Search Results (1,963)

Poor aqueous solubility and consequently low bioavailability of various NSAIDs (non-steroidal anti-inflammatory drugs) usually result in high and multiple dosing with potentially serious side effects. Therefore, systems for the effective transport of NSAIDs through the GIT (gastrointestinal tract), ensuring enhanced bioavailability, remain in high demand. In the present work, we studied chitosan (CS) hydrogel lyophilizates as carrier systems for a model NSAID, namely ibuprofen (IBU). The CS-IBU lyophilizates were prepared from homogeneous or heterogeneous CS-IBU hydrogels to assess their influence on the resulting lyophilizate microstructure and IBU dissolution profiles. To gain a complex view of the CS-IBU behavior and its practical consequences, dissolution profiles of free IBU (reference) and CS-associated IBU (CS-IBU) were examined and compared to each other at variable pH (1.2 and 6.5) in two separate dissolution systems and in one discontinuous dissolution system mimicking GIT conditions. The results of dissolution experiments were supported by kinetic model data. This study demonstrated that the dissolution of IBU from the CS-IBU lyophilizates is affected by two main pH-dependent competitive effects; i.e., dissolved CS acts as an IBU solubilizer and the undissolved CS matrix serves as an IBU trap, which could be used in the rational design of innovative stimuli (pH)-responsive oral dosage forms of IBU. Full article

Data in sickle cell disease (SCD) shows that 30–50% of adults have chronic pain. Chronic pain is a life-shaping condition that is often inadequately managed with chronic opioid therapy (COT). Buprenorphine–naloxone (bup-nal) is an alternative (to COT), yet patient perspectives on its use remain limited. This is a phenomenological qualitative study exploring how adults with SCD experience chronic pain and consider treatment options including COT and bup-nal. Semi-structured interviews were conducted with 26 adults with SCD and chronic pain who were offered pain management with bup-nal, including individuals who declined or discontinued treatment. Participants described pain as a constant, embodied experience around which they structured their daily functioning, relationships, well-being, and treatment. Those receiving bup-nal consistently reported improved daily functioning, greater independence, enhanced mood, and reduced healthcare utilization. In contrast, participants not receiving bup-nal described ongoing pain interference, reduced daily functioning, and continued reliance on acute care services. Importantly, participants emphasized functional improvement and stability, rather than pain elimination, as the most meaningful outcomes. These findings suggest that buprenorphine–naloxone may shift pain from a dominant, disabling experience to a more manageable condition that improves participation in work, relationships, and daily activities. To optimize management of chronic pain in SCD, it is necessary to align treatment with patient priorities. Full article

The volumetric mass transfer coefficient ($k_{L}a$) governs the design, operation, and scale-up of aerobic bioprocesses, yet its dependence on reactor geometry, impeller design, operating conditions, and fluid properties limits prediction by empirical correlations. Machine learning (ML) improves accuracy but faces two barriers in bioprocess practice: selecting the best model among many candidates requires expertise, and small, highly multicollinear data make models chosen based on test error alone prone to overfitting. Using a browser-based, no-code platform, we trained 14 regression algorithms under an identical pipeline on a published $k_{L}a$ dataset, and introduced a composite objective, the generalization-penalized error (GPE), which is the test RMSE plus the absolute train–test RMSE gap. Minimizing GPE rather than test RMSE expanded the top statistically equivalent group to include not only boosting ensembles but also simpler, interpretable models, indicating that black-box models hold no clear advantage once train–test consistency is assessed. Sensitivity analysis showed that tree models produce discontinuous responses, whereas algebraic learning via elastic net (ALVEN) yields smooth surfaces. Shapley additive explanations (SHAP) and an ontology graph, interpreted by a retrieval-augmented language-model agent, identified rotational speed and gas flow rate as dominant, reproducing the established mass transfer mechanism. The framework offers a reproducible, interpretable, expertise-light route to bioprocess model selection. Full article

