Accurate virus identification with interpretable Raman signatures by machine learning.

Rapid identification of newly emerging or circulating viruses is an important first step toward managing the public health response to potential outbreaks. A portable virus capture device, coupled with label-free Raman spectroscopy, holds the promise of fast detection by rapidly obtaining the Raman signature of a virus followed by a machine learning (ML) approach applied to recognize the virus based on its Raman spectrum, which is used as a fingerprint. We present such an ML approach for analyzing Raman spectra of human and avian viruses. A convolutional neural network (CNN) classifier specifically designed for spectral data achieves very high accuracy for a variety of virus type or subtype identification tasks. In particular, it achieves 99% accuracy for classifying influenza virus type A versus type B, 96% accuracy for classifying four subtypes of influenza A, 95% accuracy for differentiating enveloped and nonenveloped viruses, and 99% accuracy for differentiating avian coronavirus (infectious bronchitis virus [IBV]) from other avian viruses. Furthermore, interpretation of neural net responses in the trained CNN model using a full-gradient algorithm highlights Raman spectral ranges that are most important to virus identification. By correlating ML-selected salient Raman ranges with the signature ranges of known biomolecules and chemical functional groups­for example, amide, amino acid, and carboxylic acid­we verify that our ML model effectively recognizes the Raman signatures of proteins, lipids, and other vital functional groups present in different viruses and uses a weighted combination of these signatures to identify viruses.

AI Chatbots in Clinical Laboratory Medicine: Foundations and Trends.

BACKGROUND: Artificial intelligence (AI) conversational agents, or chatbots, are computer programs designed to simulate human conversations using natural language processing. They offer diverse functions and applications across an expanding range of healthcare domains. However, their roles in laboratory medicine remain unclear, as their accuracy, repeatability, and ability to interpret complex laboratory data have yet to be rigorously evaluated. CONTENT: This review provides an overview of the history of chatbots, two major chatbot development approaches, and their respective advantages and limitations. We discuss the capabilities and potential applications of chatbots in healthcare, focusing on the laboratory medicine field. Recent evaluations of chatbot performance are presented, with a special emphasis on large language models such as the Chat Generative Pre-trained Transformer in response to laboratory medicine questions across different categories, such as medical knowledge, laboratory operations, regulations, and interpretation of laboratory results as related to clinical context. We analyze the causes of chatbots' limitations and suggest research directions for developing more accurate, reliable, and manageable chatbots for applications in laboratory medicine. SUMMARY: Chatbots, which are rapidly evolving AI applications, hold tremendous potential to improve medical education, provide timely responses to clinical inquiries concerning laboratory tests, assist in interpreting laboratory results, and facilitate communication among patients, physicians, and laboratorians. Nevertheless, users should be vigilant of existing chatbots' limitations, such as misinformation, inconsistencies, and lack of human-like reasoning abilities. To be effectively used in laboratory medicine, chatbots must undergo extensive training on rigorously validated medical knowledge and be thoroughly evaluated against standard clinical practice.

Tenofovir Diphosphate Concentrations in Dried Blood Spots From Pregnant and Postpartum Adolescent and Young Women Receiving Daily Observed Pre-exposure Prophylaxis in Sub-Saharan Africa.

