Search Results (8,635)

Background/Objectives: We aimed to evaluate the association between early post-treatment eosinophil (Eo) elevation and survival outcomes in metastatic renal cell carcinoma (mRCC) treated with vascular endothelial growth factor receptor tyrosine kinase inhibitors (VEGFR-TKIs). We also assessed whether early post-treatment Eo elevation provided prognostic information beyond the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) model. Methods: This Turkish retrospective multicenter cohort study included 280 patients with mRCC who received first-line VEGFR-TKI monotherapy between 2015 and 2025. Early post-treatment Eo elevation was defined as a post-baseline Eo percentage >5% in the complete blood count obtained within 15 days before the first computed tomography assessment performed for response/progression evaluation. Progression-free survival (PFS) and overall survival (OS) were estimated using Kaplan–Meier methods and compared with log-rank tests. Cox regression models assessed independent prognostic associations. Time-dependent receiver operating characteristic (ROC) curve analyses were performed to evaluate the incremental prognostic value of early post-treatment Eo elevation beyond IMDC risk. Results: The median age was 61 years, and 205 patients (73.2%) were male. Early post-treatment Eo elevation was observed in 88 patients (31.4%). Median PFS was longer in patients with early post-treatment Eo elevation than in those without Eo elevation (15.1 vs. 10.1 months; p < 0.001). Median OS was also longer in the Eo elevation group (70.0 vs. 38.6 months; p < 0.001). In multivariable analysis, early post-treatment Eo elevation remained independently associated with improved PFS (hazard ratio [HR]: 0.60, 95% confidence interval [CI]: 0.44–0.81, p = 0.001) and OS (HR: 0.51, 95% CI: 0.33–0.80, p = 0.004). The combined IMDC plus Eo model showed higher time-dependent area under the curve values than the IMDC model alone. Conclusions: Early post-treatment Eo elevation was independently associated with improved PFS and OS, and the combined model showed greater prognostic discrimination than IMDC risk alone. Given that Eo status was assessed after treatment initiation, residual selection bias cannot be excluded, and prospective validation is required before clinical implementation. Full article

Objective: This systematic review and meta-analysis aimed to evaluate the effects of exercise interventions on cognitive function in diabetes. Methods: We conducted a comprehensive search of the Web of Science, PubMed, Scopus, EMBASE (Ovid), and SPORTDiscus databases from inception to December 2025 (registration number: CRD420251046731). The review included randomized controlled trials focusing on exercise interventions for diabetes and their impact on 4 cognitive domains: executive function, memory, attention, and global cognition. Subgroup and moderator analyses were performed based on exercise variables, including exercise type, session duration, total weekly exercise time, and intervention period. Study quality was assessed using the Cochrane Risk of Bias Tool 2 and the Physical Therapy Evidence Database scale (PEDro scale). Results: After screening 6230 records, 16 studies were included. Exercise interventions produced small-to-medium significant improvements in memory and global cognition in patients with diabetes, whereas the overall pooled effect on executive function was not statistically significant. Subgroup analyses suggested potential trends toward improvement in executive function for aerobic exercise (Hedge’s g = −0.30, p = 0.0039), single sessions ≥ 45 min (Hedge’s g = −0.40, p = 0.0463), and intervention periods ≥ 24 weeks (Hedge’s g = −0.27, p = 0.0383). For memory, aerobic (Hedge’s g = 0.27, p = 0.0131) or multicomponent exercise (Hedge’s g = 0.45, p = 0.0224) with single sessions ≤ 60 min (Hedge’s g = 0.28, p = 0.002), total weekly exercise time ≤ 120 min (Hedge’s g = 0.35, p = 0.0012), and interventions ≥ 12 weeks (Hedge’s g = 0.32, p = 0.0003) were significantly beneficial. Exercise also had a significant positive effect on global cognition (Hedge’s g = 0.49, p < 0.0001) without significant differences across exercise variables. In addition, exercise intervention showed a borderline, non-significant trend toward improved motor performance in patients with diabetes (Hedge’s g = 0.33, p = 0.0506). Conclusions: This comprehensive meta-analysis supports the beneficial effects of exercise on global cognition and memory in patients with diabetes. For memory enhancement, aerobic or multicomponent exercise protocols of ≥12 weeks, ≤120 min of total weekly exercise, and ≤60 min per session are more recommended. Although the overall pooled effect on executive function did not reach statistical significance, exploratory subgroup findings point toward potential domain-specific benefits under targeted exercise parameters (such as aerobic exercise of longer duration). These dose–response patterns should therefore be viewed as exploratory and hypothesis-generating, warranting confirmation in adequately powered, head-to-head trials before specific exercise prescriptions can be recommended. Full article

