Ultra-low-field magnetization transfer imaging at 0.055T with low specific absorption rate.

PURPOSE: To demonstrate magnetization transfer (MT) effects with low specific absorption rate (SAR) on ultra-low-field (ULF) MRI. METHODS: MT imaging was implemented by using sinc-modulated RF pulse train (SPT) modules to provide bilateral off-resonance irradiation. They were incorporated into 3D gradient echo (GRE) and fast spin echo (FSE) protocols on a shielding-free 0.055T head scanner. MT effects were first verified using phantoms. Brain MT imaging was conducted in both healthy subjects and patients. RESULTS: MT effects were clearly observed in phantoms using six SPT modules with total flip angle 3600° at central primary saturation bands of approximate offset ±786 Hz, even in the presence of large relative B0 inhomogeneity. For brain, strong MT effects were observed in gray matter, white matter, and muscle in 3D GRE and FSE imaging using six and sixteen SPT modules with total flip angle 3600° and 9600°, respectively. Fat, cerebrospinal fluid, and blood exhibited relatively weak MT effects. MT preparation enhanced tissue contrasts in T2-weighted and FLAIR-like images, and improved brain lesion delineation. The estimated MT SAR was 0.0024 and 0.0008 W/kg for two protocols, respectively, which is far below the US Food and Drug Administration (FDA) limit of 3.0 W/kg. CONCLUSION: Robust MT effects can be readily obtained at ULF with extremely low SAR, despite poor relative B0 homogeneity in ppm. This unique advantage enables flexible MT pulse design and implementation on low-cost ULF MRI platforms to achieve strong MT effects in brain and beyond, potentially augmenting their clinical utility in the future.

Chemical Exchange Saturation Transfer Magnetic Resonance Imaging for Longitudinal Assessment of Intracerebral Hemorrhage and Deferoxamine Treatment at 3T in a Mouse Model.

BACKGROUND: Noninvasive imaging of molecular alterations after intracerebral hemorrhage (ICH) could provide valuable information to guide and monitor treatments. Chemical exchange saturation transfer (CEST) magnetic resonance imaging has demonstrated promises in identifying proliferation, necrosis, and changes in cellularity in brain tumors. Here, we applied CEST magnetic resonance imaging to monitor molecular changes in hematoma without and with treatment noninvasively over 2 weeks at 3T using endogenous contrast. METHODS: CEST contrast related to proteins at 3.5 ppm (amide proton transfer) and proteins/lipids at -3.5 ppm (relayed nuclear overhauser effect [rNOE]) were examined over 14 days in a collagenase-induced ICH mouse model. Imaging findings were validated with immunohistochemistry based on the ICH neuropathology. We also examined iron-containing phantoms that mimicked iron concentrations in hematoma to ensure the iron will not attenuate the CEST contrast during disease progression. Based on the validity of the CEST contrast of hematoma, we further examined related molecular alterations under iron-chelation treatment with deferoxamine. RESULTS: We observed the temporal and spatial differences of CEST contrasts between rNOE at -3.5 ppm and amide proton transfer at 3.5 ppm, in which the core and perihematoma could be identified by rNOE on day 3 and day 14, and amide proton transfer on day 1, day 7, and day 14. Moreover, we observed a 25.7% significant reduction (P<0.05) of rNOE contrast after deferoxamine treatment to the ICH mice on day 3, which was not observable in amide proton transfer contrast. Our histology data indicated that rNOE primarily correlated with the myelin pathology, and amide proton transfer could reflect the cellularity increase at hematoma up to day 7. CONCLUSIONS: Significant rNOE changes correlated well with histologic findings, especially myelin lipids, and regional characteristics in hematoma indicate the uniqueness of CEST magnetic resonance imaging in monitoring molecular changes during ICH and treatment.

Location-Specific Hematoma Volume Cutoff and Clinical Outcomes in Intracerebral Hemorrhage.

