Browse Articles

Discover research articles across all indexed journals

AI-driven oncogenic risk stratification in pediatric autism spectrum disorder: A multi-omic machine learning framework for early cancer prevention.

Journal of Clinical Oncology Anita Gangadhar, U. Kumar, A. Singh Jul 01, 2026 DOI: 10.1200/jco.2026.44.19_suppl.12

12 Background: Children with Autism Spectrum Disorder (ASD), particularly those with co-occurring intellectual disabilities or PTEN/TSC1/2 mutations, exhibit an elevated cancer risk (OR up to 4.8). Despite this, standardized oncological screening for neurodivergent populations remains suboptimal, especially in Low-Middle Income Countries (LMICs). This study evaluated the efficacy of a multi-task deep learning framework in identifying high-risk oncogenic profiles in pediatric ASD cohorts. Methods: A retrospective analysis was conducted using a multi-omic dataset of 5,120 pediatric profiles (2018–2025) from LMIC healthcare settings. The study assessed the integration of genomic features (rare coding variants), maternal metabolic history, and environmental chemical exposures. A machine learning architecture utilizing Random Forest (RF), XGBoost, and SHAP-based Explainable AI (XAI) was implemented to determine risk-stratification accuracy and identify shared ASD-cancer biomarkers. Results: Of the 1,120 profiles analyzed, the AI framework identified high-risk oncogenic signatures in 12.4% of the cohort. The model achieved a high diagnostic accuracy with an AUC-ROC of 0.92 (p < 0.001). Genomic feature extraction highlighted 138 shared genes; patients with PTEN mutations were significantly more likely to be flagged for high risk compared to those with non-syndromic ASD (88.4% vs. 15.2%, p < 0.001). High-risk stratification was more frequent in nulliparous maternal lineages (42.1%) and those with documented early-life chemical exposure (64.5% vs. 28.3%, p < 0.01). Explainable AI (XAI) modules provided interpretable clinical pathways for 97.8% of flagged cases, demonstrating high robustness even in data-sparse environments typical of underprivileged regions. Conclusions: The current provision of cancer surveillance in ASD populations is disproportionately reactive rather than preventive. This AI-driven framework offers a scalable, objective tool for early risk stratification, bridging the gap between neurodevelopmental monitoring and oncological screening. Integrating such digital biomarkers into primary pediatric care offers an underutilized opportunity for early intervention, particularly for underprivileged populations where access to advanced genetic testing is limited.

Dose modifications, toxicity, and survival with enfortumab vedotin plus pembrolizumab (EVP) in advanced urothelial carcinoma (aUC): Real-world experience in Alberta, Canada.

Journal of Clinical Oncology Samad Sayed, Rishikesh Kumar, Martin Zarba et al. Jul 01, 2026 DOI: 10.1200/jco.2026.44.19_suppl.188

