Category: Statistics

Anal Chem. 2024 Mar 12. doi: 10.1021/acs.analchem.3c05741. Online ahead of print.

ABSTRACT

Pancreatic cancer is a highly aggressive and rapidly progressing disease, often diagnosed in advanced stages due to the absence of early noticeable symptoms. The KRAS mutation is a hallmark of pancreatic cancer, yet the underlying mechanisms driving pancreatic carcinogenesis remain elusive. Cancer cells display significant metabolic heterogeneity, which is relevant to the pathogenesis of cancer. Population measurements may obscure information about the metabolic heterogeneity among cancer cells. Therefore, it is crucial to analyze metabolites at the single-cell level to gain a more comprehensive understanding of metabolic heterogeneity. In this study, we employed a 3D-printed ionization source for metabolite analysis in both mice and human pancreatic cancer cells at the single-cell level. Using advanced machine learning algorithms and mass spectral feature selection, we successfully identified 23 distinct metabolites that are statistically significantly different in KRAS mutant human pancreatic cancer cells and mouse acinar cells bearing the oncogenic KRAS mutation. These metabolites encompass a variety of chemical classes, including organic nitrogen compounds, organic acids and derivatives, organoheterocyclic compounds, benzenoids, and lipids. These findings shed light on the metabolic remodeling associated with KRAS-driven pancreatic cancer initiation and indicate that the identified metabolites hold promise as potential diagnostic markers for early detection in pancreatic cancer patients.

PMID:38471062 | DOI:10.1021/acs.analchem.3c05741

J Clin Oncol. 2024 Mar 12:JCO2302474. doi: 10.1200/JCO.23.02474. Online ahead of print.

ABSTRACT

PURPOSE: Allogeneic hematopoietic cell transplantation (HCT) improves outcomes for patients with acute myeloid leukemia (AML) harboring an internal tandem duplication mutation of FLT3 (FLT3-ITD) AML. These patients are routinely treated with a FLT3 inhibitor after HCT, but there is limited evidence to support this. Accordingly, we conducted a randomized trial of post-HCT maintenance with the FLT3 inhibitor gilteritinib (ClinicalTrials.gov identifier: NCT02997202) to determine if all such patients benefit or if detection of measurable residual disease (MRD) could identify those who might benefit.

METHODS: Adults with FLT3-ITD AML in first remission underwent HCT and were randomly assigned to placebo or 120 mg once daily gilteritinib for 24 months after HCT. The primary end point was relapse-free survival (RFS). Secondary end points included overall survival (OS) and the effect of MRD pre- and post-HCT on RFS and OS.

RESULTS: Three hundred fifty-six participants were randomly assigned post-HCT to receive gilteritinib or placebo. Although RFS was higher in the gilteritinib arm, the difference was not statistically significant (hazard ratio [HR], 0.679 [95% CI, 0.459 to 1.005]; two-sided P = .0518). However, 50.5% of participants had MRD detectable pre- or post-HCT, and, in a prespecified subgroup analysis, gilteritinib was beneficial in this population (HR, 0.515 [95% CI, 0.316 to 0.838]; P = .0065). Those without detectable MRD showed no benefit (HR, 1.213 [95% CI, 0.616 to 2.387]; P = .575).

CONCLUSION: Although the overall improvement in RFS was not statistically significant, RFS was higher for participants with detectable FLT3-ITD MRD pre- or post-HCT who received gilteritinib treatment. To our knowledge, these data are among the first to support the effectiveness of MRD-based post-HCT therapy.

PMID:38471061 | DOI:10.1200/JCO.23.02474

Neurology. 2024 Apr 9;102(7):e209173. doi: 10.1212/WNL.0000000000209173. Epub 2024 Mar 12.

ABSTRACT

BACKGROUND AND OBJECTIVES: The association between statin use and the risk of intracranial hemorrhage (ICrH) following ischemic stroke (IS) or transient ischemic attack (TIA) in patients with cerebral microbleeds (CMBs) remains uncertain. This study investigated the risk of recurrent IS and ICrH in patients receiving statins based on the presence of CMBs.

METHODS: We conducted a pooled analysis of individual patient data from the Microbleeds International Collaborative Network, comprising 32 hospital-based prospective studies fulfilling the following criteria: adult patients with IS or TIA, availability of appropriate baseline MRI for CMB quantification and distribution, registration of statin use after the index stroke, and collection of stroke event data during a follow-up period of ≥3 months. The primary endpoint was the occurrence of recurrent symptomatic stroke (IS or ICrH), while secondary endpoints included IS alone or ICrH alone. We calculated incidence rates and performed Cox regression analyses adjusting for age, sex, hypertension, atrial fibrillation, previous stroke, and use of antiplatelet or anticoagulant drugs to explore the association between statin use and stroke events during follow-up in patients with CMBs.

