Search Results (645)

Sulfamethoxazole (SMX), a persistent sulfonamide antibiotic widespread in aquatic environments, resists conventional water treatment degradation. This work developed a NaOAc/Fe³⁺/O₃ homogeneous ozonation system for SMX abatement. Four comparative reaction groups confirmed significant synergism between trace Fe³⁺ and sodium acetate. Under optimized near-neutral conditions, 96.19% SMX was removed within 30 min with k_obs = 0.107 min⁻¹, outperforming sole O₃, O₃/NaOAc and O₃/Fe³⁺ by 11.0%, 31.0% and 21.9% respectively. Single-factor tests revealed excess Fe³⁺ or acetate suppressed catalytic activity, while alkaline conditions accelerated degradation yet aggravated iron precipitation. Radical quenching and p-CBA probing indicated that acetate coordination did not increase bulk ·OH exposure, while the stronger TEMP-derived TEMPO response after acetate addition was consistent with enhanced ¹O₂-associated oxidation, suggesting that acetate altered the relative contributions of ozone-derived oxidation pathways. Post acetate background deduction, the ternary system achieved higher TOC/COD elimination. Twelve SMX intermediates were identified via LC-MS, with three ring-opening degradation pathways proposed. QSAR toxicity evaluation indicated that most intermediates possessed lower bioconcentration and developmental risks than raw SMX. Overall, the results indicate that weak acetate ligands can alter iron-mediated ozone oxidation pathways, providing a low-dose and economical strategy for antibiotic wastewater treatment. Full article

Background/Objectives: Ceftriaxone (Ctx) is a third-generation cephalosporin widely used to treat infections caused by Gram-positive and Gram-negative bacteria. However, its clinical efficacy may be limited by rapid systemic elimination and suboptimal tissue distribution. Erythrocyte-based targeted drug delivery systems (TDDSs) have emerged as a promising approach to prolong drug circulation and enhance site-specific accumulation. This study investigated the pharmacokinetic profile and tissue distribution of ceftriaxone encapsulated in autologous erythrocytes (RBC-Ctx) compared with free ceftriaxone (Free-Ctx) following intravenous administration in rats. Methods: Ceftriaxone was encapsulated into autologous rat erythrocytes using a hypoosmotic hemolysis loading technique. Drug-loaded erythrocytes are called pharmacocytes. Adult male Wistar rats received a single intravenous injection of Free-Ctx or RBC-Ctx at an equivalent ceftriaxone dose of 340 mg/kg. Plasma samples were collected over 24 h for pharmacokinetic analysis, while the liver, spleen, lungs, kidneys, heart, pancreas, and skeletal muscle were harvested at 1 and 12 h for tissue distribution studies. Ceftriaxone concentrations were quantified by high-performance liquid chromatography with UV detection. Results: Erythrocyte encapsulation significantly modified the pharmacokinetic behavior of ceftriaxone. Compared with Free-Ctx, RBC-Ctx prolonged the elimination half-life (4.4 ± 0.6 vs. 1.8 ± 0.1 h), increased systemic exposure (AUC₀–last, 1.6 ± 0.1 vs. 1.2 ± 0.2 mg·h/mL), reduced total body clearance (218.2 ± 11.0 vs. 294.0 ± 48.8 mL/h/kg), and increased the apparent volume of distribution at steady state (688.3 ± 61.0 vs. 435.0 ± 23.1 mL/kg). In addition, RBC-Ctx was associated with a distinct relative tissue-distribution pattern of ceftriaxone, particularly in reticuloendothelial system-rich organs such as the liver and spleen, while ceftriaxone remained detectable in several tissues at 12 h after administration. In contrast, ceftriaxone concentrations following Free-Ctx declined markedly or became undetectable over the same period. Conclusions: Encapsulation of ceftriaxone into autologous erythrocytes substantially prolonged systemic circulation, enhanced drug exposure, reduced clearance, and altered the relative tissue distribution of ceftriaxone. These findings demonstrate that erythrocyte-based carriers effectively modulate ceftriaxone pharmacokinetics and tissue distribution, supporting their potential as a targeted antibiotic delivery platform for improving antimicrobial therapy, particularly for infections involving reticuloendothelial system-associated tissues. Further studies in experimental models of infection and inflammation are warranted to evaluate therapeutic efficacy under pathological conditions and to optimize this delivery strategy. Full article

