters can be iteratively updated using a text–image fusion loss, which aids in parameters’ regression, allowing the model to determine the optimal viewpoint positions for each 3D object. Then a group-view mechanism is introduced to model interdependencies across views, enabling efficient information fusion to achieve a more comprehensive 3D shape understanding. Experimental results can demonstrate that DILF outperforms state-of-the-art methods for zero-shot 3D classification while maintaining competitive performance for standard 3D classification. The code is available at https://github.com/yuzaiyang123/DILP.

Cross-Modal Contrastive Learning for Domain Adaptation in 3D Semantic Segmentation

Domain adaptation for 3D point cloud has attracted a lot of interest since it can avoid the time-consuming labeling process of 3D data to some extent. A recent work named xMUDA leveraged multi-modal data to domain adaptation task of 3D semantic segmentation by mimicking the predictions between 2D and 3D modalities, and outperformed the previous single modality methods only using point clouds. Based on it, in this paper, we propose a novel cross-modal contrastive learning scheme to further improve the adaptation effects. By employing constraints from the correspondences between 2D pixel features and 3D point features, our method not only facilitates interaction between the two different modalities, but also boosts feature representations in both labeled source domain and unlabeled target domain. Meanwhile, to sufficiently utilize 2D context information for domain adaptation through cross-modal learning, we introduce a neighborhood feature aggregation module to enhance pixel features. The module employs neighborhood attention to aggregate nearby pixels in the 2D image, which relieves the mismatching between the two different modalities, arising from projecting relative sparse point cloud to dense image pixels. We evaluate our method on three unsupervised domain adaptation scenarios, including country-to-country, day-to-night, and dataset-to-dataset. Experimental results show that our approach outperforms existing methods, which demonstrates the effectiveness of the proposed method.

Joint Self-Supervised Image-Volume Representation Learning with Intra-inter Contrastive Clustering

Collecting large-scale medical datasets with fully annotated samples for training of deep networks is prohibitively expensive, especially for 3D volume data. Recent breakthroughs in self-supervised learning (SSL) offer the ability to overcome the lack of labeled training samples by learning feature representations from unlabeled data. However, most current SSL techniques in the medical field have been designed for either 2D images or 3D volumes. In practice, this restricts the capability to fully leverage unlabeled data from numerous sources, which may include both 2D and 3D data. Additionally, the use of these pre-trained networks is constrained to downstream tasks with compatible data dimensions. In this paper, we propose a novel framework for unsupervised joint learning on 2D and 3D data modalities. Given a set of 2D images or 2D slices extracted from 3D volumes, we construct an SSL task based on a 2D contrastive clustering problem for distinct classes. The 3D volumes are exploited by computing vectored embedding at each slice and then assembling a holistic feature through deformable self-attention mechanisms in Transformer, allowing incorporating long-range dependencies between slices inside 3D volumes. These holistic features are further utilized to define a novel 3D clustering agreement-based SSL task and masking embedding prediction inspired by pre-trained language models. Experiments on downstream tasks, such as 3D brain segmentation, lung nodule detection, 3D heart structures segmentation, and abnormal chest X-ray detection, demonstrate the effectiveness of our joint 2D and 3D SSL approach. We improve plain 2D Deep-ClusterV2 and SwAV by a significant margin and also surpass various modern 2D and 3D SSL approaches.

Accuracy and Reliability of Left Atrial Appendage Morphology Assessment by Three-Dimensional Transesophageal Echocardiographic Glass Rendering Modality: A Comparative Study With Computed Tomography

