Pairs Of Scans Research Articles

High Incidence Rate of Computed Tomography-Measured Steatotic Liver Disease in Men With and Without HIV Infection.

We determined steatotic liver disease (SLD) incidence in a prospective cohort of men with HIV (MWH) and men without HIV (MWOH). Incident SLD was defined using paired noncontrast computed tomography scans performed during 2010-2013 and repeated during 2015-2017. Of 268 men, 173 MWH and 95 MWOH, 33 had incident SLD (11.1%, incidence rate 2.4 and 2.7/100 person-years for MWH and MWOH, respectively). Overall, higher abdominal visceral adipose tissue was independently associated with increased SLD risk. In MWH, increased visceral adipose tissue, insulin resistance, chronic hepatitis B, and cumulative etravirine use were associated with SLD. Metabolic factors, but not HIV, were associated with incident SLD. The high incidence rate suggests that SLD will continue to increase in PWH.

Comparative Analysis of CT Texture in Lumbar and Femur and Its Correlation with Bone Mineral Density and Content over Time: An Exploratory Study

Background: This research explores the application of morphometric texture analysis in chest Computed Tomography (CT) scans for determining Bone Mineral Content (BMC) and its temporal changes, both crucial in diagnosing osteoporosis. The study establishes an innovative approach to osteoporosis screening by leveraging Hounsfield Units (HUs) in CT scans to evaluate BMC, offering a comparison with dual-energy X-ray absorptiometry (DXA)-based BMC. Methods: A total of 806 instances (encompassing 379 individuals) were meticulously compiled from a sole institution, during the period stretching from 6 May 2012 to 30 June 2020. In this detailed analysis, each participant was subjected to a pair of chest CT scans, sequentially pursued by a DXA scan, spread over two years. Focused records of BMC values at the inaugural lumbar vertebra (L1) were secured from both the DXA and CT axial slices across all instances. A meticulous selection process pinpointed the largest trabecular section from the L1 vertebral body, whereupon 45 distinctive texture attributes were harvested utilizing gray-level co-occurrence matrix methodologies. Utilizing these amassed 45 attributes, a regression architecture was devised, aiming to forecast the precise BMC values individually. Moreover, an alternative regression framework was engaged, leveraging 90 distinct features, to gauge the BMC fluctuations observed between the duo of scans administered to each participant. Results: The precision of the cultivated regression frameworks was scrupulously assessed, benchmarking against the correlation coefficient (CC) and the mean absolute deviation (MAE) in comparison to the DXA-established references. The regression apparatus employed for estimating BMC unveiled a CC of 0.754 and an MAE of 1.641 (g), respectively. Conversely, the regression mechanism devoted to discerning the variations in BMC manifested a CC of 0.680, coupled with an MAE of 0.528 (g), respectively. Conclusion: The innovative methodology utilizing morphometric texture analysis in CT HUs offers an indirect, yet promising, approach for osteoporosis screening by providing estimations of BMC and its temporal changes. The estimations demonstrate moderate positive correlations with DXA measures, suggesting a potential alternative in circumstances where DXA scanning is limited.

Comparison of Focal Hippocampal T2‐Weighted MRI and Whole‐Brain T1‐Weighted MRI for Detection of Longitudinal Atrophy Using Deep Learning‐Based and Conventional Approaches

AbstractBackgroundLongitudinal measurement of medial temporal lobe (MTL) atrophy is shown to be effective in measuring Alzheimer’s disease (AD) progression and has been extensively studied on T1‐weighted MRI. However, there has been little work analyzing atrophy on high‐resolution T2‐weighted MRI of the hippocampal region, and it is unknown which MRI modality would yield a more effective biomarker of disease progression. We compare atrophy measurements derived from T1 and T2‐weighted MRI using a deep learning framework DeepAtrophy and a deformation‐based morphometry method ALOHA.MethodT2‐weighted MRI of 481 participants from ADNI2/3 are included in the study. DeepAtrophy was trained on 74 participants and evaluated on 407. During training, DeepAtrophy explicitly infers temporal information, such as the scan temporal order for a single scan pair and which image pair has a longer scan time interval for multiple scan pairs. During inference, DeepAtrophy predicts a “total progression score” for each scan pair. This score is expected to be higher (more progression) in individuals with more advanced AD. A similar experiment was conducted using T1‐weighted MRI from 481 ADNI2/GO participants (partially overlapping with T2‐weighted experiment). ALOHA pipelines tailored for T1‐weighted and hippocampal T2‐weighted MRI were evaluated on the same subjects and scan pairs as DeepAtrophy.ResultFor both T1 and T2‐weighted MRI, DeepAtrophy is more accurate, compared with ALOHA, in inferring the temporal order of a pair of longitudinal scans. In hypothetical one‐year and two‐years clinical trials scenarios simulated using ADNI data, DeepAtrophy has an improvement over ALOHA in detecting differences in progression between disease groups in both imaging modalities. The strongest effect for differentiating between Amyloid‐positive and Amyloid‐negative cognitively unimpaired group was obtained using DeepAtrophy progression measures on T2‐weighted MRI.ConclusionThis study supports the potential of high‐resolution focal T2‐weighted MRI of the hippocampus as a longitudinal biomarker of early AD progression suitable for future clinical trials.