BACKGROUND: Intracellular tenofovir diphosphate (TFV-DP) concentration in dried blood spots (DBSs) is used to monitor cumulative pre-exposure prophylaxis (PrEP) adherence. We evaluated TFV-DP in DBSs following daily oral PrEP (emtricitabine 200 mg/tenofovir diphosphate 300 mg) among pregnant and postpartum adolescent girls and young women (AGYW). METHODS: Directly observed PrEP was administered for 12 weeks in a pregnancy (14-24 weeks' gestation, n = 20) and postpartum (6-12 weeks postpartum, n = 20) group of AGYW aged 16-24 years in sub-Saharan Africa. Weekly DBS TFV-DP was measured by validated liquid chromatography-tandem mass spectrometry assay. Week 12 TFV-DP distributions were compared between groups with Wilcoxon test. Population pharmacokinetic models were fit to estimate steady-state concentrations and create benchmarks for adherence categories. Baseline correlates of TFV-DP were evaluated. RESULTS: Median age was 20 (IQR, 19-22) years. Of 3360 doses, 3352 (>99%) were directly observed. TFV-DP median (IQR) half-life was 10 (7-12) days in pregnancy and 17 (14-21) days postpartum, with steady state achieved by 5 and 8 weeks, respectively. Observed median (IQR) steady-state TFV-DP was 965 fmol/punch (691-1166) in pregnancy versus 1406 fmol/punch (1053-1859) postpartum (P = .006). Modeled median steady-state TFV-DP was 881 fmol/punch (667-1105) in pregnancy versus 1438 fmol/punch (1178-1919) postpartum. In pooled analysis, baseline creatinine clearance was associated with observed TFV-DP concentrations. CONCLUSIONS: TFV-DP in African AGYW was approximately one-third lower in pregnancy than postpartum. These Population-specific benchmarks can be used to guide PrEP adherence support in pregnant/postpartum African women. CLINICAL TRIALS REGISTRATION: NCT03386578.

Hepatotoxicity and Liver-Related Mortality in Women of Childbearing Potential Living With Human Immunodeficiency Virus and High CD4 Cell Counts Initiating Efavirenz-Containing Regimens.

BACKGROUND: Severe hepatotoxicity in people with human immunodeficiency virus (HIV) receiving efavirenz (EFV) has been reported. We assessed the incidence and risk factors of hepatotoxicity in women of childbearing age initiating EFV-containing regimens. METHODS: In the Promoting Maternal and Infant Survival Everywhere (PROMISE) trial, ART-naive pregnant women with HIV and CD4 count ≥ 350 cells/µL and alanine aminotransferase ≤ 2.5 the upper limit of normal were randomized during the antepartum and postpartum periods to antiretroviral therapy (ART) strategies to assess HIV vertical transmission, safety, and maternal disease progression. Hepatotoxicity was defined per the Division of AIDS Toxicity Tables. Cox proportional hazards models were constructed with covariates including participant characteristics, ART regimens, and timing of EFV initiation. RESULTS: Among 3576 women, 2435 (68%) initiated EFV at a median 121.1 weeks post delivery. After EFV initiation, 2.5% (61/2435) had severe (grade 3 or higher) hepatotoxicity with an incidence of 2.3 (95% confidence interval [CI], 2.0-2.6) per 100 person-years. Events occurred between 1 and 132 weeks postpartum. Of those with severe hepatotoxicity, 8.2% (5/61) were symptomatic, and 3.3% (2/61) of those with severe hepatotoxicity died from EFV-related hepatotoxicity, 1 of whom was symptomatic. The incidence of liver-related mortality was 0.07 (95% CI, .06-.08) per 100 person-years. In multivariable analysis, older age was associated with severe hepatotoxicity (adjusted hazard ratio per 5 years, 1.35 [95% CI, 1.06-1.70]). CONCLUSIONS: Severe hepatotoxicity after EFV initiation occurred in 2.5% of women and liver-related mortality occurred in 3% of those with severe hepatotoxicity. The occurrence of fatal events underscores the need for safer treatments for women of childbearing age.

Risk Factors for Adverse Radiographic Outcomes After Elastic Stable Intramedullary Nailing of Unstable Diaphyseal Tibia Fractures in Children.