Background: Penetrating keratoplasty (PK) is a technically demanding corneal transplant procedure frequently performed in elderly patients with substantial systemic comorbidities. In this population, an anesthetic strategy that ensures hemodynamic stability, cooperative sedation, adequate analgesia, and preserved spontaneous ventilation is highly desirable. Dexmedetomidine, a highly selective alpha2-adrenergic agonist, provides “cooperative” sedation with minimal risk of respiratory depression and additional sympatholytic benefits. Methods: This single-center retrospective observational case series included 50 consecutive patients (American Society of Anesthesiologists [ASA] II–III, age 50–90 years) undergoing PK under sub-Tenon block combined with continuous dexmedetomidine infusion. Dexmedetomidine was administered without a loading bolus at 0.7 mcg/kg/h for 10–15 min, then reduced to 0.5 mcg/kg/h, targeting a Ramsay Sedation Scale (RSS) score of 2–3. The sub-Tenon block was performed using a mixture of levobupivacaine 0.5% and lidocaine 2% (3–5 mL). Heart rate (HR), mean arterial pressure (MAP), oxygen saturation (SpO₂) and RSS were recorded in nine predefined perioperative phases. Data were analyzed descriptively. Results: The mean age was 72 ± 9 years; 52% of patients were ASA III. Hypertension was present in all patients; 30% had cardiovascular disease, 28% diabetes mellitus type II, and 30% chronic obstructive pulmonary disease. Progressive, controlled bradycardia was observed (mean HR decreased from 76 to 57 beats/min during graft transplantation), while MAP gradually decreased from hypertensive baseline values (150–160 mmHg) to an optimal intraoperative range of 115–130 mmHg, without episodes of clinically significant hypotension. SpO₂ remained stable at 98–99% throughout all phases, with no episodes of desaturation or need for airway intervention or supplemental oxygen. Target sedation (RSS 2–3) was achieved in all patients (median RSS 3), with preserved spontaneous breathing and cooperation. Sub-Tenon block-related bulging occurred in 6% of cases. No episodes of clinically significant bradycardia, malignant arrhythmia, respiratory compromise, or need to discontinue dexmedetomidine were recorded. No opioids or non-steroidal analgesics were required intraoperatively or in the early postoperative period. Conclusions: The combination of sub-Tenon block and continuous dexmedetomidine sedation without a loading bolus represents a hemodynamically stable and respiratory-safe anesthetic strategy for PK in elderly, high-risk patients. These preliminary, hypothesis-generating findings suggest that the protocol provides stable surgical conditions and a favorable safety profile, justifying future prospective randomized controlled trials to establish its comparative efficacy against general anesthesia or standard sedative regimens. Full article

Background/Objectives: To analyse postoperative pain trajectories after anterior cruciate ligament (ACL) reconstruction using data from a digital rehabilitation system, and to determine (i) whether combined ligament injuries are perceived as more painful than isolated ACL tears, (ii) which patient characteristics are associated with clinically relevant pain (visual analogue scale [VAS] > 5), and (iii) whether higher early pain is associated with later discontinuation of digital monitoring. Methods: This retrospective cohort study used routine data recorded by a validated sensor-based home rehabilitation system in patients after ACL reconstruction. This approach has previously been used to analyse functional recovery trajectories. All patients with ACL-related indications who performed at least one postoperative test were included and classified into four groups: isolated ACL rupture, ACL + meniscus, ACL + collateral ligament, and ACL + collateral ligament + meniscus. Pain during exercises and tests was recorded on a 0–10 VAS. High pain was defined as VAS > 5. Group comparison between indication types, anthropometric and activity-related variables and the proportion of high-pain events were performed using chi-square tests. Early pain (first postoperative month) was analysed in relation to the presence of later tests (≥3 months) to explore associations with discontinuation of digital monitoring. Results: Combined ligament injuries showed a significantly higher proportion of high-pain events during rehabilitation compared with isolated ACL ruptures (5.8% vs. 2.4%, overall p < 0.001). In particular, combined ligament injuries with ACL + collateral ligament rupture were associated with a greater share of VAS > 5 ratings in the early rehabilitation phases. No relevant association was observed between sex or BMI category and the occurrence of high pain, while age group showed an overall association without a consistent directional pattern. Sport activity level showed a strong relationship with high pain (p < 0.001). Early pain demonstrated a small but statistically significant negative correlation with later test participation (r = −0.15, approximately 2% of variance, p = 0.0076); however, this association was attenuated and no longer statistically significant when analysed using mixed-effects models accounting for within-patient clustering, indicating that patients with higher early pain tended to discontinue digital monitoring. Conclusions: Digital routine data after ACL reconstruction suggest that (i) combined ACL–collateral ligament injuries are perceived as more painful than isolated ACL tears, (ii) high postoperative pain is more closely related to activity level and injury pattern than to sex or BMI, while age group shows an overall but non-directional association, and (iii) higher early pain shows a weak bivariate correlation with digital disengagement that was not confirmed in mixed-effects models. Pain is therefore an exploratory marker warranting further investigation, rather than a confirmed independent predictor of adherence in app-based rehabilitation. Full article

