Category: Statistics

Sci Rep. 2021 Nov 26;11(1):22997. doi: 10.1038/s41598-021-02476-9.

ABSTRACT

We present a simple and efficient hypothesis-free machine learning pipeline for risk factor discovery that accounts for non-linearity and interaction in large biomedical databases with minimal variable pre-processing. In this study, mortality models were built using gradient boosting decision trees (GBDT) and important predictors were identified using a Shapley values-based feature attribution method, SHAP values. Cox models controlled for false discovery rate were used for confounder adjustment, interpretability, and further validation. The pipeline was tested using information from 502,506 UK Biobank participants, aged 37-73 years at recruitment and followed over seven years for mortality registrations. From the 11,639 predictors included in GBDT, 193 potential risk factors had SHAP values ≥ 0.05, passed the correlation test, and were selected for further modelling. Of the total variable importance summed up, 60% was directly health related, and baseline characteristics, sociodemographics, and lifestyle factors each contributed about 10%. Cox models adjusted for baseline characteristics, showed evidence for an association with mortality for 166 out of the 193 predictors. These included mostly well-known risk factors (e.g., age, sex, ethnicity, education, material deprivation, smoking, physical activity, self-rated health, BMI, and many disease outcomes). For 19 predictors we saw evidence for an association in the unadjusted but not adjusted analyses, suggesting bias by confounding. Our GBDT-SHAP pipeline was able to identify relevant predictors ‘hidden’ within thousands of variables, providing an efficient and pragmatic solution for the first stage of hypothesis free risk factor identification.

PMID:34837000 | DOI:10.1038/s41598-021-02476-9

Sci Data. 2021 Nov 26;8(1):305. doi: 10.1038/s41597-021-01089-1.

ABSTRACT

Statistical Parametric Mapping (SPM) is a computational approach for analysing functional brain images like Positron Emission Tomography (PET). When performing SPM analysis for different patient populations, brain PET template images representing population-specific brain morphometry and metabolism features are helpful. However, most currently available brain PET templates were constructed using the Caucasian data. To enrich the family of publicly available brain PET templates, we created Chinese-specific template images based on 116 [¹⁸F]-fluorodeoxyglucose ([¹⁸F]-FDG) PET images of normal participants. These images were warped into a common averaged space, in which the mean and standard deviation templates were both computed. We also developed the SPM analysis programmes to facilitate easy use of the templates. Our templates were validated through the SPM analysis of Alzheimer’s and Parkinson’s patient images. The resultant SPM t-maps accurately depicted the disease-related brain regions with abnormal [¹⁸F]-FDG uptake, proving the templates’ effectiveness in brain function impairment analysis.

PMID:34836985 | DOI:10.1038/s41597-021-01089-1

Sci Rep. 2021 Nov 26;11(1):23002. doi: 10.1038/s41598-021-02330-y.

ABSTRACT

Radiotherapy requires the target area and the organs at risk to be contoured on the CT image of the patient. During the process of organs-at-Risk (OAR) of the chest and abdomen, the doctor needs to contour at each CT image. The delineations of large and varied shapes are time-consuming and laborious. This study aims to evaluate the results of two automatic contouring softwares on OARs definition of CT images of lung cancer and rectal cancer patients. The CT images of 15 patients with rectal cancer and 15 patients with lung cancer were selected separately, and the organs at risk were manually contoured by experienced physicians as reference structures. And then the same datasets were automatically contoured based on AiContour (version 3.1.8.0, Manufactured by Linking MED, Beijing, China) and Raystation (version 4.7.5.4, Manufactured by Raysearch, Stockholm, Sweden) respectively. Deep learning auto-segmentations and Atlas were respectively performed with AiContour and Raystation. Overlap index (OI), Dice similarity index (DSC) and Volume difference (D_v) were evaluated based on the auto-contours, and independent-sample t-test analysis is applied to the results. The results of deep learning auto-segmentations on OI and DSC were better than that of Atlas with statistical difference. There was no significant difference in D_v between the results of two software. With deep learning auto-segmentations, auto-contouring results of most organs in the chest and abdomen are good, and with slight modification, it can meet the clinical requirements for planning. With Atlas, auto-contouring results in most OAR is not as good as deep learning auto-segmentations, and only the auto-contouring results of some organs can be used clinically after modification.

PMID:34836989 | DOI:10.1038/s41598-021-02330-y

Sci Rep. 2021 Nov 26;11(1):22996. doi: 10.1038/s41598-021-02040-5.