Contrast-enhanced computed tomography is the reference-standard imaging technique to assess left atrial appendage (LAA) morphology. The aim of this study was to evaluate the accuracy and reliability of two-dimensional and new three-dimensional (3D) transesophageal echocardiographic rendering modalities in assessing LAA morphology. Seventy consecutive patients who underwent both computed tomography and transesophageal echocardiography (TEE) were retrospectively enrolled. The traditional LAA morphology classification system (LAAcs; chicken wing, cauliflower, cactus, and windsock) and a new simplified LAAcs based on the LAA bend angle were used for the analysis. LAA morphology was independently assessed by two trained readers using three different modalities: two-dimensional TEE, 3D TEE with multiplanar reconstruction, and a new 3D transesophageal echocardiographic rendering modality with improved transparency (Glass). The new LAAcs and traditional LAAcs were compared in terms of intra- and interrater reliability. With the new LAAcs, two-dimensional TEE was fairly accurate in identifying LAA morphology (κ=0.43, P<.05), with moderate interrater (κ=0.50, P<.05) and substantial intrarater (κ=0.65, P<.005) agreement. Three-dimensional TEE showed higher accuracy and reliability: 3D TEE with multiplanar reconstruction had almost perfect accuracy (κ=0.85, P<.001) and substantial (κ=0.79, P<.001) interrater reliability, while 3D TEE with Glass had substantial accuracy (κ=0.70, P<.001) and almost perfect (κ=0.84, P<.001) interrater reliability. Intrarater agreement was almost perfect for both 3D transesophageal echocardiographic modalities (κ=0.85, P<.001). Accuracy was considerably lower when the traditional LAAcs was used, with 3D TEE with Glass being the most reliable technique (κ=0.75, P<.05). The new LAAcs showed higher inter- and intrarater reliability compared with the traditional LAAcs (interrater, κ=0.85 vs κ=0.49; intrarater, κ=0.94 vs κ=0.68; P<.05). Three-dimensional TEE is an accurate, reliable, and feasible alternative to computed tomography in assessing LAA morphology with the new LAAcs. The new LAAcs shows higher reliability rates than the traditional one.

Cross domain 2D-3D descriptor matching for unconstrained 6-DOF pose estimation

This paper presents a novel approach for cross-domain descriptor matching between 2D and 3D modalities. The 2D-3D matching is applied to localize 2D images in 3D point clouds. Direct cross-domain matching allows our technique to localize images in any type of 3D point cloud without any constraints on the nature or mechanism by which it is obtained. We propose a learning based framework, called Desc-Matcher, to directly match features between the two modalities. A dataset of 2D and 3D features with corresponding locations in images and point clouds is generated to train the Desc-Matcher. To estimate the pose of an image in any 3D cloud, keypoints and feature descriptors are extracted from the query image and the point cloud. The trained Desc-Matcher is then used to match the features from the image and the point cloud. A robust pose estimator is used to predict the location and orientation of the query image from the corresponding positions of the matched 2D and 3D features. We carried out an extensive evaluation of the proposed method for indoor and outdoor scenarios and with different types of point clouds to verify the feasibility of our approach. Experimental results show that the proposed approach can reliably estimate the 6-DOF poses of query cameras in any type of 3D point cloud with high precision. We achieved average median errors of 1.09cm/0.27∘ and 19cm/0.39∘ on the Stanford and Cambridge datasets, respectively.

Use of a Three-dimensional Smartphone Scanner in the Quantitative Analysis of Repaired Unilateral Cleft Lip during the Coronavirus Disease Pandemic

Twelve facial measurements were acquired after cleft lip surgery in three female and seven male patients (aged 11-29 months) with unilateral cleft lip using direct anthropometry and a 3D smartphone scanner. The accuracy and precision of the 3D smartphone scanner were assessed through comparative analyses (t test and Bland-Altman plot). The anthropometric data obtained using the 3D smartphone scanner matched the direct measurement data. The linear measurements did not differ significantly between two-dimensional and 3D modalities (P > 0.05). The intraobserver reliabilities of the two-dimensional smartphone scanner of the first and second observers were high (intraclass correlation coefficient 0.876-0.993 and Cronbach alpha 0.920-0.998) and moderate to high (intraclass correlation coefficient 0.839-0.996 and Cronbach alpha 0.940-0.996), respectively. Inter-observer data showed an intraclass correlation coefficient of 0.876-0.981 and a Cronbach α of 0.960-0.997. The 3D smartphone scanner is effective, efficient, economical, quick, and feasible for facial measurements of patients with unilateral cleft lip and is a viable alternative to direct two-dimensional measurements.