Sustainable management of antibiotic fermentation residues is critical to mitigating environmental and public health risks associated with pharmaceutical waste. Gentamicin fermentation residue (GFR), a challenging antibiotic byproduct rich in nitrogen and organic matter, presents both environmental risks and untapped resource potential. Pyrolysis can convert GFR into high-value products (pyrolysis oil, gas, and char) and eliminate environmental risks through high-temperature treatment. In this study, the product yields, pyrolytic kinetics, and nitrogen transformation pathways at different temperatures (400–800 °C) were investigated to explore the pyrolysis behavior and mechanism of GFR. Results revealed that temperature strongly influenced product distribution: biochar yield was dominant at low temperatures 51.2 ± 0.4% at 400 °C, oil yield peaked at 600 °C 14.6 ± 0.4%, and gaseous products became prevalent above 700 °C, reaching an estimated 78.4 ± 0.2% at 800 °C by mass-balance difference. Furthermore, isoconversional kinetic analysis (FWO and KAS) yielded apparent activation energies of 226.9–249.4 kJ/mol over α = 0.5–0.8, with R² values of 0.946–0.976, indicating conversion-dependent, multi-step devolatilization behavior. For nitrogen transformation, elemental and spectroscopic analyses showed that nitrogen transitioned from unstable pyrrolic forms in raw GFR to more stable pyridinic and graphitic forms in biochar, enhancing its potential for catalytic and environmental applications. Meanwhile, gaseous nitrogen species such as NH₃ and HCN were released at different temperature stages, and a three-stage nitrogen transformation mechanism was proposed linking the decomposition kinetics to the nitrogen migration pathways. These findings highlight pyrolysis as a promising, controllable, and sustainable method to convert antibiotic residues into valuable energy products and functional materials, with its contribution to sustainable pharmaceutical waste management contingent on proper NH₃/HCN gas treatment. Full article

The increasing accumulation of pharmaceutical residues in water environments poses serious threats concerning environmental safety and public health, mainly due to their tenacity, continuous bio-activity, and resistance to traditional wastewater treatment processing. Although many nano-bio/chemosensor systems have been reported for the monitoring and removal of pharmaceutical residues, the literature remains fragmented regarding their ability to integrate detection and remediation into a single platform. In this context, this review critically examines recent developments in nano-bio/chemosensor platforms for simultaneous detection and elimination of pharmaceutical effluents in wastewaters. Particular emphasis is placed on their functional integration, detection mechanisms, analytical performance, and removal pathways. This review covers the major pharmaceutical categories, including pharmaceutical drugs, antibiotics, hormones, perfluorinated compounds, and drugs of abuse and discusses nanostructured platforms based on metal organic frameworks (MOFs), nanochannel-based immunosensors, noble metal nanoparticles, layered double hydroxides, and hybrid composites. Detection approaches based on fluorescence modulation, electrochemical impedance, ionic current rectification, surface-enhanced Raman scattering (SERS), and colorimetric nanoenzyme activity could lead to extremely low detection limits. In addition, removal mechanisms such as adsorption, photocatalysis, advanced Fenton-induced oxidation processes, and nanoenzymes allow for high degradation efficiencies (>80–99%). Significant advantages for real-time monitoring and sustainable wastewater treatment can be achieved by multifunctional nanoplatforms that integrate detection and remediation capabilities. Finally, this review identifies current limitations and research gaps regarding practical application, matrix effects, regeneration, stability, scalability, and integration into real wastewater treatment systems and outlines future research directions towards more efficient and environmentally relevant multifunctional platforms. Full article

Background/Objectives: Q fever, caused by Coxiella burnetii, is a zoonotic disease usually treated with doxycycline, but alternative options are needed when doxycycline is unsuitable. This study evaluated inhaled liposomal ciprofloxacin (Apulmiq^®) as a post-exposure prophylaxis (PEP) compared with oral doxycycline in a common marmoset model. Methods: Marmosets were exposed to aerosolised C. burnetii and treated 24 h later with Apulmiq (0.8 mg/kg inhaled), doxycycline (6 mg/kg oral), or placebo for 7 days. Outcomes included fever, weight loss, bacterial burden, histopathology, liver enzymes, and immune responses. Results: Both treatments delayed fever onset and reduced disease severity versus placebo. Doxycycline prevented fever during treatment and significantly reduced bacterial loads in the lungs and spleen by Day 28. Apulmiq slightly delayed the onset of fever and reduced early bacteraemia but was less effective at bacterial clearance. Immune analyses showed macrophage activation and elevation of IFN-γ during acute infection, and adaptive immune responses by Day 28. Conclusions: Inhaled Apulmiq provided partial prophylactic protection but did not achieve bacterial clearance. Doxycycline reduced bacterial burden but did not completely eliminate disease. Inhaled antibiotics may have potential for use as an alternative approach for PEP treatment, although further optimisation and investigation are needed. Full article

