Category: Statistics

Neural Netw. 2022 Feb 4;148:232-253. doi: 10.1016/j.neunet.2022.01.017. Online ahead of print.

ABSTRACT

We consider restricted Boltzmann machine (RBMs) trained over an unstructured dataset made of blurred copies of definite but unavailable “archetypes” and we show that there exists a critical sample size beyond which the RBM can learn archetypes, namely the machine can successfully play as a generative model or as a classifier, according to the operational routine. In general, assessing a critical sample size (possibly in relation to the quality of the dataset) is still an open problem in machine learning. Here, restricting to the random theory, where shallow networks suffice and the “grandmother-cell” scenario is correct, we leverage the formal equivalence between RBMs and Hopfield networks, to obtain a phase diagram for both the neural architectures which highlights regions, in the space of the control parameters (i.e., number of archetypes, number of neurons, size and quality of the training set), where learning can be accomplished. Our investigations are led by analytical methods based on the statistical-mechanics of disordered systems and results are further corroborated by extensive Monte Carlo simulations.

PMID:35158159 | DOI:10.1016/j.neunet.2022.01.017

Breast. 2022 Feb 8;62:114-122. doi: 10.1016/j.breast.2022.02.002. Online ahead of print.

ABSTRACT

BACKGROUND: The quantitative relationship between HER2 copy number and prognosis in HER2 positive adjuvant setting remain controversial, and few studies have focused on adjuvant setting to illustrate the potential clinical relevance of HER2 in cfDNA. Our study aim to develop a novel method in HER2 quantification and explore the relationship between HER2 copy number in primary tumors or cfDNA and prognosis in HER2 positive early breast cancer.

METHODS: Two hundred and two patients with early breast cancer were prospectively included in a study where primary tumors, matching non-cancer breast tissue, corresponding plasma, and the plasma from 20 healthy volunteers were collected. Cox proportional hazard analysis was employed to determine the prognostic value of HER2 gene copy number in tissue and cfDNA. Tissue based nomograms and time-dependent decision curve analysis were used to evaluate the practicality of HER2 copy number stratification.

RESULTS: HER2 amplification by CNVplex demonstrated a robust concordance with FISH (concordance 89.2%). A three-tiered system of tissue and a two-tiered system of cfDNA classification were shown to be independent prognostic factors. A tissue copy number-based nomogram was fitted and further evaluation revealed a good performance in discrimination (c statistic 0.801) and calibration.

CONCLUSIONS: We first report CNVplex as a viable alternative for HER2 detection. Quantitative evaluation of HER2 presents tremendous potential for use in risk stratification. We also uncover the potential for using HER2 copy number in cfDNA as a biomarker for prognosis in a HER2 positive adjuvant setting.

PMID:35158152 | DOI:10.1016/j.breast.2022.02.002

Int J Med Inform. 2022 Feb 7;160:104714. doi: 10.1016/j.ijmedinf.2022.104714. Online ahead of print.

ABSTRACT

Histopathology reports are a primary data source for the case definition phase of a Cancer Registry. By reading the histopathology report, the operator that evaluates an oncology case can define the morphology and topography of cancer, and validate the case with the highest diagnosis base. The key problem of the Catania-Messina-Enna Integrated Cancer Registry (RTI) is that these reports are written in natural language and relevant information for cancer evaluation is only a little part of the total annual histopathological reports. In this population-based retrospective cohort study, we try to optimize the working time spent by the RTI operators in seeking and selecting the right information among the histopathology reports in the east Sicily population, by developing a binary classifier on a training set of labeled historical data and validating its outcome by a test set of labeled data created by the operators during the years. Using a machine learning algorithm we built a classification model that evaluates each free text report and returns a score that indicates the probability that it contains oncologic relevant information. The best performing algorithm, among the eight analyzed in this study, was the LightGBM that reached an F1-Score of 98.9%. Using the chosen classifier we shortened the time for case evaluation, improving the timeliness of cancer statistics.

PMID:35158153 | DOI:10.1016/j.ijmedinf.2022.104714

J Stroke Cerebrovasc Dis. 2022 Feb 11;31(4):106347. doi: 10.1016/j.jstrokecerebrovasdis.2022.106347. Online ahead of print.

