Search Results (251)

Torque Teno Virus (TTV), a common and genetically diverse component of the human virome, is not linked to any known disease but reflects immune status. Its plasma viral load has shown clinical relevance in solid organ transplant recipients, correlating it with immunosuppression when present at high levels. However, the clinical significance of TTV viral load in hematopoietic stem cell transplantation (HSCT) recipients is less understood. This systematic review aims to evaluate whether plasma TTV DNA load directly correlates with the degree of T-cell immune reconstitution after HSCT, supporting its potential role as a biomarker for immune competence. The study protocol was registered in the PROSPERO International Prospective Register of Systematic Reviews (CRD420251116208) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Twenty-one studies were included. The results showed concordant data about TTV kinetics with peak levels reaching approximately between +90 to +120 days after transplantation. Contradictory results have instead been found for the association of TTV load with immune parameters (lymphocyte counts, viral opportunistic infection, and development of acute graft versus host diseases). Even if a low-risk bias assessment was classified in most studies (67%), it was possible to identify important clinical and methodological differences between them, which might account for the different findings observed. Therefore, future larger studies with standardized protocols are needed to assess whether TTV viral load can serve as a reliable tool for guiding clinical decisions in the context of HSCT. Full article

We present an adaptive governance engine for blockchain-enabled Vehicular Ad Hoc Networks (VANETs) that regulates the latency–energy–coherence trade-off under rapid topology changes. The core contribution is an Ideal Information Cycle (an operational abstraction of information injection/validation) and a modular VANET Engine implemented as a real-time control loop in NS-3.35. At runtime, the Engine monitors normalized Shannon entropies—informational entropy S over active transactions and spatial entropy $H_{\mathrm{spatial}}$ over occupancy bins (both on $[0,1]$)—and adapts the consensus mode (latency-feasible PoW versus signature/quorum-based modes such as PoS/FBA) together with rigor parameters via calibrated policy maps. Governance is formulated as a constrained operational objective that trades per-block resource expenditure (radio + cryptography) against a Quality-of-Information (QoI) proxy derived from delay/error tiers, while maintaining timeliness and ledger-coherence pressure. Cryptographic cost is traced through counted operations, $E_{\mathrm{crypto}}=e_h{n}_{\mathrm{hash}}+e_{\mathrm{sig}}{n}_{\mathrm{sig}}$, and coherence is tracked using the LCP-normalized definition $D_{\mathrm{ledger}}\left(t\right)$ computed from the longest common prefix (LCP) length across nodes. We evaluate the framework under urban/highway mobility, scheduled partitions, and bounded adversarial stressors (Sybil identities and Byzantine proposers), using 600 s runs with 30 matched random seeds per configuration and 95% bias-corrected and accelerated (BCa) bootstrap confidence intervals. In high-disorder regimes ($S\gtrsim 0.8$), the Engine reduces total per-block energy (radio + cryptography) by more than 90% relative to a fixed-parameter PoW baseline tuned to the same agreement latency target. A consensus-first triggering policy further lowers agreement latency and improves throughput compared with broadcast-first baselines. In the emphasized urban setting under high mobility ($v=30$ m/s), the Engine keeps agreement/commit latency in the sub-100 ms range while maintaining finality typically within sub-150 ms ranges, bounds orphaning (≤10%), and reduces average ledger divergence below $0.07$ at high spatial disorder. The main evaluation is limited to $N\le 100$ vehicles under full PHY/MAC fidelity. PoW targets are intentionally latency-feasible and are not intended to provide cryptocurrency-grade majority-hash security; operational security assumptions and mode transition safeguards are discussed in the manuscript. Full article

Salmonella, a major global foodborne pathogen, is a leading cause of salmonellosis. Quantitative detection of Salmonella provides a scientific basis for establishing microbiological criteria and conducting risk assessments. The plate count method remains the primary approach for bacterial quantification, whereas the most probable number (MPN) method is commonly used for detecting low levels of bacterial contamination. However, both methods are time-consuming and labor-intensive. Validated digital polymerase chain reaction (dPCR) techniques are emerging as promising alternatives because they enable rapid, absolute quantification with high specificity and sensitivity. Herein, we developed a novel droplet dPCR (ddPCR) assay for identifying and quantifying Salmonella using invA as the target. The assay demonstrated high specificity and sensitivity, with a limit of quantification of 1.1 × 10² colony-forming units/mL in meat samples. Furthermore, the log₁₀ values obtained via ddPCR and plate counting exhibited a strong linear relationship (R² > 0.99). Mathematical modeling of growth kinetics further confirmed a high correlation between plate count and ddPCR measurements (Pearson correlation coefficient: 0.996; calculated bias factor: 0.88). Collectively, these results indicate that ddPCR is a viable alternative to the MPN method and represents a powerful tool for the quantitative risk assessment of food safety. Full article

