Macroscopic on-site evaluation based on core tissues length and weight during EBUS-TBNA using 22-gauge needles

To the Editor: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is the first-line approach for the diagnosis of mediastinal and hilar lymphadenopathy and peribronchial/peritracheal masses, and for invasive mediastinal staging.[1] However, the narrow-bore needle often limits tissue acquisition for pathology. The “macroscopic on-site evaluation” (MOSE) technique—referring to the immediate macroscopic assessment of the visible tissue core (VTC) by endoscopic ultrasound (EUS) operators—was first proposed by Iwashita et al[2] in their investigation of its diagnostic utility during EUS‑guided fine-needle aspiration (EUS-FNA). Unlike EUS-FNA, which primarily targets a single lesion, EBUS-TBNA routinely samples multiple targets, such as peribronchial/peritracheal masses and multi-station mediastinal/hilar lymph nodes, to simultaneously achieve diagnosis and nodal staging. Thus, per‑lesion specimen adequacy is essential to ensure accurate staging and procedural efficiency. Although studies have demonstrated that MOSE can be successfully adapted for EUS biopsy needles and can enhance specimen quality, evidence regarding its clinical applicability to EBUS-TBNA remains limited. Therefore, this study evaluated MOSE for assessing specimen adequacy during EBUS‑TBNA using 22‑gauge needles. This prospective observational study enrolled patients who required EBUS-TBNA for at least one mediastinal or hilar lymph node or mass measuring ≥10 mm (short-axis) on computed tomography (CT) or showing increased pathological uptake on positron emission tomography-CT (PET-CT) at the First Affiliated Hospital of Naval Medical University between November 2023 and January 2024. Exclusion criteria included: (1) contraindications to bronchoscopy; (2) cystic or mediastinal cystic lesions; (3) lesions directly visible and accessible by bronchoscopy; (4) prior established pathological diagnosis; (5) severe tracheal stenosis; (6) use of a needle other than 22-gauge; (7) history of severe mental disorder; and (8) inability to provide informed consent. The study complied with the Declaration of Helsinki and was approved by the Institutional Review Board of Shanghai Changhai Hospital (No. CHEC2023‑276). All EBUS-TBNA procedures used 22-gauge needles (NA-201SX-4022, Olympus, Tokyo, Japan).[3] At the bronchoscopist’s discretion, various suction techniques—including low negative-pressure suction, stylet slow-pull, and wet suction (in which 1–2 mL of saline or liquid-based cytology solution was pre-drawn into the negative-pressure syringe)—were employed. The number of passes was determined by the operator based on tissue yield. Rapid on-site evaluation (ROSE) was not used. Procedures were performed by expert (>1000 EBUS-TBNA procedures) or non-expert (10–200 procedures) endoscopists. The MOSE procedure was as follows. Step 1: EBUS-TBNA was performed after lesion identification. After removing the TBNA needle from the echoendoscope, the specimen was extruded into Dish A (containing 2–3 mL of liquid‑based cell preservation fluid). Step 2: Dish B (containing 2–3 mL of 10% neutral buffered formalin) was weighed before and after transferring tissue strips from Dish A. Step 3: The tissue strips in Dish B were examined under a benchtop magnifier to identify and measure VTCs. Step 4: The tissue strips and liquid from Dish B were then transferred to a formalin vial for histologic examination [Supplementary Figure 1, https://links.lww.com/CM9/C892]. The VTC was defined as a whitish, flesh-colored, or mixed red-and-white tissue fragment [Supplementary Figure 2, https://links.lww.com/CM9/C892]. Two endoscopists independently assessed each specimen, with discrepancies resolved by discussion. A definitive diagnosis was defined as one based on histologic evaluation yielding a conclusive result. For this analysis, a specimen was classified as “diagnostic” if the histological evaluation yielded a definitive diagnosis (malignant or specific benign, e.g., sarcoidosis, tuberculosis). A specimen was classified as “non-diagnostic” if the histological evaluation yielded only a non-specific result (e.g., nonspecific benign inflammation, reactive hyperplasia of lymph node). Patients with indeterminate or benign findings were followed