Artificial intelligence (AI) is transforming gastrointestinal (GI) endoscopy into a more standardized, data-driven, and workflow-integrated field. Advances in computer-assisted detection (CADe), diagnosis (CADx), quality assessment (CAQ), natural language processing (NLP), and multimodal deep learning have expanded AI applications across colonoscopy, upper endoscopy, endoscopic ultrasound (EUS), ERCP, cholangioscopy, and capsule endoscopy. These systems have demonstrated improvements in lesion detection, procedural quality assessment, workflow efficiency, and diagnostic support. However, current evidence remains largely focused on surrogate outcomes rather than patient-centered clinical benefits, while challenges related to generalizability, explainability, regulatory oversight, automation bias, and workflow integration continue to limit widespread adoption. Future progress will depend on prospective real-world validation, diverse datasets, explainable AI frameworks, and careful integration of human–AI interaction into clinical practice. Overall, AI is evolving from a supportive adjunct into an increasingly integrated component of gastrointestinal endoscopy with the potential to improve procedural quality, diagnostic consistency, and clinical efficiency. Full article

Background/Objectives: Orthodontic mini-implants have become an essential source of temporary skeletal anchorage because of their versatility, minimal invasiveness, and reduced dependence on patient compliance. However, their clinical success relies on maintaining adequate primary and secondary stability throughout orthodontic treatment. Photobiomodulation (PBM) with diode lasers has been proposed as a non-invasive adjunctive therapy capable of enhancing bone healing and remodeling, thereby improving mini-implant stability. Despite encouraging findings, the available evidence remains limited and heterogeneous. The objective was to systematically evaluate the effect of diode laser photobiomodulation on the stability of orthodontic mini-implants. Methods: A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The search was conducted in the PubMed, EBSCO, Scopus, and ScienceDirect databases to identify randomized clinical trials published in the last 10 years. Five randomized controlled trials involving 89 participants and 178 orthodontic mini-implants fulfilled the eligibility criteria. Mini-implant stability was assessed using objective methods, including resonance frequency analysis (ISQ) and Periotest values, while one study also evaluated insertion and removal torque. The methodological quality of the included studies was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool, and the certainty of the evidence was evaluated using the GRADE approach. Results: Three of the five included studies reported improvements in orthodontic mini-implant stability following diode laser photobiomodulation, whereas two studies found no significant differences compared with the control group. Considerable heterogeneity was identified regarding laser wavelength (618–940 nm), irradiation protocols, energy parameters, timing of force application, and stability assessment methods. Most studies presented a low risk of bias, although some concerns remained regarding allocation concealment and blinding. According to the GRADE assessment, the overall certainty of the evidence was considered moderate, mainly because of inconsistency across studies and methodological heterogeneity. Conclusions: Current evidence suggests that diode laser photobiomodulation may improve the stability of orthodontic mini-implants. Nevertheless, the small number of randomized clinical trials and the substantial variability in PBM protocols and outcome assessment methods preclude definitive clinical recommendations. Future well-designed randomized controlled trials using standardized irradiation parameters and uniform stability assessment methods are required to establish evidence-based clinical protocols. Full article