BACKGROUND: Major intracerebral hemorrhage (ICH) trials have largely been unable to demonstrate therapeutic benefit in improving functional outcomes. This may be partly due to the heterogeneity of ICH outcomes based on their location, where a small strategic ICH could be debilitating, thus confounding therapeutic effects. We aimed to determine the ideal hematoma volume cutoff for different ICH locations in predicting ICH outcomes. METHODS: We retrospectively analyzed consecutive ICH patients enrolled in the University of Hong Kong prospective stroke registry from January 2011 to December 2018. Patients with premorbid modified Rankin Scale score >2 or who underwent neurosurgical intervention were excluded. ICH volume cutoff, sensitivity, and specificity in predicting respective 6-month neurological outcomes (good [modified Rankin Scale score 0-2], poor [modified Rankin Scale score 4-6], and mortality) for specific ICH locations were determined using receiver operating characteristic curves. Separate multivariate logistic regression models were also conducted for each location-specific volume cutoff to determine whether these cutoffs were independently associated with respective outcomes. RESULTS: Among 533 ICHs, the volume cutoff for good outcome according to ICH location was 40.5 mL for lobar, 32.5 mL for putamen/external capsule, 5.5 mL for internal capsule/globus pallidus, 6.5 mL for thalamus, 17 mL for cerebellum, and 3 mL for brainstem. ICH smaller than the cutoff for all supratentorial sites had higher odds of good outcomes (all P<0.05). Volumes exceeding 48 mL for lobar, 41 mL for putamen/external capsule, 6 mL for internal capsule/globus pallidus, 9.5 mL for thalamus, 22 mL for cerebellum, and 7.5 mL for brainstem were at greater risk of poor outcomes (all P<0.05). Mortality risks were significantly higher for volumes that exceeded 89.5 mL for lobar, 42 mL for putamen/external capsule, and 21 mL for internal capsule/globus pallidus (all P<0.001). All receiver operating characteristic models for location-specific cutoffs had good discriminant values (area under the curve >0.8), except in predicting good outcome for cerebellum. CONCLUSIONS: ICH outcomes differed with location-specific hematoma size. Location-specific volume cutoff should be considered in patient selection for ICH trials.

Pushing the limits of low-cost ultra-low-field MRI by dual-acquisition deep learning 3D superresolution.

PURPOSE: Recent development of ultra-low-field (ULF) MRI presents opportunities for low-power, shielding-free, and portable clinical applications at a fraction of the cost. However, its performance remains limited by poor image quality. Here, a computational approach is formulated to advance ULF MR brain imaging through deep learning of large-scale publicly available 3T brain data. METHODS: A dual-acquisition 3D superresolution model is developed for ULF brain MRI at 0.055 T. It consists of deep cross-scale feature extraction, attentional fusion of two acquisitions, and reconstruction. Models for T1 -weighted and T2 -weighted imaging were trained with 3D ULF image data sets synthesized from the high-resolution 3T brain data from the Human Connectome Project. They were applied to 0.055T brain MRI with two repetitions and isotropic 3-mm acquisition resolution in healthy volunteers, young and old, as well as patients. RESULTS: The proposed approach significantly enhanced image spatial resolution and suppressed noise/artifacts. It yielded high 3D image quality at 0.055 T for the two most common neuroimaging protocols with isotropic 1.5-mm synthetic resolution and total scan time under 20 min. Fine anatomical details were restored with intrasubject reproducibility, intercontrast consistency, and confirmed by 3T MRI. CONCLUSION: The proposed dual-acquisition 3D superresolution approach advances ULF MRI for quality brain imaging through deep learning of high-field brain data. Such strategy can empower ULF MRI for low-cost brain imaging, especially in point-of-care scenarios or/and in low-income and mid-income countries.

Three-in-one protocol reduces mortality of patients with haemodynamically unstable pelvic fractures-a five year multi-centred review in Hong Kong.

PURPOSE: The mortality rate in patients with haemodynamically unstable pelvic fractures is as high as 40-60%. Despite the new advances in trauma care which are in phase in trauma centres in Hong Kong, the management of haemodynamically unstable pelvic fracture is still heterogeneous. The aim of this study is to review the results of management of haemodynamically unstable pelvic fracture patients in Hong Kong over a five year period. METHODS: This is a retrospective multi-centred cohort study of patients with haemodynamic and mechanically unstable pelvic fractures from 1 January 2010 to 31 December 2014. The primary outcome investigated is mortality of patients (including overall, 30-day, 7-day and 24-hour mortalities). RESULTS: Implementation of three-in-one pelvic damage control protocol was identified to be a significant independent predictive factor for overall, 30-day, seven-day and 24-hour mortalities. The overall in-hospital and 30-day mortality rates for patients managed with three-in-one protocol was 12.5%, while it was 11% for seven day mortality and 6% for 24 hour mortality. There were no significant differences in demographic characteristics, physiological measurements, types of pelvic fracture, severity and mechanism of injury between patients managed with or without three-in-one protocol. CONCLUSIONS: Implementation of the multidisciplinary three-in-one pelvic damage control protocol reduces mortality and therefore should be highly recommended. The results are convincing as it has eliminated the limitations of our previous single-centred trial.