The synergistic transition toward digital transformation and green development has been widely regarded as a core pathway to achieving sustainable development in knowledge production. Using balanced panel data from 39 economies covering 2018–2024, this study employed a two-way fixed-effects model to examine the associations of the energy efficiency of digital infrastructure and the energy structure with carbon intensity (CI). The findings showed that: (1) Reductions in Power Usage Effectiveness (PUE) values were significantly associated with higher macro-level CI (coefficient = −2.1564, p < 0.05), which is consistent with the possibility of a rebound effect in the digital sector. Further, time-series discontinuity tests further suggested that the surge in AI computing power, especially in 2023–2024, may have coincided with a structural shift in this relationship (Chow test, p < 0.05). (2) A Panel Threshold Regression (PTR) identified an optimal renewable energy threshold at 59.82%. Crucially, the carbon rebound effect remained highly significant across both high and low green power regimes, demonstrating that supply-side energy transition alone cannot fully absorb the exponential carbon footprint of digital expansion. Furthermore, Instrumental Variable (IV-2SLS) and Placebo Break Tests confirmed the strict validity of these findings. (3) The emission-reduction benefits related to digital knowledge spillovers appeared to be subject to time lags and a possible energy lock in effect, while current environmental policies and carbon pricing mechanisms appear to impose insufficient constraints. This study provides a crucial quantitative framework for monitoring and evaluating the environmental sustainability of the ICT sector. By highlighting the limitations of pure supply-side greening and the necessity of absolute carbon caps, our findings offer integrated policy approaches to align the exponential growth of Generative AI with global sustainable development goals. Full article

Background: Critical illness leads to profound metabolic, neuroendocrine and immune disorders that affect the prognosis of patients treated in intensive care units (ICUs). The ketogenic diet, a high-fat and low-carbohydrate eating model, is gaining increasing importance as a potential metabolic intervention in the ICU. Preliminary data suggest that the ketogenic diet (KD) may support the control of seizures in a super-refractive epileptic state (SRSE), stabilize glycemia, reduce insulin demand, and modulate the immune response in sepsis. The aim of this review was to present a synthetic presentation of the current state of knowledge regarding use of the KD in intensive care patients. Methods: The review was carried out in accordance with the guidelines of the Joanna Briggs Institute and PRISMA-ScR. PubMed, Scopus, EBSCO, Web of Science, Google Scholar and Cochrane Library databases were searched (10–19 April 2026) using the Population–Concept–Context model. Full-text observational studies, randomized trials and reviews of the use of KDs in ICU patients were included. Data extraction was performed independently by two reviewers. Results: Of the 42 publications identified, seven studies were included in the analysis. The KD was feasible and safe in both critically ill adults and children. In SRSE, most patients achieved stable ketosis within a few days, which often allowed for reduction or discontinuation of anesthetics. In sepsis, the KD led to glycemic stabilization, reduced insulin demand and reduced immune deregulation; in one study, “after day 4, none of the patients in the KD group required insulin treatment.” The KD also showed beneficial effects on cellular bioenergetics and mitochondrial function. The safety profile was acceptable and adverse reactions were manageable with appropriate monitoring. Conclusions: The KD represents a promising, non-pharmacological metabolic intervention in intensive care, particularly in the treatment of SRSE and in the stabilization of glucose metabolism in sepsis and other critical conditions. Despite the growing number of positive clinical observations, the available evidence remains limited due to small samples, heterogeneous protocols, and a lack of randomized trials. Further, well-designed prospective studies are needed to determine optimal KD implementation protocols and identify the patient populations that benefit most. Full article