Accurate and quantitative detection of boar sperm heads is essential for breeding selection and reproductive management. Manual microscopic counting is time-consuming, labor-intensive, and prone to subjective bias, while existing computer-based algorithms often struggle to recognize sperm cells accurately when they overlap or move rapidly in high-magnification microscopic images. This study proposes a lightweight boar sperm detection model, YOLO11_SRP, designed to improve small-object recognition in complex microscopic scenarios. The model integrates a lightweight StarNet backbone, a rectangular self-calibration module for enhanced spatial feature modeling, and an additional low-level detection layer optimized for tiny targets. We evaluated the model on a boar sperm microscopic image dataset and compared it with the standard YOLO11s framework. The results show that YOLO11_SRP achieves an mAP@0.5 of 91.9%, representing a 13.9% improvement over YOLO11s, while simultaneously reducing parameters by 39% and computational cost by 14.1%. These findings demonstrate that YOLO11_SRP provides efficient and accurate sperm detection, supporting the development of efficient and reliable automated sperm analysis pipelines, in which sperm head detection serves as a fundamental preprocessing step. Full article

Background/Objectives: Abaloparatide is an osteoanabolic therapy used in patients at high risk of fracture; however, the breadth of evidence across routes, comparators, and sequential strategies has not yet been comprehensively summarized. This study aimed to evaluate the efficacy and safety of abaloparatide for reducing fractures and improving bone mineral density (BMD) in adults with osteoporosis. Methods: Following PRISMA 2020, we searched PubMed, Embase, CENTRAL, and Web of Science (2016–2024) for randomized controlled trials and comparative real-world studies. Additional meta-analyses and network meta-analyses were included as contextual evidence but not pooled to avoid double-counting. Primary outcomes were vertebral, non-vertebral, and hip fractures; secondary outcomes included percentage change in BMD and safety endpoints. Random-effects models were used; heterogeneity, influence analyses, and prediction intervals were examined. Risk of bias was assessed using RoB 2 and AMSTAR 2. Results: Nine quantitative evidence sources met the criteria. Abaloparatide reduced vertebral fractures (RR 0.13–0.21) and showed moderate reductions in non-vertebral fractures. Lumbar spine BMD increased substantially, while hip and femoral neck gains were smaller and heterogeneous. Hypercalcemia risk was consistently lower compared to teriparatide. Transdermal delivery was less effective, and sequential abaloparatide → antiresorptive therapy further reduced fractures. Serious adverse events were not increased. Conclusions: Abaloparatide provides strong vertebral protection, significant BMD improvement, and shows a favorable calcemic profile, with moderate certainty for non-vertebral effects. Evidence in men and long-term safety remains limited. Full article

Background: Peri-implantitis, a condition characterized by inflammation and progressive bone loss around dental implants, presents a significant challenge in contemporary dentistry. Conventional non-surgical treatments often fail to fully eliminate bacterial biofilms, particularly on complex implant surfaces. Laser therapies have emerged as potential adjuncts due to their antimicrobial and bio-modulatory properties. However, their microbiological effectiveness and suitability for individualized patient treatment planning remain unclear. Objective: Τhis study aims to systematically assess and synthesize the microbiological effects of various laser-assisted non-surgical treatments for peri-implantitis compared to conventional mechanical debridement. Methods: This systematic review and meta-analysis followed PRISMA guidelines and was registered in PROSPERO (CRD420251035354). Randomized controlled trials (RCTs) evaluating microbiological changes following laser-assisted non-surgical treatment of peri-implantitis, with a minimum follow-up of one month, were identified through searches in multiple databases and registries up to February 2025. The ncluded studies used lasers such as diode, Er: YAG, and photodynamic therapy (PDT) either alone or as adjuncts to mechanical debridement. Outcomes of interest included bacterial counts. Risk of bias was assessed using the RoB2 tool, and certainty of evidence was evaluated via GRADE. Quantitative synthesis used random-effects meta-analysis, with standardized mean differences (SMDs) calculated. Results: Eight RCTs involving 266 patients and 335 implants were included in the systematic review. Quantitative synthesis of three pathogens (counts of Fusobacterium nucleatum, P. gingivalis, T. denticola) across three studies displayed no statistically significant differences between laser and control groups at 3 and 6 months (p > 0.05 for all comparisons). When examining individual study findings, PDT, particularly in patients with diabetes or acute abscess, showed short-term reductions in red complex bacteria (e.g., Porphyromonas gingivalis and Treponema denticola). In contrast, diode and Er: YAG lasers demonstrated inconsistent or transient effects. The quality of evidence was rated as very low according to GRADE. Conclusions: Laser-assisted therapies, especially PDT, may provide targeted microbiological benefit in selected patient groups, supporting their adjunctive use within personalized treatment planning rather than as replacements for mechanical debridement, which remains the gold standard. Further high-quality RCTs incorporating well-defined patient risk profiles, such as systemic conditions and behavioral factors, and precision treatment algorithms are needed. Full article