for 6 months or until lesion regression to confirm the final diagnosis. Categorical variables were presented as numbers and percentages. Continuous variables were described as mean ± standard deviation if normally distributed, or as median with interquartile range if non-normally distributed. Because definitive diagnosis of each lesion is essential in EBUS-TBNA, all analyses were performed on a per‑lesion basis. Receiver operating characteristic (ROC) curves were generated to evaluate the diagnostic performance of core tissue length and weight. Least absolute shrinkage and selection operator (LASSO) regression was used to identify factors associated with diagnostic accuracy, considering potential multicollinearity among candidate variables. All analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA) and R version 4.3.2 (https://cran.r-project.org/), with significance set at P <0.05 (two-tailed). A total of 204 lesions (176 lymph nodes and 28 masses) were included in the final analysis [Supplementary Figure 3, https://links.lww.com/CM9/C892]. Baseline clinical and procedural data were presented in Supplementary Table 1, https://links.lww.com/CM9/C892. Of these, 66 (32.4%) were sampled by experts and the median lesion diameter on ultrasound imaging was 14.3 (11.0, 21.4) mm [Supplementary Table 2, https://links.lww.com/CM9/C892]. Among the 204 EBUS-TBNA specimens, 8 (3.9%) were cytology‑only and 7 lacked formed tissue strips, leaving 189 for macroscopic analysis. VTCs were identified in 169 specimens (89.4%, 169/189). The median accumulated and median maximum VTC lengths were 10.0 (4.5, 16.0) mm and 4.0 (2.0, 7.0) mm, respectively, and the median weight of the formed tissue strips was 44.0 (21.5, 80.0) mg. Supplementary Table 3, https://links.lww.com/CM9/C892 summarizes the histopathologic diagnoses. EBUS-TBNA yielded definitive histologic diagnoses in 125/189 specimens (66.1%). The “diagnostic” group (n = 125) had significantly greater accumulated VTC length, maximum VTC length, and formed tissue strip weight than the “non-diagnostic” group (n = 64) (P <0.0001, P <0.0001, P = 0.0002, respectively). ROC analysis for accumulated VTC length yielded an area under the curve (AUC) of 0.8808 (95% CI: 0.8237–0.9379; P <0.0001), with a 7.75 mm cutoff providing 95.3% sensitivity and 76.1% specificity. The maximum VTC length yielded an AUC of 0.7920 (95% CI: 0.7206–0.8633; P <0.0001); at a 3.10 mm cutoff, sensitivity and specificity were 82.2% and 68.7%, respectively. For the formed tissue strip weight, the AUC was 0.7300 (95% CI: 0.6548–0.8051; P <0.0001), with a 36.5 mg cutoff giving 68.2% sensitivity and 64.2% specificity [Supplementary Figure 4, https://links.lww.com/CM9/C892]. Pairwise comparison of ROC curves using DeLong’s test indicated that the AUC for accumulated VTC length was significantly greater than that for formed tissue strip weight (P = 0.0002) and for maximum VTC length (P <0.0001). Among the 189 specimens with formed tissue strips, the predictive performance of four specimen assessment strategies was compared: (1) accumulated VTC length alone (AUC = 0.8573), (2) formed tissue strip weight alone (AUC = 0.6620), (3) both criteria met (AND rule; AUC = 0.7584), and (4) at least one criterion met (OR rule; AUC = 0.7609). Among the strategies, accumulated VTC length alone provided the highest diagnostic accuracy (AUC 0.8573), the OR rule the highest sensitivity (92.2%), and the AND rule the highest specificity (91.0%) [Supplementary Table 4, https://links.lww.com/CM9/C892]. LASSO regression was performed with six candidate variables: accumulated VTC length, diameter on ultrasound, tissue strip weight, number of passes, site of lesion, and operator experience. Employing the lambda.1se criterion to ensure model parsimony, two predictors—accumulated VTC length and diameter on ultrasound—were retained [Supplementary Figure 5, https://links.lww.com/CM9/C892]. In the subsequent multivariable logistic model fitted with these two variables, accumulated VTC length remained independently associated with diagnostic success (OR: 3.163, 95% CI: 1.343–7.451, P = 0.008), while lesion diameter on ultrasound showed a non‑significant trend (OR: 1.370, 95% CI: 0.846–2.218, P = 0.200), suggesting that it likely serves as an adjusting variable to stabilize the model rather than as a core independent predictor. This study evaluated whether MOSE during EBUS‑TBNA using 22‑gauge needles can predict specimen adequacy. In previous MOSE studies, Masutani et al[4] reported an optimal VTC length cutoff of 11 mm for EUS-guided fine-needle biopsy (FNB) using 22-gauge needles, likely attributable to the use of higher magnification microscopy, which may have increased procedural complexity. Conversely, a subsequent study employing 22-gauge Franseen needles for EUS-FNB without magnifying instruments determined an optimal cutoff of 10 mm.