Identification of high likelihood of dementia in population-based surveys using unsupervised clustering: a longitudinal analysis.

BACKGROUND: Dementia is defined as a cognitive decline that affects functional status. Longitudinal ageing surveys often lack a clinical diagnosis of dementia though measure cognition and daily function over time. We used unsupervised machine learning and longitudinal data to identify transition to probable dementia. METHODS: Multiple Factor Analysis was applied to longitudinal function and cognitive data of 15,278 baseline participants (aged 50 years and more) from the Survey of Health, Ageing, and Retirement in Europe (SHARE) (waves 1, 2 and 4-7, between 2004 and 2017). Hierarchical Clustering on Principal Components discriminated three clusters at each wave. We estimated probable or "Likely Dementia" prevalence by sex and age, and assessed whether dementia risk factors increased the risk of being assigned probable dementia status using multistate models. Next, we compared the "Likely Dementia" cluster with self-reported dementia status and replicated our findings in the English Longitudinal Study of Ageing (ELSA) cohort (waves 1-9, between 2002 and 2019, 7840 participants at baseline). RESULTS: Our algorithm identified a higher number of probable dementia cases compared with self-reported cases and showed good discriminative power across all waves (AUC ranged from 0.754 [0.722-0.787] to 0.830 [0.800-0.861]). "Likely Dementia" status was more prevalent in older people, displayed a 2:1 female/male ratio, and was associated with nine factors that increased risk of transition to dementia: low education, hearing loss, hypertension, drinking, smoking, depression, social isolation, physical inactivity, diabetes, and obesity. Results were replicated in ELSA cohort with good accuracy. CONCLUSIONS: Machine learning clustering can be used to study dementia determinants and outcomes in longitudinal population ageing surveys in which dementia clinical diagnosis is lacking.

A convolutional neural network model for EPID-based non-transit dosimetry.

PURPOSE: To develop an alternative computational approach for EPID-based non-transit dosimetry using a convolutional neural network model. METHOD: A U-net followed by a non-trainable layer named True Dose Modulation recovering the spatialized information was developed. The model was trained on 186 Intensity-Modulated Radiation Therapy Step & Shot beams from 36 treatment plans of different tumor locations to convert grayscale portal images into planar absolute dose distributions. Input data were acquired from an amorphous-Silicon Electronic Portal Image Device and a 6 MV X-ray beam. Ground truths were computed from a conventional kernel-based dose algorithm. The model was trained by a two-step learning process and validated through a five-fold cross-validation procedure with sets of training and validation of 80% and 20%, respectively. A study regarding the dependance of the amount of training data was conducted. The performance of the model was evaluated from a quantitative analysis based the ϒ-index, absolute and relative errors computed between the inferred dose distributions and ground truths for six square and 29 clinical beams from seven treatment plans. These results were also compared to those of an existing portal image-to-dose conversion algorithm. RESULTS: For the clinical beams, averages of ϒ-index and ϒ-passing rate (2%-2mm > 10% Dmax ) of 0.24 (±0.04) and 99.29 (±0.70)% were obtained. For the same metrics and criteria, averages of 0.31 (±0.16) and 98.83 (±2.40)% were obtained with the six square beams. Overall, the developed model performed better than the existing analytical method. The study also showed that sufficient model accuracy can be achieved with the amount of training samples used. CONCLUSION: A deep learning-based model was developed to convert portal images into absolute dose distributions. The accuracy obtained shows that this method has great potential for EPID-based non-transit dosimetry.

Identification of High Likelihood of Dementia in Population-Based Surveys using Unsupervised Clustering: a Longitudinal Analysis.