188 Background: EVP has demonstrated a survival benefit in advanced urothelial carcinoma (aUC) clinical trials. However real-world toxicity patterns, timing of onset, and prognostic significance of treatment-related adverse effects (AEs) remain incompletely characterized. We evaluated outcomes, toxicity, and dose modifications among patients with aUC treated with first-line EVP in routine clinical practice. Methods: We conducted a retrospective, multicenter analysis of aUC patients treated with first-line EVP in Alberta, Canada (Sep 2024–Jan 2026). AEs were graded using CTCAE v5.0. Time to AE onset, dose modifications, and treatment discontinuation were assessed. Progression-free survival (PFS) and overall survival (OS) were analyzed. Results: Sixty patients were included (median age 69 years, 82% male). Primary tumor site was bladder in 68% and upper genitourinary tract in 23%. Histology was pure urothelial in 80% and mixed in 20%. Disease was metastatic in 80% (55% visceral, 25% nodal-only); 20% had locally advanced/unresectable disease. Median follow-up was 10.3 months. Initial enfortumab vedotin (EV) dose was 1.25 mg/kg in 76% of patients and 1 mg/kg in 23%; the median number of EVP cycles was 8. Six-month PFS was significantly higher with an initial EV dose of 1.25 mg/kg compared with 1 mg/kg (78% vs 51%; HR 0.35, 95% CI 0.15–0.83; p = 0.017). The most common AEs of any grade were rash (63%), fatigue (57%), and neuropathy (47%). Grade III–IV AEs included rash (15%), fatigue (8%), and mucositis (3%) with median onset at 16, 95, and 66 days. Thromboembolic events occurred in 10%. No patients died due to treatment toxicity. Dose reductions occurred in 62%, dose delays in 45%, and treatment discontinuation in 38% (EV alone 25%, pembrolizumab alone 8%, both 5%). Median cycles before treatment discontinuation were 5 for rash and fatigue, and 6 for neuropathy. One-year OS was significantly higher among patients who developed a rash compared with those who did not (72% vs 46%; p = 0.018). No significant OS differences were observed with development of neuropathy (69% vs 56%; p = 0.18) or fatigue (61% vs 64%; p = 0.62). Conclusions: In this real-world cohort, EVP demonstrated efficacy and safety consistent with clinical trial data. Toxicity-related dose modifications were common but generally manageable. Higher initial EV dosing was associated with improved PFS, and treatment-related rash correlated with superior OS, suggesting a potential role for rash as an early biomarker of treatment response. Treatment related AEs and outcomes. Toxicity n (%) Median days to Onset (IQR) Dose Reductionsn (%) Dose discontinuationn (%) 1yr OS (log rank p-value) Skin Rash 38 (63) 18 (10-75) 22 (36) 6 (12) 72% vs 46% (p=.018) Neuropathy 28 (46) 96 (69-132) 12 (20) 7 (11.6) 69% vs 56%(p=0.18) Fatigue 34 (56) 29 (12-89) 14 (23) 7 (11.6) 61% vs 64%(p=0.62)

Linking device dynamics to neural network performance in ionically gated synaptic transistors

Scientific Reports Nithil Harris Manimaran, Huayuan Han, Cory Merkel et al. Jul 01, 2026 DOI: 10.1038/s41598-026-59993-8

Abstract Neuromorphic computing based on artificial synapses requires devices capable of gradual, repeatable, and energy-efficient conductance modulation. Ionically gated transistors are promising candidates because their ion dynamics naturally produce synaptic behavior under low-voltage operation. However, how key device characteristics—conductance range, number of accessible conductance states N , and weight-update nonlinearity β —jointly influence neural network performance remains insufficiently understood. Here, we investigate MoS 2 -based ionically gated synaptic transistors using a combined experimental and modeling framework that links device physics to hardware-aware artificial neural network (ANN) simulations across image-classification tasks of varying complexity. We show that under fixed-amplitude pulsing, increasing N introduces a fundamental trade-off: finer weight resolution is accompanied by stronger update nonlinearity. ANN simulations further reveal that, within the nonlinearity range studied here, classification accuracy is governed by a task-dependent optimal weight resolution rather than a simply maximized number of states. To overcome the nonlinear weight updates, we employ a physics-informed transient model to develop a predictive pulse-engineering algorithm and experimentally demonstrate that it can linearize synaptic weight evolution in the same device. These linearized updates improve ANN accuracy by 1.5%–5.2% for MNIST, 4.0%–5.2% for FMNIST, and 1.2%–12% for KMNIST across the tested state numbers, establishing a quantitative link between device-level dynamics and neural network performance in ionically gated synaptic transistors.

