Cytotechnician Rapid On-Site Evaluation for Cytology for Transbronchial Needle Aspiration

Abstract

To the Editor: There has been much interest in the use of endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) as a useful mediastinal staging and diagnostic technique; however, the cost has limited its widespread use.1,2 Therefore, many centers continue to use conventional transbronchial needle aspiration (TBNA), achieving a sensitivity of 78% for bulky proximal mediastinal disease in 1 series3 to good effect. Wherever available, cytopathologists have come into the endoscopy suite to provide a rapid on-site evaluation for cytopathology (ROSE) service, which has been shown to improve yield,4 but is not in widespread use because of the costs and the limited availability of cytopathologist expertise. We, therefore, evaluated the potential of a cytotechnician-delivered ROSE service to evaluate when an adequate sample was obtained (rather than making a specific diagnosis on site as in a cytopathologist-delivered ROSE service). We hypothesized that a trained cytotechnician would achieve a strongly positive correlation with an adequate TBNA sample as assessed by an independent cytopathologist. An adequate TBNA sample was defined as containing lymphocytes from a lymph node to the satisfaction of a blinded cytopathologist and also the cytotechnician. Twenty-two consecutive patients with suspected lung cancer with significant mediastinal adenopathy underwent conventional TBNA as part of their diagnostic workup as previously described3 under conscious sedation. This study was approved by the Institutional Board; moreover, it involved no change to existing clinical care or workup of the patients. All of them underwent computed tomography of the thorax and TBNA as part of flexible bronchoscopy. The cytotechnician came into the endoscopy suite and informed us when an adequate specimen had been obtained. Both the cytotechnician and cytopathologists were blinded to the suspected diagnosis. Statistical analysis was performed using GraphPad Prism version 5 software (San Diego, CA) to calculate κ statistic, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for diagnosis of mediastinal metastases on TBNA. A true-negative TBNA was only confirmed on at least 6 months of clinicoradiologic follow-up to exclude malignancy. The cytotechnician agreed with the cytopathologists in 21 of 22 cases (20 samples deemed adequate by both, 1 deemed inadequate by both). There was disagreement in 1 case, where the cytopathologist could not see evidence of an adequate sample but the cytopathologist deemed the subcarinal TBNA to show reactive lymph node features only (in patient followed up with no evidence of mitosis subsequently). The κ statistic for agreement was “very good” at 0.91 (0.79 to 1.03, 95% confidence intervals). In terms of TBNA performance, 13 of the 22 patients (8 non–small cell lung cancer—N2 nodes in 7 cases, N1 nodes in 1 case; 5 small cell lung cancer) had a diagnosis of malignancy confirmed on TBNA, whereas 3 had granulomatous disease consistent with sarcoidosis (patients were followed for at least 6 mo with serial imaging before classifying as definite benign disease). Further 4 patients had reactive nodes which remained stable or regressed on follow-up (attributed as true negative) and 2 patients had suspected false-negative TBNA with enlarged nodes which persisted (stations 3 and 10R, where surgical corroboration was not possible). Sensitivity of TBNA for mediastinal metastases was 87% with a specificity of 100% (positive results were not corroborated with mediastinoscopy). Positive predictive value was 100%, negative predictive value 78%, and accuracy 91% with a prevalence of malignancy of 68%. The distribution of nodal stations sampled was as follows: station 3 (n=2), 4R (n=1), 7 (n=17), 10R (n=4). This pilot study demonstrates that good agreement of TBNA sample adequacy can be achieved through a trained cytotechnician-delivered ROSE service. This study needs to be validated in a larger study assessing the number of samples. However, it provides a possible means of delivering ROSE TBNA without the need for cytopathologists to be in the endoscopy suite, a resource (like EBUS-TBNA) which is not available to many centers in our resource rationed health care system. Regarding the use of ROSE itself, these results compare favorably from a previously reported larger series from the same operator without ROSE. Cytotechnician-delivered ROSE conventional TBNA is a pragmatic alternative to the more expensive cytopathologist-delivered ROSE TBNA and EBUS-TBNA itself. Andrew R. L. Medford, MD, FRCP, FCCP Anilkumar Pillai, MRCP North Bristol Lung Centre, Southmead Hospital, Westbury-on-Trym, Bristol, United Kingdom