Search

Machine learning algorithms make predictions by fitting highly parameterized nonlinear functions to massive amounts of data. Yet those models are not necessarily consistent with physical laws and offer limited interpretability. Extending machine learning models by introducing scientific knowledge in the optimization problem is known as physics-based and data-driven modelling. A promising development are physics informed neural networks (PINN) which ensure consistency to both physical laws and measured data. The aim of this research is to model the time-dependent temperature profile in bulk materials following the passage of a moving laser focus by a PINN. The results from the PINN agree essentially with finite element simulations, proving the suitability of the approach. New perspectives for applications in laser material processing arise when PINNs are integrated in monitoring systems or used for model predictive control.

Design of experiments (DOE) is an established method to allocate resources for efficient parameter space exploration. Model based active learning (AL) data sampling strategies have shown potential for further optimization. This paper introduces a workflow for conducting DOE comparative studies using automated machine learning. Based on a practical definition of model complexity in the context of machine learning, the interplay of systematic data generation and model performance is examined considering various sources of uncertainty: this includes uncertainties caused by stochastic sampling strategies, imprecise data, suboptimal modeling, and model evaluation. Results obtained from electrical circuit models with varying complexity show that not all AL sampling strategies outperform conventional DOE strategies, depending on the available data volume, the complexity of the dataset, and data uncertainties. Trade-offs in resource allocation strategies, in particular between identical replication of data points for statistical noise reduction and broad sampling for maximum parameter space exploration, and their impact on subsequent machine learning analysis are systematically investigated. Results indicate that replication oriented strategies should not be dismissed but may prove advantageous for cases with non-negligible noise impact and intermediate resource availability. The provided workflow can be used to simulate practical experimental conditions for DOE testing and DOE selection., Competing Interests: Declarations. Competing interests: The authors declare no competing interests., (© 2024. The Author(s).)

Neurodevelopmental conditions, such as Autism Spectrum Disorder (ASD) and Attention Deficit Hyperactivity Disorder (ADHD), present unique challenges due to overlapping symptoms, making an accurate diagnosis and targeted intervention difficult. Our study employs advanced machine learning techniques to analyze functional magnetic resonance imaging (fMRI) data from individuals with ASD, ADHD, and typically developed (TD) controls, totaling 120 subjects in the study. Leveraging multiclass classification (ML) algorithms, we achieve superior accuracy in distinguishing between ASD, ADHD, and TD groups, surpassing existing benchmarks with an area under the ROC curve near 98%. Our analysis reveals distinct neural signatures associated with ASD and ADHD: individuals with ADHD exhibit altered connectivity patterns of regions involved in attention and impulse control, whereas those with ASD show disruptions in brain regions critical for social and cognitive functions. The observed connectivity patterns, on which the ML classification rests, agree with established diagnostic approaches based on clinical symptoms. Furthermore, complex network analyses highlight differences in brain network integration and segregation among the three groups. Our findings pave the way for refined, ML-enhanced diagnostics in accordance with established practices, offering a promising avenue for developing trustworthy clinical decision-support systems., Competing Interests: EpilabKI is funded through the Bavarian stated Ministery for Sciences and the Arts research. This does not alter our adherence to PLOS ONE policies on sharing data and materials., (Copyright: © 2024 Alves et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Importance: Boarding in the emergency department (ED) is a critical indicator of quality of care for hospitals. It is defined as the time between the admission decision and departure from the ED. As a result of boarding, patients stay in the ED until inpatient beds are available; moreover, boarding is associated with various adverse events., Study Objective: The objective of our systematic review was to determine whether ED boarding (EDB) time is associated with in-hospital mortality (IHM)., Methods: A systematic search was conducted in academic databases to identify relevant studies. Medline, PubMed, Scopus, Embase, Cochrane, Web of Science, Cochrane, CINAHL and PsychInfo were searched. We included all peer-reviewed published studies from all previous years until November 2018. Studies performed in the ED and focused on the association between EDB and IHM as the primary objective were included. Extracted data included study characteristics, prognostic factors, outcomes, and IHM. A search update in PubMed was performed in May 2019 to ensure the inclusion of recent studies before publishing., Results: From the initial 4,321 references found through the systematic search, the manual screening of reference lists and the updated search in PubMed, a total of 12 studies were identified as eligible for a descriptive analysis. Overall, six studies found an association between EDB and IHM, while five studies showed no association. The last remaining study included both ICU and non-ICU subgroups and showed conflicting results, with a positive association for non-ICU patients but no association for ICU patients. Overall, a tendency toward an association between EDB and IHM using the pool random effect was observed., Conclusion: Our systematic review did not find a strong evidence for the association between ED boarding and IHM but there is a tendency toward this association. Further well-controlled, international multicenter studies are needed to demonstrate whether this association exists and whether there is a specific EDB time cut-off that results in increased IHM., Competing Interests: AB is the CEO of Global Healthcare Network & Research Innovation Institute LLC. There are no related patents, products in development, or marketed products to declare. This does not alter our adherence to PLOS ONE policies on data and materials sharing.