The unglamourous AI wins that are quietly transforming drug discovery and development

Nature Reviews Drug Discovery Asher Mullard Jul 01, 2026 DOI: 10.1038/d41573-026-00099-1

Pan-RAS inhibitor daraxonrasib shows promise in pancreatic cancer

Nature Reviews Clinical Oncology David Killock Jul 01, 2026 DOI: 10.1038/s41571-026-01162-x

Daily briefing: What’s really happening with trust in science

Nature Flora Graham Jul 01, 2026 DOI: 10.1038/d41586-026-02084-5

Shape Memory Effect Controlled by Localized Flow of Liquid‐Like Atoms in Metallic Glass

Advanced Materials Bo Huang, Shansi Liao, Jun Yi et al. Jul 01, 2026 DOI: 10.1002/adma.73782

ABSTRACT Shape memory effect (SME) usually exists in crystalline alloys, polymers, ceramics, and their composites, induced by phase changes like Martensitic transformation and glass transition. Here, we report SME in metallic glasses (MGs) far below the glass transition temperatures. It is found that the recoverable strain of the trained MG, which reflects SME, increases with training temperature, stress, and time. The SME is interpreted as an intrinsic phenomenon of MGs, in which the local liquid‐like regions (LLRs) with viscosities of ∼10 10 Pa s and surrounding elastic matrix serve as the switching and memorizing segments, respectively. The activation of the LLRs, affected by deformation temperature and time in an Arrhenius‐type relation, involves the nonaffine displacement of solute atoms and the rearrangement of unstable medium‐range ordered (MRO) structure. In contrast, the recovery of the matrix containing more rigid MRO string‐like structure is mainly contributed by the reversible variation of short‐range atomic distance. Due to the instability of the LLRs, a transition from rejuvenation to aging appears during the shape memory (SM) process. The results reveal a new mechanism of SME linking to the dynamic structural heterogeneity of MGs and could be insightful for designing new types of glassy SM functional materials.

Harnessing Water Molecules for Strong Underwater Adhesion

Advanced Materials Biaolong Ma, Yi Yang, Jiqiang Wang et al. Jul 01, 2026 DOI: 10.1002/adma.73571

ABSTRACT Underwater adherable materials have attracted extensive attention due to their wide range of applications in marine exploration. However, current underwater adhesives still face a trade‐off between interfacial bonding and cohesive strength, limiting their underwater bonding performance. Herein, a water‐induced, toughening underwater adhesive to resolve the contradiction was reported, which can transform water molecules from barrier to reinforcement. The adhesive contains hydrophilic and hydrophobic units, which allow it to adaptively adhere to rough surfaces at as‐prepared soft state. Moreover, the strong interaction between polymer chains can be in situ activated by water molecules, leading to phase separation and enhancing mechanical properties of the adhesive in aqueous environments. This enables the adhesive to simultaneously achieve robust interfacial bonding and high energy dissipation during peeling. This strategy endows the adhesive with a high underwater interfacial toughness of 1602 ± 85 J m −2 . The adhesive is also transparent, recyclable, and long‐term stable. The underwater adhesion enhancement mechanism provides a new way to design underwater adhesive materials.

The Rho guanine-nucleotide exchange factor P-Rex2 exhibits structural and regulatory features distinct from the related RhoGEF P-Rex1

Journal of Biological Chemistry Lauren K. Anderson, Rohan Marde, Grace Muma et al. Jul 01, 2026 DOI: 10.1016/j.jbc.2026.113229

Real-world performance of a multimodal cell-free DNA multi-cancer early detection test in Asian populations.

Journal of Clinical Oncology Dang Luu Hong Nguyen, Thuy Phuong Le, Ngoc Minh Phan et al. Jul 01, 2026 DOI: 10.1200/jco.2026.44.19_suppl.14