OBJECTIVES: Elastic stable intramedullary nails (ESIN) are commonly utilized to treat unstable pediatric tibia fractures but have been associated with complications. The purpose of this study was to identify risk factors for adverse radiographic outcomes after ESIN of pediatric tibia fractures. METHODS: A retrospective review of all patients who underwent diaphyseal tibia fracture stabilization with ESIN between 2010 and 2018 at 3 pediatric level 1 trauma centers was performed. Inclusion criteria were open growth plates, no intra-articular or physeal fracture involvement, and radiographic follow-up until union. Patient demographics, injury mechanism, fracture characteristics, and implant fill relative to the medullary canal were recorded. Radiographic outcome measures included achievement of and time to union, residual angular deformity, and additional procedures. RESULTS: One hundred seventy-two patients met inclusion criteria and were followed for a mean of 1.2 years. Nonunions were observed in 3% of the patient cohort. Another 10% required >6 months to heal, but did not require further surgical intervention. Angular deformities were common with 57% having a residual deformity ≥5 degrees and 14% having a residual deformity ≥10 degrees. Of the patients with a residual deformity between 5 and 10 degrees, 3% were symptomatic, where as 26% of the patients with a residual deformity ≥10 degrees were symptomatic. Greater angular deformities were associated with open fractures, compartment syndrome, and longer time to union. Patient age, weight, tibial comminution, and canal fill were not associated with nonunions or malunions. CONCLUSIONS: ESIN of pediatric tibia fractures results in reliable healing for a majority of patients, but poses risks for residual angular deformities and delayed healing. Open fractures and compartment syndrome were associated with adverse radiographic outcomes.

Social inclusion among mental health service users: A reliability test of the Mini-SCOPE Scale.

AIM AND OBJECTIVES: To evaluate the reliability of the Mini-SCOPE scale through interitem consistency and test-retest consistency, as well as the initial correlation with outcome measures of recovery. BACKGROUND: Adapted from a UK, Social and Communities Opportunities Profile (Mini-SCOPE) is a short version of an social inclusion measurement for English-speaking persons in recovery (PIR) in Singapore. Prior concept mapping sets the stage for this reliability study. DESIGN: This study adopted a nonexperimental, pre- and postdesign to validate the psychosocial measurement tool for community services. METHOD: Convenient sampling was conducted at the various designated clinics. A total of 170 voluntary participants from psychiatric outpatient clinics were recruited for this study. It evaluated the 4-week interval test-retest reliability of the Mini-SCOPE. "AGREE" equator checklist was completed to guide the reporting of clinical practice. "See Supporting Information File S1." RESULT: Outcome demonstrated that the Mini-SCOPE scale has good strength of reliability. CONCLUSION: This study showed that the Mini-SCOPE measurement has the potential to be used for programme evaluation in mental health settings. RELEVANCE TO CLINICAL PRACTICE: Applicable to nurses and other mental health professionals to consider the social and wellness aspects of the patients in their care when planning appropriate services.

Loss of MeCP2 in the rat models regression, impaired sociability and transcriptional deficits of Rett syndrome.

Mouse models of the transcriptional modulator Methyl-CpG-Binding Protein 2 (MeCP2) have advanced our understanding of Rett syndrome (RTT). RTT is a 'prototypical' neurodevelopmental disorder with many clinical features overlapping with other intellectual and developmental disabilities (IDD). Therapeutic interventions for RTT may therefore have broader applications. However, the reliance on the laboratory mouse to identify viable therapies for the human condition may present challenges in translating findings from the bench to the clinic. In addition, the need to identify outcome measures in well-chosen animal models is critical for preclinical trials. Here, we report that a novel Mecp2 rat model displays high face validity for modelling psychomotor regression of a learned skill, a deficit that has not been shown in Mecp2 mice. Juvenile play, a behavioural feature that is uniquely present in rats and not mice, is also impaired in female Mecp2 rats. Finally, we demonstrate that evaluating the molecular consequences of the loss of MeCP2 in both mouse and rat may result in higher predictive validity with respect to transcriptional changes in the human RTT brain. These data underscore the similarities and differences caused by the loss of MeCP2 among divergent rodent species which may have important implications for the treatment of individuals with disease-causing MECP2 mutations. Taken together, these findings demonstrate that the Mecp2 rat model is a complementary tool with unique features for the study of RTT and highlight the potential benefit of cross-species analyses in identifying potential disease-relevant preclinical outcome measures.

