Assessing diagnostic concordance between rapid on-site evaluation and final histopathology in mediastinal lesions using conventional and endobronchial ultrasound-guided TBNA

Background Rapid on-site evaluation (ROSE) is a useful additional technique for fine-needle aspiration diagnosis of cytologic material. With ROSE, there is an increase in diagnostic yield either by transbronchial needle aspiration (TBNA) of mediastinal lesions or the guidance of endobronchial ultrasound (EBUS). Methods The study was cross-sectional, single-center study, which included 134 patients with undiagnosed mediastinal lymphadenopathy and/or lesions. Conventional TBNA or EBUS-TBNA were used for sampling. Preliminary ROSE results and the final diagnosis tissue biopsy were examined. Results The concordance between the ROSE result and the final histopathologic diagnosis was detected in 102 cases (76.1%). Regarding the diagnostic category, benign disease was detected in 102 cases (76.1%), while atypical cells were determined in one case only (0.7%), and malignant disease was detected in 31 cases (23.1%). The most common diagnosis during ROSE was non-necrotizing granuloma (51.5%). Regarding the relation between the concordance and either the category of the ROSE result, the performed procedure or the lesion site, no significant relation was found (P value = 0.075, 0.545 and 0.292, respectively). Conclusion The ROSE result is comparable with the final histopathologic diagnosis in affirming the adequacy of the samples and establishing the accurate diagnosis of pulmonary lesions in many different cases. This could be effective in clinical practice, as making decisions during TBNA or EBUS-TBNA can depend on ROSE. Thus, the integration of ROSE is recommended in routine practice, particularly in resource-limited settings.