Development and external validation of a machine learning model for the prediction of persistent acute kidney injury stage 3 in multi-centric, multi-national intensive care cohorts.

BACKGROUND: The aim of this retrospective cohort study was to develop and validate on multiple international datasets a real-time machine learning model able to accurately predict persistent acute kidney injury (AKI) in the intensive care unit (ICU). METHODS: We selected adult patients admitted to ICU classified as AKI stage 2 or 3 as defined by the "Kidney Disease: Improving Global Outcomes" criteria. The primary endpoint was the ability to predict AKI stage 3 lasting for at least 72 h while in the ICU. An explainable tree regressor was trained and calibrated on two tertiary, urban, academic, single-center databases and externally validated on two multi-centers databases. RESULTS: A total of 7759 ICU patients were enrolled for analysis. The incidence of persistent stage 3 AKI varied from 11 to 6% in the development and internal validation cohorts, respectively and 19% in external validation cohorts. The model achieved area under the receiver operating characteristic curve of 0.94 (95% CI 0.92-0.95) in the US external validation cohort and 0.85 (95% CI 0.83-0.88) in the Italian external validation cohort. CONCLUSIONS: A machine learning approach fed with the proper data pipeline can accurately predict onset of Persistent AKI Stage 3 during ICU patient stay in retrospective, multi-centric and international datasets. This model has the potential to improve management of AKI episodes in ICU if implemented in clinical practice.

Continuous and early prediction of future moderate and severe Acute Kidney Injury in critically ill patients: Development and multi-centric, multi-national external validation of a machine-learning model.

BACKGROUND: Acute Kidney Injury (AKI) is a major complication in patients admitted to Intensive Care Units (ICU), causing both clinical and economic burden on the healthcare system. This study develops a novel machine-learning (ML) model to predict, with several hours in advance, the AKI episodes of stage 2 and 3 (according to KDIGO definition) acquired in ICU. METHODS: A total of 16'760 ICU adult patients from 145 different ICU centers and 3 different countries (US, Netherland, Italy) are retrospectively enrolled for the study. Every hour the model continuously analyzes the routinely-collected clinical data to generate a new probability of developing AKI stage 2 and 3, according to KDIGO definition, during the ICU stay. RESULTS: The predictive model obtains an auROC of 0.884 for AKI (stage 2/3 KDIGO) prediction, when evaluated on the internal test set composed by 1'749 ICU stays from US and EU centers. When externally tested on a multi-centric US dataset of 6'985 ICU stays and multi-centric Italian dataset of 1'025 ICU stays, the model achieves an auROC of 0.877 and of 0.911, respectively. In all datasets, the time between model prediction and AKI (stage 2/3 KDIGO) onset is at least of 14 hours after the first day of ICU hospitalization. CONCLUSIONS: In this study, a novel ML model for continuous and early AKI (stage 2/3 KDIGO) prediction is successfully developed, leveraging only routinely-available data. It continuously predicts AKI episodes during ICU stay, at least 14 hours in advance when the AKI episode happens after the first 24 hours of ICU admission. Its performances are validated in an extensive, multi-national and multi-centric cohort of ICU adult patients. This ML model overcomes the main limitations of currently available predictive models. The benefits of its real-world implementation enable an early proactive clinical management and the prevention of AKI episodes in ICU patients. Furthermore, the software could be directly integrated with IT system of the ICU.

External validation of a deep-learning model to predict severe acute kidney injury based on urine output changes in critically ill patients.

OBJECTIVES: The purpose of this study was to externally validate algorithms (previously developed and trained in two United States populations) aimed at early detection of severe oliguric AKI (stage 2/3 KDIGO) in intensive care units patients. METHODS: The independent cohort was composed of 10'596 patients from the university hospital ICU of Amsterdam (the "AmsterdamUMC database") admitted to their intensive care units. In this cohort, we analysed the accuracy of algorithms based on logistic regression and deep learning methods. The accuracy of investigated algorithms had previously been tested with electronic intensive care unit (eICU) and MIMIC-III patients. RESULTS: The deep learning model had an area under the ROC curve (AUC) of 0,907 (± 0,007SE) with a sensitivity and specificity of 80% and 89%, respectively, for identifying oliguric AKI episodes. Logistic regression models had an AUC of 0,877 (± 0,005SE) with a sensitivity and specificity of 80% and 81%, respectively. These results were comparable to those obtained in the two US populations upon which the algorithms were previously developed and trained. CONCLUSION: External validation on the European sample confirmed the accuracy of the algorithms, previously investigated in the US population. The models show high accuracy in both the European and the American databases even though the two cohorts differ in a range of demographic and clinical characteristics, further underlining the validity and the generalizability of the two analytical approaches.

