Upper gastrointestinal endoscopic ultrasound (EUS)- and endobronchial ultrasound (EBUS)-guided fine-needle aspiration cytology (FNAC) are being increasingly used as a safe and non-surgical method to sample tissue from mediastinal, para-aortic, and perihilar masses and lymph nodes. EUS- and EBUS-guided FNAC have greatly enhanced our diagnostic armamentarium in pulmonary medicine and gastroenterology. The diagnosis of tuberculosis is confirmed by the finding of granulomatous inflammation and the isolation of Mycobacterium tuberculosis in culture by FNAC of mediastinal or retroperitoneal nodes. The diagnosis and staging of malignancy are also aided by these techniques.

Adequacy of tissue sampling is vital to the diagnostic yield of these procedures and has become even more crucial in an era of increasing molecular testing for genotype-targeted therapies such as evaluation of epidermal growth factor receptor (EGFR) mutations in non-small cell lung cancer. A number of factors are thought to be associated with yield from these procedures, such as high-volume centers, lesion site, lymph node size, positive PET scans, biopsy of more than two sites, needle size, number of passes, and expertise of the scopist [1–6].

On-site preparation and evaluation of smears by a cytopathologist during the procedure (rapid on-site evaluation [ROSE]) ensures that smears are adequately cellular and are likely to yield a diagnosis. Several studies have demonstrated that the diagnosis at on-site evaluation correlates well with the final cytopathological diagnosis. Multiple studies have shown that ROSE increases diagnostic yield by about 10% to 15% [7, 8]. ROSE has also been shown to reduce the number of needle passes performed [9]. Due to proximity to vital structures, risk of complications such as bleeding, and time, fewer needle passes may be beneficial. These benefits may be more pronounced in centers with trainees. This has led to the establishment of ROSE as the default practice in many institutions where EUS-FNA or EBUS-FNA is offered. A survey of endosonologists indicated that ROSE was used more often by US (98%) than by European (48%) or Asian (55%) respondents [10].

However, it is not clear whether ROSE will differentially impact diagnostic yield in pancreatic vs. lymph nodal disease. A meta-analysis of seven studies (one randomized) with a total of 1299 patients, comparing EUS-FNAC with and without ROSE, concluded that ROSE did not make a signifi-cant difference to cytological adequacy or diagnostic yield [11]. Although overall the number of needle passes was similar in the two groups, in the one study which was randomized, seven passes were made for cytology in the absence of ROSE compared to four passes in the presence of ROSE [12].

ROSE requires the availability of a trained cytopathologist on site at the time of the endoscopic procedure. On-site services may not necessarily be available in hospitals in developing countries, particularly those in the public sector. Furthermore, ROSE is associated with higher costs [12]. In this issue of the Journal, Sharma et al. describe their experience of EUS-FNAC without ROSE and indicate that the adequacy of the samples for diagnosis was very good, ranging from 73% for lymph nodes to 100% for mediastinal, renal, and suprarenal masses [13].

A number of parameters may be used to judge the effectiveness of EUS-FNAC. These include cytological adequacy, diagnostic yield, number of passes required for diagnosis, and diagnostic characteristics (number of true-positive, true-negative, false-positive, and false-negative observations). Several of these parameters are missing in the study by Sharma et al. which has mainly described diagnostic adequacy and number of passes made. Of interest, the median number of passes was two, which is lower than would be practiced in most centers.

ROSE performed by endosonographers trained in cytological interpretation was found to significantly increase diagnostic accuracy in one study [14] and found to be on par with evaluation by a cytopathologist in another study [15].

EUS- and EBUS-guided FNA have high a diagnostic yield, especially in high-volume centers, and with experienced endosonographers. We however owe it to our patients to strive for perfection. ROSE is a step in this direction and hence should be performed where available and feasible, while perhaps in other centers, evaluation of smears by the endosonographer may be a viable substitute.