Search Results (2,050)

Background: Recurrent or metastatic head and neck squamous cell carcinoma (R/M HNSCC) is an aggressive malignancy with a historically poor prognosis. Although immune checkpoint inhibitors (ICIs) targeting the PD-1/PD-L1 axis have established a new first-line standard of care, durable clinical benefit is restricted to a minority of patients, while primary or acquired resistance remains a major clinical challenge. This narrative review aims to provide a conceptual clinical–biological synthesis of immunotherapy in R/M HNSCC through an anatomical and biomarker-driven lens, with a focus on characterizing potential shared features of “exceptional responders”. Methods: A targeted narrative literature search was conducted across PubMed/MEDLINE and to identify key clinical trials and representative clinical reports of exceptional response to ICIs in R/M HNSCC. The literature was not systematically synthesized to construct a conceptual clinical–biological framework of response and resistance. Results: Landmark clinical data confirm durable survival benefits with frontline pembrolizumab-based regimens in selected PDL-1 positive populations compared to historical chemotherapy. Qualitative appraisal of illustrative cases and translational cohorts suggests a potential biological hypothesis: exceptional responders, defined as patients achieving unexpected, multi-year complete radiological, pathological, or metabolic remissions, often align with a favorable confluence of a pre-existing “hot” or inflamed tumor microenvironment, preserved antigen presentation machinery, and elevated antigenic novelty driven by viral oncoproteins (HPV, EBV) or high mutational/indel burdens. Conversely, primary and acquired resistance are conceptually associated with defects in antigen processing, immunosuppressive cellular barriers, and alternative immune checkpoints. Conclusions: Immunotherapy has fundamentally transformed R/M HNSCC management, yet exceptional single-agent responses remain rare. Rather than a definitive or proven biological biomarker, the proposed blueprint represents an integrative hypothesis highlighting the complex interplay of baseline immune inflammation, genomic features, and viral drivers. Translating these observations into broader clinical benefit will require validated biomarker-driven personalization and rationally designed combination regimens. Full article

Background: Tumor hypoxia is a major determinant of radioresistance, limiting oxygen-mediated fixation of radiation-induced DNA damage and promoting metabolic adaptation, tumor aggressiveness, and treatment failure. Nitroimidazole derivatives and related hypoxia-targeting compounds have been investigated as radiosensitizers because of their oxygen-mimetic properties and selective activation under low-oxygen conditions. Methods: A structured narrative review was conducted to evaluate recent evidence on hypoxia-targeting strategies and nitroimidazole-based radiosensitizers combined with radiotherapy. PubMed and Scopus were searched in March 2026 using terms related to radiotherapy and nitroimidazoles. Studies published within the last five years were included if they investigated hypoxia-targeting compounds, radiosensitization strategies, or dose–response relationships relevant to radiotherapy. Preclinical, computational, and clinical studies were considered. Results: Sixteen studies were included, comprising preclinical, computational, and clinical investigations. Most studies were preclinical and evaluated in vitro cell lines, murine tumor models, or xenografts across multiple tumor types, including head and neck cancer, glioblastoma, breast cancer, colorectal cancer, cervical cancer, renal carcinoma, pancreatic cancer, and ovarian-related models. Nitroimidazole-based and related hypoxia-targeting strategies generally enhanced radiation response under hypoxic conditions, with sensitizer enhancement ratio values ranging from approximately 1.09 to 1.65. In vivo studies reported reductions in tumor growth or tumor volume when radiosensitizers were combined with radiotherapy. Clinical evidence, mainly in head and neck squamous cell carcinoma, showed variable effects on locoregional outcomes. Conclusions: Current evidence supports the biological relevance of hypoxia-targeting strategies for improving radiotherapy response. Nitroimidazole-based compounds show consistent radiosensitizing effects in preclinical models, but clinical translation remains heterogeneous. Future studies should integrate hypoxia imaging, standardized endpoints, advanced drug delivery systems, and clinically relevant radiotherapy models to better define their role in modern radiation oncology. Full article