A deep-learning model to continuously predict severe acute kidney injury based on urine output changes in critically ill patients.

BACKGROUND: Acute Kidney Injury (AKI), a frequent complication of pateints in the Intensive Care Unit (ICU), is associated with a high mortality rate. Early prediction of AKI is essential in order to trigger the use of preventive care actions. METHODS: The aim of this study was to ascertain the accuracy of two mathematical analysis models in obtaining a predictive score for AKI development. A deep learning model based on a urine output trends was compared with a logistic regression analysis for AKI prediction in stages 2 and 3 (defined as the simultaneous increase of serum creatinine and decrease of urine output, according to  the Acute Kidney Injury Network (AKIN) guidelines). Two retrospective datasets including 35,573 ICU patients were analyzed. Urine output data were used to train and test the logistic regression and the deep learning model. RESULTS: The deep learning model defined an area under the curve (AUC) of 0.89 (± 0.01), sensitivity = 0.8 and specificity = 0.84, which was higher than the logistic regression analysis. The deep learning model was able to predict 88% of AKI cases more than 12 h before their onset: for every 6 patients identified as being at risk of AKI by the deep learning model, 5 experienced the event. On the contrary, for every 12 patients not considered to be at risk by the model, 2 developed AKI. CONCLUSION: In conclusion, by using urine output trends, deep learning analysis was able to predict AKI episodes more than 12 h in advance, and with a higher accuracy than the classical urine output thresholds. We suggest that this algorithm could be integrated in the ICU setting to better manage, and potentially prevent, AKI episodes.

Polydopamine Nanoparticles as an Organic and Biodegradable Multitasking Tool for Neuroprotection and Remote Neuronal Stimulation.

Oxidative stress represents a common issue in most neurological diseases, causing severe impairments of neuronal cell physiological activity that ultimately lead to neuron loss of function and cellular death. In this work, lipid-coated polydopamine nanoparticles (L-PDNPs) are proposed both as antioxidant and neuroprotective agents, and as a photothermal conversion platform able to stimulate neuronal activity. L-PDNPs showed the ability to counteract reactive oxygen species (ROS) accumulation in differentiated SH-SY5Y, prevented mitochondrial ROS-induced dysfunctions and stimulated neurite outgrowth. Moreover, for the first time in the literature, the photothermal conversion capacity of L-PDNPs was used to increase the intracellular temperature of neuron-like cells through near-infrared (NIR) laser stimulation, and this phenomenon was thoroughly investigated using a fluorescent temperature-sensitive dye and modeled from a mathematical point of view. It was also demonstrated that the increment in temperature caused by the NIR stimulation of L-PDNPs was able to produce a Ca2+ influx in differentiated SH-SY5Y, being, to the best of our knowledge, the first example of organic nanostructures used in such an approach. This work could pave the way to new and exciting applications of polydopamine-based and of other NIR-responsive antioxidant nanomaterials in neuronal research.

Zinc Oxide Nanocrystals and High-Energy Shock Waves: A New Synergy for the Treatment of Cancer Cells.

In the last years, different nanotools have been developed to fight cancer cells. They could be administered alone, exploiting their intrinsic toxicity, or remotely activated to achieve cell death. In the latter case, ultrasound (US) has been recently proposed to stimulate some nanomaterials because of the US outstanding property of deep tissue penetration and the possibility of focusing. In this study, for the first time, we report on the highly efficient killing capability of amino-propyl functionalized ZnO nanocrystals (ZnO NCs) in synergy with high-energy ultrasound shock waves (SW) for the treatment of cancer cells. The cytotoxicity and internalization of ZnO NCs were evaluated in cervical adenocarcinoma KB cells, as well as the safety of the SW treatment alone. Then, the remarkably high cytotoxic combination of ZnO NCs and SW was demonstrated, comparing the effect of multiple (3 times/day) SW treatments toward a single one, highlighting that multiple treatments are necessary to achieve efficient cell death. At last, preliminary tests to understand the mechanism of the observed synergistic effect were carried out, correlating the nanomaterial surface chemistry to the specific type of stimulus used. The obtained results can thus pave the way for a novel nanomedicine treatment, based on the synergistic effect of nanocrystals combined with highly intense mechanical pressure waves, offering high efficiency, deep and focused tissue penetration, and a reduction of side effects on healthy cells.