Background: Dementia is defined by cognitive decline that affects functional status. Longitudinal ageing surveys often lack a clinical diagnosis of dementia though measure cognitive and function over time. We used unsupervised machine learning and longitudinal data to identify transition to probable dementia. Methods: Multiple Factor Analysis was applied to longitudinal function and cognitive data of 15,278 baseline participants (aged 50 years and more) from the Survey of Health, Ageing, and Retirement in Europe (SHARE) (waves 1, 2 and 4-7, between 2004 and 2017). Hierarchical Clustering on Principal Components discriminated three clusters at each wave. We estimated probable or "Likely Dementia" prevalence by sex and age, and assessed whether dementia risk factors increased the risk of being assigned probable dementia status using multistate models. Next, we compared the "Likely Dementia" cluster with self-reported dementia status and replicated our findings in the English Longitudinal Study of Ageing (ELSA) cohort (waves 1-9, between 2002 and 2019, 7,840 participants at baseline). Findings: Our algorithm identified a higher number of probable dementia cases compared with self-reported cases and showed good discriminative power across all waves (AUC ranged from 0.754 [0.722-0.787] to 0.830 [0.800-0.861]). "Likely Dementia" status was more prevalent in older people, displayed a 2:1 female/male ratio and was associated with nine factors that increased risk of transition to dementia: low education, hearing loss, hypertension, drinking, smoking, depression, social isolation, physical inactivity, diabetes, and obesity. Results were replicated in ELSA cohort with good accuracy. Interpretation: Machine learning clustering can be used to study dementia determinants and outcomes in longitudinal population ageing surveys in which dementia clinical diagnosis is lacking.

Association of Early Norepinephrine Administration With 24-Hour Mortality Among Patients With Blunt Trauma and Hemorrhagic Shock.

Importance: Hemorrhagic shock is a common cause of preventable death after injury. Vasopressor administration for patients with blunt trauma and hemorrhagic shock is often discouraged. Objective: To evaluate the association of early norepinephrine administration with 24-hour mortality among patients with blunt trauma and hemorrhagic shock. Design, Setting, and Participants: This retrospective, multicenter, observational cohort study used data from 3 registries in the US and France on all consecutive patients with blunt trauma from January 1, 2013, to December 31, 2018. Patients were alive on admission with hemorrhagic shock, defined by prehospital or admission systolic blood pressure less than 100 mm Hg and evidence of hemorrhage (ie, prehospital or resuscitation room transfusion of packed red blood cells, receipt of emergency treatment for hemorrhage control, transfusion of >10 units of packed red blood cells in the first 24 hours, or death from hemorrhage). Blunt trauma was defined as any exposure to nonpenetrating kinetic energy, collision, or deceleration. Statistical analysis was performed from January 15, 2021, to February 22, 2022. Exposure: Continuous administration of norepinephrine in the prehospital environment or resuscitation room prior to hemorrhage control, according to European guidelines. Main Outcomes and Measures: The primary outcome was 24-hour mortality, and the secondary outcome was in-hospital mortality. The average treatment effect (ATE) of early norepinephrine administration on 24-hour mortality was estimated according to the Rubin causal model. Inverse propensity score weighting and the doubly robust approach with 5 distinct analytical strategies were used to determine the ATE. Results: A total of 52 568 patients were screened for inclusion, and 2164 patients (1508 men [70%]; mean [SD] age, 46 [19] years; median Injury Severity Score, 29 [IQR, 17-36]) presented with acute hemorrhage and were included. A total of 1497 patients (69.1%) required emergency hemorrhage control, 128 (5.9%) received a prehospital transfusion of packed red blood cells, and 543 (25.0%) received a massive transfusion. Norepinephrine was administered to 1498 patients (69.2%). The 24-hour mortality rate was 17.8% (385 of 2164), and the in-hospital mortality rate was 35.6% (770 of 2164). None of the 5 analytical strategies suggested any statistically significant association between norepinephrine administration and 24-hour mortality, with ATEs ranging from -4.6 (95% CI, -11.9 to 2.7) to 2.1 (95% CI, -2.1 to 6.3), or between norepinephrine administration and in-hospital mortality, with ATEs ranging from -1.3 (95% CI, -9.5 to 6.9) to 5.3 (95% CI, -2.1 to 12.8). Conclusions and Relevance: The findings of this study suggest that early norepinephrine infusion was not associated with 24-hour or in-hospital mortality among patients with blunt trauma and hemorrhagic shock. Randomized clinical trials that study the effect of early norepinephrine administration among patients with trauma and hypotension are warranted to further assess whether norepinephrine is safe for patients with hemorrhagic shock.

