Rapid on-site evaluation (ROSE) during endoscopic procedures, such as bronchoscopy, plays an important role in clinical practice. It saves time and cost, and reduces the need for repeated procedures due to inadequate sampling [1]. Cytological smears prepared during bronchoscopy can be assessed within minutes, allowing immediate feedback on sample adequacy. This enables the bronchoscopist to confirm or adjust the biopsy site as needed [2].

Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has become a reliable and indispensable diagnostic modality in the evaluation of lung cancer [3, 4]. Pathological work-up includes hematoxylin–eosin and immuncytology staining to determine the histologic subtype, while modern molecular techniques such as next-generation sequencing (NGS) allow simultaneous interrogation of multiple genes, facilitating precision oncology and treatment planning [5]. A major limitation in this pathway, however, is the availability of adequate tissue, particularly in hard-to-biopsy thoracic malignancies. Both EBUS-TBNA and conventional TBNA are cornerstone procedures in interventional pulmonology, yielding smears and cell blocks that are subsequently used by pathologists and molecular biologists for diagnosis and biomarker testing [6].

The presence of a cytopathologist at the time of sampling has been shown to improve diagnostic yield by assessing adequacy. However, due to increasing workload, this is rarely feasible in routine practice. Some studies have suggested that trained interventional pulmonologists can perform ROSE to assist adequacy assessment and thereby improve diagnostic yield [7–9]. Nevertheless, the workload of pulmonologists is already high, and it is often logistically impractical for the same operator to both perform the procedure and simultaneously evaluate ROSE, or to require a second pulmonologist solely for this task.

In our study, ROSE was performed by a university-trained clinical biochemist during bronchoscopic procedures. Our aim was to assess the accuracy of laboratory specialist–performed ROSE and to compare its results with the final cytopathologist diagnosis.

We analyzed 78 consecutive bronchoscopic sampling cases in which rapid on-site evaluation (ROSE) was performed by a university-trained clinical biochemist (laboratory specialist). Baseline characteristics are shown in Table 1. Cytological adequacy and final diagnosis were subsequently established by a board-certified team of cytopathologist, who served as the reference standard (a four-person team of pathologists).

Two sampling techniques were included: endobronchial ultrasound–guided transbronchial needle aspiration (EBUS-TBNA, n = 37) and conventional TBNA (cTBNA, n = 41). Demographic data were recorded, including patient age and sex.

The samples taken during the procedures were smeared onto glass slides (1 puncture was used to obtain 2 slides). One slide was sent fresh to pathology for analysis, while the other slide was air-dried, fixed with methanol, stained with Diff-Quik rapid stain, and assessed for suitability using a light microscope (Motic 2820).

A total of 78 patients were included (EBUS-TBNA: 37; cTBNA: 41). Mean age was 58.6 ± 14.1 years in the EBUS group and 60.1 ± 13.9 years in the cTBNA group.

In this study we evaluated the performance of rapid on-site evaluation (ROSE) performed by a university-trained laboratory specialist during bronchoscopic sampling, and compared the findings with the final cytopathologist report. Our results indicate that laboratory specialist–performed ROSE is highly reliable for the assessment of sample adequacy, with substantial-to-almost perfect agreement with the pathologists, while agreement on final diagnosis was moderate. This is especially important, as involvement of laboratory specialists could decrease the workload of pathologists, who are already missing from healthcare systems around the world [10, 11].

The adequacy results are particularly important from a clinical perspective. We observed 100% adequacy with EBUS-TBNA and 78% with cTBNA, consistent with previous studies demonstrating the superior technical yield of EBUS [3, 4]. The presence of a real-time adequacy assessment further enhances this efficiency by reducing the likelihood of nondiagnostic samples and minimizing the need for repeat procedures. ROSE performed by the laboratory specialist demonstrated high sensitivity (91.3%) and perfect specificity (100%) for adequacy, with an excellent positive predictive value. These results are similar, or better compared to previous studies, where ROSE was performed by a pulmonologist [12, 13]. These results confirm that adequacy can be reliably determined by trained personnel outside of pathology, offering a pragmatic solution where on-site cytopathology support is unavailable.

The moderate agreement on final diagnosis (κ = 0.55, AC1 = 0.57) is consistent with the intended role of ROSE. Similar findings have been reported in studies where pulmonologists or cytotechnologists performed ROSE: agreement with pathologists is typically excellent for adequacy but more limited for specific diagnostic categories [7–9]. ROSE is not designed to replace full cytopathologic evaluation, but to ensure that samples are adequate for subsequent processing and ancillary testing. The pattern of discordant cases in our analysis, including sarcoidosis and malignancy, reflects the inherent diagnostic limitations of on-site cytology without the use of ancillary stains and immunohistochemistry.

Our study contributes novel evidence by showing that a laboratory specialist (clinical biochemist) can effectively fulfill the role of on-site evaluator. This may be especially relevant in healthcare systems with limited cytopathology resources, where the increasing procedural workload makes it impractical to rely on pathologist presence during every bronchoscopy. Compared to solutions involving interventional pulmonologists performing ROSE themselves, which increases physician workload, involving trained laboratory staff provides a sustainable and scalable alternative.

Strengths of our study include the prospective collection of real-world data, application of multiple reliability indices (κ, AC1, diagnostic accuracy, McNemar’s test), and the direct comparison between EBUS and cTBNA. Limitations include the modest sample size, the single-center design, the lack of assessment of inter-observer variability, the lack of assessment of the impact on clinical practice, (such as the rate of repeat procedures or turnaround time), and the fact that only one laboratory specialist was evaluated, which may limit generalizability. Future research should include multi-center studies, training standardization, and assessment of the impact on clinical outcomes such as repeat procedure rates and time to diagnosis.

In conclusion, laboratory specialist–performed ROSE is a feasible and accurate alternative to pathologist on-site evaluation, ensuring high adequacy rates during bronchoscopic sampling. This approach can optimize diagnostic yield, support efficient patient management, and may help address resource limitations in interventional pulmonology and cytopathology.