Background: Alcohol use disorder (AUD) is a grave mental health condition that can result in significant health and social consequences. The medications Naltrexone and Nalmefene are indicated for the treatment of AUD, with Naltrexone having received the most extensive research attention. Methods: The majority of papers assessing universal measures of alcohol consumption employed two primary metrics: total alcohol consumption (TAC) and the number of days per month where individuals engaged in heavy drinking (HDD). Indicators pertaining to the maintenance of complete abstinence were excluded due to the absence of sufficient data. The safety of both substances was also assessed, as were the frequency of side effects and independent patient dropout. The study also incorporated practical factors of the therapy, such as the route of administration, dosage regimen, and the drug’s patient convenience, which can have a significant impact on adherence to therapy. Results: Nalmefene, administered in an “as needed” regimen, demonstrated statistically significant activity in reducing HDD and total alcohol consumption (TAC) among patients with AUD, particularly those with elevated World Health Organization (WHO) DRL risk. Preliminary findings from the ESENSE1 (Efficacy of Nalmefene in Alcohol Dependence; the first phase III study), ESENSE 2 (Efficacy of Nalmefene in Alcohol Dependence, the second phase III study), and SENSE (the final phase III long term-safety and cost-effectiveness study) studies indicate a substantial decrease in HDD and TAC following the initial month of treatment. These effects persist throughout the subsequent follow-up period. Several Japanese studies have corroborated the effectiveness of Nalmefene, demonstrating its efficacy across both short-term and long-term applications. Furthermore, these studies have substantiated its safety profile, indicating that there is no inherent risk of addiction or the emergence of withdrawal symptoms. The mild nature of adverse events (most commonly nausea and dizziness) led to a relatively low discontinuation rate of Nalmefene treatment. A subsequent study, employing a recognized methodology, corroborated the efficacy of psychosocial support in enhancing treatment outcomes. Meta-analyses demonstrate that Naltrexone exhibits comparable efficacy in reducing the frequency and severity of alcohol consumption. In select populations, the injectable form (LAI) of this pharmaceutical agent facilitates less frequent dosing, which is advantageous for the treatment process. A comparison of Nalmefene and Naltrexone reveals that the latter does not demonstrate a significant impact on the likelihood of individuals returning to heavy alcohol consumption. Conclusions: In the treatment of AUD, both naltrexone and nalmefene have been shown to yield positive outcomes, particularly in terms of reducing the HDD and TAC. According to the World Health Organization (WHO) classification, Nalmefene is indicated for individuals with a high risk of developing serious conditions. It has been demonstrated to produce rapid and sustained results while exhibiting a favorable safety profile, characterized by the absence of significant adverse effects. Naltrexone is a medication that has proven to be effective. LAI may have a positive impact on the efficacy of treatment. Full article

Purpose: Adjuvant radiotherapy (RT) for ductal carcinoma in situ (DCIS) remains a widely discussed topic. While it has been shown to significantly reduce recurrence rates, it does not improve overall survival (OS) and may therefore lead to overtreatment in some cases. This study aimed to assess patient acceptance, quality of life (QoL), and decisional regret (DR) in relation to late adverse effects. Methods: Clinical data from patients who underwent RT for DCIS between 2008 and 2020 were retrospectively collected, while follow-up questionnaires evaluated QoL and DR. High-risk DCIS was defined as tumor size ≥ 2.5 cm, poor differentiation (G3), age ≤ 50, and symptomatic diagnosis. Low-risk disease was defined as tumor size < 2 cm, G1/G2, age > 50, and mammographic diagnosis. Patients who did not meet all criteria for either category were classified as intermediate risk, therefore reflecting everyday work and clinical practice. Questionnaire data were analyzed using both scale-level and item-level approaches and Ipsilateral breast recurrence-free survival (IBRFS) rates were calculated using the Kaplan–Meier method. Results: Among 307 cases offered RT, only 24 declined, indicating high treatment acceptance; an additional 20 cases lacked sufficient documentation, did not receive RT at our center despite acceptance, or discontinued RT prematurely. During the study period, late toxicities occurred at acceptable frequencies: fibrosis in 6.1%, hyperpigmentation in 6.8%, telangiectasia in 8%, and edema in 1.5% of cases with excellent control rates of 90.0%. The response rate to the EORTC QLQ-C30 and Ottawa Decisional Regret Scale (DRS) was 35.4%, of whom 70 cases belonged to the intermediate-risk group. These respondents reported high QoL and low levels of DR. Conclusions: RT for intermediate grade DCIS is accepted and tolerated well, and neither late adverse effects nor QoL outcomes were associated with DR. These findings support recommending RT in this borderline-indication group. Full article

