Adenomatoid odontogenic tumor (AOT) is a rare benign epithelial odontogenic tumor that typically occurs in young females and is most often associated with an impacted maxillary canine. Although usually managed via transoral enucleation, endoscopic endonasal surgery has emerged as a potential alternative for selected lesions.
Case presentation
A 25-year-old woman presented with progressive right-sided nasal obstruction and clear rhinorrhea. Nasal endoscopy revealed a smooth submucosal bulge arising from the right inferior meatus that filled the right nasal cavity. Computed tomography demonstrated a well-defined expansile lesion occupying the right maxillary sinus, containing internal calcifications and an impacted canine tooth. The tumor was completely removed using an endoscopic endonasal approach. Histopathology confirmed the diagnosis of AOT. At the 18-month follow-up, endoscopic examination showed complete healing without recurrence.
Discussion
Endoscopic endonasal surgery offers advantages over transoral approaches in selected cases, including improved visualization, avoidance of oroantral complications, absence of external incisions, and suitability for large maxillary sinus lesions.
Conclusion
This case demonstrates the feasibility of complete AOT removal using an endoscopic endonasal approach. Further reports with long-term follow-up are required to clarify indications and outcomes for this technique.
{"title":"Adenomatoid odontogenic tumor mimicking nasal pathology: A case report on endoscopic endonasal management","authors":"Kawita Atipas , Sataporn Ruangprasertkul , Suvajana Atipas","doi":"10.1016/j.oor.2026.100783","DOIUrl":"10.1016/j.oor.2026.100783","url":null,"abstract":"<div><h3>Background</h3><div>Adenomatoid odontogenic tumor (AOT) is a rare benign epithelial odontogenic tumor that typically occurs in young females and is most often associated with an impacted maxillary canine. Although usually managed via transoral enucleation, endoscopic endonasal surgery has emerged as a potential alternative for selected lesions.</div></div><div><h3>Case presentation</h3><div>A 25-year-old woman presented with progressive right-sided nasal obstruction and clear rhinorrhea. Nasal endoscopy revealed a smooth submucosal bulge arising from the right inferior meatus that filled the right nasal cavity. Computed tomography demonstrated a well-defined expansile lesion occupying the right maxillary sinus, containing internal calcifications and an impacted canine tooth. The tumor was completely removed using an endoscopic endonasal approach. Histopathology confirmed the diagnosis of AOT. At the 18-month follow-up, endoscopic examination showed complete healing without recurrence.</div></div><div><h3>Discussion</h3><div>Endoscopic endonasal surgery offers advantages over transoral approaches in selected cases, including improved visualization, avoidance of oroantral complications, absence of external incisions, and suitability for large maxillary sinus lesions.</div></div><div><h3>Conclusion</h3><div>This case demonstrates the feasibility of complete AOT removal using an endoscopic endonasal approach. Further reports with long-term follow-up are required to clarify indications and outcomes for this technique.</div></div>","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"17 ","pages":"Article 100783"},"PeriodicalIF":0.0,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145977287","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Artificial intelligence (AI) is increasingly integrated into pathology, but its accuracy in immunohistochemical (IHC) marker selection remains underexplored. This study evaluated LeChat's ability to recommend IHC markers for benign and malignant salivary gland tumors, focusing on accuracy, completeness, relevance, consistency, and marker-level errors across tumor types and subtypes.
Methods
A total of 21 tumor types were selected and classified by behavior (benign vs. malignant) and histologic subtype. Expert-derived reference panels served as the gold standard. For each tumor, LeChat was queried three times using standardized prompts. Recommendations were scored across three domains: accuracy (inclusion of essential markers), completeness (inclusion of secondary markers), and relevance (absence of irrelevant markers). Composite scores (range: 3–9) and intra-tumor variability were calculated. Mann–Whitney U and Kruskal–Wallis tests assessed differences by tumor category and subtype. Marker-level analysis evaluated over- and under-recommendations.