14 Background: Multi-cancer early detection (MCED) addresses a critical unmet need by enabling simultaneous screening for multiple malignancies, particularly those lacking standard-of-care (SOC) options. SPOT-MAS is a liquid biopsy assay integrating methylomic, fragmentomic, and genomic signatures of cell-free DNA to detect signals across 10 cancer types. Building on its validation in the K-DETEK trial (NCT05227261), this study evaluates the clinical performance and feasibility of SPOT-MAS in a large-scale, real-world cohort across six Asian countries. Methods: Plasma cfDNA samples were analyzed from 84,145 individuals undergoing SPOT-MAS testing. Of these, 22,597 asymptomatic participants who completed a mandatory 12-month follow-up were included in the analysis. Participants with positive ctDNA signals underwent standardized post-test consultation and diagnostic workup, including imaging and/or histopathology, according to a predefined protocol to establish clinical status. Results: SPOT-MAS identified 94 positive cases (0.42%). Diagnostic workup confirmed 64 malignancies or precancerous lesions, including 20 cases (31.2%) diagnosed with cancers lacking SOC screening options, yielding a PPV of 68.1%. The assay demonstrated a Sensitivity of 79.0%, Specificity of 99.9%, NPV of 99.9% and Tissue-of-Origin accuracy of 80.0%. These performance metrics are consistent with the results of the K-DETEK trial. Conclusions: To the best of our knowledge, this study provides the first large-scale real-world evidence from Asia supporting the robustness and clinical utility of SPOT-MAS as a complementary screening strategy, particularly in LMICs lacking accessible national screening programs. Diagnostic performance comparison of the SPOT-MAS multi-cancer early detection test between the prospective K-DETEK trial and the expanded real-world data cohort. Test performance K-DETEK Real-World Data N = 9,024 % (95%CI) N = 22,597 % (95%CI) ctDNA signal detected 43 0.48 94 0.42 True positive 25 0.28 64 0.28 - Precancerous lesions 8 0.09 32 0.14 - Cancerous lesions 17 0.19 32 0.14 False positive 18 0.20 30 0.13 ctDNA signal not detected 8,981 99.52 22,503 99.58 True negative 8,974 99.45 22,486 99.51 False negative 7 0.08 17 0.08 - Precancerous lesions 0 0.00 7 0.03 - Cancerous lesions 7 0.08 10 0.04 Sensitivity 78.12 (61.24–88.98) 79.01 (68.93–86.46) Specificity 99.80 (99.68–99.87) 99.87 (99.81–99.91) Positive predictive value 58.14 (43.33–71.62) 68.09 (58.11–76.64) Negative predictive value 99.92 (99.84–99.96) 99.92 (99.88–99.95) Prediction accuracy of TOO 84.00 (65.35–93.60) 79.69 (68.29–87.73) Abbreviations: CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; TOO, tissue-of-origin. All 95% CIs were calculated using the Wilson score method. RWD cohort includes participants who completed ≥12 months of clinical follow-up.

Effect of VAV1 on metastasis and immune evasion in squamous cell carcinoma.

Journal of Clinical Oncology Taonong Cai Jul 01, 2026 DOI: 10.1200/jco.2026.44.19_suppl.228