This paper discusses the assimilation of significant wave height (Hs) observations from the China France Oceanography SATellite (CFOSAT) into the Global Deterministic Wave Prediction System developed by Environment and Climate Change Canada. We focus on the quantification of background errors in an effort to address the conventional, simplified, homogeneous assumptions made in previous studies using Optimal Interpolation (OI) to generate Hs analysis. A map of Best Correlation Length, L, is generated to count for the inhomogeneity in the wave field. This map was calculated from pairs of Hs forecasts of two grid points shifted in space and time from which a look-up table is derived and used to infer the spatial extent of correlations within the wave field. The wave spectra are then updated from Hs analysis using a frequency shift scheme. Results reveal significant spatial variance in the distribution of L, with notably high values located in the eastern tropical Pacific Ocean, a pattern that is expected due to the persistent swells dominating in this region. Experiments are conducted with spatially varying correlation lengths and a set correlation length of eight grid points in the analysis step. Forecasts from these analyses are validated independently with the Global Telecommunications System buoys and the Copernicus Marine Environment Monitoring Service (CMEMS) altimetry wave height observations. It is found that the proposed statistical method generally outperforms the conventional method with lower standard deviation and bias for both Hs and peak period forecasts. The conventional method has more drastic corrections on Hs forecasts, but such corrections are not robust, particularly in regions with relatively short spatial correlation length scales. Based on the analysis of the CMEMS comparison, the globally varying correlation length produces a positive increment of the Hs forecast, which is globally associated with forecast error reduction lasting up to 24 h into the forecast. Full article

Background and Objectives: The association between different anesthesia modalities and spermiogram parameters and reproductive outcomes in patients undergoing microscopic subinguinal varicocelectomy (MSV) remains unclear. In this retrospective cohort study, we aimed to compare spermiogram parameters and pregnancy rates between patients receiving general anesthesia (GA) versus spinal anesthesia (SA) for MSV with 2-year follow-up data. Materials and Methods: Male patients aged between 18–50 years, with ASA physical scores between I–III, who underwent unilateral or bilateral primary MSV, were included in the study. To minimize selection bias and balance the baseline characteristics between the GA group and SA group, we employed a propensity score matching approach, matching all 38 SA patients with 380 GA patients selected from a larger pool. Patients with complete 24-month follow-up data were included in the final analysis. The primary outcome of our study was determined as evaluating sperm count changes. Secondary outcomes included other sperm parameters (motility, morphology and semen volume), natural pregnancy rates, perioperative complications and recovery parameters. Results: The final analysis included 418 patients who met all inclusion criteria and completed the follow-up period. The study population comprised 380 patients in the GA group and 38 in the SA group. No significant difference was found between the groups in terms of sperm count. Greater improvement in sperm motility was observed in the SA group starting from the third month onwards (p = 0.027). Natural pregnancy was achieved in 16/38 (42.1%) of SA patients versus 125/380 (32.9%) of GA patients (p = 0.031). In addition, better results were obtained in terms of recovery parameters in the SA group. Other results were comparable between the groups. Conclusions: Spinal anesthesia for MSV was associated with greater improvement in sperm motility and higher natural pregnancy rates compared to general anesthesia, despite comparable sperm count improvements. These associations warrant further investigation in prospective randomized trials. Full article

