Find the papers that actually matter

Oral cancer and oral potentially malignant disorders (OPMDs) remain a significant challenge in diagnosis and therapy, primarily due to inherent limitations in early detection, targeted treatment, and postoperative rehabilitation. Conventional diagnostic and therapeutic modalities often lack sufficient sensitivity, specificity, and effectiveness in restoring oral function. Biomaterials including nanoparticles, hydrogels, and scaffolds, offer versatile solutions by virtue of their tuneable properties, biocompatibility, and versatility in drug delivery and tissue engineering. However, their clinical translation is limited by the need for personalisation and lingering efficacy concerns. Artificial intelligence (AI) has emerged as a transformative approach to advance the design, optimisation, and application of biomaterials in oral oncology. By integrating machine learning (ML) and data-driven modelling, AI enhances diagnostic accuracy through biosensing and radiomic analysis, guides the rational design of drug carriers and dosing regimens, and facilitates computer-aided scaffold fabrication for maxillofacial reconstruction. This review summarises recent advances at the intersection of AI and biomaterials in the context of oral cancer and OPMDs, highlighting innovations in early detection, targeted therapy, and postoperative repair. It also discusses current barriers, including data quality, model generalizability, and regulatory oversight, and outlines future directions for interdisciplinary research. When properly integrated, AI-enabled biomaterials hold considerable potential to deliver more precise, efficient, and patient-tailored solutions for oral cancer management.

Cyclin-dependent kinase 4 and 6 inhibitors (CDK4/6i) combined with endocrine therapy are the preferred choice for first-line treatment of patients with HR+/HER2- locally advanced/metastatic breast cancer (aBC). The CDK4/6i ribociclib in combination with an aromatase inhibitor (AI) or fulvestrant (FUL) has demonstrated significant progression-free survival (PFS) and overall survival (OS) benefits for pre- and postmenopausal aBC patients who were enrolled in the three pivotal MONALEESA trials. Following the initial approval of ribociclib in 2017, the non-interventional RIBANNA study was initiated to evaluate the effectiveness and safety of ribociclib plus AI/FUL therapy among patients with aBC in a real-world setting. Two additional treatment cohorts (endocrine monotherapy [ET] and chemotherapy [CT]) were included to extend the knowledge about current aBC treatments. A total of 2567 patients were enrolled in 279 study centers, of whom 1852 were treated with ribociclib+AI/FUL, 183 were treated with ET, and 139 were treated with CT, who were available for effectiveness analyses. Median PFS (mPFS) and median OS (mOS) on first-line treatment with ribociclib+AI/FUL were 35.0 and 76.0 months, respectively. Adjustment for differences in demographic and baseline characteristics resulted in a longer mPFS on ribociclib+AI/FUL (34.7 months) compared to ET (26.4 months) or CT (19.2 months). Adverse events (AEs) on ribociclib were consistent with those seen in the pivotal trials, and no new safety signals were observed. The RIBANNA study confirmed the PFS and OS benefit seen in the MONALEESA trials. Together with the safety data, this large real-world dataset supports the favorable risk/benefit profile of ribociclib in large scale patient populations.

State-of-the-art radiotherapy machines with integrated magnetic resonance (MR) imaging, known as MR-Linacs, provide the capability to track tumors in real time. This capability aids delivery of precise irradiation in the presence of patient motion, such as breathing, by adjusting the radiation beam. However, current solutions rely solely on geometric tracking without closing the loop by considering the actual endpoint-the delivered radiation (here called dose). Real-time dose-based adaptation within a single session remains highly challenging due to the immense dimensionality of the problem. To overcome this, we have developed a radiotherapy simulator and propose a novel reinforcement learning (RL)-based approach for real-time adaptive radiotherapy using 2D fluence, as a surrogate to 3D dose. To our knowledge, this is the first application of RL in real-time adaptive radiotherapy. Our in-silico experiments showed the feasibility of using RL to close the feedback loop, dynamically adapting to patient motion and minimizing discrepancies between delivered and intended dose in clinical cases. Our approach introduces a new treatment delivery paradigm, enabling delivery based on a reference fluence and motion without predefined machine settings.

Peptide-drug conjugates (PDCs) represent a targeted cancer therapy strategy that combines tumor-homing peptides with potent cytotoxic payloads, offering a promising alternative to antibody-drug conjugates (ADCs) through improved tissue penetration, synthetic accessibility, and tumor selectivity. Auristatins (MMAE, MMAF, etc.), which are synthetic analogues of antimitotic dolastatin 10 (Dol-10), are widely used as ADC payloads; however, their systematic evaluation in PDC formats remains limited. In this study, we investigated Ahx-DA1, an enzymatically stable derivative of microtubule inhibitor DA1, a previously reported dolastatin-10 analogue, as a payload for PDCs. Two receptor-specific peptides, HER2-targeting peptide A9 and EGFR-binding peptide P6, were conjugated to a Ahx-DA1 and evaluated in the HER2-overexpressing breast cancer BT-474 model and the EGFR-overexpressing KRAS-mutated colorectal (HCT116) and pancreatic (PANC1) models, respectively. A cell-based study of DA1-bearing PDCs revealed specific and potent cytotoxicity in cancer cell lines, with the corresponding overexpressed receptors demonstrating high target specificity. The DA1-based PDCs exhibited high stability and favorable tolerability profiles across all the tested xenograft models. In vivo studies demonstrated pronounced tumor growth inhibition by A9-DA1 in HER2+ xenograft and P6-DA1 in EGFR+ KRAS mutated colorectal and pancreatic xenograft models. Overall, our findings suggest that Ahx-DA1 is a highly effective auristatin-class payload for the development of DA1 based anticancer PDCs.

