Category: Statistics

JMIR Med Inform. 2021 May 12. doi: 10.2196/29242. Online ahead of print.

ABSTRACT

BACKGROUND: Care for children with autism spectrum disorder (ASD) can be challenging for families and medical care systems. This is especially true in Low-and-Middle-Income-countries (LMIC) like Bangladesh. To improve family-practitioner communication and developmental monitoring of children with ASD, [spell out] (mCARE) was developed. Within this study, mCARE was used to track child milestone achievement and family socio-demographic assets to inform mCARE feasibility/scalability and family-asset informed practitioner recommendations.

OBJECTIVE: The objectives of this paper are three-fold. First, document how mCARE can be used to monitor child milestone achievement. Second, demonstrate how advanced machine learning models can inform our understanding of milestone achievement in children with ASD. Third, describe family/child socio-demographic factors that are associated with earlier milestone achievement in children with ASD (across five machine learning models).

METHODS: Using mCARE collected data, this study assessed milestone achievement in 300 children with ASD from Bangladesh. In this study, we used four supervised machine learning (ML) algorithms (Decision Tree, Logistic Regression, k-Nearest Neighbors, Artificial Neural Network) and one unsupervised machine learning (K-means Clustering) to build models of milestone achievement based on family/child socio-demographic details. For analyses, the sample was randomly divided in half to train the ML models and then their accuracy was estimated based on the other half of the sample. Each model was specified for the following milestones: Brushes teeth, Asks to use the toilet, Urinates in the toilet or potty, and Buttons large buttons.

RESULTS: This study aimed to find a suitable machine learning algorithm for milestone prediction/achievement for children with ASD using family/child socio-demographic characteristics. For, Brushes teeth, the three supervised machine learning models met or exceeded an accuracy of 95% with Logistic Regression, KNN, and ANN as the most robust socio-demographic predictors. For Asks to use toilet, 84.00% accuracy was achieved with the KNN and ANN models. For these models, the family socio-demographic predictors of “family expenditure” and “parents’ age” accounted for most of the model variability. The last two parameters, Urinates in toilet or potty and Buttons large buttons had an accuracy of 91.00% and 76.00%, respectively, in ANN. Overall, the ANN had a higher accuracy (Above ~80% on average) among the other algorithms for all the parameters. Across the models and milestones, “family expenditure”, “family size/ type”, “living places” and “parent’s age and occupation” were the most influential family/child socio-demographic factors.

CONCLUSIONS: mCARE was successfully deployed in an LMIC (i.e., Bangladesh), allowing parents and care-practitioners a mechanism to share detailed information on child milestones achievement. Using advanced modeling techniques this study demonstrates how family/child socio-demographic elements can inform child milestone achievement. Specifically, families with fewer socio-demographic resources reported later milestone attainment. Developmental science theories highlight how family/systems can directly influence child development and this study provides a clear link between family resources and child developmental progress. Clinical implications for this work could include supporting the larger family system to improve child milestone achievement.

CLINICALTRIAL: We took the IRB from Marquette University Institutional Review Board on July 9, 2020, with the protocol number HR-1803022959, and titled “MOBILE-BASED CARE FOR CHILDREN WITH AUTISM SPECTRUM DISORDER USING REMOTE EXPERIENCE SAMPLING METHOD (MCARE)” for recruiting a total of 316 subjects, of which we recruited 300. (Details description of participants in Methods section).

PMID:33984830 | DOI:10.2196/29242

Eur Neuropsychopharmacol. 2021 May 10;50:64-74. doi: 10.1016/j.euroneuro.2021.04.007. Online ahead of print.