Background: Autism spectrum disorder (ASD) is associated with distinct visual attention patterns that provide insight into underlying social-cognitive mechanisms. Methods: This systematic review (PROSPERO: CRD42023429316), conducted per PRISMA guidelines, synthesizes evidence from 14 peer-reviewed studies using eye-tracking to compare oculomotor strategies in autistic children and typically developing (TD) controls. A comprehensive literature search was conducted in PubMed, Web of Science, and Science Direct up to March 2025. Study inclusion criteria focused on ASD versus TD group comparisons in individuals under 18 years, with key metrics, fixation duration and count, spatial distribution, saccadic parameters systematically extracted. Risk of bias was assessed using the QUADAS-2 tool, revealing high heterogeneity in both index tests and patient selection. Results: The results indicate that autistic children exhibit reduced fixation on socially salient stimuli, atypical saccadic behavior, and more variable spatial exploration compared to controls. Conclusions: These oculomotor differences suggest altered mechanisms of social attention and information processing in ASD. Findings suggest that eye-tracking can contribute valuable information about heterogeneous gaze profiles in ASD, providing preliminary insight that may inform future studies to develop more sensitive diagnostic tools. This review highlights visual attention patterns as promising indicators of neurocognitive functioning in ASD. Full article

Objectives: This systematic review and meta-analysis aimed to evaluate microbial adhesion and biofilm formation on additively manufactured composite-based orthodontic clear aligners compared with thermoformed aligners and other conventional polymeric materials. The influence of material composition, surface roughness, post-processing parameters, and cleaning protocols on microbial colonization was also assessed. Methods: A comprehensive search of PubMed, EMBASE, Scopus, Web of Science, and the Cochrane Library was conducted up to September 2025. Only in vitro studies investigating microbial adhesion, biofilm biomass, or microbiome changes on three-dimensional (3D)-printed aligner composites were included. Primary outcomes consisted of colony-forming units (CFU), optical density (OD) from crystal violet assays, viable microbial counts, and surface roughness. Risk of bias was assessed using the RoBDEMAT tool. Data were narratively synthesized, and a random-effects meta-analysis was performed for comparable datasets. Results: Five studies fulfilled the inclusion criteria, of which two in vitro studies were eligible for meta-analysis. Microbial adhesion and biofilm accumulation were influenced by the manufacturing technique, composite resin formulation, and surface characteristics. Certain additively manufactured aligners exhibited smoother surfaces and reduced bacterial adhesion compared with thermoformed controls, whereas others with increased surface roughness showed higher biofilm accumulation. Incorporating bioactive additives such as chitosan nanoparticles reduced Streptococcus mutans biofilm formation without compromising material properties. The meta-analysis, based on two in vitro studies, demonstrated higher OD values for bacterial biofilm on 3D-printed aligners compared with thermoformed aligners, indicating increased biofilm biomass (p < 0.05), but not necessarily viable bacterial load. Conclusions: Microbial adhesion and biofilm formation on 3D-printed composite clear aligners are governed by resin composition, additive manufacturing parameters, post-curing processes, and surface finishing. Although certain 3D-printed materials display antibacterial potential, the limited number of studies restricts the generalizability of these findings. Clinical Significance: Optimizing composite formulations for 3D printing, alongside careful post-curing and surface finishing, may help reduce microbial colonization. Further research is required before translating these findings into definitive clinical recommendations for clear aligner therapy. Full article

This study investigates whether a large language model (LLM) can perform governance-style mediation among multiple stakeholders when preferences are expressed only in categorical natural language. Building on prior conceptual work proposing an advisory governance layer for AI systems, we designed a controlled experiment comparing a language-based mediator with a numerical baseline (Borda count) across 1024 synthetic stakeholder scenarios, each executed ten times (10,240 paired decisions). Results show only 31% agreement with Borda, revealing distinct decision logic that produces equity-biased outcomes (68% improved fairness, ~25% Gini reduction, 38% higher minimum utility) at the cost of efficiency (14–20% lower mean utility). Stability analysis identified three reliability zones—stable (39%), middle (28%), and knife-edge (33%)—enabling risk-proportionate oversight. Qualitative analysis revealed that equity bias emerges from opaque pattern-matching followed by post hoc rationalization rather than systematic application of governance principles, with frequent semantic-grounding failures even in stable cases. These findings demonstrate that language-based mediation diverges fundamentally from numerical aggregation, suitable for advisory deliberation but requiring human oversight for value verification and factual accuracy. Full article

