Predictive value of the presence of Prevotella and the ratio of Porphyromonas gingivalis to Prevotella in saliva for esophageal squamous cell carcinoma

Abstract

Background: Imbalance of oral salivary microbiota has been linked to the pathogenesis of a variety of systemic diseases, and oral bacterial species have been shown to be useful biomarkers for systemic diseases.This study aimed to characterize the alterations of oral microbiota in patients with esophageal squamous cell carcinoma (ESCC) and to evaluate the diagnostic performance of oral microbial biomarkers for ESCC. Methods: The relative abundance of flora in saliva samples was analyzed by 16S rDNA sequencing, and differences in the species present in samples from ESCC patients and healthy controls (HCs) were identified by analyzing species diversity and performing LEfSe analysis. Receiver operating characteristic (ROC) curve analysis was applied to evaluate the diagnostic performance of the characteristic bacteria individually and in combination. Results: Differences in bacterial diversity indexes were observed for the saliva of ESCC patients versus HCs (P<0.05), but principal coordinate analysis did not detect a significant difference in the composition of oral microbiota between ESCC patients and HCs (P>0.05). LEfSe analysis showed that Leptotrichia, Porphyromonas (Pg), Streptococcus, Rothia, Lactobacillus and Peptostreptococcus were more abundant in ESCC patient saliva than in HC saliva, whereas Haemophilus, Alloprevotella (All), Prevotella_7, Prevotella (Pre), Prevotella_6, Pasteurellaceae and Pasteurellales were significantly less abundant in ESCC patient saliva (P<0.05). From ROC curve analysis, Pg could detect ESCC with an area under the ROC curve (AUC) of 0.599, sensitivity of 62.2%, and specificity of 70%, whereas the ratio of Pg/Pre had an AUC of 0.791, sensitivity of 93.3%, and specificity of 62.3%. Moreover, the combination of the Pg/Pre and Pg/All ratios showed further improved diagnostic performance for ESCC (AUC=0.826) and even good sensitivity and specificity for the diagnosis of early ESCC (68.2% and 86%, respectively; AUC=0.786). Conclusion: This study shows that Pg in saliva can be used as a characteristic marker of ESCC, and the ratios of Pg/Pre and Pg/All offered significantly improved diagnostic performance, especially for early ESCC.