A CenTauR scale based on 18F‐MK6240

AbstractBackgroundA standardized scale for the quantification of tau PET imaging would allow comparison and combination of data across tracers and sites resulting in large meta‐analysis and the application of universal cut‐offs. The proposed approach called CenTauR (CTR) was evaluated on 18F‐MK6240 using both global and regional anchoring.Method375 18F‐MK6240 PET & T1w scan pairs (HC Aß‐ = 179/MCI Aß+ = 98/AD Aß+ = 98) from AIBL&ADNeT were spatially normalised using SPM8 and SUVR normalised using the cerebellar cortex. PET scans were then quantified in 4 composite ROIs [Mesial‐Temporal (Me), Meta‐Temporal (MT), Temporo‐Parietal (TP) and Frontal (FT)] from a previously defined “universal” cortical mask. To anchor the 0 and 100 CTR, we only included HC Aß‐ (<10CL) and AD Aß+ (>25CL) individuals younger than 75y, with MMSE = 20, we also excluded HC participants with a quantification higher than the 75%tile in the Me and AD patients with a quantification lower than the 25%tile in the MT, resulting in 68 HC Aß‐ and 29 AD Aß+. We investigated two CTR scaling approaches using either a single transform based on the TP anchors to transform SUVR into CTR for all composites or using region‐based equations scaled with regional anchors.ResultThe mean and standard deviation (std) for the HC and AD groups are reported in Table 1. Using a single transformation resulted in the following equation: Ct = 100x(SUVR‐0.97)/2.26 Region‐based equations can be extrapolated from Table 1. Figures 1 and 2 show that regional CTR overstretches quantification in FT compared to single equation‐derived CTR. Thresholds set at 2 std above the HC were 2.38, 9.43, 7.66, 1.82, and 6.16 CTR for Me, MT, TP, FT and Global respectively, and were 7.05, 7.48, 7.66, 10.20, 7.68 CTR when using regional equation‐derived CTR.ConclusionThe region‐based approach provided slightly more consistent CTR distributions in the HC across the ROIs however, this approach tends to over‐stretch the CTR values in the frontal region.

A CenTauR scale based on 18F‐MK6240

AbstractBackgroundA standardized scale for the quantification of tau PET imaging would allow comparison and combination of data across tracers and sites resulting in large meta‐analysis and the application of universal cut‐offs. The proposed approach called CenTauR (CTR) was evaluated on 18F‐MK6240 using both global and regional anchoring.Method375 18F‐MK6240 PET & T1w scan pairs (HC Aβ‐ = 179/MCI Aβ+ = 98/AD Aβ+ = 98) from AIBL&ADNeT were spatially normalised using SPM8 and SUVR normalised using the cerebellar cortex. PET scans were then quantified in 4 composite ROIs [Mesial‐Temporal (Me), Meta‐Temporal (MT), Temporo‐Parietal (TP) and Frontal (FT)] from a previously defined “universal” cortical mask. To anchor the 0 and 100 CTR, we only included HC Aβ‐ (<10CL) and AD Aβ+ (>25CL) individuals younger than 75y, with MMSE ≥20, we also excluded HC participants with a quantification higher than the 75%tile in the Me and AD patients with a quantification lower than the 25%tile in the MT, resulting in 68 HC Aβ‐ and 29 AD Aβ+. We investigated two CTR scaling approaches using either a single transform based on the TP anchors to transform SUVR into CTR for all composites or using region‐based equations scaled with regional anchors.ResultThe mean and standard deviation (std) for the HC and AD groups are reported in Table 1.Using a single transformation resulted in the following equation:Ct = 100x(SUVR‐0.97)/2.26Region‐based equations can be extrapolated from Table 1. Figures 1 and 2 show that regional CTR overstretches quantification in FT compared to single equation‐derived CTR. Thresholds set at 2 std above the HC were 2.38, 9.43, 7.66, 1.82, and 6.16 CTR for Me, MT, TP, FT and Global respectively, and were 7.05, 7.48, 7.66, 10.20, 7.68 CTR when using regional equation‐derived CTR.ConclusionThe region‐based approach provided slightly more consistent CTR distributions in the HC across the ROIs however, this approach tends to over‐stretch the CTR values in the frontal region.

