HAL-Paris 13
Not a member yet
33893 research outputs found
Sort by
Uridine as a potentiator of aminoglycosides through activation of carbohydrate transporters
No full textInternational audienceAminoglycosides (AGs) are broad-spectrum antibiotics effective against Gram-negative bacteria, with uptake dependent on membrane potential. However, the mechanisms of AG entry remain incompletely understood. Here, we identify a previously undescribed uptake pathway via carbohydrate transporters in E. coli . By deleting or overexpressing 26 carbohydrate transporters, we found that 18 facilitated AG uptake, a mechanism conserved across several Gram-negative ESKAPEE pathogens. Using fluorescent-labeled AGs and flow cytometry, we quantified differential uptake. To enhance AG efficacy, we screened 198 carbon sources for their ability to induce transporter expression using a cmtA - gfp fusion. Uridine emerged as a strong inducer of cmtA and 12 additional AG-importing transporters. Coadministration of uridine considerably improved AG efficacy against clinical and resistant E. coli strains by enhancing drug uptake. This combination also improved outcomes in human blood ex vivo and in a murine urinary tract infection model. Given uridine’s clinical safety, it holds promise as an adjuvant to potentiate AG treatment against multidrug-resistant infections
In‐hospital and 1 year incremental prognostic value of drug abuse detection in acute heart failure
No full textInternational audienceAims The study aims to assess the in‐hospital and 1 year incremental prognostic value of recent drug abuse use, detected by a systematic urinary screening, in a consecutive cohort of patients hospitalized for acute heart failure (AHF). Methods All patients admitted for AHF with a drug abuse screening using a urinary assay were included in this prospective multicentric study (39 French centres). The outcomes were (i) in‐hospital major adverse cardiovascular events (MACEs) defined as all‐cause death, resuscitated cardiac arrest or cardiogenic shock; and (ii) 1 year MACEs defined as cardiovascular death or hospitalization for AHF. Incremental prognostic value was assessed using the C‐index, the global χ 2 and likelihood‐ratio (LR) test, the net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Results In total, 458 patients with AHF were included (mean age 68 ± 14 years, 67% male, 79% of new heart failure onset). In‐hospital and 1 year MACEs occurred, respectively, in 65 (14.2%) and 129 (28.2%) patients. Drug abuse detection was independently associated with in‐hospital MACEs [model 1—known comorbidities: odds ratio (OR) = 4.46, 95% confidence interval (CI) (1.88–10.3), P < 0.001; model 2—clinical severity: OR = 3.64, 95% CI (1.56–8.26), P = 0.002], even after propensity‐matched population analysis [OR = 3.34, 95% CI (1.32–8.70), P = 0.011], with a significant incremental prognostic value over and above traditional risk factors ( C ‐statistic improvement 0.04 with LR test P < 0.001 for both models). Patients with drug abuse detection had worse 1 year survival: HR = 1.82, 95% CI (1.13–2.92), P = 0.012. Drug abuse detection was independently associated with 1 year MACEs after adjustment with traditional prognosticators [OR = 2.54, 95% CI (1.28–4.98), P = 0.008] and propensity‐matched population analysis [OR = 2.77, 95% CI (1.98–5.21), P = 0.001], with an incremental prognostic value as well ( C ‐statistic improvement 0.02, LR test P < 0.001, positive NRI and IDI). Conclusions Drug abuse use was independently associated with a higher occurrence of both in‐hospital and 1 year MACEs with an incremental prognostic value. These results suggest a potential interest of a systematic illicit drug screening in these patients. TRIAL REGISTRATION ClinicalTrials.gov Identifier: NCT05063097
Amélioration de l’analyse d’image en hématologie avec une intelligence artificielle robuste, transparente et adaptable
No full textThe study of a blood sample is performed through several different tests. Among them, cytology analysis is the visual observation of blood cells under the microscope. The purpose of this step is to evaluate the respective proportion of different types of blood cells and to identify potential morphological anomalies. With the surge of computer vision, the analysis of images can be automated; however, algorithms are characterized by important limitations in terms of generalizability, interpretability, and adaptability. In other words, the implementation of AI models in clinical practice is restricted by their lack of robustness across images taken under different conditions, the difficulties they show when having to explain the reasons behind their predictions, and their reduced performance in tasks for which they were not trained. The objective of this thesis is to develop a robust, transparent, and deployable AI system applied to hematology. Hence, we first propose a simple color transform operation to normalize the style of images, coupled with fine-tuning strategy. This method enhances model robustness across different datasets, while preserving the visual aspect of