The effect of gastric fundus radiation dose on postoperative anastomotic leakage in esophageal cancer.

Standard-of-care treatment for locally advanced esophageal carcinoma (LAEC) includes neoadjuvant chemoradiotherapy followed by esophagectomy. A potentially catastrophic surgical complication is the development of a postoperative anastomotic leak. To date, the association with radiation dose exposure had been inconclusive. We examined the correlation between radiation exposure to the gastric fundus and risk of postoperative leakage using contemporary radiation doses and fractionation. A total of 69 consecutive patients with LAEC who underwent neoadjuvant chemoradiotherapy followed by esophagectomy in our tertiary center were prospectively followed (median, 27 months). Neoadjuvant regimen included 50.4 Gy in 28 fractions with 5-fluorouracil and cisplatin and 41.4 Gy in 23 fractions with carboplatin and paclitaxel. The gastric fundus was contoured and dosimetric and radiation technique parameters were retrospectively evaluated. Of the total number of patients, 71% and 29% had esophageal and gastroesophageal junction (GEJ) tumors, respectively. Fourteen patients (20.3%) experienced anastomotic leaks within a median of 2 days postoperatively, 78.6% of whom had lower third esophagus or GEJ primaries. Mean and minimum fundus dose did not significantly differ between those with and those without leakage (p = 0.42, p = 0.51). Mean fundus V25, V30, and V35 doses were numerically but not statistically higher in those with anastomotic leak (p = 0.58, p = 0.39, and p = 0.30, respectively). No correlation with incidence of leakage was seen between 3D and IMRT treatment modalities. In our comparatively large prospectively collected series of patients treated for LAEC, radiation dose to the gastric fundus during neoadjuvant combination therapy prior to surgery did not correlate with the risk of postoperative anastomotic leak.

Three-dimensional printing and virtual reconstruction in surgical planning of double-outlet right ventricle repair.

For more than a decade, 3-dimensional (3D) printing has been identified as an innovative tool for the surgical planning of double-outlet right ventricle (DORV). Nevertheless, lack of evidence concerning its benefits encourages us to identify valuable criteria for future prospective trials. We conducted a retrospective study involving 10 patients with DORV operated between 2015 and 2019 in our center. During a preoperative multidisciplinary heart team meeting, we harvested surgical decisions following a 3-increment step process: (1) multimodal imaging; (2) 3D virtual valvular reconstruction (3DVVR); and (3) 3D-printed heart model (3DPHM). The primary outcome was the proportion of predicted surgical strategy following each of the 3 steps, compared with the institutional retrospective surgical strategy. The secondary outcome was the change of surgical strategy through 3D modalities compared with multimodal imaging. The incremental benefit of the 3DVVR and 3DPHM over multimodal imaging was then assessed. The operative strategy was predicted in 5 cases after multimodal imaging, in 9 cases after 3DVVR, and the 10 cases after 3DPHM. Compared with multimodal imaging, 3DVVR modified the strategy for 4 cases. One case was correctly predicted only after 3DPHM inspection. 3DVVR and 3DPHM improved multimodal imaging in the surgical planning of patients with DORV. 3DVVR allowed a better appreciation of the relationships between great vessels, valves, and ventricular septal defects. 3DPHM offers a realistic preoperative view at patient scale and enhances the evaluation of outflow tract obstruction. Our retrospective study demonstrates benefits of preoperative 3D modalities and supports future prospective trials to assess their impact on postoperative outcomes.

Performance analysis of 3D brain MRI modalities for brain tumour sub-region segmentation using U-Net architecture

Performance analysis of 3D brain MRI modalities for brain tumour sub-region segmentation using U-Net architecture

Vat photopolymerization 3D printed microfluidic devices for organ-on-a-chip applications.