Deep learning-based automatic segmentation of bone graft material after maxillary sinus augmentation.

To investigate the accuracy and reliability of deep learning in automatic graft material segmentation after maxillary sinus augmentation (SA) from cone-beam computed tomography (CBCT) images. One hundred paired CBCT scans (a preoperative scan and a postoperative scan) were collected and randomly allocated to training (n = 82) and testing (n = 18) subsets. The ground truths of graft materials were labeled by three observers together (two experienced surgeons and a computer engineer). A deep learning model including a 3D V-Net and a 3D Attention V-Net was developed. The overall performance of the model was assessed through the testing data set. The comparative accuracy and inference time consumption of the model-driven and manual segmentation (by two surgeons with 3 years of experience in dental implant surgery) were conducted on 10 CBCT scans from the test samples. The deep learning model had a Dice coefficient (Dice) of 90.36 ± 2.53%, a 95% Hausdorff distance (HD) of 1.59 ± 0.82 mm, and an average surface distance (ASD) of 0.38 ± 0.11 mm. The proposed model only needed 7.2 s, while the surgeon took 19.15 min on average to complete a segmentation task. The overall performances of the model were significantly superior to those of surgeons. The proposed deep learning model yielded a more accurate and efficient performance of automatic segmentation of graft material after SA than that of the two surgeons. The proposed model could facilitate a powerful system for volumetric change evaluation, dental implant planning, and digital dentistry.

Evaluation of the 'Sip Til Send' regimen before elective caesarean delivery using bedside gastric ultrasound: A paired cohort pragmatic study.

Pre-operative fasting is routinely advocated to avoid pulmonary aspiration. The European Society of Anaesthesiology and Intensive Care (ESAIC) recommends a fasting period of 2 h for liquids before surgery. Liberal drinking policies such as the 'Sip Til Send' are a suggested alternative to maintain hydration before surgery. To compare residual gastric volumes in fully fasted nonlabouring parturients before elective caesarean delivery with the 'Sip Til Send' with water liberal drinking protocol. Our hypothesis was the 'Sip Til Send' would be noninferior to standard fasting at minimising the residual gastric volume immediately before surgery. A paired cohort prospective observational pragmatic study using gastric ultrasound, analysed by an operator blinded to the fasting status of each scan. A tertiary maternity hospital in Dublin, Ireland. The study was conducted between January and June 2023. Pregnant women about to undergo elective caesarean delivery who had followed ESAIC fasting guidelines before admission. Each participant underwent two pairs (semi-recumbent and the semi-recumbent right lateral positions) of standardised ultrasound examinations of the gastric antrum: the order of these scans was randomised. The first pair of scans occurred on admission before the 'Sip Til Send' protocol commenced, the other pair just before spinal anaesthesia for caesarean delivery, after a variable time following the 'Sip Til Send' protocol. The primary outcome was the difference in antral cross-sectional area (CSA) between the fully fasted women on admission and the same women after following the 'Sip Til Send' protocol until just before spinal anaesthesia. Fifty-eight women were randomised for the study: 55 and 54 scans in the semi-recumbent position on admission, and 55 and 54 scans in the right lateral position just before spinal anaesthesia. The mean differences (95% CI) in CSA in the semi-recumbent and RL positions were 0.07 (-0.39 to 0.53) cm 2 and 0.04 (-0.60 to 0.68) cm 2 , respectively. Since the of 95% CIs did not cross the predefined noninferiority margin of 0.88 cm 2 , 'Sip Til Send' was noninferior to fully fasting in in terms of the antral CSA. The 'Sip Til Send' protocol of liberal hydration with water was noninferior to standard fasting prior to elective caesarean delivery. NCT05783427 ClinicalTrials.gov.