Background: Acute kidney injury (AKI) is increasingly recognised in children with acute COVID-19 and multisystem inflammatory syndrome in children (MIS-C), yet the long-term renal consequences in younger paediatric populations remain unclear. Most studies focus on acute illness or mixed-age cohorts, with limited data specific to children aged 0–12 years. Objectives: This study aimed to systematically identify, evaluate, and synthesise evidence on post-acute (≥30 days) and long-term (≥90 days) kidney outcomes following SARS-CoV-2 infection or MIS-C in children aged 0–12 years, including chronic kidney disease (CKD), eGFR decline, proteinuria, haematuria, hypertension, and need for kidney replacement therapy. Methods: We searched MEDLINE, Embase, CINAHL, and PubMed (December 2019–30 November 2025), following PRISMA 2020 guidelines and a registered PROSPERO protocol (CRD420251241949). Observational studies reporting kidney outcomes ≥30 days post-infection in children aged 0–12 years were included. Risk of bias was assessed using the Newcastle–Ottawa Scale or ROBINS-I. Owing to heterogeneity and absence of ≥3 comparable datasets, a narrative synthesis was performed. Results: Seven studies met inclusion criteria (five MIS-C cohorts, two acute COVID-19 cohorts). Only a subset provided extractable data specific to children aged 0–12 years. Follow-up ranged from 30 days to 12 months; four studies reported outcomes ≥ 180 days. Across all studies, no incident CKD, sustained eGFR decline, or kidney replacement therapy were reported among children completing long-term follow-up; however, most long-term outcome data were derived from MIS-C cohorts with median ages around 8–11 years that included some adolescents, rather than exclusively children aged 0–12 years. One MIS-C study reported long-term hypertension in 14% of children. A cross-sectional Italian cohort of mild COVID-19 demonstrated hyperfiltration, proteinuria, and microhaematuria at ~3 months, though chronicity could not be assessed due to absence of baseline values. A large US EHR-based cohort identified increased CKD risk after COVID-19 in the broader < 21-year population; however, 0–12-year-specific event counts were not reported, preventing quantitative synthesis for young children. Conclusions: Evidence on long-term kidney outcomes after SARS-CoV-2 infection in children aged 0–12 years remains limited, and only a small subset of studies provided extractable, age-specific data. On the other hand, MIS-C cohorts generally show favourable renal recovery, small sample sizes, lack of control groups, and short follow-up restrict confidence in these findings. Large paediatric EHR studies suggest potential long-term renal risk in broader paediatric populations, highlighting the need for age-stratified, prospective cohorts with serial eGFR, urine studies, and blood pressure assessments. Until definitive evidence emerges, structured renal follow-up may be warranted for children with AKI or MIS-C during COVID-19. Full article

Background/Objectives: Emergency arterial embolization is a life-saving procedure typically performed via femoral access. This study evaluated the safety and efficacy of immediate femoral sheath removal following emergency embolization. Methods: A retrospective cohort study was conducted at a Level 1 trauma center (January 2022–May 2025). Adult patients undergoing emergency embolization with immediate sheath removal were included. Endpoints were reintervention (repeat embolization within 7 days) and access site complications. Multivariate logistic regression identified independent predictors of outcomes. Results: A total of 322 emergency embolization procedures in 299 patients were included. The most common indication was gastrointestinal hemorrhage (45.7%). Vascular closure devices (VCDs) were used in 92.5% of cases. The re-intervention rate was 4.0% (13/322). The overall access site complication rate was 6.2% (20/322), with a major complication rate of 0.9% (3/322). On univariate analysis, pre-procedural platelet level ≤ 80 × 10⁹/L (p = 0.034) and INR > 1.5 (p = 0.034) were significantly associated with an increased risk of complications. On multivariate analysis, pre-procedural platelets ≤ 80 × 10⁹/L was the strongest independent predictor of access site complications (OR 7.28, 95% CI 1.51–35.12; p = 0.013). Choice of vascular closure device was an independent predictor for both reintervention and complications (p < 0.05), likely reflecting bias. Conclusions: Immediate femoral sheath removal following emergency embolization is safe for most patients. However, thrombocytopenia is a significant risk factor for access site complications. A risk-stratified approach with consideration for delayed sheath removal is warranted for patients with platelet counts ≤ 80 × 10⁹/L. Full article