Leveraging re-chargeable nanobubbles on amine-functionalized ZnO nanocrystals for sustained ultrasound cavitation towards echographic imaging.

Nanoparticles able to promote inertial cavitation when exposed to focused ultrasound have recently gained much attention due to their vast range of possible applications in the biomedical field, such as enhancing drug penetration in tumor or supporting ultrasound contrast imaging. Due to their nanometric size, these contrast agents could penetrate through the endothelial cells of the vasculature to target tissues, thus enabling higher imaging resolutions than commercial gas-filled microbubbles. Herein, Zinc Oxide NanoCrystals (ZnO NCs), opportunely functionalized with amino-propyl groups, are developed as novel nanoscale contrast agents that are able, for the first time, to induce a repeatedly and over-time sustained inertial cavitation as well as ultrasound contrast imaging. The mechanism behind this phenomenon is investigated, revealing that re-adsorption of air gas nanobubbles on the nanocrystal surface is the key factor for this re-chargeable cavitation. Moreover, inertial cavitation and significant echographic signals are obtained at physiologically relevant ultrasound conditions (MI < 1.9), showing great potential for low side-effects in in-vivo applications of the novel nanoscale agent from diagnostic imaging to gas-generating theranostic nanoplatforms and to drug delivery.

Sonophotocatalytic degradation mechanisms of Rhodamine B dye via radicals generation by micro- and nano-particles of ZnO.

In this work, it is proposed an environmental friendly sonophotocatalytic approach to efficiently treat polluted waters from industrial dyes exploiting ZnO micro- and nano-materials. For the first time, we deeply investigated the generation of reactive oxygen species (ROS) under ultrasound stimulation of different ZnO structures by Electron Paramagnetic Resonance Spectroscopy (EPR). Indeed, five zinc oxide (ZnO) micro- and nano-structures, i.e. Desert Roses (DRs), Multipods (MPs), Microwires (MWs), Nanoparticles (NPs) and Nanowires (NWs), were studied for the Rhodamine B (RhB) sonodegradation under ultrasonic irradiation. The DRs microparticles demonstrated the best sonocatalytic performance (100% degradation of RhB in 180 min) and the highest OH· radicals generation under ultrasonic irradiation. Strikingly, the coupling of ultrasound and sun-light irradiation in a sonophotodegradation approach led to 100% degradation efficiency, i.e. color reduction, of RhB in just 10 min, revealing a great positive synergy between the photocatalytic and sonocatalytic mechanisms. The RhB sonophotocatalytic degradation was also evaluated at different initial dye concentrations and with the presence of anions in solution. It was demonstrated a good stability over repeated cycles of dye treatment, which probe the applicability of this technique with industrial effluents. In conclusion, sonophotocatalytic degradation synergizing sunlight and ultrasound in the presence of DRs microparticles shows a great potential and a starting point to investigate further the efficient treatment of organic dyes in wastewater.

Design, Fabrication, and In Vitro Evaluation of Nanoceria-Loaded Nanostructured Lipid Carriers for the Treatment of Neurological Diseases.

Neurodegenerative diseases comprise a large group of disorders characterized by a dramatic synaptic connections loss, occurring as a result of neurodegeneration, which is closely related to the overproduction of reactive oxygen and nitrogen species. Currently, the treatment of neurodegenerative diseases has been limited mainly because of the inability of the synthesized delivery systems to cross the blood-brain barrier and to successfully deliver their therapeutic cargo to the diseased tissue. Taking into consideration the aforementioned limitations, we designed a lipid-based nanotherapeutic vector composed of biomimetic lipids and CeO2 nanoparticles (nanoceria, NC). NC have shown to be a promising tool for the treatment of several pathological conditions ranging from cancer to neurological diseases, mainly because of their antioxidant properties, while lipid-based structures have been shown to have an inherent ability to cross the blood-brain barrier. The lipid-based nanotherapeutics were successfully fabricated using a combination of ultrasonication and high-pressure homogenization techniques, and they were fully characterized morphologically and physicochemically. Their antioxidant ability was demonstrated using electron paramagnetic resonance spectroscopy and antioxidant assays. These innovative nanotherapeutics demonstrated a higher colloidal stability with respect to free NC, preserving at the same time their antioxidant properties. Finally, the ability of the lipid carriers to cross a model of the blood-brain barrier and to be internalized by neurons, acting both as neuroprotective and pro-neurogenic agents, was demonstrated using single- and triple-culture systems.