Atherosclerotic renal artery stenosis (ARAS) is a predictor of increased morbidity and mortality. However, whether ARAS itself accelerates the arteriosclerotic process or whether ARAS is solely the consequence of atherosclerosis is unclear. We imaged renal arteries of 1561 hypertensive patients undergoing coronary angiography and followed this cohort for 9 years (range, 2.4-15.1 years; median, 31.2 months, interquartile range, 13.4/52.9 months). All patients received aspirin, renin-angiotensin system blockade, statins, and beta blockade as indicated. One hundred seventy-one patients had ARAS >50% diameter stenosis and 126 patients an arteriosclerotic plaque (ARAP) without significant stenosis. Blood pressures were not different in ARAS, ARAP, and non-ARAS patients. After adjustment for cardiovascular risk factors by propensity scores and matched pair analysis, ARAS patients had a lower ejection fraction and more coronary artery disease (CAD) than non-ARAS patients. The same was true for brain natriuretic peptide values, troponin I, and highly sensitive C-reative protein. Over 9 years, more ARAS patients died of any cause (34% vs 23%; P < .05). The prevalence of CAD in ARAP patients was higher than in non-ARAS patients and lower than in ARAS patients. The mortality of the ARAP patients at 9 years was 37%, not different from the ARAS patients. Atherosclerotic renal artery disease appears to be a marker for the severity of atherosclerosis rather than a causative factor for atherosclerosis progression., (Copyright © 2013 American Society of Hypertension. Published by Elsevier Inc. All rights reserved.)

Serum mid-regional pro-atrial natriuretic peptide (MR-proANP) and pro-adrenomedullin (MR-proADM) are novel biomarkers for acute heart failure (AHF). Like other AFH biomarkers, the performance of these tests are affected by the presence of clinical variables such as renal failure and obesity. In a substudy of the Biomarkers from Acute Heart Failure Study, we show that diabetes did not influence the performance of these markers with regards to AHF diagnosis or 90-day all cause death. However, in patients without AHF, increased MR-proADM alone was associated with the presence of diabetes.