Separation and quantification of lactate and lipid at 1.3 ppm by diffusion-weighted magnetic resonance spectroscopy.

PURPOSE: To separate the spectrally overlapped lactate and lipid signals at 1.3 ppm using diffusion-weighted magnetic resonance spectroscopy (DW-MRS) based on their large diffusivity difference. METHODS: DW-MRS was applied to the gel phantoms containing lactate and lipid droplets, and to the rat brain tumors. Lactate and lipid signals and their apparent diffusion coefficients were computed from the diffusion-weighted proton spectra. Biexponential fitting and direct spectral subtraction approaches were employed and compared. RESULTS: DW-MRS could effectively separate lactate and lipid signals both in phantoms and rat brain C6 glioma by biexponential fitting. In phantoms, lactate and lipid signals highly correlated with the known lactate concentration and lipid volume fractions. In C6 glioma, both lactate and lipid signals were detected, and the lipid signal was an order of magnitude higher than lactate signal. The spectral subtraction approach using three diffusion weightings also allowed the separation of lactate and lipid signals, yielding results comparable to those by the biexponential fitting approach. CONCLUSION: DW-MRS presents a new approach to separate and quantify spectrally overlapped molecules and/or macromolecules, such as lactate and lipid, by using the diffusivity difference associated with their different sizes or mobility within tissue microstructure. Magn Reson Med 77:480-489, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Multitask training promotes automaticity of a fundamental laparoscopic skill without compromising the rate of skill learning.

BACKGROUND: A defining characteristic of expertise is automated performance of skills, which frees attentional capacity to better cope with some common intraoperative stressors. There is a paucity of research on how best to foster automated performance by surgical trainees. This study examined the use of a multitask training approach to promote automated, robust laparoscopic skills. METHODS: Eighty-one medical students completed training of a fundamental laparoscopic task in either a traditional single-task training condition or a novel multitask training condition. Following training, participants' laparoscopic performance was tested in a retention test, two stress transfer tests (distraction and time pressure) and a secondary task test, which was included to evaluate automaticity of performance. The laparoscopic task was also performed as part of a formal clinical examination (OSCE). RESULTS: The training groups did not differ in the number of trials required to reach task proficiency (p = .72), retention of skill (ps > .45), or performance in the clinical examination (p = .14); however, the groups did differ with respect to the secondary task (p = .016). The movement efficiency (number of hand movements) of single-task trainees, but not multitask trainees, was negatively affected during the secondary task test. The two stress transfer tests had no discernable impact on the performance of either training group. CONCLUSION: Multitask training was not detrimental to the rate of learning of a fundamental laparoscopic skill and added value by providing resilience in the face of a secondary task load, indicative of skill automaticity. Further work is needed to determine the extent of the clinical utility afforded by multitask training.

Simultaneous pancreas and kidney transplantation as the standard surgical treatment for diabetes mellitus patients with end-stage renal disease.

OBJECTIVES: To review the outcome following simultaneous pancreas and kidney transplantation in patients with type 1 diabetes mellitus and end-stage renal disease, as well as those with type 2 diabetes mellitus, and to discuss the applicability of this treatment in this locality. METHODS: A systematic literature review was performed by searching the PubMed and Elsevier databases. The search terms used were "simultaneous pancreas and kidney transplantation", "diabetes", "pancreas transplant" and "SPK". Original and major review articles related to simultaneous pancreas and kidney transplantation were reviewed. Papers published in English after 1985 were included. Clinical outcomes following transplantation were extracted for comparison between different treatment methods. Outcomes of simultaneous pancreas and kidney transplant and other transplantation methods were identified and categorised into patient survival, graft survival, diabetic complications, and quality of life. Patient survivals and graft survivals were also compared. RESULTS: Currently available clinical evidence shows good outcomes for type 1 diabetes mellitus in terms of patient survival, graft survival, diabetic complications, and quality of life. For type 2 diabetes mellitus, the efficacy and application of the procedure remain controversial but the outcomes are possibly comparable with those in type 1 diabetes mellitus. CONCLUSIONS: Simultaneous pancreas and kidney transplantation is a technically demanding procedure that is associated with significant complications, and it should be regarded as a 'last resort' treatment in patients whose diabetic complications have become life-threatening or severely burdensome despite best efforts in maintaining good diabetic control through lifestyle modifications and medications.