The Synergistic Effect of Nanocrystals Combined With Ultrasound in the Generation of Reactive Oxygen Species for Biomedical Applications.

Reactive oxygen species (ROS) effects on living cells and tissues is multifaceted and their level or dose can considerably affect cell proliferation and viability. It is therefore necessary understand their role also designing ways able to regulate their amount inside cells, i.e., using engineered nanomaterials with either antioxidant properties or, for cancer therapy applications, capable to induce oxidative stress and cell death, through tunable ROS production. In this paper, we report on the use of single-crystalline zinc oxide (ZnO) round-shaped nanoparticles, yet ZnO nanocrystals (NCs) functionalized with amino-propyl groups (ZnO-NH2 NCs), combined with pulsed ultrasound (US). We show the synergistic effects produced by NC-assisted US which are able to produce different amount of ROS, as a result of inertial cavitation under the pulsed US exposure. Using Passive Cavitation Detection (PCD) and Electron Paramagnetic Resonance (EPR) spectroscopy, we systematically study which are the key parameters, monitoring, and influencing the amount of generated ROS measuring their concentration in water media and comparing all the results with pure water batches. We thus propose a ROS generation mechanism based on the selective application of US to the ZnO nanocrystals in water solutions. Ultrasound B-mode imaging is also applied, proving in respect to pure water, the enhanced ecographic signal generation of the aqueous solution containing ZnO-NH2 NCs when exposed to pulsed ultrasound. Furthermore, to evaluate the applicability of ZnO-NH2 NCs in the biomedical field, the ROS generation is studied by interposing different tissue mimicking materials, like phantoms and ex vivo tissues, between the US transducer and the sample well. As a whole, we clearly proof the enhanced capability to produce ROS and to control their amount when using ZnO-NH2 NCs in combination with pulsed ultrasound anticipating their applicability in the fields of biology and health care.

Lipid-Coated Zinc Oxide Nanoparticles as Innovative ROS-Generators for Photodynamic Therapy in Cancer Cells.

In the present paper, we use zinc oxide nanoparticles under the excitation of ultraviolet (UV) light for the generation of Reactive Oxygen Species (ROS), with the aim of further using these species for fighting cancer cells in vitro. Owing to the difficulties in obtaining highly dispersed nanoparticles (NPs) in biological media, we propose their coating with a double-lipidic layer and we evaluate their colloidal stability in comparison to the pristine zinc oxide NPs. Then, using Electron Paramagnetic Resonance (EPR) coupled with the spin-trapping technique, we demonstrate and characterize the ability of bare and lipid-coated ZnO NPs to generate ROS in water only when remotely actuated via UV light irradiation. Interestingly, our results reveal that the surface chemistry of the NPs greatly influences the type of photo-generated ROS. Finally, we show that lipid-coated ZnO NPs are effectively internalized inside human epithelial carcinoma cells (HeLa) via a lysosomal pathway and that they can generate ROS inside cancer cells, leading to enhanced cell death. The results are promising for the development of ZnO-based therapeutic systems.

Nanoparticle-assisted ultrasound: A special focus on sonodynamic therapy against cancer.

At present, ultrasound radiation is broadly employed in medicine for both diagnostic and therapeutic purposes at various frequencies and intensities. In this review article, we focus on therapeutically-active nanoparticles (NPs) when stimulated by ultrasound. We first introduce the different ultrasound-based therapies with special attention to the techniques involved in the oncological field, then we summarize the different NPs used, ranging from soft materials, like liposomes or micro/nano-bubbles, to metal and metal oxide NPs. We therefore focus on the sonodynamic therapy and on the possible working mechanisms under debate of NPs-assisted sonodynamic treatments. We support the idea that various, complex and synergistics physical-chemical processes take place during acoustic cavitation and NP activation. Different mechanisms are therefore responsible for the final cancer cell death and strongly depends not only on the type and structure of NPs or nanocarriers, but also on the way they interact with the ultrasonic pressure waves. We conclude with a brief overview of the clinical applications of the various ultrasound therapies and the related use of NPs-assisted ultrasound in clinics, showing that this very innovative and promising approach is however still at its infancy in the clinical cancer treatment.