RESULTS: In total, 16,373 patients were included (mean age 70.5 ± 12.8 years; 42.5% female). Among them, 10,812 received statins at discharge, and 4,668 had 1 or more CMBs. The median follow-up duration was 1.34 years (interquartile range: 0.32-2.44). In patients with CMBs, statin users were compared with nonusers. Compared with nonusers, statin therapy was associated with a reduced risk of any stroke (incidence rate [IR] 53 vs 79 per 1,000 patient-years, adjusted hazard ratio [aHR] 0.68 [95% CI 0.56-0.84]), a reduced risk of IS (IR 39 vs 65 per 1,000 patient-years, aHR 0.65 [95% CI 0.51-0.82]), and no association with the risk of ICrH (IR 11 vs 16 per 1,000 patient-years, aHR 0.73 [95% CI 0.46-1.15]). The results in aHR remained consistent when considering anatomical distribution and high burden (≥5) of CMBs.

DISCUSSION: These observational data suggest that secondary stroke prevention with statins in patients with IS or TIA and CMBs is associated with a lower risk of any stroke or IS without an increased risk of ICrH.

CLASSIFICATION OF EVIDENCE: This study provides Class III evidence that for patients with IS or TIA and CMBs, statins lower the risk of any stroke or IS without increasing the risk of ICrH.

PMID:38471056 | DOI:10.1212/WNL.0000000000209173

J Chem Inf Model. 2024 Mar 12. doi: 10.1021/acs.jcim.3c02054. Online ahead of print.

ABSTRACT

The protein-ligand binding free energy is a central quantity in structure-based computational drug discovery efforts. Although popular alchemical methods provide sound statistical means of computing the binding free energy of a large breadth of systems, they are generally too costly to be applied at the same frequency as end point or ligand-based methods. By contrast, these data-driven approaches are typically fast enough to address thousands of systems but with reduced transferability to unseen systems. We introduce DrΔG-Net (or simply Dragnet), an equivariant graph neural network that can blend ligand-based and protein-ligand data-driven approaches. It is based on a 3D fingerprint representation of the ligand alone and in complex with the protein target. Dragnet is a global scoring function to predict the binding affinity of arbitrary protein-ligand complexes, but can be easily tuned via transfer learning to specific systems or end points, performing similarly to common 2D ligand-based approaches in these tasks. Dragnet is evaluated on a total of 28 validation proteins with a set of congeneric ligands derived from the Binding DB and one custom set extracted from the ChEMBL Database. In general, a handful of experimental binding affinities are sufficient to optimize the scoring function for a particular protein and ligand scaffold. When not available, predictions from physics-based methods such as absolute free energy perturbation can be used for the transfer learning tuning of Dragnet. Furthermore, we use our data to illustrate the present limitations of data-driven modeling of binding free energy predictions.

PMID:38470995 | DOI:10.1021/acs.jcim.3c02054

Haemophilia. 2024 Mar 12. doi: 10.1111/hae.14988. Online ahead of print.

ABSTRACT

INTRODUCTION: Activity wristbands have been shown to be effective in relation to self-monitoring activity levels and increasing exercise adherence. However, previous reports have been based on short-term follow-ups in people with haemophilia (PWH).

AIM: (1) To evaluate compliance with physical activity (PA) recommendations in PWH during a 1-year follow-up period using activity wristbands to record daily steps and intensity; (2) To determine the effect of PA self-monitoring on clinical outcomes.

METHODS: A prospective observational study was conducted in 27 adults with severe haemophilia undergoing prophylactic treatment. The Fitbit Charge HR was used to track daily PA for an entire year. The participants were encouraged to try to reach a goal of 10,000 steps/day and to track their progress. The pre- and post-evaluation included quality of life (A36 Hemophilia-QoL Questionnaire), joint health (Haemophilia Joint Health Score), functionality (Timed Up and Go test), and muscle strength.

RESULTS: A total of 323.63 (95%CI: 194-364) valid days (i.e., > 2000 steps) were recorded. The annual average number of steps per day taken by participants was 10,379. Sixteen (59%) PWH reached 10,000 steps/day at baseline and 17 (63%) at 1 year follow-up, with no significant differences (x² = .33; p = .56). A statistically significant improvement was observed in daily moderate activity time (p = .012) and in the ‘physical health’ quality of life subscale (mean difference: 2.15 points; 95%CI: .64-3.65; p = .007).

CONCLUSION: Our results suggest that patients with severe haemophilia who self-managed their PA can improve their long-term quality of life in the domain of physical health and also the daily time spent in moderate-intensity PA.

PMID:38470981 | DOI:10.1111/hae.14988

Clin Orthop Relat Res. 2024 Mar 12. doi: 10.1097/CORR.0000000000003018. Online ahead of print.