Use of viscoelastic haemostatic assay in emergency and elective surgery.

OBJECTIVES: To review the current evidence for the use of viscoelastic haemostatic assays in different surgical settings including trauma, cardiac surgery, liver transplantation, as well as the monitoring of antiplatelet agents and anticoagulants prior to surgery. DATA SOURCES: PubMed database. STUDY SELECTION: Key words for the literature search were "thromboelastography" or "ROTEM" in combination with "trauma", "antiplatelet", "cardiac surgery", "liver transplantation" or "anticoagulants". DATA EXTRACTION: Original and major review articles related to the use of viscoelastic haemostatic assays. DATA SYNTHESIS: Haemostatic function is a critical factor determining patient outcomes in emergency or elective surgery. The increasing use of antiplatelet agents and anticoagulants has potentially increased the risks of haemorrhages and the need for transfusion. Conventional coagulation tests have limitations in detecting haemostatic dysfunctions in subgroups of patients and are largely ineffective in diagnosing hyperfibrinolysis. The viscoelastic haemostatic assays are potentially useful point-of-care tools that provide information on clot formation, clot strength, and fibrinolysis, as well as to guide goal-directed transfusion and antifibrinolytic therapy. They may also be used to monitor antiplatelet and anticoagulant therapy. However, standardisation of techniques and reference ranges is required before these tests can be widely used in different clinical settings. CONCLUSIONS: Viscoelastic haemostatic assays, as compared with conventional coagulation tests, are better for detecting coagulopathy and are the only tests that can provide rapid diagnosis of hyperfibrinolysis. Goal-directed administration of blood products based on the results of viscoelastic haemostatic assays was associated with reduction in allogeneic blood product transfusions in trauma, cardiac surgery, and liver transplantation cases. However, there is currently no evidence to support the routine use of viscoelastic haemostatic assays for monitoring platelet function prior to surgery.

Effects of progesterone vs. dexamethasone on brain oedema and inflammatory responses following experimental brain resection.

BACKGROUND: Dexamethasone (DEXA) is commonly used to reduce brain swelling during neurosurgical procedures. DEXA, however, has many side-effects that can increase the risks of post-operative complications. In contrast, progesterone (PRO) has fewer side-effects and has been found to be neuroprotective on traumatic brain injury (TBI). Whether PRO may be used as an alternative to DEXA during routine procedures has not been fully explored. OBJECT: To compare the effects of DEXA and PRO on surgical brain injury (SBI). METHODS: Seventy-five adult male Sprague Dawley rats were randomized into five groups: (1) SBI + drug vehicle (peanut oil, 1 ml kg(-1)); (2) SBI + DEXA (1 mg kg(-1)); (3) SBI + low-dose PRO (10 mg kg(-1)); (4) SBI + high-dose PRO (20 mg kg(-1)); and (5) sham SBI + drug vehicle. Magnetic resonance imaging study and assessments of brain water content (BWC), blood-brain barrier (BBB) permeability, cellular inflammatory responses and matrix metalloproteinase 9 (MMP-9) expression were conducted. RESULTS: This model consistently resulted in increased BWC and BBB disruption. PRO reduced astrocyte and microglia responses and attenuated brain oedema with preservation of BBB. A significant down-regulation of MMP-9 expression occurred in the PRO 20 group. CONCLUSIONS: PRO is as effective as DEXA in reducing brain oedema and inflammation following SBI; 10 mg kg(-1) of PRO was demonstrated to be more effective in relieving acute cellular inflammatory responses.

Effect of clinical context on simulator-based assessment of blood pressure taking - a pilot randomized study.