Patient-specific driver gene prediction and risk assessment through integrated network analysis of cancer omics profiles.

Extensive and multi-dimensional data sets generated from recent cancer omics profiling projects have presented new challenges and opportunities for unraveling the complexity of cancer genome landscapes. In particular, distinguishing the unique complement of genes that drive tumorigenesis in each patient from a sea of passenger mutations is necessary for translating the full benefit of cancer genome sequencing into the clinic. We address this need by presenting a data integration framework (OncoIMPACT) to nominate patient-specific driver genes based on their phenotypic impact. Extensive in silico and in vitro validation helped establish OncoIMPACT's robustness, improved precision over competing approaches and verifiable patient and cell line specific predictions (2/2 and 6/7 true positives and negatives, respectively). In particular, we computationally predicted and experimentally validated the gene TRIM24 as a putative novel amplified driver in a melanoma patient. Applying OncoIMPACT to more than 1000 tumor samples, we generated patient-specific driver gene lists in five different cancer types to identify modes of synergistic action. We also provide the first demonstration that computationally derived driver mutation signatures can be overall superior to single gene and gene expression based signatures in enabling patient stratification and prognostication. Source code and executables for OncoIMPACT are freely available from http://sourceforge.net/projects/oncoimpact.

Illness management and recovery program for mental health problems: reducing symptoms and increasing social functioning.

AIMS AND OBJECTIVES: To evaluate the effectiveness of the Illness Management and Recovery Program in comparison with the current standard of care in terms of reduction of symptoms, rehospitalisation rates and social functioning in Asia. BACKGROUND: Focus of treatment for Mental Health had been shifted from mere management of symptoms to that of achievement of recovery. In the recovery process, strategies to achieve higher level of functioning were used (Psychiatric Services 2014, 65, 171). However, two main factors hindered clients from attaining recovery: first, the lack of Mental Health resources in the community and second the negative attitudes of healthcare professionals towards mental illness (American Journal of Psychiatric Rehabilitation 2012, 15, 131). Hence, it is essential to implement an effective programme that will train mental health professionals to use more effective techniques and materials in helping the clients to better integrate into society by achieving skills in their attempt to work towards recovery. DESIGN: This study adopts a time series experimental quantitative design. METHODS: Fifty participants who consented to the study were randomly assigned to two groups. Participants in the experimental group received the experimental management and recovery programme, while the control group received standard care management by the community psychiatric nurses for a period of 12 months. RESULTS: Participants in the experimental group reported significantly lower number of admissions, shorter length of stay, lower Brief Psychiatric Rating Scale scores, and also reported significantly higher scores on both the Illness Management and Recovery Scale and the Global Assessment Scale. CONCLUSION: This study demonstrated the effectiveness of IMR in helping Asian people with mental illness to not only reduce symptoms and hospitalisations but also improve social functioning. They have benefitted from the program although they are living in a different cultural setting from where IMR was developed. RELEVANCE TO CLINICAL PRACTICE: The success of this study has raised the standard of care in the community intervention and led more people to their recovery.

Epigenome-wide DNA methylation landscape of melanoma progression to brain metastasis reveals aberrations on homeobox D cluster associated with prognosis.

Melanoma brain metastasis (MBM) represents a frequent complication of cutaneous melanoma. Despite aggressive multi-modality therapy, patients with MBM often have a survival rate of <1 year. Alteration in DNA methylation is a major hallmark of tumor progression and metastasis; however, it remains largely unexplored in MBM. In this study, we generated a comprehensive DNA methylation landscape through the use of genome-wide copy number, DNA methylation and gene expression data integrative analysis of melanoma progression to MBM. A progressive genome-wide demethylation in low CpG density and an increase in methylation level of CpG islands according to melanoma progression were observed. MBM-specific partially methylated domains (PMDs) affecting key brain developmental processes were identified. Differentially methylated CpG sites between MBM and lymph node metastasis (LNM) from patients with good prognosis were identified. Among the most significantly affected genes were the HOX family members. DNA methylation of HOXD9 gene promoter affected transcript and protein expression and was significantly higher in MBM than that in early stages. A MBM-specific PMD was identified in this region. Low methylation level of this region was associated with active HOXD9 expression, open chromatin and histone modifications associated with active transcription. Demethylating agent induced HOXD9 expression in melanoma cell lines. The clinical relevance of this finding was verified in an independent large cohort of melanomas (n = 145). Patients with HOXD9 hypermethylation in LNM had poorer disease-free and overall survival. This epigenome-wide study identified novel methylated genes with functional and clinical implications for MBM patients.