Developing efficient and eco-friendly catalysts for peroxymonosulfate (PMS) activation to degrade persistent antibiotics in water remains a major challenge. Herein, acid-base etched montmorillonite is employed to activate PMS for sulfamethoxazole (SMX) degradation under neutral to weakly alkaline conditions. Compared with pristine montmorillonite and acid-etched montmorillonite (A-Mon), base-etched montmorillonite (B-Mon) achieves the highest removal efficiency, eliminating 92% of SMX within 120 min. The degradation rate constant of the B-Mon/PMS system is 0.19 min⁻¹, which is 2.7 times that of the pristine montmorillonite/PMS system. Reactive species analysis reveals that the B-Mon/PMS system increases the singlet oxygen (¹O₂) concentration by 10.3-fold compared to PMS alone. Electron spin resonance (ESR), temperature-programmed desorption (TPD), and quenching experiments demonstrate that medium-weak Lewis basic sites on the B-Mon surface play a pivotal role in PMS activation and SMX degradation. Moreover, these medium-weak Lewis basic sites facilitate ¹O₂ generation via the self-decomposition mechanism of PMS. Stability tests confirm that B-Mon exhibits excellent cycling stability and anti-interference capability. This work provides a deeper mechanistic insight into the development of environmentally friendly aluminum-based catalysts for practical remediation of complex water. Full article

Background: Bacterial biofilms play a key role in causing periprosthetic joint infection (PJI). Current PJI management strategies commonly combine systemic antibiotic therapy with localized delivery such as antibiotic-loaded cement or beads to achieve effective tissue penetration and high antimicrobial concentrations at the implant site. We hypothesized that antibiotic-loaded calcium sulfate beads would effectively eradicate early Staphylococcus aureus biofilms but exhibit reduced efficacy against mature biofilms formed in synovial fluid (SF). This study evaluated the efficacy of vancomycin combined with gentamicin (V+G) or tobramycin (V+T) against GFP-expressing S. aureus biofilms grown in the presence of SF. Methods: Biofilms were established on clinically relevant orthopedic implant materials such as titanium (Ti), stainless steel (316L), and polyethylene (PE). Early (1-day) and mature (3-day) biofilms were treated with calcium sulfate beads loaded with V+G or V+T. Antimicrobial efficacy was quantified by colony-forming unit (CFU) enumeration. Minimum inhibitory concentration (MIC) testing was performed on surviving populations to assess potential development of antibiotic resistance. Results: Both antibiotic combinations resulted in no detectable viable bacteria in the 1-day biofilm across all tested materials following treatment, demonstrating strong reduction in early biofilm burden below the detection limit. Treatment of mature biofilms resulted in a 4–5 log reduction. MIC analysis of representative bacteria surviving post-treatment indicated no substantial increase in resistance. Conclusions: These findings show that while locally delivered antibiotic combinations eliminate early biofilms, mature biofilms demonstrate significant tolerance. MIC analysis showed little or no change in vancomycin, gentamicin, and tobramycin susceptibility after treatment, indicating no evident increase in planktonic antibiotic resistance among the surviving isolates. Together, these results highlight the importance of early intervention and the need for strategies that disrupt biofilm structure or improve antibiotic penetration to enhance PJI treatment outcomes. Full article

The use of invertebrates for sewage sludge bioconversion offers a sustainable strategy for waste valorization, yet species-specific effects on microbial communities and biosafety remain unclear. Here, we compared the impacts of the earthworm Eisenia andrei (Ea) and the mealworm Tenebrio molitor (Tm) on the bacteriomes and mycobiomes of sewage sludge (ss) using 16S rRNA and ITS amplicon sequencing. Gut passage in both Ea and Tm markedly altered bacterial and fungal composition relative to ss, but produced distinct community profiles with differential shifts across multiple taxa. Both invertebrates reduced bacterial richness by ~40%, while fungal responses diverged: Ea largely preserved mycobiome richness despite reduced evenness, whereas Tm caused a near-complete collapse (~83% Amplicon Sequence Variant loss). Beta diversity analyses revealed clear, non-overlapping separation among ss, Ea, and Tm for both microbial domains. Tm frass showed strong enrichment of clinically relevant bacterial pathogens, while Ea casts exhibited no such enrichment. For fungi, Ea reshaped pathogen composition, whereas Tm largely eliminated fungal pathogens through broad community collapse. Both treatments reduced predicted antibiotic resistance gene abundance, but functional profiles differed, with Ea showing greater functional stability. These findings demonstrate that microbiome restructuring during bioconversion is species-dependent, with contrasting ecological and biosafety implications for downstream environmental use. Full article