228 Background: Cancer progression is shaped by continuous adaptation to environmental constraints. While metastatic dissemination and immune evasion are often studied independently, whether tumors dynamically switch between these programs under therapeutic pressure remains unclear. Methods: Using a genome-wide CRISPR–Cas9 screen in murine squamous cell carcinoma (SCC), combined with in vivo immune checkpoint blockade models, extracellular vesicle profiling, proteomics, and functional immunological assays, we interrogated molecular determinants governing tumor state transitions under immune pressure. Results: We identify VAV1 as a critical molecular switch that bifurcates tumor evolutionary trajectories. Elevated VAV1 expression drives epithelial–mesenchymal transition and metastatic competence, defining an invasive tumor state. However, upon immune checkpoint blockade and CD8⁺ T cell–mediated cytotoxic stress, tumor cells rapidly downregulate VAV1 and transition into an immune-resistant state. Mechanistically, VAV1 constrains extracellular trafficking of GSDMD-N through an HSPB1-dependent complex. Loss of VAV1 unleashes exosomal export of GSDMD-N into the tumor microenvironment, where it induces neutrophil pyroptosis and robust neutrophil extracellular trap (NET) formation. NET accumulation suppresses CD8⁺ T cell effector function and promotes T cell exhaustion, thereby establishing a protective immunosuppressive niche that blunts therapeutic efficacy. These findings reveal that tumor cells do not merely survive immune pressure—they reprogram their microenvironment to neutralize it. Conclusions: Our findings reveal that VAV1 functions as a strategic regulator of tumor evolutionary behavior. During early tumorigenesis, elevated VAV1 expression confers metastatic competence, enabling rapid dissemination in a manner analogous to an aggressive expansion strategy. In contrast, under immune checkpoint blockade, selective pressure imposed by activated CD8⁺ T cells drives VAV1 downregulation, reprogramming tumor cells toward an immune-resistant state characterized by GSDMD-N–mediated NET formation and microenvironmental suppression. Thus, VAV1 operates as a molecular decision node that coordinates a switch between metastatic expansion and immune defense. Rather than being passive targets of therapy, tumor cells dynamically recalibrate their signaling circuitry to optimize survival under changing constraints. This adaptive capacity underscores the evolutionary sophistication of cancer and identifies the GSDMD-N–NETs axis as a tractable vulnerability for disrupting tumor immune resilience.

Adsorption and Sulfur-Selective Photooxidation of Cysteine on Anatase TiO <sub>2</sub> (101)

Journal of the American Chemical Society Miguel Blanco Garcia, Daniele Perilli, Chiara Daldossi et al. Jul 01, 2026 DOI: 10.1021/jacs.6c07370

Multimodal MRI-based quantification of tumoral spatial heterogeneity for preoperative upgrade prediction of ductal carcinoma in situ

Scientific Reports Nianshi Song, Lingting Kong, Binying Zhuo et al. Jul 01, 2026 DOI: 10.1038/s41598-026-60371-7

Abstract This retrospective study aimed to develop a multimodal habitat-based model integrating clinicopathological factors, radiomics features, and intratumoral heterogeneity (ITH) score for predicting upstaging in patients with ductal carcinoma in situ (DCIS). A total of 389 DCIS patients from two institutions were included. Lesions were segmented on DCE-MRI, DWI, and ADC images with peritumoral expansion, and intratumoral and peritumoral habitats were identified using Gaussian mixture model (GMM) clustering. Radiomics features were extracted from each habitat, and ITH scores were calculated. Independent predictors of upstaging were identified using multivariate logistic regression, and multiple models were constructed and evaluated. Age, nuclear grade, palpable mass, and DCE-ITH score were independent risk factors for DCIS upstaging. Habitat-based models derived from intratumoral and peritumoral regions demonstrated good predictive performance, while the combined model integrating habitat features, ITH score, and clinicopathological variables achieved the best performance, with AUCs of 0.945 and 0.869 in the training and test cohorts, respectively. This multimodal combined model may facilitate individualized preoperative risk assessment and support clinical decision-making in DCIS patients.

Upcoming FDA approval decisions in Q3 2026

Nature Reviews Drug Discovery Swetha Kota Jul 01, 2026 DOI: 10.1038/d41573-026-00103-8

Liquid biopsy to inform adjuvant decisions during neoadjuvant chemotherapy — a full-circle moment for nasopharyngeal cancer

Nature Reviews Clinical Oncology W. K. Jacky Lam, Brigette B. Y. Ma Jul 01, 2026 DOI: 10.1038/s41571-026-01157-8

Enthalpy‐Driven Topological Programming of (TPMS)‐Like Carbon Networks

Advanced Materials Jiacheng Ma, Zhengwang Liu, Pengyuan Zhu et al. Jul 01, 2026 DOI: 10.1002/adma.73743

