Role of rapid onsite evaluation in enhancing diagnostic yield of CT-guided core needle lung biopsies.

OBJECTIVE: To evaluate the adequacy of Rapid Onsite Evaluation (ROSE) in relation to diagnostic yield of CT-guided percutaneous lung core needle biopsies (CNB) at a tertiary care cancer center. METHODS: A total of 943 CT-guided percutaneous CNB performed in the year 2022 were retrospectively reviewed. ROSE of specimens provided immediate feedback regarding the sample adequacy. The biopsies were divided into 3 groups based on ROSE results: 1) adequate on first pass; 2) adequate after repeat sampling and 3) inadequate. Patient, lesion, and procedure characteristics were analyzed in relation to ROSE adequacy using a multinomial logistic regression model. RESULTS: Median lesion size was 1.8 cm (IQR: 1.2-2.8). The first-pass adequacy rate was 64.8% (611/943), increasing to 82.4% (777/943) after seven passes, with a clear plateau after three passes; 17.9% (169/943) samplings were persistently inadequate. Complete histopathologic processing resulted in an overall diagnostic yield of 89.7% (846/943). Critically, 59.2% of biopsies deemed inadequate by ROSE ultimately yielded a diagnostic result after complete pathologic processing. ROSE demonstrated a sensitivity of 88.2%, specificity of 71.1%, positive predictive value of 96.4%, and negative predictive value of 40.8%. Larger lesion size and absence of perilesional hemorrhage were associated with higher odds of first-pass adequacy on multivariable analysis. CONCLUSION: Lesion size and perilesional hemorrhage are independent predictors of ROSE- adequacy for lung CNB. The low negative predictive value (40.8%) and the finding that 59.2% of ROSE-inadequate biopsies ultimately yielded a diagnostic result highlight that ROSE inadequacy should not be interpreted as synonymous with biopsy failure.