Tag: statistics

Top Cogn Sci. 2021 Nov 18. doi: 10.1111/tops.12580. Online ahead of print.

ABSTRACT

My first 30-odd years of research in cognitive science has been driven by an attempt to balance two facts about human thought that seem incompatible and two corresponding ways of understanding information processing. The facts are that, on one hand, human memories serve as sophisticated pattern recognition devices with great flexibility and an ability to generalize and predict as long as circumstances remain sufficiently familiar. On the other hand, we are capable of deploying an enormous variety of representational schemes that map closely onto articulable structure in the world and that support explanation even in unfamiliar circumstances. The contrasting ways of modeling such processes involve, first, more and more sophisticated associative models that capture progressively higher-order statistical structure and, second, more powerful representational languages for other sorts of structure, especially compositional and causal structure. My efforts to rectify these forces have taken me from the study of memory to induction and category knowledge to causal reasoning. In the process, I have consistently appealed to dual systems of thinking. I have come to realize that a key reason for our success as cognizers is that we rely on others for most of our information processing needs; we live in a community of knowledge. We make use of others both intuitively-by outsourcing much of our thinking without knowing we are doing it-and by deliberating with others.

PMID:34792846 | DOI:10.1111/tops.12580

J Clin Psychiatry. 2021 Nov 16;83(1):21m13918. doi: 10.4088/JCP.21m13918.

ABSTRACT

Background: Recent guidelines and systematic reviews suggest that disorder-specific psychotherapeutic interventions are the first choice in the treatment of borderline personality disorder (BPD). The aim of this study is to examine brain activity changes in BPD patients (DSM-5) who received a revised BPD-adapted interpersonal psychotherapy (IPT-BPD-R) compared with patients on the waiting list.

Methods: Forty-three patients with a BPD diagnosis (DSM-5) were randomly assigned to IPT-BPD-R (n = 22 patients) or the waiting list with clinical management (n = 21 patients) for 10 months. Both groups were tested before and after treatment with the Social and Occupational Functioning Assessment Scale (SOFAS), the Clinical Global Impressions-Severity of Illness scale (CGI-S), the Borderline Personality Disorder Severity Index (BPDSI), the Barratt Impulsiveness Scale-version 11 (BIS-11), and the Autobiographical Interview. Both groups underwent pre- and posttreatment functional magnetic resonance imaging (fMRI) testing. The fMRI task consisted of the presentation of resolved and unresolved life events compared to a neutral condition. All structural and functional images were analyzed using Statistical Parametric Mapping 12 software, which interfaces with MATLAB. Clinical data were analyzed using analysis of variance for repeated measures. Patients were recruited between September 2017 and April 2019.

Results: In clinical results, for the 4 rating scales, a significant between-subject effect was found in favor of the IPT-BPD-R-treated group (CGI-S: P = .011; BPDSI: P = .009; BIS-11: P = .033; SOFAS: P = .022). In fMRI results, posttreatment versus pretreatment for the contrast unresolved life event versus neutral condition showed significantly decreased right temporoparietal junction (rTPJ: x = 45, y = -51, z = 36) (P = .043) and right anterior cingulate cortex (rACC: x = -4, y = 37, z = 8) activity (P = .021).

Conclusions: IPT-BPD-R appears to be effective in treating BPD symptoms, and these clinical effects are reflected in the functional changes observed with fMRI. Brain areas that showed modulation of their activity are the rTPJ and rACC, which are involved in mentalization processes that are fundamental to BPD pathology.

Trial Registration: Australian New Zealand Clinical Trials Registry (ANZCTR) code: ACTRN12619000078156.

PMID:34792871 | DOI:10.4088/JCP.21m13918

J Magn Reson Imaging. 2021 Nov 18. doi: 10.1002/jmri.27999. Online ahead of print.

ABSTRACT

BACKGROUND: The associations between diffusion kurtosis imaging (DKI)-derived parameters and clinical prognostic factors of breast cancer have not been fully evaluated; this knowledge may have implications for outcome prediction and treatment strategies.