The rapid development of artificial intelligence (AI) has introduced new possibilities for transforming educational assessment processes. Among these developments, AI-assisted grading systems have attracted increasing attention due to their potential to improve efficiency, consistency, and scalability of student evaluation. The present study examines the role of artificial intelligence in student grading by comparing AI-generated scores with human teacher evaluations and by exploring teachers’ perceptions regarding the use of AI in educational assessment. The research adopts a quantitative comparative design. Student-written responses were independently evaluated by teachers and AI systems, and the resulting scores were statistically analyzed to examine the level of agreement between the two grading approaches. In addition, a structured questionnaire was administered to teachers to investigate their attitudes toward AI-assisted grading. The findings indicate that while some AI systems produce scores comparable to human evaluators, others exhibit statistically significant differences, highlighting variability across models. Furthermore, AI systems were found to produce more consistent grading outcomes in relation to the corresponding human evaluators. Nevertheless, teachers recognized the potential of AI to reduce the time required for assessment tasks. However, concerns related to fairness, transparency, and the interpretation of complex student responses remain important considerations. Overall, the results suggest that artificial intelligence can effectively support educational assessment when implemented within hybrid evaluation models that combine automated analysis with human pedagogical oversight. Full article

Rapid and accurate identification of Candida species is critical for guiding antifungal therapy, especially with the emergence of intrinsically resistant pathogens. However, diagnostics using culture-based methods remain slow and labor-intensive, limiting timely treatment decisions. This systematic review evaluated the diagnostic performance and clinical applicability of nucleic acid amplification tests (NAATs) for Candida species identification using a PubMed search completed on 23 June 2025. A total of 888 records were screened, 333 full-text articles were assessed, and 158 studies were included based on criteria including comparison with standard diagnostic methods, diagnostic performance reporting, and involvement of clinical samples. PCR-based approaches were the most widely used, including conventional, nested, multiplex, real-time, and droplet digital PCR. Isothermal methods such as loop-mediated isothermal amplification (LAMP) and recombinase polymerase amplification (RPA) were increasingly represented. Common molecular targets included the ITS and 18S/28S rDNA regions. The risk of bias assessment was completed with the QUADAS-2 tool. Diagnostic performance varied depending on assay design, specimen type, gene target, and reference method. Excellent specificity and low limits of detection were achieved, especially with isothermal platforms offering the shortest turnaround times and greatest potential for point-of-care implementation. Multiplex assays were particularly advantageous for detecting mixed-species samples, while highly specific assays were optimal for distinguishing clinically important species such as Candidozyma auris, Nakaseomyces glabratus, and Pichia kudriavzevii. Overall, NAATs represent a promising diagnostic tool for Candida species identification, but broader clinical adoption will require improved standardization, validation across diverse patient populations, and clearer interpretation of fungal burden in the context of colonization versus infection. Full article

Cloud and quality screening removes approximately 65% of daily TROPOMI tropospheric NO₂ pixels, creating structured data gaps that coincide with meteorological conditions driving pollution extremes. Standard gap-filling methods—kriging, Random Forests and other machine learning methods—act as statistical smoothers that systematically suppress extreme concentrations and ignore the Missing Not At Random (MNAR) character of cloud-induced missingness. Here we present a physically informed framework that treats urban NO₂ as a forced nonlinear dynamical system and reconstructs missing satellite observations through geometric navigation on a shadow manifold rather than statistical interpolation. The framework integrates five components: (i) Multivariate State-Space Reconstruction (MSSR) using multiview embeddings of continuous ground-based NO₂, O₃, and ERA5 meteorology, grounded in Stark’s forced-system embedding theorem; (ii) Short-Time Regime-Conditioned Convergent Cross Mapping (ST-RC-CCM) with a spatial-mismatch negative control for falsifiable causal validation; (iii) Inverse Probability Weighting (IPW) to correct the clear-sky sampling bias; (iv) trajectory-matrix denoising via Singular Spectrum Analysis (SSA) and Robust PCA; (v) topology-inspired fidelity metrics—Manifold Overlap Ratio (MOR) and Dynamic Trend Capture (DTC)—that penalize smoothing artefacts. The physical basis for this coupling is the shared dynamical history of surface and column NO₂: tropospheric NO₂ has a photochemical lifetime of 1–4 h near urban emission sources, comparable to the boundary layer mixing timescale, ensuring that surface and column concentrations are jointly governed by the same emission–photolysis–transport attractor. The planetary boundary layer height (PBLH), solar zenith angle (SZA), and surface O₃—all included as MSSR coordinates—are the dominant physical drivers of the instantaneous surface-to-column scaling, and their joint trajectory in state space constitutes the physically grounded basis for analogue selection. The framework is validated on a synthetic forced Lorenz-96 system, then applied to five European primary cities spanning contrasting regimes (Sofia, Milano, Stuttgart, Kraków, Hamburg) plus five N1 spatial-mismatch control stations (Plovdiv, Genova, Frankfurt, Warszawa, Berlin)—ten urban-background stations across four countries—with structured ablations (A0-A4V-A4K). Across >3600 evaluations, MOR_ext distributions for EDM and non-EDM methods are non-overlapping by a factor exceeding 5× (EDM minimum 0.59 vs. non-EDM maximum 0.10; median non-EDM MOR_ext ≤ 0.05 at every city × mask combination), while EDM achieves MOR_ext up to 0.915 (Milano Po Valley). Under a fair-comparison benchmark that withholds ground-level NO₂ from Random Forest, EDM’s RMSE advantage remains robust at a median of 3.9× (RF_FULL) and increases to 4.2× (RF_METEO), confirming that the performance gap is physical rather than an information artefact. A three-level temporal validation—within-window pseudo-cloud masking, cross-year transfer (full 2022 holdout and DJF 2023/24), and a COVID-19 out-of-distribution test—demonstrates robustness beyond standard train/test splits, with CCM library-length convergence confirmed for 60/60 ablations (p < 0.001) across all ten stations. Spatial-mismatch tests confirm local dynamical specificity at all five primary–control pairs (Δρ = 0.090–0.210), with seasonal modulation driven by orographic and synoptic mechanisms. These results establish manifold-based gap filling as a dynamically informative complement to statistical approaches, particularly in topographically confined, stagnation-prone basins where preserving extreme-event geometry is essential for exposure assessment. Full article

