Background Cystatins are epithelial protease inhibitors that participate in sinonasal immunity and inflammation. Nasal mucus-derived exosomes (NMDEs) are small vesicles secreted by epithelial cells that carry protein cargo reflective of their host cell. NMDEs have been used as a noninvasive biomarker source to study chronic rhinosinusitis with nasal polyps (CRSwNP) proteomics with superior sensitivity to whole mucus. The purpose of this study was to noninvasively quantify exosomal cystatins in a heterogenous population to determine their utility in predicting phenotype and disease severity. Methods This was an Institutional Review Board-approved study in which NMDEs were purified from 105 patients undergoing sinonasal surgery by ultracentrifugation. Demographic and clinical variables were collected and phenotypes were assigned a priori. Linear discriminant analysis was executed based on normalized Cystatin values as phenotype predictor variables. Unsupervised cluster analysis was performed using Ward's linkage followed by Duda/Hart Je(2)/Je(1) index cluster stopping rules. Analysis of variance (ANOVA), Welch's test, and Fisher's exact tests were used for continuous and categorical variables. Results NMDE Cystatin-2 expression segregated by phenotype (mean +/- standard error [SEM]): control (23.4 +/- 4.2 pg/mu g, n = 32); CRS without NP (CRSsNP) (56.6 +/- 8.3 pg/mu g, n = 33); and CRSwNP (130.5 +/- 16.7 pg/mu g, n = 40) (p < 0.0001). Seven clusters were identified among patients where the highest NMDE Cystatin-2 levels clustered with asthma, tissue eosinophilia, and aspirin-exacerbated respiratory disease (AERD). Conclusion Cystatin levels in NMDEs predict CRS phenotype and disease severity. As a "liquid biopsy," noninvasive NMDE collection offers a promising opportunity to study disease pathophysiology, discriminate disease states, and potentially reveal novel therapeutic targets.