BACKGROUND: Blood pressure measurement is an essential clinical skill that can readily be assessed in objective structured clinical examination (OSCE). While the use of simulators can enhance test validity and reliability, the given clinical context may also affect student performance. AIMS: To investigate the impact of variations in clinical context on blood pressure measurement in a simulator-based OSCE. METHOD: We randomized 162 first-year medical students into four groups that received different lead-in statements before measuring blood pressure on a manikin simulator. These statements described hypothetical patients with different likelihoods of having systemic hypertension. RESULTS: The lead-in that described the highest likelihood of hypertension was associated with significantly higher reported readings and lower accuracy. The lead-in that suggested normality yielded the best performance. CONCLUSION: Student performance in simulator-based OSCE may be affected by the clinical context provided. However, we argue that construct validity should be viewed in light of the application of a test, in that patients may also present with different cues and likelihoods of having hypertension. Variations in construct design should be further explored to enhance the training and assessment of clinical competence that reflects the unpredictability encountered in daily clinical practice.

Loss of brain-enriched miR-124 microRNA enhances stem-like traits and invasiveness of glioma cells.

miR-124 is a brain-enriched microRNA that plays a crucial role in neural development and has been shown to be down-regulated in glioma and medulloblastoma, suggesting its possible involvement in brain tumor progression. Here, we show that miR-124 is down-regulated in a panel of different grades of glioma tissues and in all of the human glioma cell lines we examined. By integrated bioinformatics analysis and experimental confirmation, we identified SNAI2, which is often up-regulated in glioma, as a direct functional target of miR-124. Because SNAI2 has been shown to regulate stem cell functions, we examined the roles of miR-124 and SNAI2 in glioma cell stem-like traits. The results showed that overexpression of miR-124 and knockdown of SNAI2 reduced neurosphere formation, CD133(+) cell subpopulation, and stem cell marker (BMI1, Nanog, and Nestin) expression, and these effects could be rescued by re-expression of SNAI2. Furthermore, enhanced miR-124 expression significantly inhibited glioma cell invasion in vitro. Finally, stable overexpression of miR-124 and knockdown of SNAI2 inhibited the tumorigenicity and invasion of glioma cells in vivo. These findings reveal, for the first time, that the tumor suppressor activity of miR-124 could be partly due to its inhibitory effects on glioma stem-like traits and invasiveness through SNAI2.

Deep learning enabled fast 3D brain MRI at 0.055 tesla.

In recent years, there has been an intensive development of portable ultralow-field magnetic resonance imaging (MRI) for low-cost, shielding-free, and point-of-care applications. However, its quality is poor and scan time is long. We propose a fast acquisition and deep learning reconstruction framework to accelerate brain MRI at 0.055 tesla. The acquisition consists of a single average three-dimensional (3D) encoding with 2D partial Fourier sampling, reducing the scan time of T1- and T2-weighted imaging protocols to 2.5 and 3.2 minutes, respectively. The 3D deep learning leverages the homogeneous brain anatomy available in high-field human brain data to enhance image quality, reduce artifacts and noise, and improve spatial resolution to synthetic 1.5-mm isotropic resolution. Our method successfully overcomes low-signal barrier, reconstructing fine anatomical structures that are reproducible within subjects and consistent across two protocols. It enables fast and quality whole-brain MRI at 0.055 tesla, with potential for widespread biomedical applications.

Hyperoxia resensitizes chemoresistant human glioblastoma cells to temozolomide.

Temozolomide (TMZ) is standard chemotherapy for glioblastoma multiforme (GBM). Intratumoral hypoxia is common in GBM and may be associated with the development of TMZ resistance. Oxygen therapy has previously been reported to potentiate the effect of chemotherapy in cancer. In this study, we investigated whether hyperoxia can enhance the TMZ-induced cytotoxicity of human GBM cells, and whether and how it would resensitize TMZ-resistant GBM cells to TMZ. TMZ-sensitive human GBM cells (D54-S and U87-S) were treated with TMZ to develop isogenic subclones of TMZ-resistant cells (D54-R and U87-R). All cell lines were then exposed to different oxygen levels (1, 21, 40, or 80 %), with or without concomitant TMZ treatment, before assessment of cell cytotoxicity and morphology. Cell death and survival pathways elicited by TMZ and/or hyperoxia were elucidated by western blotting. Our results showed that TMZ sensitivity of both chemo-sensitive and resistant cells was enhanced significantly under hyperoxia. At the cell line-specific optimum oxygen concentration (D54-R, 80 %; U87-R, 40 %), resistant cells had the same response to TMZ as the parent chemosensitive cells under normoxia via the caspase-dependent pathway. Both TMZ and hyperoxia were associated with increased phosphorylation of ERK p44/42 MAPK (Erk1/2), but to a lesser extent in D54-R cells, suggesting that Erk1/2 activity may be involved in regulation of hyperoxia and TMZ-mediated cell death. Overall, hyperoxia enhanced TMZ toxicity in GBM cells by induction of apoptosis, possibly via MAPK-related pathways. Induced hyperoxia is a potentially promising approach for treatment of TMZ-resistant GBM.