Marine Bioinspired Underwater Contact Adhesion.

Marine mussels and barnacles are sessile biofouling organisms that adhere to a number of surfaces in wet environments and maintain remarkably strong bonds. Previous synthetic approaches to mimic biological wet adhesive properties have focused mainly on the catechol moiety, present in mussel foot proteins (mfps), and especially rich in the interfacial mfps, for example, mfp-3 and -5, found at the interface between the mussel plaque and substrate. Barnacles, however, do not use Dopa for their wet adhesion, but are instead rich in noncatecholic aromatic residues. Due to this anomaly, we were intrigued to study the initial contact adhesion properties of copolymerized acrylate films containing the key functionalities of barnacle cement proteins and interfacial mfps, for example, aromatic (catecholic or noncatecholic), cationic, anionic, and nonpolar residues. The initial wet contact adhesion of the copolymers was measured using a probe tack testing apparatus with a flat-punch contact geometry. The wet contact adhesion of an optimized, bioinspired copolymer film was ∼15.0 N/cm(2) in deionized water and ∼9.0 N/cm(2) in artificial seawater, up to 150 times greater than commercial pressure-sensitive adhesive (PSA) tapes (∼0.1 N/cm(2)). Furthermore, maximum wet contact adhesion was obtained at ∼pH 7, suggesting viability for biomedical applications.

Genome-wide characterization of circulating tumor cells identifies novel prognostic genomic alterations in systemic melanoma metastasis.

BACKGROUND: Circulating tumor cells (CTC) have been found in patients with metastatic melanoma and are associated with advanced melanoma stage and poor patient outcome. We hypothesize that CTC harbor genomic changes critical in the development of distant systemic metastasis. Here, we present the first genome-wide copy-number aberration (CNA) and loss of heterozygosity (LOH)-based characterization of melanoma CTC. METHODS: CTC were isolated from peripheral blood monocytes of 13 melanoma patients with regional metastasis stage IIIB/C using antibodies against melanoma-associated cell surface gangliosides. RESULTS: We characterized 251 CNA in CTC. Comparative analysis demonstrated >90% concordance in single-nucleotide polymorphism profiles between paired CTC and tumor metastases. In particular, there were notable recurring CNA across patients. In exploratory studies, the presence of several top CTC-associated CNA was verified in distant metastasis (stage IV) from 27 patients, suggesting that certain genomic changes are propagated from regional metastasis to CTC and to distant systemic metastases. Lastly, an exploratory biomarker panel derived from 5 CTC-associated CNA [CSMD2 (CUB and Sushi multiple domains 2), 1p35.1; CNTNAP5 (contactin associated protein-like 5), 2q14.3; NRDE2 (NRDE-2, necessary for RNA interference, domain containing), 14q32.11; ADAM6 (ADAM metallopeptidase domain 6, pseudogene), 14q32.33; and TRPM2 (transient receptor potential cation channel, subfamily m, member 2), 21q22.3] conferred prognostic utility for melanoma recurrence [hazard ratio (HR), 1.14; CI, 1.00-1.44; P = 0.0471] and death (HR, 2.86; CI, 1.23-14.42; P = 0.0014) in 35 patients with stage IIIB/C melanoma, with a 5-year disease-free survival of 13% vs 69% (P = 0.0006) and overall survival of 28% vs 94% between high-risk and low-risk groups defined by the biomarker panel, respectively. CONCLUSIONS: This study provides the first detailed CNA-based profile of melanoma CTC and illustrates how CTC may be used as a novel approach for identification of systemic metastasis.