Results
Across 63 total prompts, LeChat achieved a mean accuracy of 1.56, completeness of 1.59, and relevance of 2.35. Only 3.2 % of prompts included all essential markers, and 17.5 % achieved composite scores ≥7. No responses included both primary and secondary markers. Performance did not differ significantly between benign and malignant tumors (p = 0.405), nor across histologic subtypes (p = 0.988). Tumor-level variability was highest for basaloid squamous cell carcinoma and lowest for pleomorphic adenoma. Marker-level analysis identified 14 false positives (e.g., CD20, IgG4) and 9 false negatives (e.g., SOX10, β-catenin). A moderate correlation was observed between marker frequency and accuracy (r = 0.48, p < 0.01).
Conclusion
LeChat's IHC recommendations were generally relevant but frequently incomplete and inconsistent, particularly for malignant and histologically complex tumors. Despite its potential for assisting in straightforward cases, current performance remains inferior to expert standards. These findings highlight the limitations of general-purpose LLMs in pathology workflows and emphasize the need for domain-specific refinement before clinical integration.
{"title":"LeChat in oral pathology: Assessment of its immunohistochemical marker selection accuracy in salivary gland tumors","authors":"Maria Cuevas-Nunez , Cosimo Galletti , Javier Flores-Fraile , Cosimo Galletti , Shokoufeh Shahrabi Farahani , Wilmer Rodrigo Díaz-Castañeda , Daniele Portelli , Luca Fiorillo , Vini Mehta , Maria-Teresa Fernández-Figueras","doi":"10.1016/j.oor.2025.100781","DOIUrl":"10.1016/j.oor.2025.100781","url":null,"abstract":"<div><h3>Background</h3><div>Artificial intelligence (AI) is increasingly integrated into pathology, but its accuracy in immunohistochemical (IHC) marker selection remains underexplored. This study evaluated LeChat's ability to recommend IHC markers for benign and malignant salivary gland tumors, focusing on accuracy, completeness, relevance, consistency, and marker-level errors across tumor types and subtypes.</div></div><div><h3>Methods</h3><div>A total of 21 tumor types were selected and classified by behavior (benign vs. malignant) and histologic subtype. Expert-derived reference panels served as the gold standard. For each tumor, LeChat was queried three times using standardized prompts. Recommendations were scored across three domains: accuracy (inclusion of essential markers), completeness (inclusion of secondary markers), and relevance (absence of irrelevant markers). Composite scores (range: 3–9) and intra-tumor variability were calculated. Mann–Whitney U and Kruskal–Wallis tests assessed differences by tumor category and subtype. Marker-level analysis evaluated over- and under-recommendations.</div></div><div><h3>Results</h3><div>Across 63 total prompts, LeChat achieved a mean accuracy of 1.56, completeness of 1.59, and relevance of 2.35. Only 3.2 % of prompts included all essential markers, and 17.5 % achieved composite scores ≥7. No responses included both primary and secondary markers. Performance did not differ significantly between benign and malignant tumors (p = 0.405), nor across histologic subtypes (p = 0.988). Tumor-level variability was highest for basaloid squamous cell carcinoma and lowest for pleomorphic adenoma. Marker-level analysis identified 14 false positives (e.g., CD20, IgG4) and 9 false negatives (e.g., SOX10, β-catenin). A moderate correlation was observed between marker frequency and accuracy (r = 0.48, p < 0.01).</div></div><div><h3>Conclusion</h3><div>LeChat's IHC recommendations were generally relevant but frequently incomplete and inconsistent, particularly for malignant and histologically complex tumors. Despite its potential for assisting in straightforward cases, current performance remains inferior to expert standards. These findings highlight the limitations of general-purpose LLMs in pathology workflows and emphasize the need for domain-specific refinement before clinical integration.</div></div>","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"17 ","pages":"Article 100781"},"PeriodicalIF":0.0,"publicationDate":"2026-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145927135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-17DOI: 10.1016/j.oor.2025.100780
D. Morgan , E. Schueddig , A. Segura , T. Bodine , C. Lominska , R.T. Morse , P. Neupane , D. Pei , D. Koestler , R. Kimple , U. Duvvuri , S. Thomas , R. Nallani , L. Shnayder , K. Kakarala , A.M. Bur , E. Yilmaz , G.N. Gan
Purpose
Recurrence in early-stage laryngeal carcinoma after radiation therapy often necessitates a total salvage laryngectomy (TSL). Understanding the genetic landscape that influences cancer progression after radiation therapy may help future treatments. We investigated early-stage laryngeal carcinoma patients treated with radiotherapy to identify mutation patterns that predispose to disease recurrence.