the cells, and remaining easy-to-apply and computationally affordable. Second, we design a multimodal architecture that leverages both cell images and morphological features, extracted beforehand. This network allows scoring the importance of each feature for model prediction, which explains the main decision criteria. Third, we explore reduced supervision setup, by training the model with both labeled and unlabeled images. The categories of white blood cells are hierarchically organized to align with the differentiation of the cell and allow flexible annotation. The resulting image embeddings are more adaptable for various tasks and provide a better representation of cell biology. Finally, we broaden the point of view to highlight how recent progress in AI paves the way for a more global understanding of the patient. Our contributions addressed not only some of the generalizability, interpretability, and adaptability issues but also satisfactorily apply to further research. In particular, they could be relevant for current important topics such as the reduction of batch effect for foundation models, the improvement of explainability in large multimodal models, and the integration of biomedical knowledge in self-supervised models.L'étude d'un échantillon sanguin peut être réalisée au moyen de plusieurs tests. Parmi ceux-ci, on appelle cytologie l'observation des cellules sanguines au microscope. L'objectif de ce procédé est d'évaluer la quantité de chaque type de cellules sanguines et de repérer de potentielles anomalies morphologiques. Les récentes avancées en vision par ordinateur ont permis d'automatiser cette étape, cependant, les algorithmes posent d'importants problèmes de généralisabilité, d'expliquabilité et d'adaptabilité. Autrement dit, l'implémentation en pratique clinique est limitée par le manque de robustesse des modèles sur des images venant de conditions d'acquisition différentes, par les difficultés à expliquer les raisons qui ont poussé leur prédiction et par les compétences limitées du modèle sur des tâches pour lesquelles il n'a pas été entraîné. L'objectif de cette thèse est de développer un système qui soit robuste, transparent et déployable en conditions cliniques. Pour cela, nous proposons tout d'abord de standardiser le style des images à l'aide d'une transformation colorimétrique simple combinée à de l'ajustement fin sur différents jeux de données. La procédure permet d'améliorer la robustesse du modèle tout en préservant l'aspect visuel de la cellule et en limitant le coût de calcul ainsi que le temps d'annotation pour l'expert. Ensuite, nous développons un réseau basé sur une architecture multimodale qui prendrait en compte à la fois les images complètes et les caractéristiques morphologiques de la cellule, extraites au préalable. Cette structure de modèle nous permet de quantifier le rôle de chaque caractéristique dans la prédiction, et ainsi de comprendre les critères principaux de décision. Enfin, nous élaborons une approche semi-supervisée, où le modèle utilise des images d'annotées et non annotées. Les catégories de cellules sont représentées de manière hiérarchique, témoignant de la différentiation du globule blanc. Cela présente le double avantage de donner plus de flexibilité au système d'annotation et de générer des représentations d'images plus versatiles, qui peuvent ensuite être réutilisées pour différentes tâches. En dernier lieu, nous élargissons la perspective pour mettre en lumière comment les progrès récents de l'intelligence artificielle nous permettent d'envisager le patient sous un angle plus global. Nos contributions non seulement répondent à certains enjeux de généralisabilité, d'interprétabilité et d'adaptabilité, mais sont aussi prometteuses pour de futures recherches. Elles sont notamment pertinentes pour des problématiques actuelles de réduction de l'effet de batch dans les modèles de fondation, d'amélioration de l'explicabilité des larges modèles multimodaux et d'introduction de connaissance biomédicale dans des modèles peu supervisés
Re-thinking the Metaverse Toward Ecological Futures: A-cosmic Immersive Virtual Worlds?
No full textInternational audienc
Coopetitive Dynamics and Innovation in Services: Introduction to Drivers and Outcomes in theInsurance Sector
No full textInternational audienc
Lamb waves nonlinear imaging of impact damages in laminate composite plates
No full textInternational audienceComposite materials are increasingly used in the aeronautics industry, as they are lighter than traditional materials for equivalent mechanical strength. However, they are subject to impacts, leading to damages sometimes very difficult to detect with traditional ultrasonic Structural Health Monitoring methods. In this paper, a baseline free method making use of a low frequency vibration combined with a high frequency guided waves imaging technique is tested for real impact damages on Carbon Fiber Reinforced Composite laminate plates typically used in aeronautics
The digital transformation of the french non-life insurance sector: "ai: a resurgence of the barras reverse cycle?