B-cell lymphoma (BCL) is a hematological system malignant tumor with a relatively high incidence, and PI3Kδ and BTK play an important role in the development of BCL. In the preliminary investigation, we found that when the PI3K inhibitor and the BTK inhibitor were used in combination, the therapeutic effect was greater than that of single-drug administration at both cell and animal levels. Therefore, dual-target inhibitors of PI3Kδ and BTK were expected to potentially achieve improved therapeutic window for BCL. Here, we designed and synthesized 30 compounds, among which compound 27 showed high inhibitory activity against both targets at the kinase level (IC50-PI3Kδ = 9.0 nM, IC50-BTK = 17.3 nM). Furthermore, at the cellular level, the inhibitory activity of 27 against JeKo-1 and H9 cells (IC50-JeKo-1 = 1.6 μM, IC50-H9 = 5.8 μM) was comparable to or exceeded that of the positive drug alone and in combination. Western blot analysis confirmed that compound 27 potently suppressed phosphorylation of BTK, PI3Kδ and their downstream effectors. In addition, compound 27 showed reduced cytotoxicity in H9c2 cardiomyocytes (LD50 = 247.3 μM) compared to the positive. Preliminary pharmacokinetic studies in rats revealed favorable plasma exposure profiles. These preliminary results collectively identified compound 27 as a promising lead candidate for further development against BCL.

Previous studies have suggested that exposure to carcinogenic polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs) pollutants may increase the risk of colon cancer, their underlying molecular mechanisms remain unclear. In this study, we employed a multidisciplinary approach integrating network toxicology, machine learning, molecular docking, molecular dynamics (MD) simulations and in vivo experiments to investigate how PCDD/Fs may promote colon carcinogenesis. Machine learning algorithms converged on MMP7 as a core target, MMP7 expression was upregulated in colon cancer tissues and was associated with immune cell infiltration. Molecular docking and MD simulations further suggested stable interactions between the five representative PCDD/F congeners and the target proteins (MMP7, SRC, and HSP90AA1), supporting their potential involvement in disease progression. Consistent with these in silico findings, exposure of mice to 24 μg/kg TCDF significantly increased the expression of Mmp7 and Hsp90aa1 in murine colonic tissues, increased the levels of proinflammatory cytokines Ifn-γ, Il-1β, and Il-6, and downregulated the expression of Mucin 2 (MUC2). Connectivity Map analysis based on the PCDD/F-related gene signature identified five candidate compounds targeting MMP7 and HSP90AA1, of which four HSP90 inhibitors (tanespimycin, alvespimycin, NVP-AUY922 and AT-13387) showed negative connectivity scores, suggesting potential to reverse the pollutant-induced expression profile.

Radiomics-based machine learning models hold promise for clinical decision support, yet their deployment may be limited by the lack of transparent, prediction-level reliability assessment, especially under distributional shift. Existing uncertainty estimation methods mainly operate in probability space and may fail to identify unreliable predictions when test samples differ structurally or functionally from the training data. To address this gap, we propose the Radiomics Applicability Domain ANalysis - Classification Framework (RADANCF), a diagnostic approach for assessing the reliability of individual predictions in radiomics classification. RADANCF integrates six binary reliability criteria spanning two domains: data representativeness (A-C), describing the relationship between test samples and the training data manifold, and model behavior (D-F), capturing local inconsistencies in predictive responses. Criteria violations are aggregated into ordered reliability categories summarized using a qualitative traffic-light scheme. The framework was evaluated on six public radiomics datasets using five machine learning classifiers, resulting in 900 model configurations trained under a dissimilarity-based stratified partitioning strategy designed to challenge model generalization. Analyses included prediction-level error modeling, multiway ANOVA, correlation analysis between criteria, and assessment of frequently violated criterion combinations. External validation was performed on an independent cohort of 2689 prostate cancer patients from the ProCAncer-I project. Prediction error was significantly associated with RADANCF category, although the relationship was not strictly monotonic, with intermediate categories showing the largest error contributions. RADANCF criteria were largely complementary, as shown by low pairwise Spearman correlations (only 7.5% of cases with correlations higher than 0.5; p < 0.001). Multiway ANOVA confirmed RADANCF category as a significant factor after controlling for dataset and model effects (p < 10⁻¹²). Specific combinations of broken criteria-particularly A, B, C, and E-were significantly overrepresented among higher-error predictions (Wilcoxon test, p < 0.001). In external validation, correct predictions appeared across all traffic-light categories, confirming the diagnostic and risk-oriented nature of RADANCF. RADANCF provides a transparent, per-prediction diagnostic framework for assessing reliability in radiomics classification under distributional shift. By jointly accounting for data representativeness and model behavior, it complements traditional performance and uncertainty metrics and supports more cautious model deployment in radiomics-based models.