Background/Objectives: Post-stroke spasticity is a common and disabling complication that impairs functional recovery. Acupuncture is used as a complementary intervention in stroke rehabilitation, but evidence for specific techniques, including acupoint-to-acupoint penetrating acupuncture (AAPA), remains limited. This study evaluated the adjunctive effects and safety of AAPA applied to limb acupoints in patients with post-stroke spasticity receiving conventional care. Methods: A systematic review and meta-analysis of randomized controlled trials was conducted. Nine databases and five clinical trial registries were searched from inception to June 2026. Trials compared AAPA plus conventional care with conventional care alone. The primary outcome was post-treatment spasticity severity. Risk of bias was assessed using RoB 2, and evidence certainty using GRADE. Results: Sixteen studies involving 1170 participants were included. Adjunctive AAPA significantly reduced Modified Ashworth Scale scores (SMD = −0.91, 95% CI: −1.23 to −0.59). Motor function, gait, balance, activities of daily living, and neurological deficits were also improved. However, most studies had a high overall risk of bias, no trial used a sham acupuncture comparator, and substantial heterogeneity was observed for several outcomes. The certainty of evidence for all GRADE-assessed clinical outcomes was very low. Conclusions: The findings suggest possible add-on benefits of AAPA when used with conventional post-stroke care, but the current evidence should be interpreted cautiously and is insufficient to support definitive clinical recommendations. Rigorous multicenter trials in diverse rehabilitation settings with standardized outcomes, systematic safety assessment, and longer follow-up are needed. Full article

Objective: This study aimed to systematically evaluate risk prediction models for new vertebral fracture after vertebral augmentation in elderly patients with osteoporotic vertebral compression fractures (OVCFs) and to summarize their modeling methods, predictors, predictive performance, and methodological quality. Methods: This systematic review was prospectively registered in the Open Science Framework (OSF; DOI:10.17605/OSF.IO/S259R). PubMed, Embase, Web of Science, Cochrane Library, CNKI, Wanfang, and VIP were searched from inception to January 2026. Studies reporting model development, internal validation, external validation, or model updating for prediction models of new vertebral fracture after vertebral augmentation in elderly patients with OVCFs were included. Two reviewers independently screened the literature and extracted data. Methodological quality and risk of bias were assessed using the Prediction Model Risk of Bias Assessment Tool (PROBAST). Results: A total of 24 studies containing 29 prediction models were included, all of which were conducted in China. The incidence of new vertebral fracture after vertebral augmentation ranged from 10.60% to 38.08%. Logistic regression and nomograms were the most common modeling method and presentation format, respectively. The AUC values of the models ranged from 0.648 to 0.990. Only three studies conducted external validation. Bone mineral density, bone cement leakage, and anti-osteoporosis treatment were the leading predictive factors. PROBAST assessment confirmed that all included studies had a high overall risk of bias, with the statistical analysis domain being the most problematic area. Fourteen studies reported decision curve analysis (DCA), suggesting potential clinical utility within the development datasets. However, no study evaluated real-world clinical impact, and the overall high risk of bias and limited external validation substantially restrict confidence in clinical applicability. Conclusions: Although several models demonstrated moderate-to-good discrimination performance, methodological limitations, high overall risk of bias, and insufficient external validation substantially limit confidence in their clinical applicability. These models are not reliable for routine clinical use and should currently be regarded as research tools. Standardized study design, bias control, adequate external validation, and optimized statistical strategies are urgently required to establish credible prediction models for clinical practice. Full article