ABSTRACT

BACKGROUND: Estimating the risk of revision after arthroplasty could inform patient and surgeon decision-making. However, there is a lack of well-performing prediction models assisting in this task, which may be due to current conventional modeling approaches such as traditional survivorship estimators (such as Kaplan-Meier) or competing risk estimators. Recent advances in machine learning survival analysis might improve decision support tools in this setting. Therefore, this study aimed to assess the performance of machine learning compared with that of conventional modeling to predict revision after arthroplasty.

QUESTION/PURPOSE: Does machine learning perform better than traditional regression models for estimating the risk of revision for patients undergoing hip or knee arthroplasty?

METHODS: Eleven datasets from published studies from the Dutch Arthroplasty Register reporting on factors associated with revision or survival after partial or total knee and hip arthroplasty between 2018 and 2022 were included in our study. The 11 datasets were observational registry studies, with a sample size ranging from 3038 to 218,214 procedures. We developed a set of time-to-event models for each dataset, leading to 11 comparisons. A set of predictors (factors associated with revision surgery) was identified based on the variables that were selected in the included studies. We assessed the predictive performance of two state-of-the-art statistical time-to-event models for 1-, 2-, and 3-year follow-up: a Fine and Gray model (which models the cumulative incidence of revision) and a cause-specific Cox model (which models the hazard of revision). These were compared with a machine-learning approach (a random survival forest model, which is a decision tree-based machine-learning algorithm for time-to-event analysis). Performance was assessed according to discriminative ability (time-dependent area under the receiver operating curve), calibration (slope and intercept), and overall prediction error (scaled Brier score). Discrimination, known as the area under the receiver operating characteristic curve, measures the model’s ability to distinguish patients who achieved the outcomes from those who did not and ranges from 0.5 to 1.0, with 1.0 indicating the highest discrimination score and 0.50 the lowest. Calibration plots the predicted versus the observed probabilities; a perfect plot has an intercept of 0 and a slope of 1. The Brier score calculates a composite of discrimination and calibration, with 0 indicating perfect prediction and 1 the poorest. A scaled version of the Brier score, 1 – (model Brier score/null model Brier score), can be interpreted as the amount of overall prediction error.

RESULTS: Using machine learning survivorship analysis, we found no differences between the competing risks estimator and traditional regression models for patients undergoing arthroplasty in terms of discriminative ability (patients who received a revision compared with those who did not). We found no consistent differences between the validated performance (time-dependent area under the receiver operating characteristic curve) of different modeling approaches because these values ranged between -0.04 and 0.03 across the 11 datasets (the time-dependent area under the receiver operating characteristic curve of the models across 11 datasets ranged between 0.52 to 0.68). In addition, the calibration metrics and scaled Brier scores produced comparable estimates, showing no advantage of machine learning over traditional regression models.

CONCLUSION: Machine learning did not outperform traditional regression models.

CLINICAL RELEVANCE: Neither machine learning modeling nor traditional regression methods were sufficiently accurate in order to offer prognostic information when predicting revision arthroplasty. The benefit of these modeling approaches may be limited in this context.

PMID:38470976 | DOI:10.1097/CORR.0000000000003018

J Patient Saf. 2024 Mar 13. doi: 10.1097/PTS.0000000000001214. Online ahead of print.

ABSTRACT

OBJECTIVES: Accidental falls are among the leading hospitals’ adverse events, with incidence ranging from 2 to 20 events per 1.000 days/patients. The objective of this study is to assess the relationship between in-hospital falls and the score of 3 DEPendence and Clinical-Social Fragility indexes.

METHODS: A monocentric case-control study was conducted by retrieving data of in-hospital patients from the electronic health records.

RESULTS: Significant differences between the mean scores at the hospital admission and discharge were found. The BRASS scale mean (SD) values at the admission and at the discharge were also significantly higher in cases of in-hospital falls: at the admission 10.2 (±7.7) in cases versus 7.0 (±8.0) in controls (P = 0.003); at the discharge 10.0 (±6.4) versus 6.7 (±7.5) (P = 0.001). Barthel index mean (SD) scores also presented statistically significant differences: at the admission 60.3 (±40.6) in cases versus 76.0 (±34.8) in controls (P = 0.003); at discharge 51.3 (±34.9) versus 73.3 (±35.2) (P = 0.000).Odds ratios were as follows: for Barthel index 2.37 (95% CI, 1.28-4.39; P = 0.003); for Index of Caring Complexity 1.45 (95% CI, 0.72-2.91, P = 0. 255); for BRASS index 1.95 (95% CI, 1.03-3.70, P = 0.026). With BRASS index, the area under the curve was 0.667 (95% CI, 0.595-0.740), thus indicating a moderate predictive power of the scale.