ABSTRACT Deterministic control over pore topology remains a central bottleneck in porous carbons, limiting the ability to translate molecular design into predictable electromagnetic attenuation and coupled thermal functions. Herein, we introduce an enthalpy‐driven topological programming paradigm in which the bond‐enthalpy of energetic N‐N’ fragments acts as a quantitative dial to steer self‐propagating reconstruction of coordination frameworks into a continuous sequence of (TPMS)‐like bicontinuous architectures. This programmable topology simultaneously establishes impedance‐matched, multi‐scattering pathways for wave ingress and concentrates heterogeneous interfaces that promote coupled dielectric and magnetic dissipation via vortex‐like magnetic textures and interfacial charge accumulation. As a result, the optimized Co@1,2,3,4‐NC delivers a minimum reflection loss of −53.97 dB with an effective absorption bandwidth of 7.84 GHz at 15 wt% loading. Beyond electromagnetic performance, the same bicontinuous topology suppresses heat transport by intensifying phonon scattering across hierarchical boundaries, enabling an ultralight and hydrophobic aerogel prototype that integrates electromagnetic shielding with thermal insulation. More broadly, bond‐enthalpy‐encoded topology control provides a transferable route to program bicontinuous porous networks across material chemistries, bridging thermodynamic driving forces with topological invariants for multifunctional matter.

Ultrasonic Vibration‐Driven Cold Upcycling of Waste Plastics: Enabling High‐Efficiency Seawater Distillation

Advanced Materials Shengyu Zhao, Shike Huang, Shuzhe Zhang et al. Jul 01, 2026 DOI: 10.1002/adma.73562

ABSTRACT Plastic pollution has become a pervasive environmental crisis, intensifying greenhouse gas emissions, contaminating soils, and potentially aggravating cardiovascular risks. Prevailing recycling strategies such as incineration, landfilling, and mechanical processing remain constrained by inefficiency, contamination, and excessive energy demand, resulting in limited reusability of the recovered materials. Therefore, developing an eco‐friendly, high‐performance technology that transforms plastic waste into valuable resources is imperative. Herein, we report a novel waste plastic upcycling strategy utilizing ultrasonic vibration cold fabrication to convert waste polyethylene terephthalate into carbon materials. The polyethylene terephthalate was treated via a combination of ultrasonic vibration and dissolution methods to produce highly carbonized, porous materials with high photothermal efficiency that enable high‐performance seawater distillation. High‐frequency stress treatment promotes rapid polymer chain scission, deoxygenation, and aromatization, facilitating heat‐source‐free carbonization. We further revealed that porous carbonized materials exhibit exceptional broadband photothermal conversion with an efficiency exceeding 95%. In seawater distillation, the material generated 2.19 kg m −2 h −1 of solar steam and maintained stability in high‐salinity conditions. Leveraging the flexible fabrication capability of these porous carbons, we constructed a solar evaporation device and demonstrated a freshwater yield of 6.02 kg m −2 in 10 h of outdoor testing, which is 5.2 times that of natural evaporation.

Titin cleavage in living cardiomyocytes induces sarcomere disassembly but does not trigger cell proliferation

Journal of Biological Chemistry Maria Rosaria Pricolo, Miguel A. López-Unzu, Natalia Vicente et al. Jul 01, 2026 DOI: 10.1016/j.jbc.2026.113167

Perioperative analgesic approaches in hepatobiliary cancer malignancies: Experience from a tertiary care centre.

Journal of Clinical Oncology Surabhi Sandill, Seema Mishra Jul 01, 2026 DOI: 10.1200/jco.2026.44.19_suppl.156