Methods
The patient cohort had 35 early-stage laryngeal head and neck squamous cell carcinoma (HNSCC) patients (T1 = 15, T2 = 20) treated with radiotherapy (RT). We stratified patients: Responders (no recurrence, N = 14), non-responders (recurrence within 12 months, N = 21), and non-responders undergoing TSL (N = 18). We employed whole exome sequencing to characterize gene mutations using the Genome Analysis Toolkit (GATK) for variant detection and impact on critical biological pathways and post-radiotherapy patient outcomes. Pathway analysis Ingenuity and Reactome Pathway Analysis tools were used to explore the mutated gene pathways.
Results
– Differential mutation analysis was performed in the respective groups to find driver genes. In the pre-treatment samples, KCNT2 and AGAP6 mutations were exclusively found in non-responders (OR = 0, P = 0.005 and OR = 0, P = 0.027, respectively), while ADAMTS7 mutations were solely present in responders (OR = inf, P = 0.019). PLEC mutations were more prevalent in responders (OR = 11.6, P = 0.006). Pathway analysis revealed that significant genes were involved in the RND2 GTPase cycle, protein O-glycosylation-related diseases, and apoptotic pathways. Post-treatment analysis in patients undergoing TSL had enrichment of mutations in apoptosis regulation pathways.
Conclusions
– The study reveals that mutations in apoptosis controlling genes were predominantly represented in the larynx non-responder patients both in the pre-radiotherapy and post-TSL populations.
目的早期喉癌放射治疗后复发往往需要全喉切除术（TSL）。了解影响放射治疗后癌症进展的基因景观可能有助于未来的治疗。我们调查了早期喉癌放疗患者，以确定易导致疾病复发的突变模式。方法35例早期喉头颈部鳞状细胞癌（HNSCC）患者（T1 = 15, T2 = 20）行放射治疗（RT）。我们对患者进行了分层：反应者（无复发，N = 14），无反应者（12个月内复发，N = 21），以及接受TSL的无反应者（N = 18）。我们采用全外显子组测序来表征基因突变，使用基因组分析工具包（GATK）进行变异检测和对关键生物学途径和放疗后患者预后的影响。通路分析使用Ingenuity和Reactome通路分析工具探索突变基因通路。结果-在各自的组中进行差异突变分析以寻找驱动基因。在预处理样本中，KCNT2和AGAP6突变仅存在于无应答者中（OR = 0, P = 0.005和OR = 0, P = 0.027），而ADAMTS7突变仅存在于应答者中（OR = inf, P = 0.019）。应答者中PLEC突变更为普遍（OR = 11.6, P = 0.006）。通路分析显示，重要基因参与RND2 GTPase循环、蛋白o糖基化相关疾病和凋亡通路。TSL患者治疗后分析细胞凋亡调控通路突变富集。结论-该研究表明，在放疗前和tsl后的喉部无反应患者中，细胞凋亡控制基因的突变主要代表。
{"title":"Role of molecular driver mutations in recurrence of early-stage laryngeal cancer following narrow field definitive radiotherapy","authors":"D. Morgan , E. Schueddig , A. Segura , T. Bodine , C. Lominska , R.T. Morse , P. Neupane , D. Pei , D. Koestler , R. Kimple , U. Duvvuri , S. Thomas , R. Nallani , L. Shnayder , K. Kakarala , A.M. Bur , E. Yilmaz , G.N. Gan","doi":"10.1016/j.oor.2025.100780","DOIUrl":"10.1016/j.oor.2025.100780","url":null,"abstract":"<div><h3>Purpose</h3><div>Recurrence in early-stage laryngeal carcinoma after radiation therapy often necessitates a total salvage laryngectomy (TSL). Understanding the genetic landscape that influences cancer progression after radiation therapy may help future treatments. We investigated early-stage laryngeal carcinoma patients treated with radiotherapy to identify mutation patterns that predispose to disease recurrence.