ABSTRACT

PURPOSE: Anticoagulation (AC) is the main preventive strategy for ischemic stroke in atrial fibrillation (AF) patients. We aim to investigate the association of prior AC with thrombus composition and clinical outcome in AF patients with acute ischemic stroke (AIS).

MATERIALS AND METHODS: From January 2019 to December 2020, consecutive AIS patients with AF treated with mechanical thrombectomy (MT) in our center were included in this analysis. Retrieved thrombi were stained with hematoxylin and eosin (H&E) and Martius Scarlet blue (MSB). The relative fractions of red blood cell (RBC), white blood cell (WBC), fibrin, and platelet were quantitatively analyzed. Procedural and clinical outcomes were compared between patients with and without prior AC.

RESULTS: A total of 133 patients were enrolled in this study, with 39 in AC group and 94 in non-AC (NAC) group. Thrombi in AC group contained more fibrins (36% vs 20%, p<0.001), more platelets (36% vs 24%, p<0.001) and fewer RBCs (25% vs 54%, p<0.001). No difference was detected in terms of successful recanalization evaluated with modified Thrombolysis in Cerebral Infarction scale (mTICI 2b-3, 97% vs 86%, p=0.065), functional independence at 90 days with modified Rankin Score (mRS 0-2, 44% vs 33%, p=0.246).

CONCLUSION: Thrombi retrieved from AF patients with prior AC contained more fibrins, more platelets and fewer RBCs compared with those of NAC patients. A trend of higher successful reperfusion rate was observed in AC patients but failed to reach statistical significance.

PMID:35158148 | DOI:10.1016/j.jstrokecerebrovasdis.2022.106347

Cancer Treat Rev. 2022 Feb 7;104:102355. doi: 10.1016/j.ctrv.2022.102355. Online ahead of print.

ABSTRACT

Neuroblastoma survivors have an increased risk of unfavorable long-term health outcomes, of which developing subsequent neoplasms is one of the most serious. We aimed to provide an overview of the current knowledge on the risk of subsequent neoplasms in neuroblastoma survivors. We conducted a systematic literature search in Medline/Pubmed (01-01-1945-13-01-2022) to identify studies that reported on ≥ 100 neuroblastoma survivors and assessed subsequent neoplasms as an outcome. We identified 410 potentially eligible articles, of which we eventually included 13 reports. All articles described retrospective cohorts with sizes varying from 145 to 5,987 neuroblastoma survivors. Within these cohorts 0.7% – 17.2% of the survivors developed a subsequent neoplasm. A wide variety of types of subsequent malignant and non-malignant neoplasms were observed, of which thyroid carcinoma and acute myeloid leukemia were most frequently reported. The risk of developing a subsequent neoplasm was 2.8 to 10.4 times higher in neuroblastoma survivors than in the general population. Although no statistically significant risk factors for subsequent neoplasms were observed in multivariable analyses, high-risk group survivors, women and those treated with radiotherapy seemed to have a higher risk. In conclusion, the studies in this systematic review consistently show that neuroblastoma survivors are at elevated risk of developing subsequent neoplasms. Future research should further explore risk factors for subsequent neoplasms in neuroblastoma survivors, so future treatment protocols and follow-up care can be improved.

PMID:35158111 | DOI:10.1016/j.ctrv.2022.102355

Nurse Educ Today. 2022 Feb 8;111:105298. doi: 10.1016/j.nedt.2022.105298. Online ahead of print.

ABSTRACT

BACKGROUND: Although students are well prepared theoretically, they lack real-life practical skills because they have not faced an adequate number of emergencies such as neonatal resuscitation.

OBJECTIVES: This study was conducted with the objective of determining the impact of integrating serious game simulation (SGS) into neonatal resuscitation training on the neonatal resuscitation related knowledge, skills, satisfaction with training, and self confidence in learning of nursing students.

DESIGN: The study is a randomized controlled, pre-test post-test design and single-blind study.