156 Background: Peri-operative analgesic management in hepatobiliary cancer surgeries is crucial for enhancing patient recovery and outcomes. This study presents insights from a tertiary care centre practice, highlighting the effective analgesic strategies tailored to the complex needs of patients with hepatobiliary malignancies. Methods: Retrospective observational study of hepatobiliary cancer patients operated between 1st June, 2020 and 31st May, 2025. Data collected using medical records and hospital database. The aim of our study was to assess perioperative analgesic strategies employed in hepatobiliary malignancy surgeries. Results: A total of 412 patients were operated for hepatobiliary malignancies between 1st June, 2020 and 31st May, 2025. The mean age was 52 years with 51.2% male patients. Most of the patients belonged to ASA-II (50.48%) and ASA-III (33.98%). The most common surgery performed was radical cholecystectomy (31.7%), whippers procedure (29.8%), palliative procedures (14.6%). Commonly employed intra-operative anaesthesia plan was general with neuraxial anaesthesia (69.17%) - epidural anaesthesia (54.77%). Lower NRS scores (both on rest and movement) on post operative day 1, 2 were observed in patients who received neuraxial block (p&lt;0.05) in comparison to those who receive only GA or GA with regional block. Common complications seen with neuraxial analgesia were vomiting, urinary retention, pruritis (p&gt;0.05). The postoperative outcomes including hospital stay, surgical complications and 30-day mortality showed no statistical difference between different perioperative analgesic modalities (p&gt;0.1). Conclusions: In our experience at a tertiary care centre, epidural analgesia emerged as a highly effective perioperative analgesic approach in patients undergoing surgery for hepatobiliary malignancies. It provided superior pain control post-operatively and contributed to early recovery and improved postoperative outcomes.

Efficacy of exercise-based interventions for the management of chemotherapy-induced peripheral neuropathy in patients with breast cancer: A systematic review and meta-analysis.

Journal of Clinical Oncology Melisha Koirala, Aakash Pandit Jul 01, 2026 DOI: 10.1200/jco.2026.44.19_suppl.53

53 Background: Chemotherapy-induced peripheral neuropathy (CIPN) is a common and dose-limiting side effect in breast cancer patients, particularly those receiving taxane-based regimens. CIPN manifests as sensory, motor, and autonomic dysfunction, significantly impacting quality of life and physical function. While pharmacological treatments often yield limited efficacy or adverse effects, exercise-based interventions have emerged as a potential non-pharmacological strategy. This meta-analysis aims to evaluate the efficacy of exercise in reducing the symptom burden of CIPN in breast cancer patients. Methods: A systematic review and meta-analysis of randomized controlled trials (RCTs) were conducted. Eligible studies investigated the impact of various exercise modalities (e.g., aerobic, resistance, or sensorimotor training) compared to standard care on CIPN symptoms in breast cancer patients. Primary outcomes were sensory, motor, and autonomic symptom scores, as well as overall CIPN burden. Data were pooled using a random-effects model. Standardized Mean Differences (SMD) and 95% Confidence Intervals (CI) were calculated to account for the use of diverse assessment scales. Heterogeneity was assessed using the I 2 statistic. Results: Eight RCTs involving 722 patients were included. Exercise interventions resulted in a statistically significant reduction in sensory symptoms (SMD -1.61; 95% CI -2.59 to -0.63; P = 0.001) and motor symptoms (SMD -1.81; 95% CI -2.96 to -0.65; P = 0.002), although heterogeneity was high for both (I 2 = 96% and 95%, respectively). Autonomic symptoms were also significantly reduced (SMD -0.69; 95% CI -1.10 to -0.28; P = 0.001), with notably low heterogeneity (I 2 = 0%). Furthermore, the overall CIPN burden was significantly lower in the exercise group compared to controls (SMD -1.34; 95% CI -2.16 to -0.53; P = 0.001; I 2 = 91%). While effect sizes varied across studies, the direction of effect consistently favored exercise interventions. Conclusions: Exercise-based interventions significantly alleviate sensory, motor, and autonomic symptoms of CIPN and reduce the overall neuropathy burden in breast cancer patients. These findings support the integration of structured exercise programs as a core component of supportive care for patients at risk of or experiencing CIPN. Given the high heterogeneity observed in sensory and motor outcomes, future research should focus on identifying the optimal exercise dose and modality for specific CIPN phenotypes.