Abstract 19109: AI-Enabled Automated Cardiac Chambers Volumetry in ECG-Gated Coronary Calcium Scans vs. Non-Gated Lung Cancer Screening Scans

Introduction: We have developed an AI-enabled automated volume measurement of cardiac chambers (AutoChamber) that works on ECG-gated non-contrast coronary artery calcium scans and correlates well with contrast enhanced coronary CT angiography scans. Automated LA and LV volumes have shown to predict atrial fibrillation (Afib) and congestive heart failure. Hypothesis: In this study, we hypothesize that AutoChamber volumetry results in ECG-gated CAC scans correlated well with those of non-gated full-chest lung cancer screening scans in the same individuals. Methods: We have studied 169 cases of paired ECG-gated cardiac CT scans and non-gated lung scans. Mean±SD for age was 62±10 with 52% female. All caseswere asymptomatic and were scanned for preventive health assessment. AutoChamber was run on all cases by an independent operator who was not involvedin data analysis. P value was calculated using a two-tailed test of significance with α=0.05. Results: AutoChamber in cardiac scans vs lung scans reported volume (Mean±SD) for left atrium (LA) as 67.2 and 70.1, left ventricle (LV) 102.8 and 105.5 right atrium (RA) 81.9 and 85.7 right ventricle (RV) 140.7 and 134.2 left ventricle wall (LVW) 113.0 and 109.7 respectively (P < 0.0001). Correlations between cardiac and lung scans for each cardiac chamber: LA (R2= 0.85), LV (R 2= 0.93), RA (R2= 0.91), RV (R2= 0.92), and left ventricular wall (LVW) R2= 0.95. (P <0.0001 for all). Conclusions: AutoChamber volumetry results are similar in ECG-gated cardiac scans vs non-gated full-chest lung scans. This AI-enabled automated tool is promising as it can provide added value to patients undergoing coronary calcium and lung cancer screening scans flagging enlarged cardiac chambers at risk of future Afib and heart failure.

CT-Based Micromotion Analysis After Locking Plate Fixation of AO Type C Distal Radius Fractures

BackgroundVolar locking plate fixation (VLPF) is the most common method for operative fixation of distal radius fractures (DRF). The dorsal ulnar corner (DUC) can be difficult to stabilize as the fragment is small and not exposed when using the volar approach. The purpose of this study was to study fracture fragment migration after VLPF of AO type C DRF, using a volume registration technique of paired CT scans with special focus on the DUC fragment.Materials and MethodsThis pilot study included ten patients with AO type C DRF, all operated with VLPF. The primary outcome was radiographic outcome. Postoperative and 1-year scans were compared and analyzed. Fragment migration was assessed with CT-based micromotion analysis (CTMA), a software technique used for volume registration of paired CT scans.ResultsAll plates were stable over time. Two patients showed signs of screw movement (0.2–0.35 mm and 0.35– > 1 mm respectively). Postoperative reduction was maintained, and there was no fragment migration at the 1-year follow-up except for one case with increased dorsal tilt. The DUC fragment was found in 8/10 cases, fixated in 7/8 cases, and not dislocated in any case at the 1-year follow-up.ConclusionThe CTMA results indicate that variable-angle VLPF after AO type C DRF can yield and maintain a highly stable reduction of the fracture fragments. The DUC fragment remained stable regardless of the number of screws through the fragment. CT volume registration can be a valuable tool in the detailed assessment of fracture fragment migration following volar plate fixation of DRFs.

Spatial-Intensity Transforms for Medical Image-to-Image Translation.

Image-to-image translation has seen major advances in computer vision but can be difficult to apply to medical images, where imaging artifacts and data scarcity degrade the performance of conditional generative adversarial networks. We develop the spatial-intensity transform (SIT) to improve output image quality while closely matching the target domain. SIT constrains the generator to a smooth spatial transform (diffeomorphism) composed with sparse intensity changes. SIT is a lightweight, modular network component that is effective on various architectures and training schemes. Relative to unconstrained baselines, this technique significantly improves image fidelity, and our models generalize robustly to different scanners. Additionally, SIT provides a disentangled view of anatomical and textural changes for each translation, making it easier to interpret the model's predictions in terms of physiological phenomena. We demonstrate SIT on two tasks: predicting longitudinal brain MRIs in patients with various stages of neurodegeneration, and visualizing changes with age and stroke severity in clinical brain scans of stroke patients. On the first task, our model accurately forecasts brain aging trajectories without supervised training on paired scans. On the second task, it captures associations between ventricle expansion and aging, as well as between white matter hyperintensities and stroke severity. As conditional generative models become increasingly versatile tools for visualization and forecasting, our approach demonstrates a simple and powerful technique for improving robustness, which is critical for translation to clinical settings. Source code is available at github.com/clintonjwang/spatial-intensity-transforms.