ABSTRACT

Multiple N-methyl-D-aspartate (NMDA) receptor enhancing agents have had promising effects on cognition among patients with dementia. However, the results remain inconsistent. This exploratory meta-analysis investigated the effectiveness of NMDA receptor enhancing agents for cognitive function. PubMed, the Cochrane Central Register of Controlled Trials, and the Cochrane Database of Systematic Reviews were searched for randomized controlled trials (RCTs). Controlled trials assessing add-on NMDA receptor enhancing agent treatment in patients with dementia and using cognition rating scales were eligible and pooled using a random-effect model for comparisons. The standardized mean difference (SMD) was calculated in each study from the effect size; positive values indicated that NMDA receptor enhancing agent treatment improved cognitive function. Funnel plots and the I2 statistic were evaluated for statistical heterogeneity. Moderators were evaluated using meta-regression. We identified 14 RCTs with 2224 participants meeting the inclusion criteria. Add-on NMDA receptor enhancing agents had small positive significant effects on overall cognitive function among patients with dementia (SMD = 0.1002, 95% CI 0.0105-0.1900, P = 0.02860). Subgroup meta-analysis showed patients with Alzheimer’s Disease and trials using the Alzheimer Disease Assessment Scale-cognitive subscale as the primary outcome had small positive significant effects (SMD = 0.1042, 95% CI 0.0076-0.2007, P = 0.03451; SMD = 0.1267, 95% CI 0.0145-0.2388, P = 0.2686). This exploratory meta-analysis showed a very small, positive, and significant effect on overall cognition function in patients with dementia. Studies with larger samples are needed to evaluate different cognitive domains and phases of dementia.

PMID:34836972 | DOI:10.1038/s41598-021-02040-5

Sci Rep. 2021 Nov 26;11(1):22970. doi: 10.1038/s41598-021-00083-2.

ABSTRACT

Several vaccines with different efficacies and effectivenesses are currently being distributed across the world to control the COVID-19 pandemic. Having enough doses from the most efficient vaccines in a short time is not possible for all countries. Hence, policymakers may propose using various combinations of available vaccines to control the pandemic with vaccine-induced herd immunity by vaccinating a fraction of the population. The classic vaccine-induced herd-immunity threshold suggests that we can stop spreading the disease by vaccinating a fraction of the population. However, that classic threshold is defined only for a single vaccine and may be invalid and biased when we have multi-vaccine strategies for a disease or multiple variants, potentially leading policymakers to suboptimal vaccine-allocation policies. Here, we determine which combination of multiple vaccines may lead to herd immunity. We show that simplifying the problem and considering the vaccination of the population as a single-vaccine strategy whose effectiveness is the sample mean of all effectivenesses would not be ideal, because many multi-vaccine strategies with a smaller herd-immunity threshold can be proposed. We show that the herd-immunity threshold may vary due to changes in vaccine-uptake proportions. Moreover, we propose methods to determine the optimal combination of multiple vaccines in order to achieve herd immunity and apply our results to the issue of multiple variants. In addition, we determine a condition for reaching herd immunity in the presence of new emerging variants of concern. We show by example that new variants could influence our estimation of the vaccination reproduction number. It follows that the herd-immunity threshold must be updated not only when multi-vaccine strategies are used but also when multiple variants coexist in the population.

PMID:34836984 | DOI:10.1038/s41598-021-00083-2

Sci Data. 2021 Nov 26;8(1):304. doi: 10.1038/s41597-021-01081-9.

ABSTRACT

Remote sensing satellite imagery has the potential to monitor and understand dynamic environmental phenomena by retrieving information about Earth’s surface. Marine ecosystems, however, have been studied with less intensity than terrestrial ecosystems due, in part, to data limitations. Data on sea surface temperature (SST) and Chlorophyll-a (Chlo-a) can provide quantitative information of environmental conditions in coastal regions at a high spatial and temporal resolutions. Using the exclusive economic zone of coastal regions as the study area, we compiled monthly and annual statistics of SST and Chlo-a globally for 2003 to 2020. This ready-to-use dataset aims to reduce the computational time and costs for local-, regional-, continental-, and global-level studies of coastal areas. Data may be of interest to researchers in the areas of ecology, oceanography, biogeography, fisheries, and global change. Target applications of the database include environmental monitoring of biodiversity and marine microorganisms, and environmental anomalies.

PMID:34836949 | DOI:10.1038/s41597-021-01081-9

Nat Commun. 2021 Nov 26;12(1):6920. doi: 10.1038/s41467-021-26556-6.

ABSTRACT

Bone marrow is a preferred metastatic site for multiple solid tumours and is associated with poor prognosis and significant morbidity. Accumulating evidence indicates that cancer cells colonise specialised niches within the bone marrow to support their long-term propagation, but the precise location and mechanisms that mediate niche interactions are unknown. Using breast cancer as a model of solid tumour metastasis to the bone marrow, we applied large-scale quantitative three-dimensional imaging to characterise temporal changes in the bone marrow microenvironment during disease progression. We show that mouse mammary tumour cells preferentially home to a pre-existing metaphyseal domain enriched for type H vessels. Metastatic lesion outgrowth rapidly remodelled the local vasculature through extensive sprouting to establish a tumour-supportive microenvironment. The evolution of this tumour microenvironment reflects direct remodelling of the vascular endothelium through tumour-derived granulocyte-colony stimulating factor (G-CSF) in a hematopoietic cell-independent manner. Therapeutic targeting of the metastatic niche by blocking G-CSF receptor inhibited pathological blood vessel remodelling and reduced bone metastasis burden. These findings elucidate a mechanism of ‘host’ microenvironment hijacking by mammary tumour cells to subvert the local microvasculature to form a specialised, pro-tumorigenic niche.