PURPOSE: To determine associations between quantitative diffusion parameters derived from DKI and diffusion-weighted imaging (DWI) and the prognostic factors and molecular subtypes of breast cancer.

STUDY TYPE: Retrospective.

POPULATION: A total of 383 women (mean age, 53.8 years; range, 31-82 years) with breast cancer who underwent preoperative breast MRI including DKI and DWI.

FIELD STRENGTH/SEQUENCE: A 3.0 T; DKI using a spin-echo echo-planar imaging (EPI) sequence (b values: 200, 500, 1000, 1500, and 2000 sec/mm² ), DWI using a readout-segmented EPI sequence (b values: 0 and 1000 sec/mm² ) and dynamic contrast-enhanced breast MRI.

ASSESSMENT: Two radiologists (J.Y.K. and H.S.K. with 9 years and 1 year of experience in MRI, respectively) independently measured kurtosis, diffusivity, and apparent diffusion coefficient (ADC) values of breast cancer by manually placing a regions of interest within the lesion. Diffusion measures were compared according to nodal status, grade, and molecular subtypes.

STATISTICAL TESTS: Kruskal-Wallis test, Mann-Whitney U test with Bonferroni correction, receiver operating characteristic (ROC) analysis, and multivariate logistic regression analysis. (Statistical significance level of P < 0.05).

RESULTS: All diffusion measures showed significant differences according to axillary nodal status and histological grade. Kurtosis showed significant differences among molecular subtypes. The luminal subtype (median 1.163) showed a higher kurtosis value compared to the HER2-positive (median 0.962) or triple-negative subtypes (median 1.072). ROC analysis for differentiating HER2-positive from luminal subtypes revealed that kurtosis yielded the highest area under the curve of 0.781. In multivariate analyses, kurtosis remained a significant factor associated with differentiation between HER2-positive and luminal (odds ratio [OR] = 0.993), triple-negative and luminal (OR = 0.995), and HER2-positive and triple-negative subtypes (OR = 0.994).

DATA CONCLUSION: Quantitative diffusion parameters derived from DKI and DWI are associated with prognostic factors for breast cancer. Moreover, DKI-derived kurtosis can help distinguish between the molecular subtypes of breast cancer.

EVIDENCE LEVEL: 4 TECHNICAL EFFICACY: 3.

PMID:34792837 | DOI:10.1002/jmri.27999

Vet Comp Oncol. 2021 Nov 18. doi: 10.1111/vco.12786. Online ahead of print.

ABSTRACT

Standardized veterinary neuroimaging response assessment methods for brain tumours are lacking. Consequently, a response assessment in veterinary neuro-oncology (RAVNO) system which uses the sum product of orthogonal lesion diameters on 1-image section with the largest tumour area, has recently been proposed. In this retrospective study, 22 pre-treatment magnetic resonance imaging (MRI) studies from 18 dogs and four cats with suspected intracranial neoplasia were compared by a single observer to 32 post-treatment MRIs using the RAVNO system and two volumetric methods based on tumour margin or area delineation with HOROS and 3D Slicer software, respectively. Intra-observer variability was low, with no statistically significant differences in agreement index between methods (mean AI ± SD, 0.91 ± 0.06 for RAVNO; 0.86 ± 0.08 for HOROS; and 0.91 ± 0.05 for 3D slicer), indicating good reproducibility. Response assessments consisting of complete or partial responses, and stable or progressive disease, agreed in 23 out of 32 (72%) MRI evaluations using the three methods. The RAVNO system failed to identify changes in mass burden detected with volumetric methods in 6 cases. 3D Slicer differed from the other two methods in 3 cases involving cysts or necrotic tissue as it allowed for more accurate exclusion of these structures. The volumetric response assessment methods were more precise in determining changes in absolute tumour burden than RAVNO but were more time-consuming to use. Based on observed agreement between methods, low intra-observer variability, and decreased time constraint, RAVNO might be a suitable response assessment method for the clinical setting.

PMID:34792828 | DOI:10.1111/vco.12786

J Magn Reson Imaging. 2021 Nov 18. doi: 10.1002/jmri.27995. Online ahead of print.