[5] Notably, our identified accumulated VTC length cutoff of 7.75 mm is well within the range reported in prior EUS literature. A recent meta-analysis highlighted substantial inter-observer variability in the macroscopic evaluation of VTCs. Given these limitations, we explored the diagnostic value of formed tissue strip weight. We found that specimens with formed tissue strips weighing ≥36.5 mg demonstrated a significantly higher diagnostic yield, albeit with relatively low sensitivity (45.7%). Notably, 4 of 7 specimens lacking formed tissue strips nonetheless yielded definitive histologic diagnoses. These fragmented, difficult-to-isolate specimens warrant further investigation to clarify the relationship between fragment weight and diagnostic success. We also identified several macroscopic characteristics distinct from typical EUS-FNA or EUS-FNB specimens. First, anthracotic pigment deposition—likely related to tuberculosis, smoking history, or environmental exposure—was strongly associated with benign lesions, consistent with previous Korean studies.[6] Second, metastatic carcinoma, tuberculosis, and infectious diseases frequently exhibited focal degenerative changes with necrotic material within lymph nodes. Such necrotic foci may macroscopically resemble VTCs, potentially leading to misinterpretation; however, necrotic tissue can generally be distinguished by its friable, “melting” consistency. Finally, some specimens contained transparent or translucent tissue strips that proved, on histology, to be coagulated lymphatic fluid [Supplementary Figure 6, https://links.lww.com/CM9/C892]. This study has several limitations. First, it was an exploratory, single-center investigation with a modest sample size. Second, per-pass tissue quantity was not recorded, though this could help optimize the number of passes. Third, although the measured weight of formed tissue strips may have been slightly influenced by residual cell preservation fluid, this step likely minimized the confounding effect of blood clots. Fourth, nonspecific benign diagnoses were intentionally classified as “non‑diagnostic” to avoid incorporation bias which may have misclassified some cases with adequate specimen quality but intrinsically nonspecific benign pathology. Fifth, the study cohort predominantly comprised common benign and malignant pulmonary lesions. The cutoff established in this study may therefore not be generalizable to rare mediastinal diseases (e.g., lymphoma, immunoglobulin G4‑related disease, Castleman disease), which often require larger specimens for definitive diagnosis. Finally, the optimal VTC length for different needle types and for molecular or immunohistochemical analyses remains to be determined. In summary, this study confirms that MOSE provides a reliable real‑time assessment of specimen adequacy during EBUS‑TBNA performed with a 22‑gauge needle. An accumulated VTC length of ≥7.75 mm accurately predicts diagnostic yield for common benign and malignant pulmonary lesions. MOSE is a simple, reliable method for real‑time EBUS‑TBNA specimen assessment, particularly where ROSE is unavailable. Acknowledgements The authors thank Shanghai Municipal Hospital Respiratory and Critical Care Medicine Specialist Alliance. We also thank Professor Xiaofei Ye from the Department of Military Health Statistics, Faculty of Health Services, Naval Medical University, for his valuable assistance with the statistical review of this manuscript. Funding This study was supported by the Military Key Clinical Specialties Medical New Technologies Innovation Development Project (No. ZXJS2024B05), Collaborative Innovation Cluster Project of Shanghai Municipal Health Commission (No. 2020CXJQ03), and National Natural Science Foundation of China (No. 82270112). Conflicts of interest None.