Background: The immune landscape in head and neck squamous cell carcinoma (HNSCC) has been widely investigated. However, the crucial role played by the lymph node microenvironment in metastatic and non-metastatic HNSCC remains unknown. This systematic review aims to discuss the immunological crosstalk between the tumor and the nodal microenvironment and to describe the distribution of immune cells in metastatic and non-metastatic lymph nodes. Methods: A systematic review was conducted according to the PRISMA guidelines. PubMed, Scopus, and the Cochrane Library were searched for studies published between 1990 and 2025, with the final search performed in December 2025. Prospective and retrospective studies evaluating immune cell infiltration in metastatic and non-metastatic cervical lymph nodes were included, whereas studies focusing exclusively on non-cellular biomarkers and non-English publications were excluded. The risk of bias was assessed using the Newcastle–Ottawa Scale. The results were synthesized narratively, and no meta-analysis was performed. Results: The screening process identified 608 articles, of which 27 met our predefined inclusion criteria. These studies focused on macrophages, dendritic cells, neutrophils, natural killer cells, T helper cells, cytotoxic T cells, regulatory T cells, B cells, total lymphocytes, and surface markers. This systematic review provides a well-structured analysis of current knowledge on the impact of the innate and adaptive immune systems on the response against cancer cells and the recruitment of immune cells in lymph nodes. The most significant findings highlight the crucial role of antigen presentation in the antitumor response, particularly through the recruitment and activation of dendritic cells and subcapsular sinus macrophages in tumor-draining lymph nodes. It also presents in a fairly comprehensible manner that the density of mature dendritic cells, cytotoxic T cells, and B cells is higher in non-metastatic lymph nodes. Discussion: The lymph node microenvironment is highly enriched with immune cell infiltration, and their distribution between metastatic and non-metastatic lymph nodes can contribute to a better understanding of the underlying pathological processes. Particular attention should be given to the innate immune system cells and their implication in antigen presentation. Our findings suggest that identifying immunological profiles of lymph nodes may provide a rationale for treatment de-escalation protocols and raise the question of lymph node preservation in antitumor immune responses. However, the heterogeneity and bias assessment of the included studies warrant a cautious interpretation of these findings. Another important limitation is the limited number of studies comparing the immune microenvironment of primary tumors and lymph nodes. Full article

Background: Head and neck cutaneous squamous cell carcinoma (HNcSCC) presents intersecting oncologic, functional, and reconstructive challenges. Although numerous clinicopathologic factors have been associated with recurrence, metastasis, and survival, their relationship with reconstructive complexity and patient-centered outcomes remains insufficiently studied. This review aimed to examine how established oncologic risk factors might inform reconstructive planning while distinguishing measured reconstructive evidence from hypothesis-generating clinical inferences. Methods: A structured PubMed/MEDLINE search conducted through 15 July 2026 was used to identify the literature addressing clinicopathologic prognostic factors in HNcSCC. Twenty-nine prognostic publications were retained for structured charting. Additional reconstructive, functional, aesthetic, and patient-reported outcome sources were identified through reference-list screening and were used solely for narrative contextualization. Because no dedicated multi-database systematic search of reconstructive outcomes was performed, the article is presented as a structured narrative review and hypothesis-generating research framework rather than a systematic review of reconstructive evidence. Results: The prognostic literature reported associations between adverse oncologic outcomes and factors including tumor size and depth, perineural invasion, lymphovascular invasion, poor differentiation, immunosuppression, recurrent disease, positive margins, nodal involvement, and extranodal extension. However, most of these studies did not measure post-excision defect characteristics, reconstructive technique, wound complications, functional recovery, scar quality, aesthetic outcomes, or patient-reported outcomes. Direct reconstructive evidence was limited and predominantly derived from site-specific, mixed-histology, technical, or methodological publications. Consequently, clinicopathologic factors should be regarded as potential upstream variables for future investigation rather than validated predictors of reconstructive outcomes. Conclusions: Current evidence supports oncologic risk stratification more strongly than prediction of reconstructive difficulty or patient-centered outcomes in HNcSCC. Prospective studies should jointly measure patient, tumor, treatment-field, defect, reconstructive, functional, aesthetic, and patient-reported variables. The proposed framework is intended to guide such research and is not a validated clinical prediction model. Full article