ABSTRACT

The specific role of white matter (WM) microstructure in parkinsonism among patients with schizophrenia spectrum disorders (SSD) is largely unknown. To determine whether topographical alterations of WM microstructure contribute to parkinsonism in SSD patients, we examined healthy controls (HC, n=16) and SSD patients with and without parkinsonism, as defined by Simpson-Angus Scale total score of ≥4 (SSD-P, n=33) or <4 (SSD-nonP, n=62). We used whole brain tract-based spatial statistics (TBSS), tractometry (along tract statistics using TractSeg) and graph analytics (clustering coefficient (CCO), local betweenness centrality (BC)) to provide a framework of specific WM microstructural changes underlying parkinsonism in SSD. Using these methods, post hoc analyses showed (a) decreased fractional anisotrophy (FA), as measured via tractometry, in the corpus callosum, corticospinal tract and striato-fronto-orbital tract, and (b) increased CCO, as derived by graph analytics, in the left orbitofrontal cortex (OFC) and left superior frontal gyrus (SFG), in SSD-P patients when compared to SSD-nonP patients. Increased CCO in the left OFC and SFG was associated with SAS scores. These findings indicate the prominence of OFC alterations and aberrant connectivity with fronto-parietal regions and striatum in the pathogenesis of parkinsonism in SSD. This study further supports the notion of altered “bottom-up modulation” between basal ganglia and fronto-parietal regions in the pathobiology of parkinsonism, which may reflect an interaction between movement disorder intrinsic to SSD and antipsychotic drug-induced sensorimotor dysfunction.

PMID:33984810 | DOI:10.1016/j.euroneuro.2021.04.007

Phys Med. 2021 May 10;85:79-86. doi: 10.1016/j.ejmp.2021.05.004. Online ahead of print.

ABSTRACT

PURPOSE: The robustness against setup and motion uncertainties of gated four-dimensional restricted robust optimization (4DRRO) was investigated for hypofractionated carbon ion radiotherapy (CIRT) of lung tumors.

METHODS: CIRT plans of 9 patients were optimized using 4DRRO strategy with 3 mm setup errors, 3% density errors and 3 breathing phases related to the gate window. The prescription was 60 Gy(RBE) in 4 fractions. Standard spots (SS) were compared to big spots (BS). Plans were recalculated on multiple 4DCTs acquired within 3 weeks from treatment simulation and rigidly registered with planning images using bone matching. Warped dose distributions were generated using deformable image registration and accumulated on the planning 4DCTs. Target coverage (D98%, D95% and V95%) and dose to lung were evaluated in the recalculated and accumulated dose distributions.

RESULTS: Comparable target coverage was obtained with both spot sizes (p = 0.53 for D95%). The mean lung dose increased of 0.6 Gy(RBE) with BS (p = 0.0078), still respecting the dose constraint of a 4-fraction stereotactic treatment for the risk of radiation pneumonitis. Statistically significant differences were found in the recalculated and accumulated D95% (p = 0.048 and p = 0.024), with BS showing to be more robust. Using BS, the average degradations of the D98%, D95% and V95% in the accumulated doses were -2.7%, -1.6% and -1.5%.

CONCLUSIONS: Gated 4DRRO was highly robust against setup and motion uncertainties. BS increased the dose to healthy tissues but were more robust than SS. The selected optimization settings guaranteed adequate target coverage during the simulated treatment course with acceptable risk of toxicity.

PMID:33984821 | DOI:10.1016/j.ejmp.2021.05.004

Phys Med. 2021 May 10;85:87-97. doi: 10.1016/j.ejmp.2021.03.039. Online ahead of print.

ABSTRACT

PURPOSE: Oxygen extraction fraction (OEF) can be a factor to identify brain tissue’s disability in epileptic patients. This study aimed to assess the OEF’s level measurement in refractory epileptic patients (REPs) using a quantitative susceptibility mapping (QSM) method and to determine whether the OEF parameters change.

METHODS: QSM-OEF maps of 26 REPs and 16 healthy subjects were acquired using 3T MRI with a 64-channel coil. Eighteen regions-of-interest (ROIs) were chosen around the cortex in one appropriate slice of the brain and the mean QSM-OEF for each ROI was obtained. The correlations of QSM-OEF among different clinical characteristics of the disease, as well as between the patients and normal subjects, were also investigated.

RESULTS: QSM-OEF was shown to be significantly higher in REPs (44.9 ± 5.8) than that in HS (41.9 ± 6.2) (p < 0.05). Mean QSM-OEF was statistically lower in the ipsilateral side (44.5 ± 6.6) compared to the contralateral side (46.4 ± 6.8) (P < 0.01). QSM-OEF was illustrated to have a strong positive correlation with the attack duration (r = 0.6), and a moderate negative correlation with the attack frequency (r = -0.3). Using an optimized support vector machine algorithm, we could predict the disease in subjects having abnormal OEF values in the brain-selected-ROIs with sensitivity, specificity, AUC, and the precision of 0.96, 1, 0.98, and 1, respectively.

