Carlos M Ardila, Eliana Pineda-Vélez, Anny M Vivares-Builes
{"title":"Transcriptomic and Metagenomic Biomarkers in Peri-Implantitis: A Systematic Review, Diagnostic Meta-Analysis, and Functional Meta-Synthesis.","authors":"Carlos M Ardila, Eliana Pineda-Vélez, Anny M Vivares-Builes","doi":"10.3390/medsci13030187","DOIUrl":null,"url":null,"abstract":"<p><p><b>Background/Objectives:</b> Evidence from transcriptomic and histopathologic studies has revealed that peri-implantitis lesions are characterized by deeper inflammatory infiltration, increased immune cell accumulation, and distinctive molecular signatures. This systematic review aimed to evaluate the diagnostic and pathophysiological potential of transcriptomic, metagenomic, and bioinformatic biomarkers in peri-implantitis by integrating findings from bioinformatics and machine learning-based studies. The dual objective was to identify biologically relevant markers and assess the accuracy of predictive models, addressing diagnostic gaps in peri-implant disease management. <b>Methods:</b> Eligible designs included cross-sectional, case-control, and cohort studies. Literature searches were conducted across PubMed, EMBASE, Scielo, and Scopus, with independent screening, data extraction, and quality assessment. Functional meta-synthesis was used to thematically organize biomarkers and pathways, while diagnostic meta-analysis pooled ROC-AUC values to assess model performance. <b>Results:</b> Eleven studies met the inclusion criteria. Functional synthesis revealed five recurring biomarker themes: innate and adaptive immune responses, immune cell infiltration, fibroblast activation, and ceRNA regulation. A meta-analysis of six studies reported a pooled AUC of 0.91 (95% CI: 0.88-0.93) with I<sup>2</sup> = 0%, indicating no heterogeneity, supporting the reliability of ML-based models in distinguishing peri-implantitis from healthy conditions. Sources of variation included differences in validation strategies and data preprocessing. <b>Conclusions:</b> Integrating transcriptomic, metagenomic, and bioinformatic biomarkers with machine learning may enable earlier and more accurate diagnosis of peri-implantitis. The identified biomarkers highlight molecular and microbial pathways linked to inflammation and tissue remodeling, underscoring their potential as diagnostic indicators and therapeutic targets with translational relevance.</p>","PeriodicalId":74152,"journal":{"name":"Medical sciences (Basel, Switzerland)","volume":"13 3","pages":""},"PeriodicalIF":4.4000,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12452457/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Medical sciences (Basel, Switzerland)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.3390/medsci13030187","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"Medicine","Score":null,"Total":0}
引用次数: 0
Abstract
Background/Objectives: Evidence from transcriptomic and histopathologic studies has revealed that peri-implantitis lesions are characterized by deeper inflammatory infiltration, increased immune cell accumulation, and distinctive molecular signatures. This systematic review aimed to evaluate the diagnostic and pathophysiological potential of transcriptomic, metagenomic, and bioinformatic biomarkers in peri-implantitis by integrating findings from bioinformatics and machine learning-based studies. The dual objective was to identify biologically relevant markers and assess the accuracy of predictive models, addressing diagnostic gaps in peri-implant disease management. Methods: Eligible designs included cross-sectional, case-control, and cohort studies. Literature searches were conducted across PubMed, EMBASE, Scielo, and Scopus, with independent screening, data extraction, and quality assessment. Functional meta-synthesis was used to thematically organize biomarkers and pathways, while diagnostic meta-analysis pooled ROC-AUC values to assess model performance. Results: Eleven studies met the inclusion criteria. Functional synthesis revealed five recurring biomarker themes: innate and adaptive immune responses, immune cell infiltration, fibroblast activation, and ceRNA regulation. A meta-analysis of six studies reported a pooled AUC of 0.91 (95% CI: 0.88-0.93) with I2 = 0%, indicating no heterogeneity, supporting the reliability of ML-based models in distinguishing peri-implantitis from healthy conditions. Sources of variation included differences in validation strategies and data preprocessing. Conclusions: Integrating transcriptomic, metagenomic, and bioinformatic biomarkers with machine learning may enable earlier and more accurate diagnosis of peri-implantitis. The identified biomarkers highlight molecular and microbial pathways linked to inflammation and tissue remodeling, underscoring their potential as diagnostic indicators and therapeutic targets with translational relevance.