Protein alterations associated with temozolomide resistance in subclones of human glioblastoma cell lines.

Temozolomide (TMZ) is the standard chemotherapeutic agent for human malignant glioma, but intrinsic or acquired chemoresistance represents a major obstacle to successful treatment of this highly lethal group of tumours. Obtaining better understanding of the molecular mechanisms underlying TMZ resistance in malignant glioma is important for the development of better treatment strategies. We have successfully established a passage control line (D54-C10) and resistant variants (D54-P5 and D54-P10) from the parental TMZ-sensitive malignant glioma cell line D54-C0. The resistant sub-cell lines showed alterations in cell morphology, enhanced cell adhesion, increased migration capacities, and cell cycle arrests. Proteomic analysis identified a set of proteins that showed gradual changes in expression according to their 50% inhibitory concentration (IC(50)). Successful validation was provided by transcript profiling in another malignant glioma cell line U87-MG and its resistant counterparts. Moreover, three of the identified proteins (vimentin, cathepsin D and prolyl 4-hydroxylase, beta polypeptide) were confirmed to be upregulated in high-grade glioma. Our data suggest that acquired TMZ resistance in human malignant glioma is associated with promotion of malignant phenotypes, and our reported molecular candidates may serve not only as markers of chemoresistance but also as potential therapeutic targets in the treatment of TMZ-resistant human malignant glioma, providing a platform for future investigations.

Postcooling But Not Precooling Benefits Motor Recovery by Suppressing Cell Death After Surgical Spinal Cord Injury in Rats.

BACKGROUND: Surgical spinal cord injury (SSCI) is often inevitable in patients with intramedullary lesions. Although regional hypothermia (RH) has been demonstrated neuroprotective, the value of priming RH in SSCI has never been studied. Herein, the authors investigated the impact of pre- and post-RH on neurologic recovery in a clinically relevant model. METHODS: An SSCI model was established at T10. RH was conducted by focal 4oC saline perfusion; room temperature (RT) saline was used as controls. Animals were randomized into 6 groups: SHAM-RT/RH, Pre-RT/RH, and Post-RT/RH. Motor and sensory functions were evaluated using the Basso, Beattie, and Bresnahan rating scale and Plantar test 2 weeks after surgery. TUNEL assay and Fluoro-Jade C staining were conducted to examine the cell death, and the alterations of apoptotic markers including total and cleaved casepase 3, Bcl-2, and Bax, as well as the pyroptotic proteins including NLRP3, ASC, and caspase 1, were determined. RESULTS: RH perfusion successfully created an intramedullary hypothermia approximately at 24oC, while RT controls remained above 30oC. Animals receiving postinjury RH had the least cell death and the best motor performance, while pre-RH showed the most dead cells and worst hind limb movements. Immunoblotting depicted that post-RH suppressed both apoptotic and pyroptotic death as the cleaved/total caspase 3, Bcl-2/Bax ratio, and NLRP3/ASC/caspase 1 signaling were inhibited. Priming cooling, on the contrary, elevated pyroptosis and did not affect apoptosis significantly. CONCLUSIONS: Priming RH before surgical incision could not be supported as it caused excessive cell death. In contrast, instant introduction of RH is beneficial in rescuing neurologic function.

Development and Validation of Indicators for Population Injury Surveillance in Hong Kong: Development and Usability Study.