Organs-on-a-chip, or OoCs, are microfluidic tissue culture devices with micro-scaled architectures that repeatedly achieve biomimicry of biological phenomena. They are well positioned to become the primary pre-clinical testing modality as they possess high translational value. Current methods of fabrication have facilitated the development of many custom OoCs that have generated promising results. However, the reliance on microfabrication and soft lithographic fabrication techniques has limited their prototyping turnover rate and scalability. Additive manufacturing, known commonly as 3D printing, shows promise to expedite this prototyping process, while also making fabrication easier and more reproducible. We briefly introduce common 3D printing modalities before identifying two sub-types of vat photopolymerization - stereolithography (SLA) and digital light processing (DLP) - as the most advantageous fabrication methods for the future of OoC development. We then outline the motivations for shifting to 3D printing, the requirements for 3D printed OoCs to be competitive with the current state of the art, and several considerations for achieving successful 3D printed OoC devices touching on design and fabrication techniques, including a survey of commercial and custom 3D printers and resins. In all, we aim to form a guide for the end-user to facilitate the in-house generation of 3D printed OoCs, along with the future translation of these important devices.

Right Ventricle Involvement by Glycogen Storage Cardiomyopathy (PRKAG2): Standard and Advanced Echocardiography Analyses.

PRKAG2 syndrome is a rare, early-onset autosomal dominant inherited disease. We aimed to describe the right ventricle (RV) echocardiographic findings using two and three-dimensional (2D and 3D) modalities including myocardial deformation indices in this cardiomyopathy. We also aimed to demonstrate whether this technique could identify changes in RV function that could distinguish any particular findings. Thirty patients with genetically proven PRKAG2 (R302Q and H401Q), 16 (53.3%) males, mean age 39.1 ± 15.4 years, underwent complete echocardiography examination. RV-focused, 4-chamber view was acquired for 2D and 3D measurements. Student's t or Wilcoxon-Mann-Whitney tests were used to compare numerical variables between 2 groups, and p < 0.05 was considered significant. Twelve patients (40%) had a pacemaker implanted for 12.4 ± 9.9 years. RV free wall mean diastolic thickness was 7.9 ± 2.9 mm. RV 4-chamber longitudinal strain (RV4LS), including the free wall and interventricular septum, was -17.3% ± 6.7%, and RV free wall longitudinal strain (RVFWLS) was -19.1% ± 8.5%. The RVFWLS apical ratio measured 0.63 ± 0.15. Mean RV 3D ejection fraction (EF) was 42.6% ± 10.9% and below normal limits in 56.7% of patients. Positive correlation occurred between RV 3DEF, RV4LS, and RVFWLS, especially for patients without a pacemaker (p = 0.006). RV involvement in PRKAG2 syndrome is frequent, occurring in different degrees. Echocardiography is a valuable tool in detecting RV myocardial abnormalities in this condition. The use of 2D RV4LS, RVFWLS, and 3DEF offers reliable indicators of RV systolic dysfunction in this rare, challenging cardiomyopathy.

New multiple sclerosis lesion segmentation and detection using pre-activation U-Net.

Automated segmentation of new multiple sclerosis (MS) lesions in 3D MRI data is an essential prerequisite for monitoring and quantifying MS progression. Manual delineation of such lesions is time-consuming and expensive, especially because raters need to deal with 3D images and several modalities. In this paper, we propose Pre-U-Net, a 3D encoder-decoder architecture with pre-activation residual blocks, for the segmentation and detection of new MS lesions. Due to the limited training set and the class imbalance problem, we apply intensive data augmentation and use deep supervision to train our models effectively. Following the same U-shaped architecture but different blocks, Pre-U-Net outperforms U-Net and Res-U-Net on the MSSEG-2 dataset, achieving a Dice score of 40.3% on new lesion segmentation and an F1 score of 48.1% on new lesion detection. The codes and trained models are publicly available at https://github.com/pashtari/xunet.