Anxiety disorders are highly prevalent conditions associated with substantial psychosocial burden, and low-level laser therapy auriculotherapy (LLLT-AT) has emerged as a non-invasive intervention with potential neuromodulatory effects for anxiety management. This systematic review aimed to evaluate the effects of LLLT-AT on anxiety symptoms across diverse populations. The review was conducted according to PRISMA 2020 guidelines and prospectively registered in PROSPERO (CRD420251159376). Randomized controlled trials comparing LLLT-AT with sham, placebo, or active interventions were identified through searches in PubMed, Web of Science, Scopus, CINAHL, MEDLINE, ScienceDirect, Cochrane Library, PEDro, and Google Scholar. The literature search was last updated on 2 July 2026. Risk of bias was assessed using RoB 2 and certainty of evidence using GRADE. Six RCTs involving 386 participants were included. Meta-analysis suggested significant reductions in post-treatment anxiety symptoms favoring LLLT-AT compared with control interventions (SMD = −0.65; 95% CI: −1.17 to −0.13), although substantial heterogeneity was identified (I² = 78.9%). Significant reductions were also observed for pain intensity and disability, while adverse eventst analyses demonstrated lower risks of ear pain, hyperemia, and itchy ear following LLLT-AT. However, sensitivity analyses excluding studies with high risk of bias yielded a non-significant effect for anxiety outcomes (SMD = −0.35; 95% CI: −0.94 to 0.25). Interpretation of the findings should be considered cautiously due to substantial heterogeneity and the very low-to-low certainty of evidence for anxiety outcomes. Across all outcomes, the certainty of evidence ranged from very low to low, with very low certainty for anxiety outcomes and low certainty for pain intensity, disability, and adverse eventsoutcomes (ear pain, hyperemia, and itchy ear). LLLT-AT may represent a promising non-invasive complementary intervention for anxiety-related symptoms; however, these findings should be interpreted cautiously given the limited evidence base and substantial heterogeneity. Further high-quality randomized controlled trials with standardized protocols and longer follow-up assessments are required. Full article

Background: Obesity is a chronic condition that compromises health and quality of life, with physical exercise representing a relevant therapeutic strategy in its management. Functional training (FT) has been adopted as a modality that integrates neuromuscular and cardiorespiratory stimuli, promoting both physical and psychosocial benefits. However, gaps remain in the literature regarding the effects of FT on body composition in individuals with excess weight. In this context, a systematic review was conducted to analyze the current state of scientific knowledge on this approach. Methods: Randomized controlled clinical trials investigating FT were selected, including interventions classified as high-intensity functional training (HIIT-F), multicomponent programs, and CrossFit-like interventions when they fulfilled the predefined FT eligibility criteria. Participants were required to be between 18 and 65 years of age, physically independent, and in generally good health. The systematic search was conducted in the PubMed/MEDLINE, Embase, Web of Science, Cochrane Library, PEDro, and LILACS databases. The risk of bias (RoB 2.0) and the certainty of the evidence (GRADE) were assessed. Results: A total of 2,972 records were identified. Seven studies met the eligibility criteria, comprising 296 participants. Two randomized controlled trials (61 participants) were eligible for meta-analysis. The pooled analysis demonstrated a significant reduction in body fat percentage (%BF) in favor of FT/HIIT-F compared with non-exercising controls (mean difference = −4.44 percentage points; 95% confidence interval −7.08 to −1.80; p = 0.001). Risk of bias assessment showed moderate methodological limitations across studies. None were classified as low risk. Three had high risk due to missing outcome data and deviations from intended interventions, whereas four showed some concerns related to allocation concealment and analysis reporting. Bias in outcome measurement was consistently low. Conclusions: Current evidence suggests that FT/HIIT-F may be associated with reductions in %BF among adults with overweight or obesity. However, these findings should be interpreted with caution because the certainty of the evidence was rated as very low according to the GRADE approach, the quantitative synthesis included only two studies, and important methodological limitations were identified. As the available data are predominantly derived from female participants, the applicability of these findings to men remains uncertain. Full article