Exchange-rate forecasting requires modelling non-stationary dynamics, heavy-tailed shocks, and complex temporal dependencies. However, forecasting performance in emerging-market currencies is fundamentally constrained by intrinsic dynamical instability, while most existing approaches are evaluated primarily through predictive accuracy rather than the predictability limits of the underlying system. This paper develops a predictability-aware framework that combines nonlinear dynamical diagnostics with a Lévy-driven neural stochastic differential equation model. Drift and diffusion are parameterized by neural networks and driven by $\alpha$-stable Lévy motion, enabling the representation of non-Gaussian fluctuations, abrupt shocks, and regime changes. To learn under discontinuous dynamics, we introduce a structurally constrained training objective based on a strong-form discretization of the underlying SDE. To characterise intrinsic predictability, we employ phase-space reconstruction and maximal Lyapunov exponent estimation. These diagnostics are interpreted as finite-sample measures of trajectory divergence and effective instability in a stochastic system, rather than evidence of low-dimensional deterministic chaos—a distinction motivated by well-documented limitations of chaos testing in financial data. Experiments on multiple West African currency pairs demonstrate competitive short-horizon forecasting performance relative to econometric and neural baselines while providing a principled framework for analysing predictability degradation under heavy-tailed stochastic dynamics. Across currencies and model classes, forecasting accuracy deteriorates beyond horizons comparable to the estimated Lyapunov time, suggesting that forecast degradation reflects intrinsic dynamical instability rather than model-specific limitations. The results support the view that reliable exchange-rate prediction is fundamentally a short-horizon problem and illustrate how stochastic dynamical modelling and predictability diagnostics can be combined to characterise forecasting limits in heavy-tailed financial systems. Full article

Accurately tracking the spatial and temporal variations of water vapor is indispensable for weather forecasting and climate adaptation, yet remains challenging due to the sparse coverage and discontinuity of ground-based observations. Satellite remote sensing, particularly from geostationary satellites like Meteosat Third Generation Imager-1 (MTG-I1), offers continuous, high-resolution data. To the best of our knowledge, MTG-I1 is the first geostationary satellite equipped with a near-infrared (NIR) spectral band specifically designed for detecting water vapor. To address the lack of precipitable water vapor (PWV) data derived from the Flexible Combined Imager (FCI) onboard MTG-I1, a novel semi-empirical (SE) algorithm optimized for PWV retrieval is proposed. Validation against ground-based PWV measurements using an initial test set and a temporally independent test set yielded relative errors of no more than 0.10, indicating stable retrieval performance outside the model-development period. The FCI-derived PWV retrievals were also more accurate than the corresponding MODIS PWV data. Compared to the traditional radiative transfer model (RTM)-based retrieval method, the SE method shows greater adaptability to systematic differences between the observed and RTM-simulated FCI reflectance. After correcting for radiometric degradation, the RTM-based algorithm achieves a 41% reduction in absolute error and a 47% reduction in relative error, bringing its accuracy in line with the SE algorithm. Overall, the proposed SE algorithm demonstrates superior robustness and adaptability, and can provide more reliable remote sensing PWV data to support weather forecasting and climate research. Full article

On 4 February 1975, the Haicheng Ms 7.3 earthquake occurred in the Liaodong Uplift, northeastern China. To investigate its seismogenic structure and deep geological environment, we acquired broadband magnetotelluric data along two intersecting profiles across the epicentral region and performed three-dimensional inversion. Two orthogonal electrical sections were then extracted from the resulting 3D resistivity model to image the crustal structure beneath the Haicheng earthquake area. The model reveals that the northern segment of the Tanlu fault corresponds to a major electrical discontinuity between the Xialiaohe Basin and the Liaodong Uplift, suggesting that it may represent a deep-seated fault zone extending into the lithosphere. Beneath the Liaodong Uplift, a prominent mid-crustal low-resistivity layer is developed, and a synform conductive body is resolved beneath the source region. The Haicheng mainshock and relocated aftershocks are mainly distributed along the interface between this conductive body and the overlying high-resistivity upper crust. In addition, the Haichenghe–Dayanghe fault is imaged as a conductive zone that connects the mid-crustal conductor with shallower crustal levels. These electrical features suggest that deep crustal fluids, possibly related to Pacific Plate subduction and craton destruction, may have migrated upward along fault zones, weakened the seismogenic fault system, and promoted earthquake nucleation. Compared with the volcanic regions of the Jilin–Heilongjiang orogenic belt, where conductive anomalies extend into the upper mantle, the Haicheng region is characterized mainly by intracrustal conductors. This contrast highlights the role of crustal-scale conductive structures in the seismogenic environment of the Haicheng earthquake and provides geophysical constraints for comparing earthquake- and volcano-related deep processes in northeastern China. Full article

