The Impact of Rapid On-site Evaluation on Diagnostic Performance of Computed Tomography-Guided Core Needle Biopsy in Lung Cancer.

Abstract

Rapid on-site-evaluation (ROSE) is a technique aimed at improving the diagnostic performance of computed tomography (CT)-guided core needle biopsy (CNB) in lung cancer. The aim of this retrospective study was to investigate the impact of ROSE on the rate of nondiagnostic specimens and on accuracy computed on diagnostic specimens. During a 3-year period, 417 CT-guided CNBs were performed at our center. The biopsies were retrospectively classified into 2 groups: 141 procedures were assisted by ROSE and 276 were not. All of them were reviewed for clinical, procedural, and pathological data. Pathology results were classified as diagnostic (positive or negative for malignancy) or nondiagnostic. The results were compared with the final diagnosis after surgery or clinical follow-up. Nondiagnostic rate, sensitivity/specificity/negative predictive value/positive predictive value for the ROSE and non-ROSE groups were calculated. Finally, procedural complications and the adequacy of the specimens for the molecular analysis were recorded. The study evaluated 417 CNBs (mean patients' age 71 years, 278 men). Nondiagnostic rates with and without ROSE were 4% (6/142) and 11% (29/276), respectively (P = 0.028). Sensitivity/specificity/negative predictive value/positive predictive value with and without ROSE did not show statistically significant differences, and no difference in major/minor complication rates was observed between the 2 groups. The adequacy of specimen for subsequent molecular analysis was 100% with (42/42) and 82% without ROSE (51/62). Rapid on-site-evaluation reduced the rate of nondiagnostic specimens by 50% with no change in complication rates or accuracy and increased by 20% the chances of a successful subsequent molecular analysis.