CONCLUSIONS: The results of this study revealed that QSM-OEF of the REPs’ brain is higher than that of HS, which indicates that QSM-OEF is associated with disease activity.

PMID:33984822 | DOI:10.1016/j.ejmp.2021.03.039

Comput Biol Med. 2021 Apr 30;134:104452. doi: 10.1016/j.compbiomed.2021.104452. Online ahead of print.

ABSTRACT

BACKGROUND: Metabolic syndrome (MetS) is prevalent in patients receiving atypical antipsychotic drugs (AADs), but there are few effective interventions. The Traditional Chinese herbal decoction Liu-Yu-Tang (LYT) has achieved clinical improvement for AAD-induced MetS, but its pharmacological mechanism remains unclear.

METHOD: A network pharmacology-based method was utilized in this study. First, the TCMSP and SwissTargetPrediction database were used to acquire plasma-absorbed components and putative targets of LYT, respectively. Second, an interaction network between shared targets of LYT and MetS was constructed using STRING online tool. Topological analyses were performed to extract hub gene targets. Finally, we did a pathway analysis of gene targets using the Kyoto Encyclopedia of Genes and Genomes (KEGG) to find biological pathways of LYT.

RESULTS: We obtained 655 putative targets of LYT, 434 known targets of AADs, and 1577 MetS-related gene targets. There are 232 shared targets between LYT and MetS. Interaction network construction and topological analysis yielded 60 hub targets, of which 18 were major hub targets, among which IL-6, IL-8, TNF, PI3K, MAPK, and NF-κB (RELA) are the most important in LYT’s treatment of AAD-induced MetS. Pathway enrichment analysis revealed a statistically high significance of the AGE-RAGE signaling pathway in diabetic complications, lipid and atherosclerosis and the insulin resistance pathway.

CONCLUSIONS: LYT may control activities of the pro-inflammatory cytokines IL-6, IL-8, TNF and the important signal transduction molecules PI3K, MAPKs, and NF-κB (RELA), regulating metabolic disturbance-related pathways like the AGE-RAGE signaling pathway in diabetic complications, lipid and atherosclerosis, and the insulin resistance pathway, generating therapeutic effects for AAD-induced MetS.

PMID:33984751 | DOI:10.1016/j.compbiomed.2021.104452

Clin Neurol Neurosurg. 2021 Apr 27;206:106669. doi: 10.1016/j.clineuro.2021.106669. Online ahead of print.

ABSTRACT

BACKGROUND: Standardized uptake values (SUVs) are important indexes for evaluating the accuracy of disease diagnoses achieved via fluoro-18 deoxyglucose (F-18 FDG) positron emission tomography/computed tomography (PET/CT) and positron emission tomography/magnetic resonance imaging (PET/MRI). The purpose of this study is to describe normal cerebral FDG uptake in the pediatric population and compare SUVmax/mean results for brain images obtained from PET/CT and PET/MRI in neurologically healthy pediatric examinees.

METHODS: This study included 20 patients who were < 18 years of age and were without intracranial malignancy and/or brain disorders. Patients underwent either PET/CT imaging (n = 10) or PET/MRI imaging (n = 10) after 70-80 min of F-18 FDG injection. The SUVmax and SUVmean for various brain regions were calculated and compared between sides and imaging modalities using with appropriate statistical tests.

RESULTS: The median SUVmax/SUVmean values of the right-sided frontal, parietal, temporal, and occipital lobes were 8.63/ 6.18, 8.85 / 6.97, 6.88 / 4.99, and 11.06 / 7.02 in PET/CT, respectively, and 11.45 / 8.59, 10.16 / 8.47, 8.82 / 6.6, and 11.71 / 8.25 in PET/MRI, respectively. The median SUVmax/SUVmean values of the left-sided frontal, parietal, temporal, and occipital lobes were 9.05 / 6.86, 8.03 / 6.62, 6.49 / 4.77, and 10.6 / 7.73 in PET/CT, respectively, and 10.7 / 8.16, 11.06 / 7.88, 8.13 / 6.09, and 10.96 / 9.22 in PET/MRI, respectively.