SETTINGS AND PARTICIPANTS: This study was conducted on 90 undergraduate nursing students (SGS based training group = 45, control group = 45) enrolled in the fifth semester at the Faculty of Nursing.

METHODS: The students were allocated with simple randomization method to intervention and control groups. The training program prepared on the basis of neonatal resuscitation algorithm used a neonatal resuscitation serious game simulation method. At the same time, the serious game simulation method was used as a pre-test and post-test skill assessment tool. Support was obtained from a statistician in evaluation of the data and the data were analyzed using the SPSS (Statistical Package for Social Sciences) for Windows 25.0 program.

RESULTS: Post-test measurements indicated a statistically significant positive difference in the ventilation and chest compression performing skills of the intervention group compared to the control group (p = .011, p = .020, respectively). A considerable increase was found in the knowledge and skills level of both groups, after the neonatal resuscitation training (p < .05). The score averages of the Student Satisfaction and Self-Confidence in Learning Scale and its sub-dimensions were high for both groups.

CONCLUSIONS: It was concluded that the serious game simulation application used in neonatal resuscitation training was effective in raising the students’ ventilation and compression performing skills.

PMID:35158135 | DOI:10.1016/j.nedt.2022.105298

Epilepsia Open. 2022 Feb 14. doi: 10.1002/epi4.12585. Online ahead of print.

ABSTRACT

Studies of epilepsy surgery outcomes are often small and thus underpowered to reach statistically valid conclusions. We hypothesized that ordinal logistic regression would have greater statistical power than binary logistic regression when analyzing epilepsy surgery outcomes. We reviewed 10 manuscripts included in a recent meta-analysis which found that mesial temporal sclerosis (MTS) predicted better surgical outcome after a stereotactic laser amygdalohippocampectomy (SLAH). We extracted data from 239 patients from eight studies which reported four discrete Engel surgical outcomes after SLAH, stratified by the presence or absence of MTS. The rate of freedom from disabling seizures (Engel I) was 64.3% (110/171) for patients with MTS compared to 44.1% (30/68) without MTS. The statistical power to detect MTS as a predictor for better surgical outcome after a SLAH was 29% using ordinal regression, which was significantly more than the 13% power using binary logistic regression (paired t-test, p<0.001). Only 120 patients are needed to achieve 80% power to detect MTS as a predictor using ordinal regression, compared to 210 patients that are needed to achieve 80% power using binary logistic regression. Ordinal regression should be considered when analyzing ordinal outcomes (such as Engel surgical outcome), especially for datasets with small sample sizes.

PMID:35156772 | DOI:10.1002/epi4.12585

Chem Asian J. 2022 Feb 14. doi: 10.1002/asia.202200076. Online ahead of print.

ABSTRACT

A halogenated bowl-shaped guest, corannulene, was encapsulated in a cylindrical host, [4]cyclochrysenylene, to form a bowl-in-tube complex, which mimicked supramolecular complexes between bowl guests and carbon nanotubes. As was the case with carbon nanotubes, the cylindrical space of [4]cyclochrysenylene trapped multiple corannulene molecules in an array, and 1:2 complexes were commonly obtained with the corannulene guests with various halogen substituents (F, Cl, Br and I). Careful statistical analyses of isothermal titration calorimetry titration data succeeded in revealing the stoichiometry, and the molecular structures of the 1:2 complexes were further clarified by X-ray crystallographic analyses. Two fluorinated corannulene guests were stacked perpendicular to the cylinder axis, while two chlorinated guests were stacked with inclined orientations. The structural difference resulted in a large difference in the cooperativity of the two-stage association in solution: fluorinated corannulene guests showed negative cooperativity for the 1:2 complexation, and the other, larger halogenated corannulene guests showed positive cooperativity.

PMID:35156775 | DOI:10.1002/asia.202200076

J Cosmet Dermatol. 2022 Feb 14. doi: 10.1111/jocd.14845. Online ahead of print.