</div></div><div><h3>Methods</h3><div>The patient cohort had 35 early-stage laryngeal head and neck squamous cell carcinoma (HNSCC) patients (T1 = 15, T2 = 20) treated with radiotherapy (RT). We stratified patients: Responders (no recurrence, N = 14), non-responders (recurrence within 12 months, N = 21), and non-responders undergoing TSL (N = 18). We employed whole exome sequencing to characterize gene mutations using the Genome Analysis Toolkit (GATK) for variant detection and impact on critical biological pathways and post-radiotherapy patient outcomes. Pathway analysis Ingenuity and Reactome Pathway Analysis tools were used to explore the mutated gene pathways.</div></div><div><h3>Results</h3><div>– Differential mutation analysis was performed in the respective groups to find driver genes. In the pre-treatment samples, KCNT2 and AGAP6 mutations were exclusively found in non-responders (OR = 0, P = 0.005 and OR = 0, P = 0.027, respectively), while ADAMTS7 mutations were solely present in responders (OR = inf, P = 0.019). PLEC mutations were more prevalent in responders (OR = 11.6, P = 0.006). Pathway analysis revealed that significant genes were involved in the RND2 GTPase cycle, protein O-glycosylation-related diseases, and apoptotic pathways. Post-treatment analysis in patients undergoing TSL had enrichment of mutations in apoptosis regulation pathways.</div></div><div><h3>Conclusions</h3><div>– The study reveals that mutations in apoptosis controlling genes were predominantly represented in the larynx non-responder patients both in the pre-radiotherapy and post-TSL populations.</div></div>","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"17 ","pages":"Article 100780"},"PeriodicalIF":0.0,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145927047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-17DOI: 10.1016/j.oor.2025.100779
Amelia Fogle , Celeste Kim , Brian J. Boyce , Jennifer H. Gross , Harry M. Baddour , Nicole C. Schmitt , Azeem S. Kaka
{"title":"Submental island flap for reconstruction of preauricular defects: A narrative review","authors":"Amelia Fogle , Celeste Kim , Brian J. Boyce , Jennifer H. Gross , Harry M. Baddour , Nicole C. Schmitt , Azeem S. Kaka","doi":"10.1016/j.oor.2025.100779","DOIUrl":"10.1016/j.oor.2025.100779","url":null,"abstract":"","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"17 ","pages":"Article 100779"},"PeriodicalIF":0.0,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145927136","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-16DOI: 10.1016/j.oor.2025.100778
Klementyna Dmowska , Nikola Leszczyńska , Piotr Nogal , Małgorzata Leszczyńska , Mariusz Kasprzyk , Mateusz Banaszewski , Jacek Banaszewski
Correct diagnosis of cervical cystic lesion might prove difficult - while in most cases those are benign, malignant diseases must be taken into consideration. We would like to present a case of a 66-year-old female patient with folliculoma metastasis to cervical lymph nodes - primarily recognised as branchial cleft cyst.