Early detection of pancreatic cancer remains a critical challenge in oncology, with current diagnostic methods often failing to identify the disease until advanced stages. However, diagnostic accuracy alone may be insufficient for clinical adoption as regulatory frameworks and clinical workflows increasingly demand transparent, interpretable AI systems. This study investigates Fourier Transform Infrared (FTIR) spectroscopy combined with machine learning for non-invasive pancreatic cancer detection using urine and blood biofluids, augmented by a language-model-assisted transparency framework to bridge spectral feature attributions and biochemical interpretation. Five datasets were evaluated: urine ATR-FTIR (61.7% balanced accuracy), urine transmission FTIR (74.8%), filtered blood (<10 kDa; 89.8%), and two matched urine-blood fusion datasets. Transmission-mode urine combined with filtered blood achieved the highest performance (96.9% balanced accuracy), exceeding either biofluid alone. To support transparency, we developed an LLM-augmented explainability pipeline incorporating Monte Carlo Tree Search (MCTS) for structured hypothesis exploration, a curated retrieval-augmented knowledge base (RAG), and reliability-gated explanations that acknowledge disagreement between feature attribution methods. Explainability methods showed substantial disagreement (mean Spearman ρ = 0.23-0.28), motivating a tiered strategy: wavenumber-level interpretation when methods agree (ρ ≥ 0.3, with knowledge base verification) and zone-level interpretation otherwise. These results highlight both the potential and current limitations of transparent spectroscopic diagnostics.

Iodine-131 (131I) therapy is a cornerstone of nuclear medicine for thyroid diseases and certain cancers. This review evaluates the transition from empirical to personalized dosimetry, focusing on image-based and AI-driven strategies. Following PRISMA guidelines, we analyzed studies from 1980 to 2025 across PubMed, Scopus, and other databases. We examine the core dosimetric frameworks used in 131I therapy, from fixed-activity protocols and the Medical Internal Radiation Dose (MIRD) schema to more sophisticated voxel-based techniques like Dose Point Kernels (DPKs), Voxel S-Values (VSVs), and full Monte Carlo simulations. Each method is discussed in terms of its strengths, limitations, and relevance to different clinical scenarios, especially in the context of metastatic thyroid cancer and neuroblastoma. Advanced methods such as voxel-based dosimetry and Monte Carlo simulations improve dose accuracy, while AI enhances segmentation and real-time planning. Personalized approaches show promise in optimizing therapeutic outcomes and minimizing toxicity, paving the way for standardized theranostic protocols. As the field increasingly incorporates AI, hybrid imaging, and personalized modeling, the objective is to deliver the right dose to the right patient at the right time with minimal risk. Standardizing these innovations across clinical practice will be key to ensuring safer, more effective treatments in the era of theranostics and personalized medicine.

Patients with recurrent/metastatic nasopharyngeal carcinoma (RM-NPC) do not yet have a strong recommended regimen after failure of first-line systemic therapy. This study retrospectively analyzed the efficacy of immunotherapy regimens in RM-NPC that failed at least first-line therapy. From February 2014 to August 2023, a total of 794 patients with RM-NPC were included in this study, of which 75 patients received anti-programmed cell death protein-1 (PD-1) only (P), 130 patients received anti-PD-1 plus antiangiogenic therapy (AP), 210 patients received anti-PD-1 plus single-agent chemotherapy (C1P), 276 patients received anti-PD-1 plus two-agent chemotherapy (C2P), and 103 patients received anti-PD-1 plus antiangiogenic plus chemotherapy (CAP). Progression-free survival (PFS), overall survival (OS), and adverse events were analyzed. In the inverse probability of treatment weighting (IPTW) cohort, median follow-up time was 28.7 months, median PFS in the P, AP, C1P, C2P, and CAP were 3.6, 8.5, 7.6, 12.7, and 16.3 months, respectively. PFS was significantly better in the CAP, C2P, C1P, and AP than P (p values of <.001, <.001, .026, and .002). Median OS in the P, AP, C1P, C2P, and CAP were 35.8, 46.6, 50.9, 50.6, and 47.2 months, respectively. Grade 3-4 treatment-related adverse events (TRAEs) occurred in 288 patients, with rates significantly higher in the C1P than AP (34.3% vs. 16.9%) and higher in the C2P than CAP (54.0% vs. 42.7%). These results showed that, in RM-NPC patients who failed at least first-line therapy, anti-PD-1 combination therapy was associated with improved PFS compared with anti-PD-1 monotherapy.