Daniela Gompelmann1,2, Konstantina Kontogianni3,4, Nilab Sarmand3, Philine Kaukel3, Johannes Krisam5, Ralf Eberhardt3,4, Felix J F Herth3,4. 1. Pneumology and Critical Care medicine, Thoraxklinik, University of Heidelberg, Heidelberg, Germany, daniela.gompelmann@directbox.com. 2. Department of Internal Medicine II, Division of Pulmonology, Medical University of Vienna, Vienna, Austria, daniela.gompelmann@directbox.com. 3. Pneumology and Critical Care medicine, Thoraxklinik, University of Heidelberg, Heidelberg, Germany. 4. Translational Lung Research Center Heidelberg (TLRCH), German Center for Lung Research, Heidelberg, Germany. 5. Institute of Medical Biometry and Informatics, University of Heidelberg, Heidelberg, Germany.
Abstract
BACKGROUND: Endobronchial ultrasound elastography that provides information on tissue stiffness may help distinguish malignant from benign mediastinal and hilar lymph nodes. OBJECTIVES: In this prospective trial, we assessed the diagnostic value of elastographic images and the interobserver agreement in its evaluation. METHOD: Elastographic images from 77 lymph nodes in 65 patients were reviewed by 3 pneumologists. The elastographic image was classified based on the predominant colour: predominantly green, intermediary, and predominantly blue. With 2 or 3 interobserver matches, the corresponding elastographic image was correlated with the pathological result obtained from endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) and/or other invasive procedures. RESULTS: All 3 reviewers had agreement in classifying elastographic images in 45% (35/77). Overall, the interobserver agreement among the 3 readers for classifying elastographic pattern was found to be moderate (Fleiss Kappa index = 0.519; 95% CI = [0.427; 0.611]). On cytological/histological evaluation, 55 lymph nodes were malignant and 22 were benign. In classifying "green" as benign and "blue" as malignant, the sensitivity and specificity were 71% (95% CI = [54%; 85%]) and 67% (95%-CI = [35%; 90%]), respectively. CONCLUSIONS: Elastography will not replace invasive EBUS-TBNA due to a moderate interobserver agreement and insufficient sensitivity and specificity. However, elastography will, maybe, present an additional feature to identify malignant lymph nodes in the context of clinical, radiological, and cytological results.
BACKGROUND: Endobronchial ultrasound elastography that provides information on tissue stiffness may help distinguish malignant from benign mediastinal and hilar lymph nodes. OBJECTIVES: In this prospective trial, we assessed the diagnostic value of elastographic images and the interobserver agreement in its evaluation. METHOD: Elastographic images from 77 lymph nodes in 65 patients were reviewed by 3 pneumologists. The elastographic image was classified based on the predominant colour: predominantly green, intermediary, and predominantly blue. With 2 or 3 interobserver matches, the corresponding elastographic image was correlated with the pathological result obtained from endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) and/or other invasive procedures. RESULTS:All 3 reviewers had agreement in classifying elastographic images in 45% (35/77). Overall, the interobserver agreement among the 3 readers for classifying elastographic pattern was found to be moderate (Fleiss Kappa index = 0.519; 95% CI = [0.427; 0.611]). On cytological/histological evaluation, 55 lymph nodes were malignant and 22 were benign. In classifying "green" as benign and "blue" as malignant, the sensitivity and specificity were 71% (95% CI = [54%; 85%]) and 67% (95%-CI = [35%; 90%]), respectively. CONCLUSIONS: Elastography will not replace invasive EBUS-TBNA due to a moderate interobserver agreement and insufficient sensitivity and specificity. However, elastography will, maybe, present an additional feature to identify malignant lymph nodes in the context of clinical, radiological, and cytological results.