Comparative study of fluorescence ROSE quantitative detection of Helicobacter pylori and traditional...

To establish a quantitative and viable-nonviable differentiation system for Helicobacter pylori (Hp) in gastric mucosal biopsy specimens using fluorescence rapid on-site evaluation (F-ROSE), to determine the minimum bacterial count threshold for histopathological positivity, to reveal the bacterial load mechanism underlying the high specificity, low sensitivity, and high false-negative rate of traditional histopathology, and to provide an objective basis for real-time quantitative diagnosis and precise stratification of Hp infection under endoscopy.

A total of 160 gastroscopic biopsy specimens were enrolled and simultaneously examined by F-ROSE quantitative assay and traditional histopathological staining. F-ROSE used acridine orange-ethidium bromide (AO-EB) dual staining to count total, viable, and nonviable bacteria under ×400 high-power field (HPF). Histopathological results were interpreted in a double-blind manner. No traditional gold standard was set; the latent true infection status was inferred by latent class analysis (LCA) combined with Bayesian model, based on which diagnostic efficacy was calculated and the critical bacterial count for histopathological positivity was determined.

A total of 102 true positive and 58 true negative cases were inferred. The positive rate of F-ROSE (61.25%) was significantly higher than that of histopathology (41.88%, P < 0.001). Using true infection status as the reference, F-ROSE showed a sensitivity of 94.1%, specificity of 93.1%, and AUC of 0.943; histopathology showed a sensitivity of 65.7%, specificity of 89.7%, and AUC of 0.782. The mean total bacterial count in F-ROSE was (42.6 ± 9.3)/HPF in histopathologically positive specimens, significantly higher than that in true-positive but histopathologically negative specimens (P < 0.001). A total bacterial count ≥22/HPF was the critical value for histopathological positivity, below which the histopathological positive rate was only 9.3%. Traditional histopathology could not distinguish viable from nonviable bacteria and its positivity was related only to total bacterial count, not to bacterial viability. Patients with a viable bacterial ratio > 60% had significantly more severe dyspeptic symptoms.

Traditional histopathology has high specificity but low sensitivity; it reliably detects Hp only at a bacterial load ≥22/HPF and misses most low-level infections, while being unable to differentiate viable and nonviable bacteria. F-ROSE enables real-time quantification, viability differentiation, and detection of low-level infection, showing promising preliminary diagnostic performance in this single-center Phase I study. These findings require validation in larger multicenter trials with independent reference standards before clinical implementation.