Immune checkpoint inhibitors (ICIs) targeting PD-1/PD-L1 have improved outcomes across squamous cell carcinomas (SCCs) of the head and neck, lung, esophagus, and skin, yet durable responses remain confined to a subset of patients in every subtype. Objective response rates vary substantially across SCCs despite overlapping genomic alterations, comparable tumor mutational burden, and high PD-L1 expression, indicating that tumor-intrinsic biomarkers alone do not explain this variability. Growing evidence points to the tumor immune microenvironment, and in particular the myeloid compartment, as a critical determinant of immunotherapy responsiveness. In this review, we synthesize current evidence on myeloid-mediated immune regulation across SCC subtypes, focusing on tumor-associated macrophages, myeloid-derived suppressor cells/tumor-associated neutrophils, and dendritic cells, and the mechanisms by which these populations impair antigen presentation, restrict T cell infiltration, and sustain immunologically “cold” tumor states. We further examine therapeutic strategies aimed at reprogramming rather than simply depleting suppressive myeloid populations, including radiation therapy, STING agonism, and myeloid-targeted agents (CSF1R, PI3Kγ, and CXCR2 inhibition), each of which has shown encouraging preclinical and early clinical activity in combination with ICI. Collectively, this evidence supports a model in which the myeloid compartment functions as an actionable, convergent determinant of ICI resistance across SCC subtypes, rather than merely a passive biomarker. We propose that through the integration of spatial and single-cell profiling of myeloid states with clinical history it will be possible to predict response to immune checkpoint therapy and personalize myeloid-directed combination strategies, though the specific biomarkers needed to match individual patients to a given myeloid-targeted approach remain to be defined. We further discuss the toxicity considerations associated with both immune checkpoint blockade and radiation-based combination approaches, the early-phase status of most myeloid-targeted agents currently in clinical development, and the extent to which mechanistic insight, derived predominantly from HNSCC, generalizes to squamous cell carcinomas arising at other anatomic sites. Full article

Liquid biopsy, the analysis of tumor-derived material in blood, saliva, and other body fluids, is increasingly explored for the diagnosis, surveillance, and molecular characterization of head and neck squamous cell carcinoma (HNSCC). Its performance, however, is not uniform across the disease, and the reason is fundamentally molecular. human papillomavirus (HPV)-positive oropharyngeal cancers carry viral oncogenes that are absent from the host genome and therefore provide a near ideal, tumor specific circulating marker, whereas HPV-negative tumors are driven by a heterogeneous somatic landscape that offers no single universal target. In this narrative review, we adopt a molecular perspective. We first examine the biological origin of circulating tumor DNA and of the other analytes that liquid biopsy can interrogate including circulating tumor HPV DNA, viral transcripts, microRNAs, extracellular vesicles, and methylation signatures. We then consider how analytical platforms, from droplet digital PCR to next generation and ultrasensitive whole-genome sequencing, translate these molecules into measurements. Only afterward do we discuss the clinical questions, organized by clinical objective rather than by individual study: diagnosis and early detection, prognosis and risk stratification, treatment response monitoring, minimal residual disease and surveillance, and biomarker guided de-escalation in HPV-positive disease. Twelve registered clinical trials, involving approximately 1183 patients, are presented as illustrations of these questions. We close on the biological and technical gaps that still separate promising signals from clinical practice, and on the multi analyte and dynamic strategies most likely to bridge them. At present, liquid biopsy should be regarded as a complementary tool rather than as a replacement for established clinicopathological assessment. Full article