SAM: Self-Supervised Learning of Pixel-Wise Anatomical Embeddings in Radiological Images

Radiological images such as computed tomography (CT) and X-rays render anatomy with intrinsic structures. Being able to reliably locate the same anatomical structure across varying images is a fundamental task in medical image analysis. In principle it is possible to use landmark detection or semantic segmentation for this task, but to work well these require large numbers of labeled data for each anatomical structure and sub-structure of interest. A more universal approach would learn the intrinsic structure from unlabeled images. We introduce such an approach, called Self-supervised Anatomical eMbedding (SAM). SAM generates semantic embeddings for each image pixel that describes its anatomical location or body part. To produce such embeddings, we propose a pixel-level contrastive learning framework. A coarse-to-fine strategy ensures both global and local anatomical information are encoded. Negative sample selection strategies are designed to enhance the embedding's discriminability. Using SAM, one can label any point of interest on a template image and then locate the same body part in other images by simple nearest neighbor searching. We demonstrate the effectiveness of SAM in multiple tasks with 2D and 3D image modalities. On a chest CT dataset with 19 landmarks, SAM outperforms widely-used registration algorithms while only taking 0.23 seconds for inference. On two X-ray datasets, SAM, with only one labeled template image, surpasses supervised methods trained on 50 labeled images. We also apply SAM on whole-body follow-up lesion matching in CT and obtain an accuracy of 91%. SAM can also be applied for improving image registration and initializing CNN weights.

Three-dimensional visualization of planta clathrin-coated vesicles at ultrastructural resolution

Biological systems are the sum of their dynamic three-dimensional (3D) parts. Therefore, it is critical to study biological structures in 3D and at high resolution to gain insights into their physiological functions. Electron microscopy of metal replicas of unroofed cells and isolated organelles has been a key technique to visualize intracellular structures at nanometer resolution. However, many of these methods require specialized equipment and personnel to complete them. Here, we present novel accessible methods to analyze biological structures in unroofed cells and biochemically isolated organelles in 3D and at nanometer resolution, focusing on Arabidopsis clathrin-coated vesicles (CCVs). While CCVs are essential trafficking organelles, their detailed structural information is lacking due to their poor preservation when observed via classical electron microscopy protocols experiments. First, we establish a method to visualize CCVs in unroofed cells using scanning transmission electron microscopy tomography, providing sufficient resolution to define the clathrin coat arrangements. Critically, the samples are prepared directly on electron microscopy grids, removing the requirement to use extremely corrosive acids, thereby enabling the use of this method in any electron microscopy lab. Secondly, we demonstrate that this standardized sample preparation allows the direct comparison of isolated CCV samples with those visualized in cells. Finally, to facilitate the high-throughput and robust screening of metal replicated samples, we provide a deep learning analysis method to screen the “pseudo 3D” morphologies of CCVs imaged with 2D modalities. Collectively, our work establishes accessible ways to examine the 3D structure of biological samples and provide novel insights into the structure of plant CCVs.

