Abstract
INTRODUCTION: Xerostomia (dry mouth) and dysgeusia (taste changes) are subjective oral symptoms affecting radiation-treated head and neck cancer (HNC) patients. Our objective was to determine longitudinal and cross-sectional changes in oral microbiome profiles and predicted molecular pathways. METHODS: Saliva and tongue swab samples from HNC patients (N = 25) were collected at baseline (BL) and 18-months (18(mos)) post-radiation therapy (RT). 16S-rRNA gene V1-V3 next generation sequencing was performed. Subjective xerostomia and dysgeusia scores were determined using EORTC QLQ-H&N ranking scale. Patient groups were classified by worsening symptom severity (high > 2.5 vs low < 2.5) at BL and 18(mos). For xerostomia, longitudinal comparisons included high vs low scores at BL and 18(mos) (N = 26, N = 52; N = 24, N = 48), cross-sectional comparisons included BL high vs low (N = 2 vs 23; N = 4 vs 46) and 18(mos) high vs low (N = 15 vs 10; N = 30 vs 20). For dysgeusia, longitudinal comparisons included high vs low scores at BL and 18(mos) (N = 16; N = 32; N = 34, N = 68), and cross-sectional comparisons included BL high vs low (N = 2 vs 23; N = 4 vs 46) and 18(mos) high vs low (N = 12 vs 13; N = 24 vs 26). α-diversity significance was determined via Mann–Whitney U-test and β-diversity was determined via PERMANOVA. PiCrust2 was used to predict bacterial pathways alterations. MaAsLin2 program was used to determine microbial and pathway associations between timepoint and scoring. ROC curve analysis was used to identify candidate taxa biomarkers. RESULTS: α-diversity did not differ among comparisons. Longitudinal β-diversity comparisons were all significant (p < 0.05). All, but one cross-sectional comparison, were not significant. MaAsLin2 determined 32 significant taxonomic and 14 MetaCyc pathway features. Haemophilus parainfluenzae (padj = 0.01) and Gemella sanguinis (padj = 0.006) were significant taxa distinguishing high and low score dysgeusia groups, per multivariate analysis. Streptococcus salivarius and Ralstonia pickettii were consistently most abundant contributors to pathway alterations. Significant pathways included PWY-7431 (Timepoint; p = 0.001) and KDO-NAGLIPASYN-PWY (BL dysgeusia; p = 0.00035). CONCLUSIONS: Oral microbial composition and functional changes occur post-RT that may contribute to xerostomia and dysgeusia. Significant species H. parainfluenzae, G. sanguinis, S. salivarius, and R. pickettii may be involved in xerostomia and taste change post-RT and may be further investigated for possible characterization as biomarkers at the functional level for dry mouth and taste change. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12866-026-04885-0.