Extracellular vesicles (EVs) are crucial mediators of intercellular communication in the tumor microenvironment (TME) which facilitate the transfer of bioactive molecules including functional mitochondria and their integral components. This review summarizes the emerging role of EV-mediated mitochondrial transfer in cancer progression. We delineate the mechanisms governing the packaging of mitochondria and their constituents into EVs and subsequently highlight their multifaceted functions across various malignancies, including breast cancer, prostate cancer, blood malignancies, head and neck squamous cell carcinoma, digestive system cancers, etc. Mitochondrial cargo, such as intact mitochondria, mitochondrial DNA (mtDNA), and RNA (mtRNA), are shown to reconfigure metabolism, enhance bioenergetics, promote proliferation and invasion, induce drug resistance, and remodel TME by suppressing antitumor immunity. While previous reviews have predominantly focused on the role of mitochondrial transfer in individual cancers or specific systemic diseases, we made a comprehensive overview encompassing diverse cancer types. These findings suggest that EV-mediated mitochondrial cargo transfer represents a biological intercellular communication mechanism with implications for tumor progression and therapeutic resistance. It is worth noting that we also apply standardized evidence-grading frameworks (C1–C4) across cancer types to provide a critical assessment of the current evidence and identify key methodological gaps that must be addressed in future studies. Collectively, this review underscores the significance of EV-mediated mitochondrial transfer as an important biological process in cancer, presenting it as a promising frontier for novel diagnostic and therapeutic interventions. Full article

Background/Objectives: The emergence of radioresistance is a major challenge in cancer treatment. Herein, we evaluated the radiosensitizing effects and associated molecular mechanisms of a small molecule, berberine, in HNSCC. Methods: HNSCC cells were treated with berberine, ionizing radiation (IR), and their combination. Radiosensitivity was assessed using methods for clonogenic survival, proliferation, cell death, apoptosis, cell cycle distribution, DNA damage, and associated mechanisms. Spheroid models were employed to examine tumor growth and stemness-related markers. Therapeutic efficacy was further evaluated using the syngeneic MOC2 tumor syngraft mouse model. Results: The combination of berberine and IR significantly decreased the colony-forming ability, cell proliferation, and survival in UM-SCC-22B and MOC2 cells. Correspondingly, a decrease in the pro-survival signaling, EGFR, ERK1/2, mTOR, and STAT-3 was noted. The combination treatment increased sub-G1 cell population and apoptotic cell death with enhanced DNA damage, which was also associated with downregulation of DNA repair proteins. Spheroids showed disintegration and altered morphology in the combination treatment. This combination also inhibited Wnt/β-catenin signaling and reduced the expression of genes associated with stemness-related pluripotency. Berberine further enhanced DNA damage and reduced stemness in EGFR-knockdown cells. The berberine-IR combination not only inhibited the growth of syngeneic MOC2 tumors but also mitigated the toxicity associated with IR. The in vitro findings of reduced expression of stemness and DNA damage repair genes were also observed in tumors. Conclusions: Our findings indicated that berberine enhances the radiation response of HNSCC through perturbation of pro-survival signaling, DDR signaling and reduction of stemness-associated features, thereby enhancing the therapeutic efficacy. Full article

Conventional histopathological diagnostics rely on morphological assessment of stained tissue sections, requiring extensive sample preparation and subjective expert interpretation. Mid-infrared (MIR) imaging offers a complementary approach by providing spatially resolved, label-free access to the intrinsic biochemical composition of tissue without exogenous contrast agents. This work introduces a preprocessing and analysis pipeline for multiphotometric MIR data, applied to formalin-fixed, paraffin-embedded tissue sections from two patients with histopathologically confirmed head and neck squamous cell carcinoma. Combining differential scattering correction, sub-pixel channel registration, and automated tissue segmentation with unsupervised K-Means clustering, the pipeline achieves label-free discrimination of biochemically distinct tissue compartments. K-Means clustering identified four distinct clusters, of which three corresponded to tissue compartments with protein-to-lipid ratios tentatively consistent with epithelial, tumor-associated, and stromal compartments. The resulting cluster maps showed partial spatial correspondence with mIF reference stainings targeting epithelial and stromal markers, supporting the potential of this approach for label-free tissue characterization in digital pathology. Full article