ABSTRACT

INTRODUCTION: Botulinum toxin A (BoNT-A) has been increasingly used in the cosmetic dermatology daily routine. The treatment areas of BoNT-A are glabellar lines, forehead wrinkles, periorbital lines (crow’s feet), eyebrow lifting, perioral lines, platysmal bands, masseter hyperthrophy and hyperhydrosis. The aim of this study was to reveal the epidemiologic characteristics of the patients treated with BoNT-A and the frequencies of treatment areas.

PATIENT AND METHODS: The patients were grouped into five categories: facial wrinkles, platysmal bands, masseter hypertrophy, axillary hyperhydrosis and palmoplantar hyperhydrosis. Each patient received either onabotulinum toxin-A (diluted with 2 cc saline) or abobotulinum toxin-A (diluted with 3 cc saline) treatment. The patients’ age, gender and the total numbers of treatment received in each area were noted. SPSS version 21 was used for the statistical analysis.

RESULTS: A total of 1614 patients were included: 165 (6.5%) male and 1509 (93.5%) female. There is no difference between genders in terms of treatment area. (p=0.855). The number of treatment sessions is independent of the patients gender. (p=0.703). The mean age of the patients was 46.6 years. There is a relationship between the patients age and treatment area (sig 0.000). There is no relationship between the number of repetitive treatments and the patient’s age (p=0.081). The number of repetitive treatments is dependent to the treatment area (p< 0.001).

CONCLUSION: Facial wrinkles, platysmal bands and masseter hypertrophy are the most commonly treated areas. The treatment area is dependent on the age as platysmal bands are more commonly treated in older patients, however no relationship between gender and treatment area was found. BoNT-A treatment for hyperhydrosis is not as commonly repeated as for that of cosmetic concerns.

PMID:35156765 | DOI:10.1111/jocd.14845

J Magn Reson Imaging. 2022 Feb 14. doi: 10.1002/jmri.28113. Online ahead of print.

ABSTRACT

BACKGROUND: There has been a growing interest in exploring the applications of stretched-exponential (SEM) and intravoxel incoherent motion (IVIM) models of diffusion-weighted imaging (DWI) in breast imaging, with the focus on differentiation of breast lesions. However, the use of SEM and IVIM models to predict early response to neoadjuvant chemotherapy (NACT) has received less attention.

PURPOSE: To investigate the value of monoexponential, SEM, and IVIM models to predict early response to NACT in patients with primary breast cancer.

STUDY TYPE: Prospective.

POPULATION: Thirty-seven patients with primary breast cancer (aged 46 ± 11 years) due to undergo NACT.

FIELD STRENGTH/SEQUENCES: A 1.5-T MR scanner, T₁ -weighted three-dimensional spoiled gradient-echo, two-dimensional single-shot spin-echo echo-planar imaging sequence (DWI) at six b-values (0-800 s mm^-2 ).

ASSESSMENT: Tumor volume, apparent diffusion coefficient, tissue diffusion (D_t ), pseudo-diffusion coefficient (D_p ), perfusion fraction (f), distributed diffusion coefficient, and alpha (α) were extracted, following volumetric sampling of the tumors, at three time-points: pretreatment, post one and three cycles of NACT.

STATISTICAL TESTS: Mann-Whitney test, receiver operating characteristic (ROC) curve. Statistical significance level was P < 0.05.

RESULTS: Following NACT, 17 patients were determined to be pathological responders and 20 nonresponders. Tumor volume was significantly larger in nonresponders at each MRI time-point and demonstrated reasonable performance in predicting response (area under the ROC curve [AUC] = 0.83-0.87). No significant differences between groups were found in the diffusion coefficients at each time-point (P = 0.09-1). The parameters α (SEM), f, and f × D_p (IVIM) were able to differentiate between response groups after one cycle of NACT (AUC = 0.73, 0.72, and 0.74, respectively).

CONCLUSION: Diffusion coefficients derived from the monoexponential, SEM, and IVIM models did not predict pathological response. However, the IVIM-derived parameters f and f × D_p and the SEM-derived parameter α were able to predict response to NACT in breast cancer patients following one cycle of NACT.

LEVEL OF EVIDENCE: 2 TECHNICAL EFFICACY STAGE: 2.

PMID:35156741 | DOI:10.1002/jmri.28113