CONCLUSIONS: These results showed that there was no statistically significant difference in SUVs values between the two brain imaging modalities except from SUVmax value of left-sided parietal lobe and no asymmetric radiopharmaceutical uptake between the left and right brain regions or cerebellums in each modality, suggested that in brain imaging, PET/MRI can be used reliably instead of PET/CT.

PMID:33984753 | DOI:10.1016/j.clineuro.2021.106669

Vet Microbiol. 2021 May 4;258:109100. doi: 10.1016/j.vetmic.2021.109100. Online ahead of print.

ABSTRACT

The purpose of this experimental study was to evaluate the efficacy of a new trivalent vaccine containing porcine circovirus types 2a/b (PCV2a/b) and Mycoplasma hyopneumoniae. Pigs were administered the vaccine intramuscularly as either at 3 and 24 days of age with 1.0 mL or at 21 days of age with 2.0 mL according to the manufacturer’s recommendations. The pigs were challenged at 42 days of age with either PCV2d (intranasal route) or M. hyopneumoniae (intratracheal route), or both. No statistical differences were observed between the one-dose and two-dose experiments based on clinical (growth performance), immunological (protective immunity), microbiological (viremia and laryngeal swab), and pathological (pulmonary and lymphoid lesion) outcomes. Pigs in vaccinated/challenged and unvaccinated/unchallenged groups showed significant difference in growth performance compared to pigs in the unvaccinated/challenged group in both dosage experiments. Vaccinated pigs elicited a significant amount of protective immunity for PCV2d-specific neutralizing antibodies and interferon-γ secreting cells (IFN-γ-SC) as well as M. hyopneumoniae-specific IFN-γ-SC significantly post-challenge compared to unvaccinated/challenged pigs. Vaccination and challenge reduced the viral load amount of PCV2d in the blood and reduced the M. hyopneumoniae load in laryngeal swab, while simultaneously reducing both pulmonary and lymphoid lesion severity when compared to unvaccinated/challenged pigs. Trivalent vaccination provided good protection against a single PCV2d challenge, single M. hyopneumoniae challenge, and a PCV2d/M. hyopneumoniae dual challenge.

PMID:33984792 | DOI:10.1016/j.vetmic.2021.109100

Neural Netw. 2021 Apr 30;142:57-72. doi: 10.1016/j.neunet.2021.04.036. Online ahead of print.

ABSTRACT

Even when neural networks are widely used in a large number of applications, they are still considered as black boxes and present some difficulties for dimensioning or evaluating their prediction error. This has led to an increasing interest in the overlapping area between neural networks and more traditional statistical methods, which can help overcome those problems. In this article, a mathematical framework relating neural networks and polynomial regression is explored by building an explicit expression for the coefficients of a polynomial regression from the weights of a given neural network, using a Taylor expansion approach. This is achieved for single hidden layer neural networks in regression problems. The validity of the proposed method depends on different factors like the distribution of the synaptic potentials or the chosen activation function. The performance of this method is empirically tested via simulation of synthetic data generated from polynomials to train neural networks with different structures and hyperparameters, showing that almost identical predictions can be obtained when certain conditions are met. Lastly, when learning from polynomial generated data, the proposed method produces polynomials that approximate correctly the data locally.

PMID:33984736 | DOI:10.1016/j.neunet.2021.04.036

J Stroke Cerebrovasc Dis. 2021 May 10;30(7):105844. doi: 10.1016/j.jstrokecerebrovasdis.2021.105844. Online ahead of print.

ABSTRACT

OBJECTIVES: We aimed to analyse the relationship between specific inflammatory biomarkers’ levels and the temporal pattern of cerebral venous thrombosis (CVT) symptoms.

MATERIALS AND METHODS: We performed a retrospective study of adult CVT patients admitted between Jan 01 2006 and Dec 31 2019. We excluded patients with infection at admission, autoimmune, inflammatory or haematological disorders. We evaluated serum inflammatory biomarkers at admission: C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), absolute neutrophil count, absolute lymphocyte count, platelet count, monocyte count, neutrophile-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), platelet-to-lymphocyte ratio (PLR), bilirubin and monocyte-to-HDL ratio (M-HDLR). These were evaluated according to the time from symptom onset (acute, subacute or chronic).