ABSTRACT

BACKGROUND: Automated segmentation using convolutional neural networks (CNNs) have been developed using four-dimensional (4D) flow magnetic resonance imaging (MRI). To broaden usability for congenital heart disease (CHD), training with multi-institution data is necessary. However, the performance impact of heterogeneous multi-site and multi-vendor data on CNNs is unclear.

PURPOSE: To investigate multi-site CNN segmentation of 4D flow MRI for pediatric blood flow measurement.

STUDY TYPE: Retrospective.

POPULATION: A total of 174 subjects across two sites (female: 46%; N = 38 healthy controls, N = 136 CHD patients). Participants from site 1 (N = 100), site 2 (N = 74), and both sites (N = 174) were divided into subgroups to conduct 10-fold cross validation (10% for testing, 90% for training).

FIELD STRENGTH/SEQUENCE: 3 T/1.5 T; retrospectively gated gradient recalled echo-based 4D flow MRI.

ASSESSMENT: Accuracy of the 3D CNN segmentations trained on data from single site (single-site CNNs) and data across both sites (multi-site CNN) were evaluated by geometrical similarity (Dice score, human segmentation as ground truth) and net flow quantification at the ascending aorta (Qs), main pulmonary artery (Qp), and their balance (Qp/Qs), between human observers, single-site and multi-site CNNs.

STATISTICAL TESTS: Kruskal-Wallis test, Wilcoxon rank-sum test, and Bland-Altman analysis. A P-value <0.05 was considered statistically significant.

RESULTS: No difference existed between single-site and multi-site CNNs for geometrical similarity in the aorta by Dice score (site 1: 0.916 vs. 0.915, P = 0.55; site 2: 0.906 vs. 0.904, P = 0.69) and for the pulmonary arteries (site 1: 0.894 vs. 0.895, P = 0.64; site 2: 0.870 vs. 0.869, P = 0.96). Qs site-1 medians were 51.0-51.3 mL/cycle (P = 0.81) and site-2 medians were 66.7-69.4 mL/cycle (P = 0.84). Qp site-1 medians were 46.8-48.0 mL/cycle (P = 0.97) and site-2 medians were 76.0-77.4 mL/cycle (P = 0.98). Qp/Qs site-1 medians were 0.87-0.88 (P = 0.97) and site-2 medians were 1.01-1.03 (P = 0.43). Bland-Altman analysis for flow quantification found equivalent performance.

DATA CONCLUSION: Multi-site CNN-based segmentation and blood flow measurement are feasible for pediatric 4D flow MRI and maintain performance of single-site CNNs.

LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY: Stage 2.

PMID:34792835 | DOI:10.1002/jmri.27995

J Rural Health. 2021 Nov 18. doi: 10.1111/jrh.12635. Online ahead of print.

ABSTRACT

INTRODUCTION: Broadband access is a “super determinant of health.” Understanding the spatial distribution and predictors of access may help target government programs and telehealth applications. Our aim was to examine broadband access across geography and sociodemographic characteristics using American Community Survey (ACS) data.

METHODS: We used 5-year ACS estimates from 2014 to 2018 to evaluate broadband access across contiguous US census tracts. Rural-Urban Commuting Area (RUCA) codes were categorized as metropolitan, micropolitan, small town, and isolated rural. We performed bivariate analyses to determine differences by RUCA categories and meeting the Healthy People 2020 (HP2020) objective (83.2% broadband access) or not. We conducted spatial statistics and spatial regression analyses to identify clusters of broadband access and sociodemographic factors associated with broadband access.

RESULTS: No RUCA grouping met the HP2020 objective; 80.6% of households had broadband access, including 82.0% of metropolitan, 73.9% of micropolitan, 70.7% of small town, and 70.0% of isolated rural households. Areas with high percentages of Black residents had lower broadband access, particularly in isolated rural tracts (54.9%). Low access was spatially clustered in the Southeast, Southwest, and northern plains. In spatial regression models, poverty and education were most strongly associated with broadband access, while the proportion of American Indian/Alaska Native population was the strongest racial/ethnic factor.