In conventional anaerobic digestion (AD) of livestock manure, temperature regulation faces an inherent paradox: thermophilic conditions facilitate pathogen inactivation but inhibit methanogenesis, whereas mesophilic conditions enable efficient methane production but are less effective at eliminating pathogens. This study systematically analyzed the methanogenic performance, microbial physiology, and the dynamics of biosafety factors during AD of chicken manure at 4 °C, 35 °C, and 55 °C. Additionally, a ceramic membrane (CM) coupled with ozone micro-nano bubbles (O₃-MNBs) backwashing was configured for advanced digestate purification. Contrary to conventional understanding, mesophilic conditions (35 °C) simultaneously achieved optimal methanogenic efficiency (277.83 mL/gVS) and the efficient removal of viruses (66.65%) and antibiotic resistance genes (ARGs) (>77.11%), supported by a more diverse microbial community. While thermophilic conditions (55 °C) inactivated certain viruses, methanogenic efficiency was significantly inhibited by thermal stress. Moreover, the residual plasmid-borne ARGs were 3.04-fold higher than under mesophilic conditions. The integration of CM filtration with O₃-MNBs backwashing effectively retained and inactivated pathogens while reducing the membrane fouling rate by 83.97%. The integration of mesophilic digestion, membrane filtration, and O₃-MNBs processes synergistically achieved efficient energy recovery and robust pathogen control, providing technical support for the safe treatment and high-value resource utilization of manure. Full article

Staphylococcus aureus frequently colonizes the skin and upper airways and produces a broad repertoire of immunomodulatory molecules. In asthma, the most consistent evidence concerns staphylococcal enterotoxins (SEs), which can act both as superantigens and as allergens, and IgE sensitization to SEs (SE-sIgE). SE-sIgE is associated with severe asthma, type 2 inflammation, chronic rhinosinusitis with nasal polyps (CRSwNP), exacerbations, and, in some longitudinal studies, persistent airflow obstruction. However, this relationship is not necessarily causal: SE-sIgE may reflect exposure, an immune response, or a biologically active endotype, whereas colonization, local toxin production, and systemic sensitization are not equivalent. SEs simultaneously bind class II MHC molecules and Vbeta regions of the T-cell receptor, activating large fractions of T lymphocytes; they also promote IL-4, IL-5, and IL-13 production, polyclonal B-cell activation, local IgE synthesis, mast-cell degranulation, eosinophilia, and IL-8/neutrophil circuits. Alpha-toxin (Hla) and SEB can damage the epithelial barrier, facilitating allergen penetration and alarmin signalling. These observations support an interaction model in which dysbiosis, barrier dysfunction, and type 2 immunity mutually reinforce one another along the nasobronchial axis. Corticosteroids and antibiotics may modify selected nodes in this circuit, but current evidence is insufficient to recommend decolonization or antitoxin therapy in stable asthma. Biologics interrupt downstream pathways potentially fuelled by toxins: omalizumab neutralizes free IgE; mepolizumab and benralizumab reduce the eosinophilic axis; dupilumab blocks IL-4/IL-13 signalling; and tezepelumab acts upstream on TSLP. Nevertheless, randomized trials stratified by SE-sIgE are lacking, and no evidence demonstrates that these treatments eliminate colonization or toxin production. SE-sIgE therefore appears to be a promising biomarker, particularly in severe asthma with CRSwNP, but it is not yet an autonomous criterion for biologic selection. A broader barrier-organ analysis also identifies nasal, cutaneous, and intestinal colonization as distinct ecological states; atopic dermatitis as a complementary model of toxin-amplified type 2 inflammation; and biofilms and extracellular vesicles as candidate mechanisms of persistent toxin delivery. These data increase biological plausibility but remain indirect for asthma. Full article