Background: Contemporary adults face fragmented schedules, sedentary digital work, cognitive load, logistical barriers, and unstable routines that can make conventional session-based exercise difficult to sustain. Exercise snacks, vigorous intermittent lifestyle physical activity (VILPA), movement breaks, low-volume high-intensity interval training (HIIT), workplace exercise, and digitally mediated physical activity may reflect a broader redesign of exercise for real-world lifestyles. Objective: To develop a conceptual framework for lifestyle-congruent exercise models: approaches that intentionally modify dose, timing, setting, delivery, autonomy, or routine integration to reduce lifestyle-related friction while preserving a plausible physiological or behavioral stimulus. Methods: This evidence-based narrative review purposively identified sources through targeted searches, citation tracking, and manual verification across exercise science, sedentary behavior, workplace health, behavioral medicine, and digital health. Sources were selected for conceptual relevance, model-family representation, evidence-type diversity, and contribution to framework development, and then charted by model family, population, context, exercise architecture, delivery, outcomes, and implementation features. No systematic protocol, formal risk-of-bias assessment, or certainty grading were applied. Results: Thirty-one sources informed the framework. Five design logics were identified: dose compression, temporal distribution, routine embedding, sedentary substitution, and delivery mediation. These logics show that brief or time-efficient models are not interchangeable; they differ in intentionality, context, behavioral demand, safety considerations, implementation logic, and dominant friction addressed. Conclusions: Lifestyle-congruent exercise models are complementary strategies, not replacements for structured aerobic and muscle-strengthening exercise. Their value lies in matching exercise architecture to lifestyle-related friction while supporting progression toward guideline-consistent activity. Future research should test safety, equity, scalability, and long-term maintenance across diverse populations. Full article

This paper examines how the emergence of over-parameterized artificial intelligence models and the phenomenon of double descent challenge the classical assumption that simpler models generalize better. Traditional predictive analytics relied on parsimonious models grounded in the bias-variance trade-off, where increasing complexity was expected to produce overfitting. However, recent advances in deep learning demonstrate that highly over-parameterized models can achieve superior generalization after surpassing the interpolation threshold. This paradigm shift has enabled systems such as AlphaFold, Aurora, Delphi-2M, and recommenders to model complex, high-dimensional relationships through contextual attention rather than global feature selection. The paper argues that higher education analytics remains largely reductionist, relying on limited variables such as GPA, demographics, and course completion rates to identify “at-risk” students. While interpretable, these approaches often fail to capture the dynamic and multidimensional nature of student success. In response, this study proposes a transition toward over-parameterized personalization, where students’ academic and behavioral histories are modeled as longitudinal high-dimensional sequences. Drawing parallels to commercial recommendation systems such as Amazon, Netflix, and YouTube, the paper explores how higher education can move from generalized early-warning systems toward adaptive “n-of-1” interventions. Importantly, the paper is conceptual rather than empirical: it develops a research agenda and a set of testable propositions, and it identifies the evaluation designs—temporally valid prediction protocols and causal intervention studies—by which the promise of over-parameterized personalization in higher education should be assessed before any claim of superiority can be made. Full article