PMID:34836954 | DOI:10.1038/s41467-021-26556-6

BMJ Glob Health. 2021 Nov;6(11):e007405. doi: 10.1136/bmjgh-2021-007405.

ABSTRACT

Global surveys have built-in gender-related biases associated with data missingness across the gender dimensions of people’s lives, imbalanced or incomplete representation of population groups, and biased ways in which gender information is elicited and used. While increasing focus is being placed on the integration of sex-disaggregated statistics into national programmes and on understanding effects of gender-based disparities on the health of all people, the data necessary for elucidating underlying causes of gender disparities and designing effective intervention programmes continue to be lacking. Approaches exist, however, that can reasonably address some shortcomings, such as separating questions of gender identification from biological sex. Qualitative research can elucidate ways to rephrase questions and translate gendered terms to avoid perpetuating historical gender biases and prompting biased responses. Non-health disciplines may offer lessons in collecting gender-related data. Ultimately, multidisciplinary global collaborations are needed to advance this evolving field and to set standards for how we measure gender in all its forms.

PMID:34836912 | DOI:10.1136/bmjgh-2021-007405

Nat Commun. 2021 Nov 26;12(1):6923. doi: 10.1038/s41467-021-27292-7.

ABSTRACT

Nationwide nonpharmaceutical interventions (NPIs) have been effective at mitigating the spread of the novel coronavirus disease (COVID-19), but their broad impact on other diseases remains under-investigated. Here we report an ecological analysis comparing the incidence of 31 major notifiable infectious diseases in China in 2020 to the average level during 2014-2019, controlling for temporal phases defined by NPI intensity levels. Respiratory diseases and gastrointestinal or enteroviral diseases declined more than sexually transmitted or bloodborne diseases and vector-borne or zoonotic diseases. Early pandemic phases with more stringent NPIs were associated with greater reductions in disease incidence. Non-respiratory diseases, such as hand, foot and mouth disease, rebounded substantially towards the end of the year 2020 as the NPIs were relaxed. Statistical modeling analyses confirm that strong NPIs were associated with a broad mitigation effect on communicable diseases, but resurgence of non-respiratory diseases should be expected when the NPIs, especially restrictions of human movement and gathering, become less stringent.

PMID:34836947 | DOI:10.1038/s41467-021-27292-7

BMJ Open. 2021 Nov 26;11(11):e056436. doi: 10.1136/bmjopen-2021-056436.

ABSTRACT

OBJECTIVE: Patients treated with long-term opioid therapy (LTOT) are known to have compromised immune systems and respiratory function, both of which make them particularly susceptible to the SARS-CoV-2 virus. The objective of this study was to assess the risk of developing severe clinical outcomes among COVID-19 non-cancer patients on LTOT, compared with those without LTOT.

DESIGN AND DATA SOURCES: A retrospective cohort design using electronic health records in the TriNetX research database.

PARTICIPANTS AND SETTING: 418 216 adults diagnosed with COVID-19 in January-December 2020 from 51 US healthcare organisations: 9558 in the LTOT and 408 658 in the control cohort. They did not have cancer diagnoses; only a small proportion might have been treated with opioid maintenance for opioid use disorder.

RESULTS: Patient on LTOT had a higher risk ratio (RR) than control patients to visit an emergency department (RR 2.04, 95% CI 1.93 to 2.16) and be hospitalised (RR 2.91, 95% CI 2.69 to 3.15). Once admitted, LTOT patients were more likely to require intensive care (RR 3.65, 95% CI 3.10 to 4.29), mechanical ventilation (RR 3.47, 95% CI 2.89 to 4.15) and vasopressor support (RR 5.28, 95% CI 3.70 to 7.53) and die within 30 days (RR 1.96, 95% CI 1.67 to 2.30). The LTOT group also showed increased risk (RRs from 2.06 to 3.98, all significant to 95% CI) of more-severe infection (eg, cough, dyspnoea, fever, hypoxaemia, thrombocytopaenia and acute respiratory distress syndrome). Statistically significant differences in several laboratory results and other vital signs appeared clinically negligible.

CONCLUSION: COVID-19 patients on LTOT were at higher risk of increased morbidity, mortality and healthcare utilisation. Interventions to reduce the need for LTOT and to increase compliance with COVID-19 protective measures may improve outcomes and reduce healthcare cost in this population. Prospective studies need to confirm and refine these findings.

PMID:34836910 | DOI:10.1136/bmjopen-2021-056436