Home OCT Imaging for Newly Diagnosed Neovascular Age-Related Macular Degeneration: A Feasibility Study

To assess the feasibility of daily Home OCT imaging among patients with neovascular age-related macular degeneration (nAMD). Prospective observational study. Participants with ≥ 1 eye with previously untreated nAMD and visual acuity 20/20 to 20/320. Participants meeting the ocular eligibility criteria were considered for enrollment; those who provided consent received a Notal Vision Home OCT device. Participants were instructed to scan both eyes daily. Retina specialists managed treatment according to their standard practice, without access to the Home OCT data. The presence of fluid detected by a reading center (RC) from in-office OCT scans was compared with fluid volumes measured by the Notal OCT Analyzer (NOA) on Home OCT images. Proportion of participants meeting ocular eligibility criteria who participated in daily scanning, frequency and duration of scanning, proportion of scans eligible for fluid quantification, participant experience with the device, agreement between the RC and NOA fluid determinations, and characteristics of fluid dynamics. Among 40 participants meeting ocular eligibility criteria, 14 (35%) initiated self-scanning. Planned travel (n= 7, 17.5%) and patient-reported inadequate cell reception for the upload of images (n= 5, 12.5%) were the most frequent reasons for not participating. Considering scans of the study eye only, the mean (standard deviation) was 6.3 (0.6) for weekly scanning frequency and 47 (17) seconds for scan duration per eye. Among 2304 scans, 86.5% were eligible for fluid quantification. All participants agreed that scanning became easier over time, and only 1 did not want to continue daily scanning. For 35 scan pairs judged as having fluid by in-office OCT, the NOA detected fluid on 31 scans (89%). For 14 scan pairs judged as having no fluid on in-office OCT, the NOA did not detect fluid on 10 scans (71%). Daily fluid patterns after treatment initiation varied considerably between patients. For patients with nAMD who initiated home scanning, frequency and quality of scanning and accuracy of fluid detection were sufficient to assess the monitoring of fluid at home. Accommodations for travel and Wi-Fi connectivity could improve uptake of the Home OCT device. Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

A real-world clinical validation for AI-based MRI monitoring in multiple sclerosis

Modern management of MS targets No Evidence of Disease Activity (NEDA): no clinical relapses, no magnetic resonance imaging (MRI) disease activity and no disability worsening. While MRI is the principal tool available to neurologists for monitoring clinically silent MS disease activity and, where appropriate, escalating treatment, standard radiology reports are qualitative and may be insensitive to the development of new or enlarging lesions. Existing quantitative neuroimaging tools lack adequate clinical validation. In 397 multi-center MRI scan pairs acquired in routine practice, we demonstrate superior case-level sensitivity of a clinically integrated AI-based tool over standard radiology reports (93.3% vs 58.3%), relative to a consensus ground truth, with minimal loss of specificity. We also demonstrate equivalence of the AI-tool with a core clinical trial imaging lab for lesion activity and quantitative brain volumetric measures, including percentage brain volume loss (PBVC), an accepted biomarker of neurodegeneration in MS (mean PBVC −0.32% vs −0.36%, respectively), whereas even severe atrophy (>0.8% loss) was not appreciated in radiology reports. Finally, the AI-tool additionally embeds a clinically meaningful, experiential comparator that returns a relevant MS patient centile for lesion burden, revealing, in our cohort, inconsistencies in qualitative descriptors used in radiology reports. AI-based image quantitation enhances the accuracy of, and value-adds to, qualitative radiology reporting. Scaled deployment of these tools will open a path to precision management for patients with MS.

Prostate volume analysis in image registration for prostate cancer care: a verification study

Combined magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT) may enhance diagnosis, aid surgical planning and intra-operative orientation for prostate biopsy and radical prostatectomy. Although PET-MRI may provide these benefits, PET-MRI machines are not widely available. Image fusion of Prostate specific membrane antigen PET/CT and MRI acquired separately may be a suitable clinical alternative. This study compares CT-MR registration algorithms for urological prostate cancer care. Paired whole-pelvis MR and CT scan data were used (n = 20). A manual prostate CTV contour was performed independently on each patients MR and CT image. A semi-automated rigid-, automated rigid- and automated non-rigid registration technique was applied to align the MR and CT data. Dice Similarity Index (DSI), 95% Hausdorff distance (95%HD) and average surface distance (ASD) measures were used to assess the closeness of the manual and registered contours. The automated non-rigid approach had a significantly improved performance compared to the automated rigid- and semi-automated rigid-registration, having better average scores and decreased spread for the DSI, 95%HD and ASD (all p < 0.001). Additionally, the automated rigid approach had similar significantly improved performance compared to the semi-automated rigid registration across all accuracy metrics observed (all p < 0.001). Overall, all registration techniques studied here demonstrated sufficient accuracy for exploring their clinical use. While the fully automated non-rigid registration algorithm in the present study provided the most accurate registration, the semi-automated rigid registration is a quick, feasible, and accessible method to perform image registration for prostate cancer care by urologists and radiation oncologists now.