Multimodal and multicontrast image fusion via deep generative models

Recently, it has become progressively more evident that classic diagnostic labels are unable to accurately and reliably describe the complexity and variability of several clinical phenotypes. This is particularly true for a broad range of neuropsychiatric illnesses such as depression and anxiety disorders or behavioural phenotypes such as aggression and antisocial personality. Patient heterogeneity can be better described and conceptualized by grouping individuals into novel categories, which are based on empirically-derived sections of intersecting continua that span both across and beyond traditional categorical borders. In this context, neuroimaging data (i.e. the set of images which result from functional/metabolic (e.g. functional magnetic resonance imaging, functional near-infrared spectroscopy, or positron emission tomography) and structural (e.g. computed tomography, T1-, T2- PD- or diffusion weighted magnetic resonance imaging) carry a wealth of spatiotemporally resolved information about each patient's brain. However, they are usually heavily collapsed a priori through procedures which are not learned as part of model training, and consequently not optimized for the downstream prediction task. This is due to the fact that every individual participant usually comes with multiple whole-brain 3D imaging modalities often accompanied by a deep genotypic and phenotypic characterization, hence posing formidable computational challenges.In this paper we design and validate a deep learning architecture based on generative models rooted in a modular approach and separable convolutional blocks (which result in a 20-fold decrease in parameter utilization) in order to a) fuse multiple 3D neuroimaging modalities on a voxel-wise level, b) efficiently convert them into informative latent embeddings through heavy dimensionality reduction, c) maintain good generalizability and minimal information loss. As proof of concept, we test our architecture on the well characterized Human Connectome Project database (n = 974 healthy subjects), demonstrating that our latent embeddings can be clustered into easily separable subject strata which, in turn, map to different phenotypical information (including organic, neuropsychological, personality variables) which was not included in the embedding creation process.The ability to extract meaningful and separable phenotypic information from brain images alone can aid in creating multi-dimensional biomarkers able to chart spatio-temporal trajectories which may correspond to different pathophysiological mechanisms unidentifiable to traditional data analysis approaches. In turn, this may be of aid in predicting disease evolution as well as drug response, hence supporting mechanistic disease understanding and also empowering clinical trials.

어텐션 모듈과 기하학적 데이터 증강을 통한 X-ray 영상 내 해부학적 랜드마크 검출 성능 향상

Recently, deep learning-based automated systems for identifying and detecting landmarks have been proposed. In order to train such a deep learning-based model without overfitting, a large amount of image and labeling data is required. Conventionally, an experienced reader manually identifies and labels landmarks in a patient’s image. However, such measurement is not only expensive, but also has poor reproducibility, so the need for an automated labeling method has been raised. In addition, in the X-ray image, since various human tissues on the path through which the photons pass are displayed, it is difficult to identify the landmark compared to a general natural image or a 3D image modality image. In this study, we propose a geometric data augmentation technique that enables the generation of a large amount of labeling data in X-ray images. In addition, the optimal attention mechanism for landmark detection was presented through the implementation and application of various attention techniques to improve the detection performance of 16 major landmarks in the skull. Finally, among the major cranial landmarks, markers that ensure stable detection are derived, and these markers are expected to have high clinical application potential.

Comparison of Diagnostic Test Accuracy of Cone-Beam Breast Computed Tomography and Digital Breast Tomosynthesis for Breast Cancer: A Systematic Review and Meta-Analysis Approach.

Background: Cone-beam breast computed tomography (CBBCT) and digital breast tomosynthesis (DBT) remain the main 3D modalities for X-ray breast imaging. This study aimed to systematically evaluate and meta-analyze the comparison of diagnostic accuracy of CBBCT and DBT to characterize breast cancers. Methods: Two independent reviewers identified screening on diagnostic studies from 1 January 2015 to 30 December 2021, with at least reported sensitivity and specificity for both CBBCT and DBT. A univariate pooled meta-analysis was performed using the random-effects model to estimate the sensitivity and specificity while other diagnostic parameters like the area under the ROC curve (AUC), positive likelihood ratio (), and negative likelihood ratio () were estimated using the bivariate model. Results: The pooled sensitivity specificity, and and AUC at 95% confidence interval are 86.7% (80.3–91.2), 87.0% (79.9–91.8), 6.28 (4.40–8.96), 0.17 (0.12–0.25) and 0.925 for the 17 included studies in DBT arm, respectively, while, 83.7% (54.6–95.7), 71.3% (47.5–87.2), 2.71 (1.39–5.29), 0.20 (0.04–1.05), and 0.831 are the pooled sensitivity specificity, and and AUC for the five studies in the CBBCT arm, respectively. Conclusions: Our study demonstrates that DBT shows improved diagnostic performance over CBBCT regarding all estimated diagnostic parameters; with the statistical improvement in the AUC of DBT over CBBCT. The CBBCT might be a useful modality for breast cancer detection, thus we recommend more prospective studies on CBBCT application.