BACKGROUND: Injury is an increasingly pressing global health issue. An effective surveillance system is required to monitor the trends and burden of injuries. OBJECTIVE: This study aimed to identify a set of valid and context-specific injury indicators to facilitate the establishment of an injury surveillance program in Hong Kong. METHODS: This development of indicators adopted a multiphased modified Delphi research design. A literature search was conducted on academic databases using injury-related search terms in various combinations. A list of potential indicators was sent to a panel of experts from various backgrounds to rate the validity and context-specificity of these indicators. Local hospital data on the selected core indicators were used to examine their applicability in the context of Hong Kong. RESULTS: We reviewed 142 articles and identified 55 indicators, which were classified into 4 domains. On the basis of the ratings by the expert panel, 13 indicators were selected as core indicators because of their good validity and high relevance to the local context. Among these indicators, 10 were from the construct of health care service use, and 3 were from the construct of postdischarge outcomes. Regression analyses of local hospitalization data showed that the Hong Kong Safe Community certification status had no association with 5 core indicators (admission to intensive care unit, mortality rate, length of intensive care unit stay, need for a rehabilitation facility, and long-term behavioral and emotional outcomes), negative associations with 4 core indicators (operative intervention, infection rate, length of hospitalization, and disability-adjusted life years), and positive associations with the remaining 4 core indicators (attendance to accident and emergency department, discharge rate, suicide rate, and hospitalization rate after attending the accident and emergency department). These results confirmed the validity of the selected core indicators for the quantification of injury burden and evaluation of injury-related services, although some indicators may better measure the consequences of severe injuries. CONCLUSIONS: This study developed a set of injury outcome indicators that would be useful for monitoring injury trends and burdens in Hong Kong.

A low-cost and shielding-free ultra-low-field brain MRI scanner.

Magnetic resonance imaging is a key diagnostic tool in modern healthcare, yet it can be cost-prohibitive given the high installation, maintenance and operation costs of the machinery. There are approximately seven scanners per million inhabitants and over 90% are concentrated in high-income countries. We describe an ultra-low-field brain MRI scanner that operates using a standard AC power outlet and is low cost to build. Using a permanent 0.055 Tesla Samarium-cobalt magnet and deep learning for cancellation of electromagnetic interference, it requires neither magnetic nor radiofrequency shielding cages. The scanner is compact, mobile, and acoustically quiet during scanning. We implement four standard clinical neuroimaging protocols (T1- and T2-weighted, fluid-attenuated inversion recovery like, and diffusion-weighted imaging) on this system, and demonstrate preliminary feasibility in diagnosing brain tumor and stroke. Such technology has the potential to meet clinical needs at point of care or in low and middle income countries.

3D dissimilar-siamese-u-net for hyperdense Middle cerebral artery sign segmentation.

The hyperdense middle cerebral artery sign (HMCAS) representing a thromboembolus has been declared as a vital CT finding for intravascular thrombus in the diagnosis of acute ischemia stroke. Early recognition of HMCAS can assist in patient triage and subsequent thrombolysis or thrombectomy treatment. A total of 624 annotated head non-contrast-enhanced CT (NCCT) image scans were retrospectively collected from multiple public hospitals in Hong Kong. In this study, we present a deep Dissimilar-Siamese-U-Net (DSU-Net) that is able to precisely segment the lesions by integrating Siamese and U-Net architectures. The proposed framework consists of twin sub-networks that allow inputs of left and right hemispheres in head NCCT images separately. The proposed Dissimilar block fully explores the feature representation of the differences between the bilateral hemispheres. Ablation studies were carried out to validate the performance of various components of the proposed DSU-Net. Our findings reveal that the proposed DSU-Net provides a novel approach for HMCAS automatic segmentation and it outperforms the baseline U-Net and many state-of-the-art models for clinical practice.

Quantification of injury burden using multiple data sources: a longitudinal study.

Quantification of injury burden is vital for injury prevention, as it provides a guide for setting policies and priorities. This study generated a set of Hong Kong specific disability weights (DWs) derived from patient experiences and hospital records. Patients were recruited from the Accident and Emergency Department (AED) of three major trauma centers in Hong Kong between September 2014 and December 2015 and subsequently interviewed with a focus on health-related quality of life at most three times over a 12-month period. These patient-reported data were then used for estimation of DWs. The burden of injury was determined using the mortality and inpatient data from 2001 to 2012 and then compared with those reported in the UK Burden of Injury (UKBOI) and global burden of diseases (GBD) studies. There were 22,856 mortality cases and 817,953 morbidity cases caused by injuries, in total contributing to 1,027,641 disability-adjusted life years (DALYs) in the 12-year study timeframe. Estimates for DALYs per 100,000 in Hong Kong amounted to 1192, compared with 2924 in UKBOI and 3459 in GBD. Our findings support the use of multiple data sources including patient-reported data and hospital records for estimation of injury burden.