CONCLUSIONS: The use of only Conley scale-despite its sensitivity and specificity-is not enough to fully address this need because of the multiple and heterogeneous factors that predispose to in-hospital falls. Therefore, the combination of multiple tools should be recommended.

PMID:38470963 | DOI:10.1097/PTS.0000000000001214

J Patient Saf. 2024 Mar 13. doi: 10.1097/PTS.0000000000001220. Online ahead of print.

ABSTRACT

OBJECTIVE: This article aims to assess the reproducibility of Manufacturer and User Facility Device Experience (MAUDE) data-driven studies by analyzing the data queries used in their research processes.

METHODS: Studies using MAUDE data were sourced from PubMed by searching for “MAUDE” or “Manufacturer and User Facility Device Experience” in titles or abstracts. We manually chose articles with executable queries. The reproducibility of each query was assessed by replicating it in the MAUDE Application Programming Interface. The reproducibility of a query is determined by a reproducibility coefficient that ranges from 0.95 to 1.05. This coefficient is calculated by comparing the number of medical device reports (MDRs) returned by the reproduced queries to the number of reported MDRs in the original studies. We also computed the reproducibility ratio, which is the fraction of reproducible queries in subgroups divided by the query complexity, the device category, and the presence of a data processing flow.

RESULTS: As of August 8, 2022, we identified 523 articles from which 336 contained queries, and 60 of these were executable. Among these, 14 queries were reproducible. Queries using a single field like product code, product class, or brand name showed higher reproducibility (50%, 33.3%, 31.3%) compared with other fields (8.3%, P = 0.037). Single-category device queries exhibited a higher reproducibility ratio than multicategory ones, but without statistical significance (27.1% versus 8.3%, P = 0.321). Studies including a data processing flow had a higher reproducibility ratio than those without, although this difference was not statistically significant (42.9% versus 17.4%, P = 0.107).

CONCLUSIONS: Our findings indicate that the reproducibility of queries in MAUDE data-driven studies is limited. Enhancing this requires the development of more effective MAUDE data query strategies and improved application programming interfaces.

PMID:38470959 | DOI:10.1097/PTS.0000000000001220

Proc Natl Acad Sci U S A. 2024 Mar 19;121(12):e2319496121. doi: 10.1073/pnas.2319496121. Epub 2024 Mar 12.

ABSTRACT

Without the ability to control or randomize environments (or genotypes), it is difficult to determine the degree to which observed phenotypic differences between two groups of individuals are due to genetic vs. environmental differences. However, some have suggested that these concerns may be limited to pathological cases, and methods have appeared that seem to give-directly or indirectly-some support to claims that aggregate heritable variation within groups can be related to heritable variation among groups. We consider three families of approaches: the “between-group heritability” sometimes invoked in behavior genetics, the statistic [Formula: see text] used in empirical work in evolutionary quantitative genetics, and methods based on variation in ancestry in an admixed population, used in anthropological and statistical genetics. We take up these examples to show mathematically that information on within-group genetic and phenotypic information in the aggregate cannot separate among-group differences into genetic and environmental components, and we provide simulation results that support our claims. We discuss these results in terms of the long-running debate on this topic.

PMID:38470926 | DOI:10.1073/pnas.2319496121

Proc Natl Acad Sci U S A. 2024 Mar 19;121(12):e2314995121. doi: 10.1073/pnas.2314995121. Epub 2024 Mar 12.

ABSTRACT

Collective properties of complex systems composed of many interacting components such as neurons in our brain can be modeled by artificial networks based on disordered systems. We show that a disordered neural network of superconducting loops with Josephson junctions can exhibit computational properties like categorization and associative memory in the time evolution of its state in response to information from external excitations. Superconducting loops can trap multiples of fluxons in many discrete memory configurations defined by the local free energy minima in the configuration space of all possible states. A memory state can be updated by exciting the Josephson junctions to fire or allow the movement of fluxons through the network as the current through them surpasses their critical current thresholds. Simulations performed with a lumped element circuit model of a 4-loop network show that information written through excitations is translated into stable states of trapped flux and their time evolution. Experimental implementation on a high-Tc superconductor YBCO-based 4-loop network shows dynamically stable flux flow in each pathway characterized by the correlations between junction firing statistics. Neural network behavior is observed as energy barriers separating state categories in simulations in response to multiple excitations, and experimentally as junction responses characterizing different flux flow patterns in the network. The state categories that produce these patterns have different temporal stabilities relative to each other and the excitations. This provides strong evidence for time-dependent (short-to-long-term) memories, that are dependent on the geometrical and junction parameters of the loops, as described with a network model.

PMID:38470918 | DOI:10.1073/pnas.2314995121