Prevalence of detectable HIV-DNA and -RNA in cerebrospinal fluid of youth with perinatal HIV and impaired cognition on antiretroviral therapy.

OBJECTIVE: Central nervous system (CNS) HIV infection can impact cognition and may be an obstacle to cure in adolescents and young adults with perinatal HIV (AYAPHIV). IMPAACT2015 enrolled AYAPHIV on suppressive antiretroviral therapy (ART) with cognitive impairment to detect and quantify HIV in blood and cerebrospinal fluid (CSF). DESIGN: IMPAACT2015 was a U.S.-based multi-site, exploratory, observational study. METHODS: Cognitive impairment was defined as NIH Toolbox Fluid Cognition Composite score (FCCS) > 1 standard deviation below age-adjusted normative group mean. Cell-free HIV-RNA and cell-associated HIV pol/gag -DNA and 10 biomarkers of inflammation/neuronal injury were measured in paired CSF and blood. ART exposure concentrations were quantified in hair. RESULTS: Among 24 participants, 20 had successful CSF collection and 18 also met viral suppression criteria. 9/18 (50%) were female sex-at-birth, 14/18 (78%) were Black. Median (range) age was 20 years (13-27), time on ART 18.3 years (8.0-25.5), and FCCS 68 (53-80). HIV-DNA was detected in PBMCs from all participants. In CSF, 2/18 (11%, 95% CI: 1.4-34.7%) participants had detectable cell-free HIV-RNA, while HIV gag or pol -DNA was detectable in 13/18 (72%, 95% CI: 47-90%). Detectable HIV-DNA in CSF was associated with male sex-at-birth (p = 0.051), lower CD4 count at enrollment (p = 0.016), and higher PBMC HIV pol -DNA copies (p = 0.058). Hair antiretroviral concentrations and biomarkers were not associated with CSF HIV-DNA detection. CONCLUSIONS: We found a high proportion of AYAPHIV with neurocognitive impairment had CSF cells harboring HIV-DNA during long-term virologic suppression. This evidence of persistent HIV-DNA in CSF suggests that the CNS should be considered in treatment and cure studies.

Immune correlates of herpes zoster in HIV-infected children and youth.

To identify immunologic factors that modulate the risk of herpes zoster (HZ), we compared varicella-zoster virus (VZV)-specific and nonspecific T-cell subpopulations of 47 HIV-infected children before they developed HZ with those of 141 VZV-positive HZ-negative matched controls. Compared with controls, HZ cases had lower VZV-specific CD8(+) CD107a(+) cell percentages independently of CD4(+) percentages or HIV loads, suggesting that VZV-specific cytotoxic T cells are protective against HZ. In contrast, high nonspecific regulatory and activated T cells were associated with an increased risk of HZ.

Semi-supervised body parsing and pose estimation for enhancing infant general movement assessment.

General movement assessment (GMA) of infant movement videos (IMVs) is an effective method for early detection of cerebral palsy (CP) in infants. We demonstrate in this paper that end-to-end trainable neural networks for image sequence recognition can be applied to achieve good results in GMA, and more importantly, augmenting raw video with infant body parsing and pose estimation information can significantly improve performance. To solve the problem of efficiently utilizing partially labeled IMVs for body parsing, we propose a semi-supervised model, termed SiamParseNet (SPN), which consists of two branches, one for intra-frame body parts segmentation and another for inter-frame label propagation. During training, the two branches are jointly trained by alternating between using input pairs of only labeled frames and input of both labeled and unlabeled frames. We also investigate training data augmentation by proposing a factorized video generative adversarial network (FVGAN) to synthesize novel labeled frames for training. FVGAN decouples foreground and background generation which allows for generating multiple labeled frames from one real labeled frame. When testing, we employ a multi-source inference mechanism, where the final result for a test frame is either obtained via the segmentation branch or via propagation from a nearby key frame. We conduct extensive experiments for body parsing using SPN on two infant movement video datasets; on these partially labeled IMVs, we show that SPN coupled with FVGAN achieves state-of-the-art performance. We further demonstrate that our proposed SPN can be easily adapted to the infant pose estimation task with superior performance. Last but not least, we explore the clinical application of our method for GMA. We collected a new clinical IMV dataset with GMA annotations, and our experiments show that our SPN models for body parsing and pose estimation trained on the first two datasets generalize well to the new clinical dataset and their results can significantly boost the convolutional recurrent neural network (CRNN) based GMA prediction performance when combined with raw video inputs.