CONCLUSIONS: Rural areas had less broadband access with the greatest disparities experienced among geographically isolated areas with larger Black and American Indian/Alaska Native populations, more poverty, and lower educational attainment, following well-known social gradients in health. Resources and initiatives should target these areas of greatest need.

PMID:34792815 | DOI:10.1111/jrh.12635

J Prosthodont. 2021 Nov 18. doi: 10.1111/jopr.13454. Online ahead of print.

ABSTRACT

PURPOSE: To investigate the accuracy of additive manufacturing (AM) by means of internal fit of fixed dental prostheses (FDPs) fabricated with two AM technologies using different resins and printing modes (validated vs non-validated) compared to milling and direct manual methods.

MATERIAL AND METHODS: Sixty 3-unit interim FDPs replacing the first mandibular molar were divided into 6 groups (n = 10): manual (Protemp 4), milled (Telio-CAD), and AM groups were subdivided based on AM technology (direct light processing (Rapidshape P30 [RS]) and stereolithography (FormLabs 2 [FL])) and the polymer type (P-Pro-C&B [St] and SHERAprint-cb [Sh]) (RS-St, RS-Sh, FL-St, FL-Sh). Validated (RS-Sh and RS-St) or non-validated (FL-St and FL-Sh) modes were adopted for AM. The specimens were scanned to 3D align (GOM inspect) according to the triple scan method. The internal space between the FDPs and preparation surfaces in four sites (marginal, axial, occlusal, and total) was measured using equidistant surface points (GOM Inspect). Statistical analysis was done using Kruskal Wallis and Dunn post-hoc tests. (α = .05).

RESULTS: One AM group (FL-Sh) and milling exhibited better adaptation compared to manual and RS-St at molar site (P<.05). FDPs with St resin (FL-St and RS-St) displayed bigger marginal space than milled, FL-Sh, and RS-Sh. The non-validated printing mode showed better mean space results (P<.05) with higher predictability and repeatability (P<.001).

CONCLUSIONS: The AM interim FDPs tested provided valid alternatives to the milled ones in regard to their accuracy results. The printing mode, resin, and the AM technology used significantly influenced the manufacturing accuracy of interim FDPs, particularly at the marginal area. The non-validated printing mode with lower-cost 3D printers is a promising solution for clinical applications. This article is protected by copyright. All rights reserved.

PMID:34792821 | DOI:10.1111/jopr.13454

Med Phys. 2021 Nov 18. doi: 10.1002/mp.15362. Online ahead of print.

ABSTRACT

PURPOSE: Conventional model observers (MO) in CT are often limited to a uniform background or varying background that is random and can be modelled in an analytical form. It is unclear if these conventional MOs can be readily generalized to predict human observer performance in clinical CT tasks that involve realistic anatomical background. Deep-learning-based model observers (DL-MO) have recently been developed, but have not been validated for challenging low contrast diagnostic tasks in abdominal CT. We consequently sought to validate a DL-MO for a low-contrast hepatic metastases localization task.

METHODS: We adapted our recently-developed DL-MO framework for the liver metastases localization task. Our previously-validated projection-domain lesion- / noise-insertion techniques were used to synthesize realistic positive and low-dose abdominal CT exams, using the archived patient projection data. Ten experimental conditions were generated, which involved different lesion sizes / contrasts, radiation dose levels, and image reconstruction types. Each condition included 100 trials generated from a patient cohort of 7 cases. Each trial was presented as liver image patches (160×160×5 voxels). The DL-MO performance was calculated for each condition and was compared with human observer performance, which was obtained by 3 sub-specialized radiologists in an observer study. The performance of DL-MO and radiologists was gauged by the area under localization receiver-operating-characteristic curves. The generalization performance of the DL-MO was estimated with the repeated 2-fold cross-validation method over the same set of trials used in human observer study. A multi-slice Channelized Hoteling Observers (CHO) was compared with the DL-MO across the same experimental conditions.

