Assessing Rapid On-Site Evaluation (ROSE) for Adequacy of Endobronchial Ultrasound Transbronchial Needle Aspiration (EBUS-TBNA) Procedures Compared to A Direct-to- Cellblock Approach As a Response to COVID-19 Pandemic

ABSTRACT

Background: Rapid on-site evaluation (ROSE) of endobronchial ultrasound transbronchial needle aspiration (EBUS-TBNA) has been used as standard to improve diagnostic yield and to triage samples for additional testing. A method has been prompted to leveraging the expertise of cytotechnologists only or the utility of tele cytology. However, due to the COVID-19 pandemic, our institution had to discontinue ROSE for EBUS-TBNA, and instead adopted a direct-to-cellblock approach. This study aims to determine if removing ROSE affected the adequacy of samples obtained from the EBUS-TBNA cases, and if this new approach should be employed continuously for an extended period without jeopardizing the standard of care to the patient since we are still in the pandemic. Design: We collected data of total 1903 EBUS-TBNA cases from 734 patients. 1097 cases were performed from 1/2019 to 3/2020 with ROSE on 452 patients;806 cases were performed between 4/2020 to 12/2020 without ROSE on 282 patients, and the sample was sent directly to be processed as a cell block. Data variables collected included patient's age and gender, specimen sites and frequencies, and final cytopathological diagnosis. Calculated standardized residuals and fitted multivariate generalized linear model were performed at Moffitt Cancer Center & Research Institute Biostats Core using software SAS 9.4. Results: On average, holding specimen site variable constant, a biopsy from a patient with ROSE is 0.936 (=exp -0.066) times less likely to be reported as 'Satisfactory' as compared to a biopsy taken from a patient without ROSE, and this difference is not statistically significant (p = 0.785). Among all 6 categories of diagnosis, biopsies reported as 'Diagnostic for malignancy' are significantly different between ROSE and no ROSE groups with the standardized residual of 1.80 (p = 0.036) & only other significantly different category was 'Other';however, a very small number of biopsies were reported in this category (3.5%). Conclusions: The inadequacy rate of EBUS-TBNA was 6.4% higher on average in cases with ROSE when compared to a directto- cell block approach in this data set, however this difference was not statistically significant. This suggests that ROSE could be discontinued in favor of a direct-to-cell block approach. Although additional studies are needed, the discontinuation of ROSE for EBUS-TBNA would decrease the time and cost burden on cytopathology service while providing the same quality of care to the patient. (Table Presented).