Background/Objectives: Pretreatment sarcopenia predicts adverse outcomes in head and neck squamous cell carcinoma (HNSCC). The conventional L3 skeletal muscle index (SMI) requires abdominal imaging, whereas the predictive value of the more accessible C3 muscle index for acute treatment-related complications remains uncertain. We compared pretreatment L3 SMI and the C3 index for identifying patients at risk of severe oral intake impairment, radiation interruption, and a longer treatment-to-discharge interval during definitive chemoradiotherapy. Methods: This retrospective single-center study included 55 patients with HNSCC treated with 80 mg/m² cisplatin every three weeks and 66 Gy intensity-modulated radiotherapy. Pretreatment computed tomography (CT) was analyzed using deep-learning-based automated segmentation. Firth-penalized logistic and multivariable linear regression were adjusted for age, sex, performance status, body mass index, and stage, with estimates standardized per 1-standard-deviation (SD) increase. Results: L3 SMI was associated with severe oral intake impairment (adjusted odds ratio per SD, 0.248; 95% CI, 0.053–0.891) and the treatment-to-discharge interval (adjusted β, −6.23 days; 95% CI, −11.89 to −0.58); the C3 index was not associated with any outcome. The standardized coefficients differed for oral intake impairment (p = 0.014) but not for other outcomes. Each binary outcome comprised only 10 events. Conclusions: L3 SMI was associated with nutritional vulnerability, whereas the fully automated C3 index was not; in an exploratory direct comparison, the two indices differed for severe oral intake impairment. These findings are hypothesis-generating and require confirmation in larger, multicenter cohorts. Full article

Senescence-associated secretory phenotype (SASP) signaling, along with key markers such as P16^INK4a/CDKN2A and LMNB1, has not been systematically studied in circulating lymphocyte subsets in head and neck squamous cell carcinoma (HNSCC). This study aimed to evaluate SASP-related genes and senescence markers in peripheral CD4⁺, CD8⁺, and CD19⁺ cells and assess their clinical relevance. Expression of IL-6, IL-1β, TNFα, CXCL1, P16^INK4a/CDKN2A, and LMNB1 was measured by RT-qPCR in sorted lymphocytes from 58 HNSCC patients at baseline, 31 post-treatment, and 13 controls. Statistical analyses included nonparametric tests, correlation analyses, and survival models (Kaplan–Meier, Cox regression). In the results, LMNB1 was significantly upregulated in all lymphocyte subsets of HNSCC patients. IL-6, CXCL1, and IL-1β were elevated in CD4⁺ T cells. A coordinated co-expression network involving IL-6, CXCL1, IL-1β, P16^INK4a/CDKN2A, and LMNB1 was observed. Clinically, IL-6 in CD8⁺ T cells was associated with higher nodal stage and worse survival, while CXCL1 in CD19⁺ B cells independently predicted survival. No differences were found between pre- and post-treatment samples. Circulating lymphocytes in HNSCC display coordinated expression of selected senescence-associated genes, with IL-6 and CXCL1 as candidate prognostic biomarkers linked to tumor progression that warrant further validation. Full article