Background/Objectives: Augmented renal clearance (ARC) may reduce blood concentrations of renally eliminated antibiotics and increase the risk of treatment failure in sepsis. We investigated the risk factors for ARC and the effect of inflammation on ARC duration. Methods: This retrospective cohort study included 282 patients treated for sepsis. The primary endpoint was the ARC incidence, while the secondary endpoints were the relationship between the ARC duration and the time courses of inflammatory marker levels. Patient characteristics, disease severity, laboratory data, and treatment data were collected from medical records. Risk factors for ARC were identified using logistic regression. The effect of the inflammatory response on the ARC duration was evaluated using the Mann–Whitney U test. Results: Among 174 patients (median age, 78 years [range: 29–95]; 94 [54.0%] male; the median intensive care unit length of stay was 6 days [1–76 days]) with ARC identified in 15.5% of patients. Significant risk factors for ARC were age < 75 years (odds ratio, 4.748), albumin levels < 3.0 g/dL (3.672), serum potassium levels < 3.6 mmol/L (5.196), and sequential organ failure assessment scores ≤ 9 (3.803). Patients with an ARC duration of ≥ 8 days had significantly higher white blood cell counts and C-reactive protein levels at ARC incidence than those with an ARC duration < 8 days. Conclusions: Age, albumin, serum potassium, and sequential organ failure assessment scores were risk factors for ARC in patients with sepsis. Persistent inflammation may prolong ARC. These findings can aid in choosing antibiotic dosing regimens based on renal function in patients with sepsis at high risk of ARC. Full article

Background: This study aimed to comprehensively investigate the complex association between the use of fluoroquinolone (FQ) antibiotics and cancer risk, with a specific focus on patients with interstitial lung disease (ILD)—a unique clinical population characterized by a high inflammatory burden and a high susceptibility to infections. Methods: We conducted a large-scale retrospective cohort study using a high-quality clinical database. A total of 7906 matched patients (3953 pairs) were included after propensity score matching (PSM). Three complementary statistical models were applied: the standard Cox proportional hazards model, the time-dependent Cox regression model, and the Fine–Gray competing-risks model, to provide a multidimensional assessment of cancer risk. Results: A total of 7906 matched patients (3953 pairs) were followed. After strictly defining the index date to eliminate immortal time bias, FQ exposure was associated with an increased risk of all-cause cancer in the standard Cox model (adjusted HR 1.45; 95% CI, 1.20–1.76) and the competing risk model (adjusted SHR 1.28; 95% CI, 1.06–1.55). Site-specific analyses revealed elevated risks for certain malignancies, notably prostate cancer. Importantly, when modeled as a continuous variable, the cumulative dose of fluoroquinolones showed no significant dose–response relationship with overall cancer risk (adjusted HR 0.99; 95% CI, 0.99–1.00). Conclusions: After correcting for immortal time bias, the previously hypothesized protective effect of fluoroquinolones on cancer risk was not observed. The increased risk observed in categorical models, coupled with a lack of a continuous dose–response, strongly suggests that these findings are driven by confounding by indication and reverse causation (i.e., frequent infections masking undiagnosed malignancies or reflecting severe underlying ILD), rather than a direct pharmacological effect. Full article

With the widespread application of new chemicals, the concentration of emerging contaminants (ECs), such as antibiotics, per- and polyfluoroalkyl substances (PFAS), microplastics, and new pesticides, in aquatic environments is on the rise. ECs such as those examined in the studies exhibit high toxicity, persistence, and a propensity for bioaccumulation, which can lead to significant risks for ecosystems and human health. Traditional water treatment technologies exhibit limited removal capabilities for ECs, whereas micro-nano bubbles (MNBs) exhibit great potential in the field of ECs treatment, due to their unique physicochemical properties. This article systematically reviews the research progress on the treatment of ECs using MNBs combined with other technologies, including physical methods (adsorption enhancement), chemical methods (ozonation, persulfate oxidation, photocatalysis, and material catalysis) and biological methods (microbial synergy). This review summarizes the research progress and mechanisms of MNBs combination technologies, outlining the critical knowledge gaps and future research perspectives to advance the rational design and engineering application of MNBs for ECs’ elimination in water treatment. Full article