Development of neonatal-specific sequences for portable ultralow field magnetic resonance brain imaging: a prospective, single-centre, cohort study

Magnetic Resonance (MR) imaging is key for investigation of suspected newborn brain abnormalities. Access is limited in low-resource settings and challenging in infants needing intensive care. Portable ultralow field (ULF) MRI is showing promise in bedside adult brain imaging. Use in infants and children has been limited as brain-tissue composition differences necessitate sequence modification. The aim of this study was to develop neonatal-specific ULF structural sequences and test these across a range of gestational maturities and pathologies to inform future validation studies. Prospective cohort study within a UK neonatal specialist referral centre. Infants undergoing 3T MRI were recruited for paired ULF (64mT) portable MRI by convenience sampling from the neonatal unit and post-natal ward. Key inclusion criteria: 1) Infants with risk or suspicion of brain abnormality, or 2) preterm and term infants without suspicion of major genetic, chromosomal or neurological abnormality. Exclusions: presence of contra-indication for MR scanning. ULF sequence parameters were optimised for neonatal brain-tissues by iterative and explorative design. Neuroanatomic and pathologic features were compared by unblinded review, informing optimisation of subsequent sequence generations in a step-wise manner. Main outcome: visual identification of healthy and abnormal brain tissues/structures. ULF MR spectroscopy, diffusion, susceptibility weighted imaging, arteriography, and venography require pre-clinical technical development and have not been tested. Between September 23, 2021 and October 25, 2022, 102 paired scans were acquired in 87 infants; 1.17 paired scans per infant. Median age 9 days, median postmenstrual age 40+2 weeks (range: 31+3-53+4). Infants had a range of intensive care requirements. No adverse events observed. Optimised ULF sequences can visualise key neuroanatomy and brain abnormalities. In finalised neonatal sequences: T2w imaging distinguished grey and white matter (7/7 infants), ventricles (7/7), pituitary tissue (5/7), corpus callosum (7/7) and optic nerves (7/7). Signal congruence was seen within the posterior limb of the internal capsule in 10/11 infants on finalised T1w scans. In addition, brain abnormalities visualised on ULF optimised sequences have similar MR signal patterns to 3T imaging, including injury secondary to infarction (6/6 infants on T2w scans), hypoxia-ischaemia (abnormal signal in basal ganglia, thalami and white matter 2/2 infants on T2w scans, cortical highlighting 1/1 infant on T1w scan), and congenital malformations: polymicrogyria 3/3, absent corpus callosum 2/2, and vermian hypoplasia 3/3 infants on T2w scans. Sequences are susceptible to motion corruption, noise, and ULF artefact. Non-identified pathologies were small or subtle. On unblinded review, optimised portable MR can provide sufficient contrast, signal, and resolution for neuroanatomical identification and detection of a range of clinically important abnormalities. Blinded validation studies are now warranted. The Bill and Melinda Gates Foundation, the MRC, the Wellcome/EPSRC Centre for Medical Engineering, the MRC Centre for Neurodevelopmental Disorders, and the National Institute for Health Research (NIHR) Biomedical Research Centres based at Guy's and St Thomas' and South London & Maudsley NHS Foundation Trusts and King's College London.

Organ-aware CBCT enhancement via dual path learning for prostate cancer treatment.

Cone-beam computed tomography (CBCT) plays a crucial role in the intensity modulated radiotherapy (IMRT) of prostate cancer. However, poor image contrast and fuzzy organ boundaries pose challenges to precise targeting for dose delivery and plan reoptimization for adaptivetherapy. In this work, we aim to enhance pelvic CBCT images by translating them to high-quality CT images with a particular focus on the anatomical structures important for radiotherapy. We develop a novel dual-path learning framework, covering both global and local information, for organ-aware enhancement of the prostate, bladder and rectum. The global path learns coarse inter-modality translation at the image level. The local path learns organ-aware translation at the regional level. This dual-path learning architecture can serve as a plug-and-play module adaptable to other medical image-to-image translationframeworks. We evaluated the performance of the proposed method both quantitatively and qualitatively. The training dataset consists of unpaired 40 CBCT and 40 CT scans, the validation dataset consists of 5 paired CBCT-CT scans, and the testing dataset consists of 10 paired CBCT-CT scans. The peak signal-to-noise ratio (PSNR) between enhanced CBCT and reference CT images is 27.22 ± 1.79, and the structural similarity (SSIM) between enhanced CBCT and the reference CT images is 0.71 ± 0.03. We also compared our method with state-of-the-art image-to-image translation methods, where our method achieves the best performance. Moreover, the statistical analysis confirms that the improvements achieved by our method are statistically significant. The proposed method demonstrates its superiority in enhancing pelvic CBCT images, especially at the organ level, compared to relevant methods.