RESULTS: We included 78 patients with CVT (mean age 41 ± 13 years). Neutrophil count (p = 0.017), monocyte (p = 0.024), CRP (p = 0.004), NLR (p<0.001) and LMR (p = 0.004) showed significant variation with CVT duration. Acute onset CVT exhibited higher absolute neutrophil count and NLR but lower LMR. The subacute group had higher monocyte values, and the chronic phase patients displayed higher LMR, but lower CRP. ESR, PLR and M-HDLR showed a tendency to decrease in the chronic phase. We did not observe any statistical difference between the duration of symptoms and levels of bilirubin.

CONCLUSIONS: CVT patients present a differential inflammatory pattern along the time course of the disease: higher NLR and lower LMR in acute phase, and higher LMR and lower CRP level during the chronic phase. These differences may help to ascertain the onset of poorly defined symptoms and provide input regarding anticoagulation management.

PMID:33984744 | DOI:10.1016/j.jstrokecerebrovasdis.2021.105844

Int Immunopharmacol. 2021 May 10;96:107742. doi: 10.1016/j.intimp.2021.107742. Online ahead of print.

ABSTRACT

BACKGROUND: CD19-specific chimeric antigen receptor T-cell (CAR-T) therapy has shown promising clinical outcomes in relapsed/refractory acute B-cell lymphoblastic leukemia (R/R B-ALL) patients. However, some patients did not respond to this therapy or relapsed after remission. Regulatory T cells (Tregs) have shown great importance in promoting tumor escape, but little is known about their role in R/R B-ALL patients with CAR-T therapy. Our previous study has proved that higher Tregs before infusion was an independent high-risk factor for relapse-free survival (RFS). To further clarify the relationship between Tregs and the efficacy of CAR-T therapy, the present study tested the levels of CD4⁺CD25⁺CD127^low Tregs in peripheral blood (PB) of R/R B-ALL patients at different stages of CD19 CAR-T therapy, and evaluate their impact on the efficacy and prognosis of CAR-T therapy.

METHODS: From November 2015 to May 2019, 47 R/R B-ALL patients successfully received CD19 CAR-T therapy at our institution and followed up for at least 1 month. Among them, one patient did not tested for Tregs, so 46 patients enrolled in this study. We collected clinical information of them and dynamically detected the frequency of CD4⁺CD25⁺CD127^low Tregs within CD4 + T cells at different time points (before infusion and at 1 week after infusion) by flow cytometry, and validated the relationship of circulating Tregs with clinical efficacy, OS, and recurrence of CAR-T therapy.

RESULTS: Circulating Tregs of R/R B-ALL patients in pre-infusion group (median 6.67%) and in 1 week after infusion group (median 6.80%) were all higher than that of the healthy control group (median 5.04%), with statistical significance (P < 0.05). The frequencies of Tregs in not remission (NR) group at baseline (pre-infusion) and at 1 week after infusion were all significantly higher than those in remission group. With cut-off values of 11.54% (before infusion) and 13.56% (1 week after infusion), the specificity for Tregs were 94.6% and 100% , respectively. In remission group, 11 patients underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) after achieving remission by Sino 19 cell therapy. No significant differences of Tregs expression were found between transplantation and non-transplantation groups. Time-dependent Cox model showed that transplantation group had lower risk of relapse and death when compared with non-transplantation group (HR = 0.664 for RFS and HR = 0.364 for OS), however, no statistical significances were found (P = 0.403 and 0.106, respectively). Higher Tregs before infusion and at 1 week after infusion were significant associated with shorter RFS and OS by Kaplan-Meier analysis. Multivariate analysis showed that higher Tregs at 1 week after infusion was the independently factor for poor RFS (P = 0.032) and shorter OS (P = 0.025) in R/R B-ALL patients with CD19 CAR-T therapy. Besides, Tregs levels before and at 1 week after infusion were negatively correlated with the persistence time of Sino 19 cell.

CONCLUSION: Higher circulating Tregs, especially 1 week after CD19 CAR-T cell infusion, was a poor predict indicator for CD19 CAR-T therapy in R/R B-ALL patients.

PMID:33984717 | DOI:10.1016/j.intimp.2021.107742