The Easterlin paradox and recent distributional reassessments suggest that average effects obscure how subjective disadvantage is generated and reproduced over time. We propose the Social Comparison System (SCS), a framework that represents subjective well-being (SWB) as an internal state and relative income rank as an external conditioning variable within a feedback structure, with three structural properties: threshold activation, state-dependent gain, and rank-conditioned attractor displacement. The properties are recovered through a sample-isolated three-stage framework integrating tree-based machine learning, forest-based heterogeneity estimation, panel-data estimation, and hierarchical Bayesian Markov modeling on a balanced four-wave panel of the China Family Panel Studies (CFPS; 8099 individuals; 32,396 person-wave observations). Stage 1 locates a discrete predictive discontinuity in relative income rank between rank 2 and rank 3 (SHAP jump = 0.383, permutation p < 0.001). Stage 2 carries this boundary into a disjoint validation panel and recovers a negative rank-by-prior-SWB interaction (β = −0.036) and a 2.30-fold larger conditional effect in low- than in high-prior-SWB strata. Stage 3 recovers a 22.6-percentage-point gap in the rank-conditioned occupancy of the lowest within-wave SWB quartile between low- and high-rank subsystems, which under a first-order Markov approximation corresponds to a long-run stationary gap, robust to alternative state-space discretizations. Throughout this paper, relative income rank is treated as a conditioning variable, and the rank-conditioned patterns are interpreted as associational; the long-run quantities are reported under a first-order dynamical approximation rather than as identified causal or fully validated long-run effects. Persistent subjective disadvantage is therefore characterized by unequal dynamics of activation, amplification, and escape, rather than by unequal resources alone. This reframing provides a methodological template for identifying rank-conditioned feedback structures in social-systems data. Full article

Slope stability in open-pit mining is not a static condition but evolves continuously as excavation progresses and geomechanical conditions change. In this study, an integrated approach combining ground-based radar monitoring, satellite-based InSAR time-series analysis, and numerical stability modeling was applied to evaluate slope behavior in a large-scale open-pit copper mine with complex geological and structural characteristics. Radar data revealed progressive and episodic deformation concentrated in specific slope sectors, while InSAR observations showed that deformation continued at lower rates after the main movement phase, providing a longer-term perspective of slope response. Stability analyses using limit equilibrium and finite element methods indicate that the slope operates close to a limit equilibrium condition, particularly under saturated scenarios where factors of safety approach critical levels and strain localization becomes more pronounced. The results show a clear link between observed deformation patterns and calculated stability conditions, with structural discontinuities and groundwater playing a dominant role in controlling slope behavior. Based on these findings, an integrated workflow is proposed that links monitoring data with stability assessment, enabling the identification of critical zones and supporting the evaluation of slope conditions during ongoing mining operations. This approach contributes to more reliable decision-making and supports safer and more sustainable open-pit mining practices. Full article

Monitoring inland surface water quality is essential for water security, ecosystem conservation, public health, and sustainable water resource management. Although in situ measurements remain indispensable, they are often limited by high costs, restricted spatial coverage, low temporal frequency, and discontinuous monitoring networks. This study presents a systematic literature review, guided by the PRISMA 2020 framework, of empirical studies published between 2021 and 2025 on the integration of artificial intelligence (AI) and remote sensing (RS) for inland surface water quality monitoring. Searches were conducted in the Web of Science database, resulting in a final corpus of 367 peer-reviewed articles. Preliminary bibliometric characterization and qualitative content analysis were performed to identify sensors, platforms, AI paradigms, algorithms, estimated parameters, validation strategies, limitations, challenges, trends, and research gaps. The results show rapid growth in the field, with Sentinel-2 and Landsat-8 as the most recurrent sensors and multispectral data as the dominant spectral source. Machine learning approaches, especially Random Forest, Artificial Neural Networks, XGBoost, and Support Vector Machine, predominated, while deep learning, multi-source integration, hybrid models, and Explainable AI emerged as relevant trends. AI–RS integration shows strong potential to complement conventional monitoring, but persistent challenges remain regarding in situ data dependence, limited external and temporal validation, model transferability, generalization, uncertainty reporting, validation robustness, and interpretability. Full article