RESULTS: The performance of DL-MO was highly correlated to that of radiologists (Pearson’s correlation coefficient: 0.987; 95% C.I.: 0.942, 0.997). The performance level of DL-MO was comparable to that of the grouped radiologists, i.e. the mean performance difference was -3.3%. The CHO performance was poorer than the grouped radiologist performance, before internal noise could be added. The correlation between CHO and radiologists was weaker (Pearson’s correlation coefficient: 0.812, and 95% C.I.: [0.378, 0.955]), and the corresponding performance bias (-29.5%) was statistically significant.

CONCLUSION: The presented study demonstrated the potential of using the DL-MO for image quality assessment in patient abdominal CT tasks. This article is protected by copyright. All rights reserved.

PMID:34792800 | DOI:10.1002/mp.15362

Histopathology. 2021 Nov 18. doi: 10.1111/his.14600. Online ahead of print.

ABSTRACT

BACKGROUND: A number of randomized controlled trials (RCT) have compared different techniques to improve lymph node yield (LNY) in colorectal cancer specimens but data on comparative effectiveness are sparse. Our aim was to compare the relative effectiveness and rank all available techniques.

METHODS: A systematic search of Embase, Cochrane, PubMed and Scopus was performed for randomized trials. Pairwise meta-analysis performed if more than two homogeneous studies were available for each comparison. Network meta-analysis was used to rank and compare all available techniques.

RESULTS: Fifteen studies fulfilled the inclusion criteria. Techniques that were compared included methylene blue (MB), GEWF, Carnoy solution (CS), patent blue (PB), formalin, fat clearing (FC) and their combinations. The overall quality of studies was found to be fair. In pairwise meta-analysis MB had a higher lymph node yield weighted mean difference [WMD] 13.67 [4.83-22.51], P<0.01, lower number of specimens with less than 12 lymph nodes log Odds Ratio= -1.88(-2.8, -0.91), P<0.01 and higher LNY in patients with prior chemoradiotherapy (WMD 9.11 [3.15,15.08], p=0.02) as compared to formalin. Evaluation of the network plot revealed a well-connected network. In network meta-analysis MBFC had a higher LNY with [Mean Difference (MD) 13 and 95% credible interval (CI) (2.09- 23.91)] as compared to formalin. MBFC probability of being the best technique for LNY was 91.4%. In network meta-analysis MB did not have a statistically significant difference when compared to formalin.

CONCLUSIONS: MBFCS seems to be the most effective technique for LNY. Further studies are required to make safe conclusions for outcomes such positive lymph nodes and upstaging.

PMID:34792803 | DOI:10.1111/his.14600

Med Biol Eng Comput. 2021 Nov 18. doi: 10.1007/s11517-021-02469-w. Online ahead of print.

ABSTRACT

Nowadays, retinal optical coherence tomography (OCT) plays an important role in ophthalmology and automatic analysis of the OCT is of real importance: image denoising facilitates a better diagnosis and image segmentation and classification are undeniably critical in treatment evaluation. Synthetic OCT was recently considered to provide a benchmark for quantitative comparison of automatic algorithms and to be utilized in the training stage of novel solutions based on deep learning. Due to complicated data structure in retinal OCTs, a limited number of delineated OCT datasets are already available in presence of abnormalities; furthermore, the intrinsic three-dimensional (3D) structure of OCT is ignored in many public 2D datasets. We propose a new synthetic method, applicable to 3D data and feasible in presence of abnormalities like diabetic macular edema (DME). In this method, a limited number of OCT data is used during the training step and the Active Shape Model is used to produce synthetic OCTs plus delineation of retinal boundaries and location of abnormalities. Statistical comparison of thickness maps showed that synthetic dataset can be used as a statistically acceptable representative of the original dataset (p > 0.05). Visual inspection of the synthesized vessels was also promising. Regarding the texture features of the synthesized datasets, Q-Q plots were used, and even in cases that the points have slightly digressed from the straight line, the p-values of the Kolmogorov-Smirnov test rejected the null hypothesis and showed the same distribution in texture features of the real and the synthetic data. The proposed algorithm provides a unique benchmark for comparison of OCT enhancement methods and a tailored augmentation method to overcome the limited number of OCTs in deep learning algorithms.

PMID:34792759 | DOI:10.1007/s11517-021-02469-w