{"title":"A 66-year-old female patient with folliculoma metastasis to cervical lymph nodes. Case report","authors":"Klementyna Dmowska , Nikola Leszczyńska , Piotr Nogal , Małgorzata Leszczyńska , Mariusz Kasprzyk , Mateusz Banaszewski , Jacek Banaszewski","doi":"10.1016/j.oor.2025.100778","DOIUrl":"10.1016/j.oor.2025.100778","url":null,"abstract":"<div><div>Correct diagnosis of cervical cystic lesion might prove difficult - while in most cases those are benign, malignant diseases must be taken into consideration. We would like to present a case of a 66-year-old female patient with folliculoma metastasis to cervical lymph nodes - primarily recognised as branchial cleft cyst.</div></div>","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"17 ","pages":"Article 100778"},"PeriodicalIF":0.0,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145792004","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-11DOI: 10.1016/j.oor.2025.100774
Thangapandi Marudhupandi
This letter expresses appreciation for the article “Prognostic significance of systemic inflammatory markers in oropharyngeal squamous cell carcinoma” by Horackiewicz et al. in Oral Oncology reports. The study's key finding that serum albumin and neutrophil-to-lymphocyte ratio are independent prognostic factors for overall survival in oropharyngeal squamous cell carcinoma (OPSCC) is lauded for its clinical relevance and robust analysis of a large, diverse cohort. Constructive suggestions for future research include the necessity of prospective external validation, incorporating additional clinical and molecular variables, and leveraging artificial intelligence/machine learning to develop more dynamic and precise prognostic models. This forward-looking perspective aims to further enhance patient stratification and personalized treatment in OPSCC.
{"title":"Letter to the editor: Prognostic significance of systemic inflammatory markers in oropharyngeal squamous cell carcinoma","authors":"Thangapandi Marudhupandi","doi":"10.1016/j.oor.2025.100774","DOIUrl":"10.1016/j.oor.2025.100774","url":null,"abstract":"<div><div>This letter expresses appreciation for the article “Prognostic significance of systemic inflammatory markers in oropharyngeal squamous cell carcinoma” by Horackiewicz et al. in Oral Oncology reports. The study's key finding that serum albumin and neutrophil-to-lymphocyte ratio are independent prognostic factors for overall survival in oropharyngeal squamous cell carcinoma (OPSCC) is lauded for its clinical relevance and robust analysis of a large, diverse cohort. Constructive suggestions for future research include the necessity of prospective external validation, incorporating additional clinical and molecular variables, and leveraging artificial intelligence/machine learning to develop more dynamic and precise prognostic models. This forward-looking perspective aims to further enhance patient stratification and personalized treatment in OPSCC.</div></div>","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"17 ","pages":"Article 100774"},"PeriodicalIF":0.0,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145792005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-06DOI: 10.1016/j.oor.2025.100773
Michael Truppe , Kurt Schicho , Michael Figl , Simone Holawe , Christos Perisanidis
Objective
To test whether a locally hosted, retrieval-augmented clinical decision framework (CDF) improves recognition of cocaine-induced midline destructive lesions (CIMDL)—a malignancy mimicker—during dental teleconsultation, compared with a cloud large language model (LLM) without retrieval.
Materials and methods
We conducted a paired feasibility evaluation using 401 simulated consultations derived from Sigmund Freud's historical lesion narrative. The local CDF used mistral-7b-instruct-v0.2 with retrieval-augmented generation (RAG) over a curated CIMDL corpus; the comparator was GPT-4o accessed without retrieval. Identical prompts and clinical context were used in both arms. The primary endpoint was inclusion of CIMDL in the differential diagnosis, adjudicated against published criteria. Paired outcomes were compared with McNemar's test; odds ratios (OR) and 95 % confidence intervals (CI) were computed from discordant pairs.
Results
On the MAIN-prompt dataset, the local RAG model identified CIMDL more often than the cloud comparator (discordant pairs: 44 [local-only] vs 19 [cloud-only]; χ2 with Yates correction = 9.14, p = 0.0025; OR = 2.28, 95 % CI 1.34–3.89).
Conclusions
Under realistic privacy constraints, a RAG-enabled local framework improved recognition of a key cancer mimicker during dental triage relative to a cloud LLM without retrieval.
Clinical relevance
Enhanced early recognition of CIMDL may prevent misclassification as oral SCC and expedite appropriate referral; local, privacy-preserving AI is feasible for chairside decision support.
