<b>Salivary Biomarker Panel for Detecting Periodontitis</b>

<b>Aim: </b>This study tested the hypothesis that a salivary biomarker panel (i.e., consisting of 2 to 6 biomarker features) could yield diagnostic accuracy of 95% for periodontitis, allowing for earlier disease detection.<b>Materials and Methods:</b> Concentrations of 11 protein biomarkers by immunoassays and oral microbiome species by 16s rRNA sequencing were evaluated in saliva samples from 28 healthy adults and 28 Type 2 diabetic patients with periodontitis. Data were analyzed for discrimination of health or periodontitis using 5-fold cross-validation logistic regression, receiver operator characteristics (ROC), and odds ratios.<b>Results:</b> Bacteria showed better predictive value than individual salivary proteins. Two bacteria (<i>Porphyromonas gingivalis</i> and <i>Mycoplasma faucium</i>) yielded sensitivities &gt;90% and <i>Treponema socranskii</i> demonstrated the top AUC (0.86). A combination of bacteria (<i>Selenomonas sputigena</i>, <i>P. gingivalis</i>, <i>Prevotella nigrescens</i>, <i>Pr. dentalis</i>) with salivary protein ratios (prostaglandin E2/tissue inhibitor of metalloproteinase-1 or macrophage inflammatory protein-1a/tissue inhibitor of metalloproteinase-1) produced robust accuracy (95%) and precision 96.5% for the detection of periodontitis.<b>Conclusions:</b> A salivary panel using bacteria and ratios of host-response biomarkers accurately discriminated oral health from periodontitis in Type 2 diabetics.**研究目的：** 本研究旨在验证一项假说：由2至6项生物标志物特征构成的唾液生物标志物组合（salivary biomarker panel），可对牙周炎实现95%的诊断准确率，从而达成疾病的早期检测。 **材料与方法：** 本研究通过免疫测定法（immunoassays）检测11种蛋白质生物标志物的浓度，并借助16S核糖体RNA测序（16S rRNA sequencing）分析口腔微生物组物种；共纳入28名健康成人与28名合并牙周炎的2型糖尿病患者的唾液样本开展评估。采用5折交叉验证逻辑回归、受试者工作特征曲线（receiver operator characteristics, ROC）及比值比（odds ratios）对数据进行分析，以区分口腔健康与牙周炎状态。 **研究结果：** 细菌的预测价值优于单个唾液蛋白。其中，牙龈卟啉单胞菌（*Porphyromonas gingivalis*）与口支原体（*Mycoplasma faucium*）这两种细菌的敏感性超过90%，索氏密螺旋体（*Treponema socranskii*）的曲线下面积（AUC）最高（0.86）。将痰月形单胞菌（*Selenomonas sputigena*）、牙龈卟啉单胞菌、变黑普雷沃菌（*Prevotella nigrescens*）、牙科普雷沃菌（*Prevotella dentalis*）与唾液蛋白质比值（前列腺素E2/金属蛋白酶组织抑制剂-1或巨噬细胞炎症蛋白-1α/金属蛋白酶组织抑制剂-1）相结合的检测组，可实现牙周炎检测的优异准确率（95%）与精确性（96.5%）。 **研究结论：** 纳入细菌与宿主应答生物标志物比值的唾液生物标志物组合，可准确区分2型糖尿病患者的口腔健康状态与牙周炎。