Objectives: To develop and validate an individualized prediction model for assessing the risk of illness uncertainty in patients with malignant tumors, using a cross-sectional design. Methods: Patients with malignant tumors treated at Peking University People’s Hospital from August 2024 to January 2025 were enrolled. The Mishel Uncertainty in Illness Scale (MUIS) was used to classify patients into high-risk and low-to-moderate risk groups. Patients were divided into a model development set and a time-stratified validation set. The development set was further randomly split into a training set and an internal validation set at a 7:3 ratio. Three machine learning algorithms Least Absolute Shrinkage and Selection Operator (LASSO), Random Forest (RF), and eXtreme Gradient Boosting (XGBoost) were employed to screen for common predictors, and a nomogram was constructed based on logistic regression. The model’s performance was evaluated using the Area Under the Receiver Operating Characteristic Curve, calibration curves, and the Hosmer–Lemeshow test. Results: A total of 966 patients were included, with 676 in the development set and 290 in the time-stratified validation set. Seven predictors were ultimately identified for the nomogram: age, education level, diagnosis, depression, anxiety, medical coping modes, and social support. Notably, anxiety was the only variable jointly identified by all three algorithms, while the other six were commonly selected by both Random Forest and XGBoost. The nomogram showed good discrimination in training (AUC = 0.763, 95% CI: 0.717–0.808) and internal validation (AUC = 0.724, 95% CI: 0.645–0.803), with well-calibrated probabilities (p > 0.05). In time-stratified validation, discrimination was acceptable (AUC = 0.663, 95% CI: 0.58–0.739), but calibration revealed risk underestimation for low-to-moderate risk group (p < 0.05). Conclusions: The nomogram shows acceptable internal performance as an exploratory concurrent classifier rather than a genuine predictor, but calibration bias and modest temporal discrimination (AUC = 0.663) indicate it is not ready for clinical use. Further optimization and multi-center prospective validation are required. Full article

Background: Metronomic chemo-endocrine therapy combining fulvestrant with metronomic VEC (vinorelbine, cyclophosphamide, and capecitabine)—the FulVEC regimen—demonstrated promising activity in an initial cohort of 38 patients with advanced ER+/HER2− breast cancer (JCM 2023). Here, we present an extended analysis of 72 consecutive patients, with a focus on a novel hypothesis: that treatment-emergent adverse events (AEs) requiring dose modification serve as a surrogate for sufficient pharmacological exposure in metronomic combination chemotherapy. Methods: Retrospective analysis of 72 consecutive patients with metastatic ER+/HER2− breast cancer treated with FulVEC at Jagiellonian University Hospital between 2018 and 2024. Efficacy endpoints included progression-free survival (PFS), overall survival (OS), and biochemical response, as assessed by CA15-3 dynamics. Patients were stratified by AE severity requiring intervention (grade 0: no modification; grade 1: dose reduction; and grade 2: treatment delay). The association between AE grade and efficacy outcomes was assessed using Spearman’s correlation, the log-rank test, and the chi-square test. Results: The median PFS was 8.5 months, and the median OS was 18.0 months. The biochemical benefit rate (any CA15-3 decline) was 81.6%. No statistically significant differences in efficacy were observed according to prior exposure to CDK4/6 inhibitors, fulvestrant, or cytotoxic components of the FulVEC regimen. A monotonic dose–response relationship was observed across AE grade categories: non-progression rates increased from 73.2% (grade 0) to 84.2% (grade 1) and 91.7% (grade 2); biochemical benefit rates from 68.4% to 90.9% and 100.0%; and median CA15-3 reduction deepened from −34% to −44% and −52%, respectively (Spearman r = 0.258 and p = 0.043 for AE grade vs. treatment duration). Formal log-rank comparisons of PFS and OS across the three AE-grade categories did not reach statistical significance (p = 0.583 and p = 0.743, respectively), reflecting the limited size of the treatment-delay subgroup (n = 12); the dose–response signal should, therefore, be regarded as exploratory. No patient required permanent treatment discontinuation due to toxicity. Conclusions: The extended FulVEC cohort confirms durable activity and a reproducible, manageable safety profile in a heavily pretreated population, including CDK4/6i-refractory patients. The exploratory, hypothesis-generating observation of a dose–response gradient between AE severity and clinical outcomes raises the possibility that treatment-emergent AEs may, in some patients, reflect adequate pharmacological exposure to the metronomic regimen. Given confounding by treatment duration and survivor bias, and the absence of pharmacokinetic data, this hypothesis requires prospective validation and does not, at this stage, support any change to current treatment practice. Full article