Home visiting impacts during the pandemic: Evidence from a randomized controlled trial of child first.

Existing research has found that home visiting programs for families with young children can improve children's development and strengthen caregiver and family well-being. However, the pandemic created numerous challenges for home visiting programs, forcing them to deliver services online or in a hybrid format to respond to pandemic-related challenges. Questions remain about the impacts of these programs when delivered at-scale via a hybrid model, especially during this uniquely challenging time. The present study reports 12-month impacts from a randomized controlled trial of Child First-an evidence-based home visiting program that provides psychotherapeutic, parent-child intervention (children ages 0-5) embedded in a coordinated system of care-when implemented as a hybrid service. This study estimates impacts within four domains: families' receipt of services, caregiver psychological well-being and parenting, child behavior, and family economic well-being. After randomly assigning families (N = 226) to receive Child First or typical community services, the research team surveyed caregivers (N = 183) about a year after study enrollment. Results from regression models with site fixed effects revealed suggestive evidence that Child First reduced caregivers' job loss, residential mobility, and self-reported substance abuse, and increased receipt of virtual services during the pandemic. There were null impacts on caregivers' psychological well-being, families' involvement with the child welfare system, children's behaviors, and other indicators of economic well-being. Implications for future research and policy are discussed. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Sinogram domain angular upsampling of sparse-view micro-CT with dense residual hierarchical transformer and attention-weighted loss.

Reduced angular sampling is a key strategy for increasing scanning efficiency of micron-scale computed tomography (micro-CT). Despite boosting throughput, this strategy introduces noise and extrapolation artifacts due to undersampling. In this work, we present a solution to this issue, by proposing a novel Dense Residual Hierarchical Transformer (DRHT) network to recover high-quality sinograms from 2×, 4× and 8× undersampled scans. DRHT is trained to utilize limited information available from sparsely angular sampled scans and once trained, it can be applied to recover higher-resolution sinograms from shorter scan sessions. Our proposed DRHT model aggregates the benefits of a hierarchical- multi-scale structure along with the combination of local and global feature extraction through dense residual convolutional blocks and non-overlapping window transformer blocks respectively. We also propose a novel noise-aware loss function named KL-L1 to improve sinogram restoration to full resolution. KL-L1, a weighted combination of pixel-level and distribution-level cost functions, leverages inconsistencies in noise distribution and uses learnable spatial weight maps to improve the training of the DRHT model. We present ablation studies and evaluations of our method against other state-of-the-art (SOTA) models over multiple datasets. Our proposed DRHT network achieves an average increase in peak signal to noise ratio (PSNR) of 17.73 dB and a structural similarity index (SSIM) of 0.161, for 8× upsampling, across the three diverse datasets, compared to their respective Bicubic interpolated versions. This novel approach can be utilized to decrease radiation exposure to patients and reduce imaging time for large-scale CT imaging projects.

Forecasting User Interests Through Topic Tag Predictions in Online Health Communities.