Orthopantomography Versus Cone Beam Computed Tomography for the Assessment of the Proximity of Posterior Maxillary Apexes with the Maxillary Sinus: A Cross-sectional Study.

The lack of knowledge of the relation of the maxillary sinus with the apexes of maxillary posterior teeth can lead to important complications during common dental procedures. This can be avoided using different imaging techniques, such as orthopantomography (OPG) and cone beam computed tomography (CBCT). The present study aims to compare the performance of OPG with CBCT in measuring the vertical distance of the apexes of posterior-superior teeth to the maxillary sinus. This study corresponded to a cross-sectional study. OPGs and CBCT scans were obtained from the same individuals, and the qualitative and quantitative vertical distance of the apexes in relation to the maxillary sinus was categorized and measured in mm. A total of 28 pairs of OPGs and CBCT scans from the same patients were obtained. About 381 roots were analysed, which included 89 upper first premolars, 51 upper second premolars, 115 upper first molars, and 126 upper second molars. Projection/protrusion was observed with more frequency in molars, specially 1º molars in both OPG (n= 75, 65.2%) and CBCT (n= 31, 27%); however, 106 more cases (27.9%) were classified as projected in the OPG compared to CBCT (p < 0.05). When comparing the performance of the OPG and CBTC for analysing all roots qualitatively, there was a 57.8% agreement between both techniques. This difference was statistically significant (p <0.0001). Statistically significant differences were also observed when comparing the millimetric differences. This study showed that OPG is not an accurate technique to observe the relationship between the maxillary sinus and the apexes of the upper posterior teeth. In those cases where precision is required when performing dental procedures in this area, CBCT should be used. When not available, the clinicians should be aware of the limitations of the OPG and add other complementary techniques.

Sparse2Noise: Low-dose synchrotron X-ray tomography without high-quality reference data

BackgroundSynchrotron radiation computed tomography (SR-CT) holds promise for high-resolution in vivo imaging. Notably, the reconstruction of SR-CT images necessitates a large set of data to be captured with sufficient photons from multiple angles, resulting in high radiation dose received by the object. Reducing the number of projections and/or photon flux is a straightforward means to lessen the radiation dose, however, compromises data completeness, thus introducing noises and artifacts. Deep learning (DL)-based supervised methods effectively denoise and remove artifacts, but they heavily depend on high-quality paired data acquired at high doses. Although algorithms exist for training without high-quality references, they struggle to effectively eliminate persistent artifacts present in real-world data. MethodsThis work presents a novel low-dose imaging strategy namely Sparse2Noise, which combines the reconstruction data from paired sparse-view CT scan (normal-flux) and full-view CT scan (low-flux) using a convolutional neural network (CNN). Sparse2Noise does not require high-quality reconstructed data as references and allows for fresh training on data with very small size. Sparse2Noise was evaluated by both simulated and experimental data. ResultsSparse2Noise effectively reduces noise and ring artifacts while maintaining high image quality, outperforming state-of-the-art image denoising methods at same dose levels. Furthermore, Sparse2Noise produces impressive high image quality for ex vivo rat hindlimb imaging with the acceptable low radiation dose (i.e., 0.5 Gy with the isotropic voxel size of 26 μm). ConclusionsThis work represents a significant advance towards in vivo SR-CT imaging. It is noteworthy that Sparse2Noise can also be used for denoising in conventional CT and/or phase-contrast CT.

