Diabetes mellitus (DM), the most prevalent metabolic disorder globally, has been extensively investigated due to its significant clinical implications. Among its most common complications is diabetic peripheral neuropathy (DPN), which predominantly manifests as bilateral limb pain, numbness, and paresthesia. In advanced cases, DPN may progress to foot ulceration and potentially necessitate limb amputation. The precise etiology and pathogenesis of diabetic neuropathy remain incompletely understood; however, sustained hyperglycemia, dysregulated lipid metabolism, and impaired insulin signaling are recognized as key triggers for the cascade of pathophysiological alterations in DPN. Under these metabolic disturbances, the peripheral nervous system’s structural and functional integrity—encompassing myelinated and unmyelinated axons, neuronal somata, neurovascular units, and glial cells—is progressively undermined. This review synthesizes recent mechanistic advances in DPN research, centering on the sigma-1 receptor’s modulatory orchestration, SGC dysfunction and GalCer metabolic dysregulation in DPN pathogenesis, inhibition of neuroinflammation and apoptosis by microRNA-146a, nerve damage due to Ca²⁺ overload, and dorsal root ganglia neuronal apoptosis resulting from SFA, and summarize possible interventions based on these mechanisms. A systems-level understanding of how these mechanisms collectively drive neuropathy pathogenesis is critical for advancing both early diagnostic tools and pathway-selective therapeutics.

In recent years, advancements in gene structure prediction have been significantly driven by the integration of deep learning technologies into bioinformatics. Transitioning from traditional thermodynamics and comparative genomics methods to modern deep learning-based models such as CDSBERT, DNABERT, RNA-FM, and PlantRNA-FM prediction accuracy and generalization have seen remarkable improvements. These models, leveraging genome sequence data along with secondary and tertiary structure information, have facilitated diverse applications in studying gene functions across animals, plants, and humans. They also hold substantial potential for multi-application in early disease diagnosis, personalized treatment, and genomic evolution research. This review combines traditional gene structure prediction methods with advancements in deep learning, showcasing applications in functional region annotation, protein-RNA interactions, and cross-species genome analysis. It highlights their contributions to animal, plant, and human disease research while exploring future opportunities in cancer mutation prediction, RNA vaccine design, and CRISPR gene editing optimization. The review also emphasizes future directions, such as model refinement, multimodal integration, and global collaboration. By offering a concise overview and forward-looking insights, this article aims to provide a foundational resource and practical guidance for advancing nucleic acid structure prediction research.

Recent increases in overweight and obesity have established Metabolic–Bariatric Surgery (MBS) as a principal intervention for durable weight reduction and metabolic improvement. Given the elevated perioperative complication risk among patients with obesity, there is a growing imperative to enhance the precision of preoperative management. This review synthesizes evidence from 2020–2025 on the application of artificial intelligence (AI) to bariatric surgery preoperative assessment, focusing on machine learning (ML), deep learning (DL), and large language models (LLMs). We summarize AI applications across three domains: preoperative risk prediction and individual assessment, patient stratification and procedure selection, and preoperative education and surgical training. The evidence suggests that integrating AI into routine preoperative workflows enables individualized, quantitative estimation of high-risk complications and weight-loss prognosis, thereby optimizing risk management; it can also support patient stratification and procedure matching to facilitate patient-centered shared decision-making; and NLP- and vision-based tools promote standardization and visualization of knowledge for patients and surgeons. Overall, AI is driving preoperative assessment toward greater individualization, interpretability, and multidisciplinary coordination, but its clinical generalizability requires validation in multicenter, prospective studies.

Renal cell carcinoma (RCC) is distinguished by a highly inflamed tumor microenvironment (TME) that offers both opportunities and challenges for immunotherapy. This review synthesizes current insights into the immunological landscape of RCC, highlighting robust cluster of differentiation 8‑positiv (CD8⁺) T-cell infiltration, unconventional antigen sources such as endogenous retroviruses and frameshift neoantigens, and the heterogeneity of immune niches revealed by single-cell and spatial profiling. We then examine the clinical impact and mechanisms of immune checkpoint inhibitors —including programmed cell death protein 1 (PD‑1), programmed death‑ligand 1 (PD‑L1), and cytotoxic T‑lymphocyte‑associated protein 4 (CTLA‑4)—tumor vaccines, cellular therapies such as chimeric antigen receptor T cell (CAR‑T) therapy and tumor‑infiltrating lymphocytes (TILs) and bispecific antibody constructs, emphasizing advances in dosing, engineering, and combination regimens. Combination strategies—including dual checkpoint blockade, integration with anti-angiogenic tyrosine kinase inhibitors, radiotherapy, metabolism-targeted agents such as adenosine and poly (ADP‑ribose) polymerase (PARP) inhibitors, and hypoxia modulators—are reviewed for their capacity to overcome resistance and remodel the microenvironment. We further explore intrinsic and acquired resistance mechanisms, the immunosuppressive roles of myeloid and stromal elements, and emerging biomarker approaches spanning genomic, transcriptomic, spatial, and circulating analytes. Finally, we discuss current limitations—such as variable clinical response, toxicities, and biomarker gaps—and outline future prospects, including personalized combination regimens, next-generation engineered cell products, and artificial intelligence (AI)-driven precision monitoring. Together, these insights chart a path toward more effective, individualized immunotherapy in RCC.

Chronic neuropathic pain affects 7–10% of the global population and imposes a substantial socioeconomic burden, yet existing pharmacological options remain inadequate in efficacy and burdened by systemic toxicity. Antibody-drug conjugates (ADCs) — engineered to deliver neuromodulatory payloads selectively to peripheral nociceptors — represent a compelling but unexplored therapeutic modality; no clinical-stage pain ADC currently exists. To the best of our knowledge based on a structured literature search (January 2000–March 2025) and targeted updates through 6 June 2026, this review provides a systematic and comprehensive conceptual framework mapping AI methodologies validated in oncology ADC development onto the unique challenges of chronic neuropathic pain. The framework spans five pipeline stages: multi-omics integration and graph neural networks for nociceptor-selective target prioritization (Nav1.7, TRPV1, P2X3, TrkA, CGRP receptor, ASIC3); structure prediction and protein language models for antibody engineering against transmembrane pain targets; generative AI for neuromodulatory payload design; deep learning for neural tissue pharmacokinetics and neurotoxicity prediction; and AI-driven patient stratification for precision clinical trial design. A unique three-layer optimization challenge — neural selectivity, analgesic potency, and systemic safety — distinguishes pain ADC AI from oncology ADC AI and defines the core design constraints of this review. Clinical precedent from FDA-approved anti-CGRP pathway antibodies (erenumab, fremanezumab, galcanezumab, eptinezumab) and Phase III anti-NGF antibody trials establishes that peripheral nociceptor-targeting biologics are both feasible and clinically active, providing the biological foundation for the pain-ADC concept. Realizing this framework requires curated pain-ADC datasets, interpretable AI, and closed-loop platforms incorporating dorsal root ganglion (DRG) organoids.