目的测试本地托管的检索增强临床决策框架（CDF）与无检索的云大语言模型（LLM）相比，是否能提高牙科远程会诊期间可卡因诱导的中线破坏性病变（CIMDL）（一种恶性模拟物）的识别。材料和方法我们使用401次模拟咨询进行了配对可行性评估，这些咨询来自西格蒙德·弗洛伊德的历史病变叙述。本地CDF使用mistral-7b-instruct-v0.2与检索增强生成（RAG）对一个策划的CIMDL语料库；比较器采用gpt - 40访问，无需检索。两组使用了相同的提示和临床背景。主要终点是将CIMDL纳入鉴别诊断，根据已公布的标准进行判定。配对结果与McNemar检验进行比较；从不一致的对中计算比值比（OR）和95%置信区间（CI）。结果在MAIN-prompt数据集中，本地RAG模型识别CIMDL的频率高于云比较器（不一致对：44对[local-only] vs 19对[cloud-only]； χ2与Yates校正= 9.14,p = 0.0025; OR = 2.28, 95% CI 1.34-3.89）。结论：在现实的隐私约束下，相对于没有检索的云LLM， ragg支持的本地框架在牙科分诊过程中提高了对关键癌症模拟物的识别。临床相关性：加强对CIMDL的早期识别可以防止误诊为口腔SCC，并加快适当的转诊；本地的、保护隐私的人工智能对于主席决策支持是可行的。
{"title":"Artificial intelligence in oral cancer: a feasibility study informed by Freud's case","authors":"Michael Truppe , Kurt Schicho , Michael Figl , Simone Holawe , Christos Perisanidis","doi":"10.1016/j.oor.2025.100773","DOIUrl":"10.1016/j.oor.2025.100773","url":null,"abstract":"<div><h3>Objective</h3><div>To test whether a locally hosted, retrieval-augmented clinical decision framework (CDF) improves recognition of cocaine-induced midline destructive lesions (CIMDL)—a malignancy mimicker—during dental teleconsultation, compared with a cloud large language model (LLM) without retrieval.</div></div><div><h3>Materials and methods</h3><div>We conducted a paired feasibility evaluation using 401 simulated consultations derived from Sigmund Freud's historical lesion narrative. The local CDF used mistral-7b-instruct-v0.2 with retrieval-augmented generation (RAG) over a curated CIMDL corpus; the comparator was GPT-4o accessed without retrieval. Identical prompts and clinical context were used in both arms. The primary endpoint was inclusion of CIMDL in the differential diagnosis, adjudicated against published criteria. Paired outcomes were compared with McNemar's test; odds ratios (OR) and 95 % confidence intervals (CI) were computed from discordant pairs.</div></div><div><h3>Results</h3><div>On the MAIN-prompt dataset, the local RAG model identified CIMDL more often than the cloud comparator (discordant pairs: 44 [local-only] vs 19 [cloud-only]; χ<sup>2</sup> with Yates correction = 9.14, p = 0.0025; OR = 2.28, 95 % CI 1.34–3.89).</div></div><div><h3>Conclusions</h3><div>Under realistic privacy constraints, a RAG-enabled local framework improved recognition of a key cancer mimicker during dental triage relative to a cloud LLM without retrieval.</div></div><div><h3>Clinical relevance</h3><div>Enhanced early recognition of CIMDL may prevent misclassification as oral SCC and expedite appropriate referral; local, privacy-preserving AI is feasible for chairside decision support.</div></div>","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"17 ","pages":"Article 100773"},"PeriodicalIF":0.0,"publicationDate":"2025-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145712471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-20DOI: 10.1016/j.oor.2025.100770
Erik Dargel , Paul Meissner , Anne-Sophie Becker , Daniel Fabian Strüder , Björn Schneider , Annette Zimpfer , Sonja Oehmcke-Hecht , Israel Barrantes , Claudia Maletzki
Recent studies have underscored the influence of intratumoral microbiota on cancer biology, including head and neck squamous cell carcinoma (HNSCC). However, the relationship between microbial distribution, survival outcomes, and tumor metabolism remains poorly understood. In this study, we investigated the presence, diversity, and clinical relevance of intratumoral microbiota in HNSCC using immunohistochemistry and 16S rDNA amplicon sequencing.