Currently, the global incidence of head and neck squamous cell carcinoma (HNSCC) continues to increase, with a large proportion of cases being diagnosed at advanced stages. Conventional oncological treatments are often associated with poor quality of life, while the aggressive biological behavior of these tumors results in recurrence in approximately half of treated patients according to current evidence. The need for rapid, targeted diagnosis and personalized treatment has driven research toward the identification of tumor biomarkers with diagnostic, prognostic, and predictive value for treatment response. In this scoping review, we performed a comprehensive literature search to map the available evidence on tumor biomarkers in HNSCC. We identified a wide range of biomarkers, several of which exhibit overlapping roles, and classified them according to their genomic, proteomic, cytokine-related, tumor microenvironment, metabolic, and microbiota-associated characteristics. This scoping review also summarizes the current evidence regarding biomarkers involved in carcinogenesis, treatment response and therapy resistance, while highlighting their current level of clinical applicability. Although only a limited number of biomarkers have currently been implemented in routine clinical practice in head and neck cancers, the studies included in this review identified several promising candidates with the potential to be incorporated into clinical practice following further prospective validation and methodological standardization. Many additional biomarkers are still under investigation, opening new perspectives for future diagnostic and therapeutic targets and supporting the development of personalized multimodal treatment strategies for patients with HNSCC. Full article

Background/Objectives: Immune exclusion in head and neck squamous cell carcinoma (HNSCC) limits immunotherapy efficacy, yet the molecular determinants of stromal-immune organization remain incompletely characterized. The desmosomal cadherin DSG2 is highly expressed in squamous epithelium; its role in shaping the tumor microenvironment (TME) is unknown. Methods: We integrated bulk RNA-seq from 836 HNSCC patients (TCGA-HNSC n = 566, GSE65858 n = 270), single-cell RNA-seq (GSE139324, n = 26 patients, 133,308 cells), spatial transcriptomics (GSE208253, n = 12), proteomics (CPTAC-HNSCC, n = 108), and external validation cohorts (GSE41613, n = 97). CellChat ligand-receptor analysis, mediation analysis, Mendelian randomization (MR), LASSO-penalized Cox regression, HPV-stratified sensitivity analysis, and transcription factor (TF) correlation analysis were employed. Results: DSG2 exhibited epithelial-specific expression and showed consistent positive correlation with CXCL8 (IL-8; TCGA ρ = 0.228, p = 4.4 × 10⁻⁸) and myCAF activation across independent cohorts. Single-cell analysis revealed that 99.5% of CXCL8-producing cells have zero DSG2 expression, establishing the bulk correlation as compositional rather than cell-intrinsic. CellChat identified CXCL8-CXCR2 as the strongest tumor-stroma interaction in DSG2-high regions (probability = 0.821, 1.80-fold enrichment). Mediation analysis demonstrated 43.6% (95% CI [34.3–53.6%]) of DSG2’s tissue-level association with myCAF activation is mediated through CXCL8 (compositional mediation). Multi-instrument MR (IVW: Beta = −0.028, p = 0.028; I² = 0.0%) corroborated the compositional model. Protein-level validation in CPTAC-HNSCC confirmed DSG2-CD8A inverse correlation (Spearman ρ = −0.35, p = 2.2 × 10⁻⁴). Pan-squamous meta-analysis confirmed negative DSG2-cytolytic activity correlations (pooled ρ = −0.213, 95% CI [−0.296, −0.128], I² = 58.6%, 4 cohorts). DSG2 correlated with TIDE score (ρ = 0.176) and TGF-β exclusion subscore (ρ = 0.428). DepMap analysis identified CXCR2 inhibitor collateral sensitivity (ρ = −0.408, p < 0.0001). An eight-gene co-expression module was validated in two independent cohorts (GSE41613: HR = 3.09, p = 0.003; GSE65858: HR = 1.57, p = 0.032). Conclusions: DSG2 marks a stromal-immune organizational state characterized by CXCL8-CXCR2 paracrine signaling, myCAF activation, and immune exclusion, conserved across squamous malignancies. DSG2-high/PD-L1-high tumors (30.4% prevalence) exhibit the worst predicted ICI response and represent a candidate population for biomarker-selected CXCR2 inhibitor trials in combination with anti-PD-1 therapy. Full article