Topologic Parametric Response Mapping Identifies Tissue Subtypes Associated with Emphysema Progression

Small airways disease (SAD) and emphysema are significant components of chronic obstructive pulmonary disease (COPD), a heterogenous disease where predicting progression is difficult. SAD, a principal cause of airflow obstruction in mild COPD, has been identified as a precursor to emphysema. Parametric Response Mapping (PRM) of chest computed tomography (CT) can help distinguish SAD from emphysema. Specifically, topologic PRM can define local patterns of both diseases to characterize how and in whom COPD progresses. We aimed to determine if distribution of CT-based PRM of functional SAD (fSAD) is associated with emphysema progression. We analyzed paired inspiratory-expiratory chest CT scans at baseline and 5-year follow up in 1495 COPDGene subjects using topological analyses of PRM classifications. By spatially aligning temporal scans, we mapped local emphysema at year five to baseline lobar PRM-derived topological readouts. K-means clustering was applied to all observations. Subjects were subtyped based on predominant PRM cluster assignments and assessed using non-parametric statistical tests to determine differences in PRM values, pulmonary function metrics, and clinical measures. We identified distinct lobar imaging patterns and classified subjects into three radiologic subtypes: emphysema-dominant (ED), fSAD-dominant (FD), and fSAD-transition (FT: transition from healthy lung to fSAD). Relative to year five emphysema, FT showed rapid local emphysema progression (-57.5%±1.1) compared to FD (-49.9%±0.5) and ED (-33.1%±0.4). FT consisted primarily of at-risk subjects (roughly 60%) with normal spirometry. The FT subtype of COPD may allow earlier identification of individuals without spirometrically-defined COPD at-risk for developing emphysema.

Synthesized 7T MPRAGE From 3T MPRAGE Using Generative Adversarial Network and Validation in Clinical Brain Imaging: A Feasibility Study.

Ultra-high field 7T MRI can provide excellent tissue contrast and anatomical details, but is often cost prohibitive, and is not widely accessible in clinical practice. To generate synthetic 7T images from widely acquired 3T images with deep learning and to evaluate the feasibility of this approach for brain imaging. Prospective. 33 healthy volunteers and 89 patients with brain diseases, divided into training, and evaluation datasets in the ratio 4:1. T1-weighted nonenhanced or contrast-enhanced magnetization-prepared rapid acquisition gradient-echo sequence at both 3T and 7T. A generative adversarial network (SynGAN) was developed to produce synthetic 7T images from 3T images as input. SynGAN training and evaluation were performed separately for nonenhanced and contrast-enhanced paired acquisitions. Qualitative image quality of acquired 3T and 7T images and of synthesized 7T images was evaluated by three radiologists in terms of overall image quality, artifacts, sharpness, contrast, and visualization of vessel using 5-point Likert scales. Wilcoxon signed rank tests to compare synthetic 7T images with acquired 7T and 3T images and intraclass correlation coefficients to evaluate interobserver variability. P < 0.05 was considered significant. Of the 122 paired 3T and 7T MRI scans, 66 were acquired without contrast agent and 56 with contrast agent. The average time to generate synthetic images was ~11.4 msec per slice (2.95 sec per participant). The synthetic 7T images achieved significantly improved tissue contrast and sharpness in comparison to 3T images in both nonenhanced and contrast-enhanced subgroups. Meanwhile, there was no significant difference between acquired 7T and synthetic 7T images in terms of all the evaluation criteria for both nonenhanced and contrast-enhanced subgroups (P ≥ 0.180). The deep learning model has potential to generate synthetic 7T images with similar image quality to acquired 7T images. 2 TECHNICAL EFFICACY: Stage 1.

Deep learning-based ultrasound transducer induced CT metal artifact reduction using generative adversarial networks for ultrasound-guided cardiac radioablation

In US-guided cardiac radioablation, a possible workflow includes simultaneous US and planning CT acquisitions, which can result in US transducer-induced metal artifacts on the planning CT scans. To reduce the impact of these artifacts, a metal artifact reduction (MAR) algorithm has been developed based on a deep learning Generative Adversarial Network called Cycle-MAR, and compared with iMAR (Siemens), O-MAR (Philips) and MDT (ReVision Radiology), and CCS-MAR (Combined Clustered Scan-based MAR). Cycle-MAR was trained with a supervised learning scheme using sets of paired clinical CT scans with and without simulated artifacts. It was then evaluated on CT scans with real artifacts of an anthropomorphic phantom, and on sets of clinical CT scans with simulated artifacts which were not used for Cycle-MAR training. Image quality metrics and HU value-based analysis were used to evaluate the performance of Cycle-MAR compared to the other algorithms. The proposed Cycle-MAR network effectively reduces the negative impact of the metal artifacts. For example, the calculated HU value improvement percentage for the cardiac structures in the clinical CT scans was 59.58%, 62.22%, and 72.84% after MDT, CCS-MAR, and Cycle-MAR application, respectively. The application of MAR algorithms reduces the impact of US transducer-induced metal artifacts on CT scans. In comparison to iMAR, O-MAR, MDT, and CCS-MAR, the application of developed Cycle-MAR network on CT scans performs better in reducing these metal artifacts.