The increasing reliance on online communities for healthcare information by patients and caregivers has led to the increase in the spread of misinformation, or subjective, anecdotal and inaccurate or non-specific recommendations, which, if acted on, could cause serious harm to the patients. Hence, there is an urgent need to connect users with accurate and tailored health information in a timely manner to prevent such harm. This article proposes an innovative approach to suggesting reliable information to participants in online communities as they move through different stages in their disease or treatment. We hypothesize that patients with similar histories of disease progression or course of treatment would have similar information needs at comparable stages. Specifically, we pose the problem of predicting topic tags or keywords that describe the future information needs of users based on their profiles, traces of their online interactions within the community (past posts, replies) and the profiles and traces of online interactions of other users with similar profiles and similar traces of past interaction with the target users. The result is a variant of the collaborative information filtering or recommendation system tailored to the needs of users of online health communities. We report results of our experiments on two unique datasets from two different social media platforms which demonstrates the superiority of the proposed approach over the state of the art baselines with respect to accurate and timely prediction of topic tags (and hence information sources of interest).

BBox-Guided Segmentor: Leveraging expert knowledge for accurate stroke lesion segmentation using weakly supervised bounding box prior.

Stroke is one of the leading causes of death and disability in the world. Despite intensive research on automatic stroke lesion segmentation from non-invasive imaging modalities including diffusion-weighted imaging (DWI), challenges remain such as a lack of sufficient labeled data for training deep learning models and failure in detecting small lesions. In this paper, we propose BBox-Guided Segmentor, a method that significantly improves the accuracy of stroke lesion segmentation by leveraging expert knowledge. Specifically, our model uses a very coarse bounding box label provided by the expert and then performs accurate segmentation automatically. The small overhead of having the expert provide a rough bounding box leads to large performance improvement in segmentation, which is paramount to accurate stroke diagnosis. To train our model, we employ a weakly-supervised approach that uses a large number of weakly-labeled images with only bounding boxes and a small number of fully labeled images. The scarce fully labeled images are used to train a generator segmentation network, while adversarial training is used to leverage the large number of weakly-labeled images to provide additional learning signals. We evaluate our method extensively using a unique clinical dataset of 99 fully labeled cases (i.e., with full segmentation map labels) and 831 weakly labeled cases (i.e., with only bounding box labels), and the results demonstrate the superior performance of our approach over state-of-the-art stroke lesion segmentation models. We also achieve competitive performance as a SOTA fully supervised method using less than one-tenth of the complete labels. Our proposed approach has the potential to improve stroke diagnosis and treatment planning, which may lead to better patient outcomes.

Elective musculoskeletal surgery is associated with postoperative weight changes in pediatric and adolescent patients.

Purpose: Patients of elective orthopedic surgeries often reduce activity levels during postoperative recovery. It is unclear whether these extended periods of modified activities lead to weight changes. The purpose of this study was to evaluate changes in body mass index percentile in pediatric patients over 2.5 years following primary musculoskeletal surgeries. Methods: Institutional records for utilized current procedural terminology codes were used to identify patients aged 21 years or younger who underwent elective surgery at a single pediatric orthopedic institution between October 2016 and December 2018. Non-primary surgeries and patients without preoperative body mass index measurements were excluded. Demographic characteristics, height, weight, and body mass index within 30 months of surgery were collected. Body mass index relative to age was calculated. Analysis of body mass index changes at follow-up intervals of 3-7, 9-18, and 24-30 months after surgery was performed for the overall sample, within surgical categories, and within preoperative weight classifications. Results: A total of 1566 patients (53.1% female, average age 12.4 years) were included. Over one-third of patients were overweight or obese at presentation. The average change in body mass index percentile relative to baseline was increased at all follow-up intervals. Values reached significance at 9-18 months (p = .002) and 24-30 months (p = .001). While underweight and normal-weight patients had increased body mass index at all three timepoints, overweight or obese patients decreased. Conclusions: Patients undergoing elective orthopedic procedures may experience significant changes in body mass index percentile postoperatively. At extremes of weight, patients experience improvement toward the mean, but most patients may undergo body mass index increases beyond what would be expected during normal growth. Level of evidence: Retrospective level III.