No full textInternational audienceThis paper studies the digitization of French non-life insurance services, and in particular the impacts of the introduction of artificial intelligence. By detecting anomalies quickly, systematizing repetitive cognitive tasks and improving forecasting, artificial intelligence is boosting the productivity of service activities, and substantially reducing the costs of certain tasks. But the integration of artificial intelligence into the service sector also brings a number of new challenges such as the financial weight of its introduction into organizations, the ethical issues related to data protection, and above all the strategic challenges associated with partnership building and the evolution of the market share following the transformation of the insurance ecosystem. On the methodological side, this research is based on a qualitative analysis. Several interviews of the key players of the French insurance ecosystem are conducted. To analyze the knowledge and practice areas, we mobilize the work of Ivanov and Webster (2019). These authors highlight the various dynamics that crystallize the issues and prospects of AI in insurance services. This conceptual framework is based on the insurer organization and identifies the activities in its value chain that can be outsourced and those that remain managed in-house. Other insurers (strict competitors and "coopetitors") and AI-providing service providers (not exclusive to the insurance sector) represent the other side of this ecosystem. The initial results show that, while the introduction of artificial intelligence appears to follow the stages described by Barras' model, it leads to significant transformations in these activities. The introduction of artificial intelligence seems to lead to profound changes in the insurance ecosystem
Circular RNA signature of aggressive CLL with t(14;19)(q32;q13). An ERIC study
No full textInternational audienceIn Chronic Lymphocytic Leukemia (CLL), t(14;19)(q32;q13), leading to the overexpression of BCL3, is found in ∼1% of cases and is associated with an aggressive disease. In this study, leveraging a large CLL patient cohort collected thanks to an international collaboration, we investigate for the first time the circular transcriptome (circRNAome) associated with the rare t(14;19), in comparison with CLL without t(14;19) and B cells of age-matched healthy donors. We described the circRNAs commonly dysregulated in CLL, including circCSNK1G3 and circEXOC6B(3–5), which were depleted, and circZNF609 and circLPAR3, which were overexpressed in malignant cells. Of importance, we disclosed the circRNA signature of CLL with t(14;19), formed by circRNAs with expression significantly altered specifically in link with this lesion, ectopically expressed like circCDK14(3–4), circCORO1C, circCLEC2D, and circEMB, or downregulated like circCEP70(3–6). Several of these molecules were previously shown to be dysregulated or play a role in cancer, whereas most of the signature circRNAs deserve further investigation. CLL patients with high circCORO1C and circCLEC2D expression had significantly worse clinical outcomes, with shorter time to first treatment and overall survival. This study disclosed new molecular features of the aggressive CLL subtype with t(14;19)
« Portes ouvertes. Déchets collectifs du quotidien »
No full textTexte dans : Gouchon M., Joxe S., Perez J., Régeard N. et Tixadou S., (coord.), 2025, « Portfolio : La mise en image des déchets dans la ville : objets et matières (partie 1) », Urbanités, Vu, mars 2025, en ligne
Sparing effects of FLASH irradiation in patient-derived lung tissue
No full textInternational audienceBackground and purpose: Radiation toxicities, such as pneumonitis and fibrosis, are major limitations affecting patients' quality of life. Developed a decade ago, FLASH radiotherapy is an innovative method that, by delivering radiation at ultrafast dose rate, reduces radiation toxicities on healthy tissue while preserving the anti-tumoral effect of radiotherapy. This so-called FLASH effect has been described in different preclinical models but has not been observed in human tissue. This study aims to determine if FLASH irradiation can induce a sparing effect on human healthy lung tissue.Materials and methods: To address this question, precision-cut lung slices (Hu-PCLS) were prepared from healthy lung samples collected from 19 lung cancer patients undergoing lobectomy. These Hu-PCLS were irradiated ex vivo at a dose of 9 Gy using the ElectronFLASH (SIT) device operated either in conventional or FLASH mode. We monitored cell division for each patient and performed RNAseq analysis to uncover some mechanistic insights.Results: Analysis of cell division 24 h after treatment with conventional or ultra-high dose rate showed a higher proportion of dividing cells in Hu-PCLS after FLASH irradiation. Consistently, RNAseq analysis from irradiated lung samples confirmed an attenuated cell cycle checkpoint inhibition, p53 pro-apoptotic genes, DNA damage, and antioxidant pathways after ultra-high dose rate compared to conventional treatment.Conclusion: Altogether, this study shows that, using freshly isolated patient-derived lung samples, cell proliferation can serve as an early marker of the normal lung response to FLASH irradiation. These findings hold great promises for future applications of FLASH radiotherapy in the clinic