Commissioning and Evaluation of an Electronic Portal Imaging Device-Based In-Vivo Dosimetry Software.

This study reports on our experience with the in-vivo dose verification software, EPIgray® (DOSIsoft, Cachan, France). After the initial commissioning process, clinical experiments on phantom treatments were evaluated to assess the level of accuracy of the electronic portal imaging device (EPID) based in-vivo dose verification. EPIgray was commissioned based on the company's instructions. This involved ion chamber measurements and portal imaging of solid water blocks of various thicknesses between 5 and 35 cm. Field sizes varied between 2 x 2 cm2 and 20 x 20 cm2. The determined conversion factors were adjusted through an additional iterative process using treatment planning system calculations. Subsequently, evaluation was performed using treatment plans of single and opposed beams, as well as intensity modulated radiotherapy (IMRT) plans, based on recommendations from the task group report TG-119 to test for dose reconstruction accuracy. All tests were performed using blocks of solid water slabs as a phantom. For single square fields, the dose at isocenter was reconstructed within 3% accuracy in EPIgray compared to the treatment planning system dose. Similarly, the relative deviation of the total dose was accurately reconstructed within 3% for all IMRT plans with points placed inside a high-dose region near the isocenter. Predictions became less accurate than < 5% when the evaluation point was outside the treatment target. Dose at points 5 cm or more away from the isocenter or within an avoidance structure was reconstructed less reliably. EPIgray formalism accuracy is adequate for an efficient error detection system with verifications performed in high-dose volumes. It provides immediate intra-fractional feedback on the delivery of treatment plans without affecting the treatment beam. Besides the EPID, no additional hardware is required. The software evaluates local point dose measurements to verify treatment plan delivery and patient positioning within 5% accuracy, depending on the placement of evaluation points.

Regulatory T Cells in Melanoma Revisited by a Computational Clustering of FOXP3+ T Cell Subpopulations.

CD4(+) T cells that express the transcription factor FOXP3 (FOXP3(+) T cells) are commonly regarded as immunosuppressive regulatory T cells (Tregs). FOXP3(+) T cells are reported to be increased in tumor-bearing patients or animals and are considered to suppress antitumor immunity, but the evidence is often contradictory. In addition, accumulating evidence indicates that FOXP3 is induced by antigenic stimulation and that some non-Treg FOXP3(+) T cells, especially memory-phenotype FOXP3(low) cells, produce proinflammatory cytokines. Accordingly, the subclassification of FOXP3(+) T cells is fundamental for revealing the significance of FOXP3(+) T cells in tumor immunity, but the arbitrariness and complexity of manual gating have complicated the issue. In this article, we report a computational method to automatically identify and classify FOXP3(+) T cells into subsets using clustering algorithms. By analyzing flow cytometric data of melanoma patients, the proposed method showed that the FOXP3(+) subpopulation that had relatively high FOXP3, CD45RO, and CD25 expressions was increased in melanoma patients, whereas manual gating did not produce significant results on the FOXP3(+) subpopulations. Interestingly, the computationally identified FOXP3(+) subpopulation included not only classical FOXP3(high) Tregs, but also memory-phenotype FOXP3(low) cells by manual gating. Furthermore, the proposed method successfully analyzed an independent data set, showing that the same FOXP3(+) subpopulation was increased in melanoma patients, validating the method. Collectively, the proposed method successfully captured an important feature of melanoma without relying on the existing criteria of FOXP3(+) T cells, revealing a hidden association between the T cell profile and melanoma, and providing new insights into FOXP3(+) T cells and Tregs.

Effects of dry heating on the progression of in vitro digestion of egg white proteins: contribution of multifactorial data analysis.