Analyzing a tissue microarray of over 150 HNSCC samples, we detected both Gram-positive and Gram-negative bacterial components, with Gram-negative bacteria being more prevalent (44.1 % vs. 18.9 %). Substantial heterogeneity was observed both within tumors and between individuals. Microbial 16S rRNA was identified in 37.4 % of primary HNSCC samples (37/99), with microbial load varying considerably across cases. Firmicutes was the most dominant phylum, followed by Bacteroidota and Proteobacteria. Microbial composition was significantly associated with HPV status, gender, age, and alcohol consumption.
Importantly, we identified distinct metabolic pathways enriched in HPV+ and HPV− tumors. The methylaspartate cycle was upregulated in HPV + cases, while L-glutamate degradation VIII was predominant in HPV− tumors. Despite these differences, the presence of intratumoral microbiota was not significantly associated with patient survival.
In conclusion, our findings confirm the presence of diverse bacterial communities within HNSCC tumors, shaped by clinical and demographic variables. While microbiota did not correlate with survival outcomes, their influence on tumor metabolism and microenvironment warrants further investigation. These insights lay the groundwork for exploring microbiome-targeted strategies as adjuncts to conventional cancer therapies.
{"title":"The tumor microbiome in head and neck cancer: Insights into the complexity, distribution, and differential regulation of metabolic pathways between HPV-positive and HPV-negative subtypes","authors":"Erik Dargel , Paul Meissner , Anne-Sophie Becker , Daniel Fabian Strüder , Björn Schneider , Annette Zimpfer , Sonja Oehmcke-Hecht , Israel Barrantes , Claudia Maletzki","doi":"10.1016/j.oor.2025.100770","DOIUrl":"10.1016/j.oor.2025.100770","url":null,"abstract":"<div><div>Recent studies have underscored the influence of intratumoral microbiota on cancer biology, including head and neck squamous cell carcinoma (HNSCC). However, the relationship between microbial distribution, survival outcomes, and tumor metabolism remains poorly understood. In this study, we investigated the presence, diversity, and clinical relevance of intratumoral microbiota in HNSCC using immunohistochemistry and 16S rDNA amplicon sequencing.</div><div>Analyzing a tissue microarray of over 150 HNSCC samples, we detected both Gram-positive and Gram-negative bacterial components, with Gram-negative bacteria being more prevalent (44.1 % vs. 18.9 %). Substantial heterogeneity was observed both within tumors and between individuals. Microbial 16S rRNA was identified in 37.4 % of primary HNSCC samples (37/99), with microbial load varying considerably across cases. <em>Firmicutes</em> was the most dominant phylum, followed by <em>Bacteroidota</em> and <em>Proteobacteria</em>. Microbial composition was significantly associated with HPV status, gender, age, and alcohol consumption.</div><div>Importantly, we identified distinct metabolic pathways enriched in HPV+ and HPV− tumors. The methylaspartate cycle was upregulated in HPV + cases, while L-glutamate degradation VIII was predominant in HPV− tumors. Despite these differences, the presence of intratumoral microbiota was not significantly associated with patient survival.</div><div>In conclusion, our findings confirm the presence of diverse bacterial communities within HNSCC tumors, shaped by clinical and demographic variables. While microbiota did not correlate with survival outcomes, their influence on tumor metabolism and microenvironment warrants further investigation. These insights lay the groundwork for exploring microbiome-targeted strategies as adjuncts to conventional cancer therapies.</div></div>","PeriodicalId":94378,"journal":{"name":"Oral Oncology Reports","volume":"16 ","pages":"Article 100770"},"PeriodicalIF":0.0,"publicationDate":"2025-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145571382","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号