UV-based water treatment technologies are suitable for removing persistent organic micropollutants and can be used as a quaternary treatment in the production of drinking water. The most commonly used light source is a low-pressure mercury vapor lamp (LPMV) that emits at 254 nm. KrCl excimer lamps emitting at 222 nm are a potential mercury-free light source for ultraviolet (UV)-based advanced oxidation processes. In addition, 222 nm UV light may be more effective in photolysis of oxidants and radical generation than 254 nm. Experiments were conducted in both ultrapure water and biologically treated domestic wastewater, using LPMV and KrCl lamps, with two radical promoters (H₂O₂ and peroxydisulfate (PDS)) to transform and eliminate the trimetoprim antibiotic. At the same electrical power, the UV photon flux of the LPMV exceeds that of the KrCl lamp by an order of magnitude. However, the molar absorbance of PDS and H₂O₂ at 222 nm was significantly higher than that measured at 254 nm, thereby compensating for the low photon flux. Although competition between trimethoprim and matrix components for 222 nm UV photons may significantly reduce efficiency, the 222 nm excimer lamp was found to be a viable alternative to the 254 nm LPM, given its advantages and disadvantages. Full article

Objectives: To assess the feasibility and reproducibility of predicting antimicrobial resistance (AMR) in Escherichia coli from MALDI-TOF mass spectrometry data using a standardized, open-source machine learning (ML) workflow, we systematically compared four ML algorithms, evaluated the impact of culture conditions, extract storage, and spectral preprocessing on model performance, and validated results through nested cross-validation with statistical significance testing. Methods: A total of 282 clinical E. coli isolates were analyzed. Two MALDI-TOF MS datasets were generated from freshly cultured extracts (T₁) and recultured isolates one year later (T₃), yielding 4468 spectra. A third dataset from the T₁ extracts stored at −20 °C for one year (T₂) was evaluated for spectral stability but excluded from primary modeling likely due to storage-induced degradation. Protein spectra (m/z 2000–15,000) were preprocessed using an in-house developed MALDI-TOF preprocessing pipeline (MTPP) comprising variance stabilization, Savitzky–Golay smoothing, SNIP baseline correction, TIC normalization, LOWESS alignment, and MAD-based peak detection (SNR ≥ 3), yielding 121 m/z features. Four classifiers—Random Forest (RF), Logistic Regression, Support Vector Machine, and Gradient Boosting—were trained to predict resistance to 11 antibiotics using nested cross-validation: outer GroupShuffleSplit (5-fold, isolate-level) for evaluation and inner GroupKFold for recursive feature elimination (RFECV) and hyperparameter tuning (RandomizedSearchCV). Classification thresholds were optimized via the precision–recall curve. Model performance was assessed using AUROC, AUPRC, F1-score, Matthews Correlation Coefficient (MCC), and bootstrap 95% confidence intervals (1000 replicates). Pairwise model comparisons were tested with McNemar’s chi-squared test. Results: Among the 12 antibiotics included in the analysis (meropenem excluded for absence of resistance), resistance prevalence ranged from 1.1% (colistin) to 59.9% (amoxicillin). Colistin was subsequently also excluded from ML modeling due to insufficient resistant isolates (n = 3), leaving 11 antibiotics for prediction. The best predictive performance was observed for ciprofloxacin (AUROC 0.76 [95% CI 0.74–0.77]; F1 0.54; MCC 0.38) and ceftazidime (AUROC 0.68 [0.65–0.71]; F1 0.36; MCC 0.29), using 13 and 37 RFECV-selected features, respectively. Amoxicillin achieved the highest F1-score (0.76), driven by high recall (0.98) but modest AUROC (0.58). No meaningful predictive signal was detected for amikacin, cefepime, or tigecycline (AUROC ≤ 0.57, F1 ≤ 0.17), attributable to extreme class imbalance, and no robust multi-peak resistance signature was detected in this dataset. McNemar’s test confirmed that RF significantly outperformed Logistic Regression for all antibiotics (p < 0.01), while Gradient Boosting performed comparably to RF for ciprofloxacin (p = 0.17) and ceftazidime (p = 0.28). Frozen extracts (T₂) produced lower spectral similarity and were excluded from model training; the aligned T₁₊₃ dataset yielded the most stable performance across metrics. Conclusions: Machine learning analysis of MALDI-TOF spectra enables reproducible AMR prediction for selected antibiotics in E. coli, with ciprofloxacin and ceftazidime showing the strongest signal. Nested isolate-level cross-validation, multi-model comparison with statistical testing, and open-source code provide a transparent, reproducible foundation for integrating ML-assisted MALDI-TOF analysis into diagnostic AMR surveillance. Extract storage at −20 °C degrades spectral quality and should be avoided in ML training workflows. Full article