The impact of dry heating on the progression of in vitro digestion of egg white proteins was investigated through application of multiple factor analysis (MFA) to electrophoresis data. Dry heating (from 1 to 10 days between 60 and 90 °C) enhanced protein unfolding and aggregation, thus generating different SDS-PAGE patterns for each sample before digestion. The progression of in vitro digestion was then modified according to the degree of protein unfolding and/or aggregation. In vitro digestion tended to decrease the heterogeneity of sample electrophoretic patterns overall but it occurred either at the very beginning of the gastric stage or throughout the gastric stage or again during the duodenal stage, depending on the heat treatment to which the sample had been subjected. At the end of digestion, three groups of samples were obtained: all samples dry heated at 60 °C and one sample dry heated for 1 day at 70 °C that were more hydrolysed than the control, samples dry heated for more than 2 days at 80 °C or 90 °C that were less hydrolysed than the control, and samples dry heated for more than 2 days at 70 °C or 1 day at 80 or 90 °C that were as hydrolysed as the control.

A comprehensive analysis of the IMRT dose delivery process using statistical process control (SPC).

The aim of this study is to introduce tools to improve the security of each IMRT patient treatment by determining action levels for the dose delivery process. To achieve this, the patient-specific quality control results performed with an ionization chamber--and which characterize the dose delivery process--have been retrospectively analyzed using a method borrowed from industry: Statistical process control (SPC). The latter consisted in fulfilling four principal well-structured steps. The authors first quantified the short-term variability of ionization chamber measurements regarding the clinical tolerances used in the cancer center (+/- 4% of deviation between the calculated and measured doses) by calculating a control process capability (C(pc)) index. The C(pc) index was found superior to 4, which implies that the observed variability of the dose delivery process is not biased by the short-term variability of the measurement. Then, the authors demonstrated using a normality test that the quality control results could be approximated by a normal distribution with two parameters (mean and standard deviation). Finally, the authors used two complementary tools--control charts and performance indices--to thoroughly analyze the IMRT dose delivery process. Control charts aim at monitoring the process over time using statistical control limits to distinguish random (natural) variations from significant changes in the process, whereas performance indices aim at quantifying the ability of the process to produce data that are within the clinical tolerances, at a precise moment. The authors retrospectively showed that the analysis of three selected control charts (individual value, moving-range, and EWMA control charts) allowed efficient drift detection of the dose delivery process for prostate and head-and-neck treatments before the quality controls were outside the clinical tolerances. Therefore, when analyzed in real time, during quality controls, they should improve the security of treatments. They also showed that the dose delivery processes in the cancer center were in control for prostate and head-and-neck treatments. In parallel, long-term process performance indices (P(p), P(pk), and P(pm)) have been analyzed. Their analysis helped defining which actions should be undertaken in order to improve the performance of the process. The prostate dose delivery process has been shown statistically capable (0.08% of the results is expected to be outside the clinical tolerances) contrary to the head-and-neck dose delivery process (5.76% of the results are expected to be outside the clinical tolerances).

Simultaneous analysis of distinct Omics data sets with integration of biological knowledge: Multiple Factor Analysis approach.

BACKGROUND: Genomic analysis will greatly benefit from considering in a global way various sources of molecular data with the related biological knowledge. It is thus of great importance to provide useful integrative approaches dedicated to ease the interpretation of microarray data. RESULTS: Here, we introduce a data-mining approach, Multiple Factor Analysis (MFA), to combine multiple data sets and to add formalized knowledge. MFA is used to jointly analyse the structure emerging from genomic and transcriptomic data sets. The common structures are underlined and graphical outputs are provided such that biological meaning becomes easily retrievable. Gene Ontology terms are used to build gene modules that are superimposed on the experimentally interpreted plots. Functional interpretations are then supported by a step-by-step sequence of graphical representations. CONCLUSION: When applied to genomic and transcriptomic data and associated Gene Ontology annotations, our method prioritize the biological processes linked to the experimental settings. Furthermore, it reduces the time and effort to analyze large amounts of 'Omics' data.