Parasitic infection is a major cause of infertility in small ruminants. This study aimed to assess the association between bioelectrical impedance analysis (BIA) measurements, testicular morphometrics, and sperm quality in parasitized goats. Thirty-eight intact mature Spanish bucks were allowed to graze on a naturally parasitically infected pasture for 3 months. Nineteen bucks were dewormed regularly (healthy group), while the other 19 bucks did not receive any anthelmintics (parasitized group). Fecal and blood samples were collected weekly to assess fecal egg count (FEC) and packed cell volume (PCV), respectively. Based on the size and morphology of the parasite eggs, they were presumptively identified as Haemonchus contortus. At the end of the grazing period, bucks were slaughtered, and testicles and epididymis were collected for analysis. In addition, BIA was applied to each testicle to measure series resistance (Rs) and reactance (Xc). Epididymal spermatozoa were retrieved and evaluated for motility, viability, morphology, and membrane and acrosome integrities. Data was analyzed using the Mann–Whitney U and Pearson Correlation Coefficient tests. The results showed that Rs (169.41 ± 1.76 Ω vs. 235.21 ± 20.21 Ω), Xc (37.55 ± 0.48 Ω vs. 52.08 ± 4.68 Ω), testicular and epididymis weights and lengths, sperm motility, viability, and membrane and acrosome integrities were lower (p < 0.0001) in parasitized than in healthy goats. Strong correlations (p < 0.001) were observed between Rs, sperm viability (r = 0.20), membrane integrity (r = 0.15), and acrosome integrity (r = 0.14), as well as between Xc and the same sperm parameters (r = 0.21, 0.18, and 0.16, respectively). In conclusion, our findings demonstrate that parasitic infection is associated with testicular health and subsequent epididymal sperm quality of goats. BIA can be utilized as an efficient tool to predict the impact of parasitic infection on testicular function in goats. Full article

Background: Retrieval-augmented generation (RAG) aims to reduce hallucinations and outdated knowledge by grounding LLM outputs in retrieved evidence, but empirical results are scattered across tasks, systems, and metrics, limiting cumulative insight. Objective: We aimed to synthesise empirical evidence on RAG effectiveness versus parametric-only baselines, map datasets/architectures/evaluation practices, and surface limitations and research gaps. Methods: This systematic review was conducted and reported in accordance with PRISMA 2020. We searched the ACM Digital Library, IEEE Xplore, Scopus, ScienceDirect, and DBLP; all sources were last searched on 13 May 2025. This included studies from January 2020–May 2025 that addressed RAG or similar retrieval-supported systems producing text output, met citation thresholds (≥15 for 2025; ≥30 for 2024 or earlier), and offered original contributions; excluded non-English items, irrelevant works, duplicates, and records without accessible full text. Bias was appraised with a brief checklist; screening used one reviewer with an independent check and discussion. LLM suggestions were advisory only; 2025 citation thresholds were adjusted to limit citation-lag. We used a descriptive approach to synthesise the results, organising studies by themes aligned to RQ1–RQ4 and reporting summary counts/frequencies; no meta-analysis was undertaken due to heterogeneity of designs and metrics. Results: We included 128 studies spanning knowledge-intensive tasks (35/128; 27.3%), open-domain QA (20/128; 15.6%), software engineering (13/128; 10.2%), and medical domains (11/128; 8.6%). Methods have shifted from DPR + seq2seq baselines to modular, policy-driven RAG with hybrid/structure-aware retrieval, uncertainty-triggered loops, memory, and emerging multimodality. Evaluation remains overlap-heavy (EM/$F_1$), with increasing use of retrieval diagnostics (e.g., Recall@k, MRR@k), human judgements, and LLM-as-judge protocols. Efficiency and security (poisoning, leakage, jailbreaks) are growing concerns. Discussion: Evidence supports a shift to modular, policy-driven RAG, combining hybrid/structure-aware retrieval, uncertainty-aware control, memory, and multimodality, to improve grounding and efficiency. To advance from prototypes to dependable systems, we recommend: (i) holistic benchmarks pairing quality with cost/latency and safety, (ii) budget-aware retrieval/tool-use policies, and (iii) provenance-aware pipelines that expose uncertainty and deliver traceable evidence. We note the evidence base may be affected by citation-lag from the inclusion thresholds and by English-only, five-library coverage. Funding: Advanced Research and Engineering Centre. Registration: Not registered. Full article