We prospectively evaluated the learning curve (LC) for laparoscopic urethrovesical anastomosis (L-UVA) in an operator-training model and program using an innovative simplified pelvic-trainer model. Over a period of 12 months, 30 LRP were performed by one urologist skilled in open surgery but inexperienced in laparoscopy. During the first 15 procedures no systematic training was done. Consequentially, a systematic simplified daily program was performed on the pelvic trainer with a videolaparoscopic unit. The training lesson consisted of intracorporal knotting and suturing, linear and circular interrupted suture anastomosis. At the end of each lesson, time and performance error scores were recorded and progression curve was plotted for each task. The performances of each training tasks were plotted against the performance of L-UVA during the LRP. The significance of progression was evaluated using logarithmic regression analysis. A steady improvement in time and accuracy of performance skill was shown during the first 20 lessons (p<0.001). These improved skill acquisitions were proportionally correlated with the time and the accuracy (water-tight) of L-UVA performance during the last 15 L-RPE. Compared to the first 15 L-RPE, where no systematic training was performed, time and accuracy of L-UVA performance in the last 15 L-RPE were improved from a mean 51 (median 48, range: 38-75) to 26 (median 24, range 18-33) min (p<0.001) and from 10 to 15 watertight anastomoses (p<0.001), respectively. Using a continuing, systematic, simplified training model the LC of L-UVA can be improved significantly in a short time.

Objective: We prospectively evaluated the learning curve (LC) of laparoscopic radical prostatectomy (L-RPE) regarding the improvements in operative times (OT) and technical difficulty in one-operator-practice as it compares with open RPE., Methods: Over 18 months, 50 L-RPE were performed by an inexperienced surgeon in laparoscopy but skilled in open surgery. Difficulty scores were obtained at the completion of each L-RPE comparing L-RPE to open RPE. OT, estimated blood loss (EBL), length of stay, and catheterization time were also obtained., Results: In the ablative part of L-RPE the median difficulty score was significantly higher (p<0.001) for the first 10 cases, decreased dramatically by case 11 becoming equivalent (p=0.3) to open RPE and by case 31 the L-RPE becomes significantly easier than open RPE (p=0.002). The difficulty scores for the urethrovesical anastomosis performance is always uniformly higher for the whole patient series (p<0.001). Median OT decreased significantly from 293 minutes in the first 10 cases to 114 minutes in the last 10 cases (p<0.001). Catheterization time and length of hospitalization, decreased significantly with the progression of the LC. EBL remained stable throughout the patient cases. Obesity, prior surgery, and extension of the procedure (lymphadenectomy, nerve-sparing) significantly increased the OT., Conclusion: Although the ablative part of L-RPE has a relatively short LC for a skilled open surgeon reflected by the rapid decrease in difficulty scores and OT by case 21, the performance of anastomosis shows a longer LC. Intensive training on anastomosis may be necessary to master this skill.

Objectives: To compare the impact of radical nephrectomy and nephron-sparing surgery (NSS) for localized renal cell carcinoma on quality of life (QOL)., Methods: Retrospectively, 357 patients who had undergone NSS (n = 158) or radical nephrectomy (n = 199) for localized renal cell carcinoma completed postal questionnaires, including measures of QOL with validated instruments (SF-36, European Organization for Research and Treatment of Cancer Quality of Life Questionnaire C30 [EORTC QLQ-C30]), the impact of the stress of cancer, fear of recurrence, and worry about having fewer than two normal kidneys. A subset of 51 patients diagnosed after 2000 were followed up prospectively for at least 1 year., Results: The mental and physical health composite scores were not significantly different from the validated norms for an age and sex-matched community sample. Although the type of operation had no influence on patients' overall QOL, all patients who underwent elective NSS showed a significantly greater score on physical function than patients treated with radical nephrectomy (P <0.001). Predictors for higher scores included elective NSS, comorbidity (assessed with standardized checklist), tumor size, and time since nephrectomy. The overall QOL scores and recovery of stress from cancer in patients treated with NSS for tumor less than 4 cm with a normal contralateral kidney were significantly superior to those who underwent NSS for tumor greater than 4 cm (P <0.05). Patients questioned after mandatory NSS were significantly more concerned about cancer recurrence., Conclusions: Patients without evidence of disease have relatively normal physical and mental health after operative treatment for localized renal cell carcinoma, independent of the kind of surgery. The QOL correlates proportionally with the size of tumor and is significantly better for patients undergoing NSS for tumor less than 4 cm with a normal contralateral kidney.