Background/Objectives: Supraglottic and hypopharyngeal carcinomas are aggressive malignancies that are often diagnosed at advanced stages. Timely recognition of these malignancies is influenced by many factors, such as endoscope quality and experience. This study evaluated the potential of artificial intelligence (AI) to support real-time detection and classification of such lesions during flexible endoscopy in the outpatient clinic. Methods: A previously developed deep learning (DL) algorithm was extended from a glottic lesion model to the unified localization and classification of glottic, supraglottic, and hypopharyngeal lesions during flexible endoscopy. Lesion frames were extracted from endoscopy videos obtained at two head and neck oncology centers and one secondary referral center between 2012 and 2024. These frames were annotated and labeled based on histopathological or clinically confirmed reference diagnoses. The primary outcome was the unified model’s performance in detection of lesions per frame. After training (70% of data), the positive predictive value (precision) and sensitivity (recall) of this model were calculated on an independent test set (30% of data), stratified by subsite and tumor (T-) classification. Secondly, the model’s binary classification performance (benign or malignant) was evaluated. Results: From 490 supraglottic and hypopharyngeal endoscopy videos, 40,059 frames with a benign or malignant lesion were extracted and added to the 56,036 glottic lesion frames in the database, comprising 1336 lesion videos and 123 healthy control videos (total n = 1459). On the test set, the model achieved a detection precision of 92.1% (95% CI: 90.9–93.2) and a recall of 73.3% (95% CI: 69.8–76.6). Detection performance increased with higher T-stage. Among correctly detected lesion instances in a malignancy-dominated test set (65.3% of lesion instances), sensitivity for malignancy detection was 94.3% (95% CI: 91.8–96.5). Combined end-to-end performance for correct malignant lesion detection and classification was estimated at 69.1% (object-level recall 73.3% × classification sensitivity 94.3%). Evaluation of 33 lesion-free videos demonstrated a mean frame-level specificity of 46.6% and a median of 19 false positive detections per video. Conclusions: This is the first study to report a DL model for real-time endoscopic detection and classification of benign and malignant laryngeal and pharyngeal lesions. The developed model showed promising lesion detection and cancer classification performance in the evaluated test set. T1 tumor detection remains an important limitation, particularly because early-stage detection is a primary aim of AI-assisted endoscopy. Further model testing is required in real-world lesion prevalence settings, where external validation and clinical usability should be investigated. Full article

Objective: While advancements in radiotherapy and systemic agents have significantly improved survival rates in head and neck squamous cell carcinoma (HNSCC), managing long-term, treatment-induced toxicities remains a critical clinical challenge. This study aimed to develop a personalized, supervised machine learning-driven predictive model for multiple chronic toxicities by integrating clinical, dosimetric, and genetic data specifically evaluating the impact of the NBN gene rs1805794 (c.553G>C) polymorphism. Methods: This study enrolled 125 patients with HNSCC who received curative-intent radiotherapy and remained disease-free during follow-up with a median of 98 months. Comprehensive clinical and dosimetric data were collected, and chronic toxicities were recorded. Peripheral blood samples were analyzed for the NBN rs1805794 polymorphism using allele-specific PCR (AS-PCR). Following feature selection, four supervised machine learning classifiers were trained and evaluated to identify the optimal model for predicting multiple chronic toxicities. Results: The XGBoost algorithm emerged as the highest performing model. Baseline clinico-dosimetric predictors of multiple chronic toxicities included PTV70 volume, the addition of concurrent chemotherapy, advanced T and N stages, and continued smoking. Integrating the NBN rs1805794 genotype into the XGBoost architecture enhances its predictive capability. The final model accurately identified patients at high risk for multiple chronic toxicities, achieving an area under the curve (AUC) of 0.78, an accuracy of 0.77, a sensitivity of 0.74, and a specificity of 0.79. Conclusions: Integrating clinical, dosimetric, and genetic data within a machine learning framework effectively predicts multiple chronic toxicities in HNSCC. This approach enabled early risk stratification, providing the potential for personalized therapy. Full article