Macroscopic on-site magnifier-based evaluation to estimate visible tissue core cut-off lengths using EUS-FNA with 22-gauge needles

To the Editor: Endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) has replaced other high-risk invasive methods and has become the primary method of tissue collection around the gastrointestinal tract.[1] To further improve the effectiveness and safety of EUS-FNA, pathologists may perform rapid on-site cytology evaluation (ROSE), but this process requires the presence of pathologists. Macroscopic on-site evaluation (MOSE) has gained attention as it resolve the shortage of pathologist resources.[2] Endoscopists can inspect the tissue core in real time and assess the adequacy of the punctured material. In this study, we conducted a single-center and single-arm exploratory study to explore the value of MOSE in assessing specimen quality during magnifier-based EUS-FNA and to confirm the cut-off length of visible tissue core (VTC) for EUS-FNA to help endosonographers independently predict the adequacy of punctured specimens. The VTC was defined as a white or incarnadine tissue core obtained from focal lesions with a relatively complete histological structure. Consecutive patients who underwent EUS-FNA for solid tumors from April 2020 to April 2021 in Second Affiliated Hospital of Soochow University were recruited. Patients with solid tumor measuring >10 mm, around or inside the gastrointestinal tract were included. If a patient had two or more lesions, individual lesions were counted separately. Patients who had hemorrhagic tendency, cystic mass, a history of digestive tract diversion surgery, and lesions with a diameter less than 10 mm, or those were deemed unsuitable for the study by investigators for any reason were excluded. All patients provided written informed consent for study enrollment and for undergoing EUS-FNA. This pilot study was approved by the Ethics Committee of the Second Affiliated Hospital of Soochow University (No. JD-LK-2020-049-01) and was registered prior to patient enrollment at the Chinese Clinical Trial Registry (https://www.chictr.org, No. ChiCTR2000031529). All patients were randomized to undergo EUS-FNA by three experienced endosonographers. The endosonographers determined the most suitable path to puncture lesions with a 22-G needle (Echo Tip Ultra; Wilson-Cook, Winston-Salem, NC, USA). The stylet was removed after the needle was advanced into the target lesion. Then, either 5 mL or 10 mL of negative pressure was applied for dry suction. Wet suction was also used if the quality of the dry suction sample after 2 passes is unsatisfactory. After puncturing and aspirating, the needle was withdrawn from the lesion, and samples were pushed into the stylet before being transferred to a petri dish. Then, the stylet was removed, and the remaining specimen was sprayed with an air syringe onto a slide to prepare the smear. MOSE is a key technique in improving the accuracy of EUS-FNA. This technique can be divided into two parts. The first part is pretreatment of puncture specimens to clear the blood clot around samples. The second part is observation of puncture specimens. Our team has developed a VTC observation table consisting a magnifying glass with a light source and a petri dish with a scale at the bottom [Supplementary Figure 1, https://links.lww.com/CM9/B853]. This observation table provides an enlarged view of the specimen, which improves the VTC detection rate. Endoscopists observed the white or flesh-colored tissue strips through a magnifying glass (×5), and the length of the longest tissue strip was measured against the scale. The puncture was completed when the sample was judged to be of adequate quality under the magnifying glass. Each individual lesion was punctured for no more than four passes. The lesion size, puncture method, and VTC length of each needle were recorded [Supplementary Figures 2 and 3, https://links.lww.com/CM9/B853]. Following on-site sample observation, all samples were sent to the pathology department for pathological examination, including cytologic examination for smears with hematoxylin–eosin (HE) staining, liquid-based cytology examination, and histopathology examination for VTC-rich samples with HE staining and immunohistochemical staining. The cell block sections were independently evaluated by two pathologists. In case where there were conflicting evaluation in puncture specimens, a third pathologist was included to ensure quality assessment. The final score was determined as the average of the three scores. The tissue integrity was scored from 0 to 3 as follows: score 0 indicated that samples were not collected; score 1 indicated that the sample did not provide histological information (no architecturally intact tissue present to yield a diagnosis); score 2 indicated that the sample contained limited histological information; and score 3 indicated adequate sample quality for histological interpretation (an architecturally intact piece of tissue with a length of at least 550 µm under the microscope). A higher score indicates a higher proportion of the core tissue structure in the sample, providing more information to the pathologist. Typical pictures of each score of pancreatic cancer according to this classification are displayed in Supplementary Figure 4, https://links.lww.com/CM9/B853. All patients were followed up after EUS-FNA, and the gold standard was the pathological results of surgery. Patients who were clinically comprehensively evaluated as inoperable or negative for EUS-FNA were followed up for at least 6 months, and the final diagnosis was determined based on the patient’s clinical process and the imaging examination. Data were expressed as number (percentage) or median (Q1, Q3), appropriately. Statistical comparisons were made using chi-squared test or Fisher’s exact test for categorical variables. Spearman’s rank-order correlation analysis was used to assess the relationship between VTC length and sample quality. Kappa test was used to analyze the consistency between the histological diagnosis of EUS-FNA and the final diagnosis. The receiver-operating characteristic (ROC) curve of the VTC length with respect to the final diagnoses was plotted, and the optimal cut-off VTC length was determined with the Youden index. Multivariate analyses were also performed to find the factors affecting the tissue diagnosis. Statistical Package SPSS 21.0 (SPSS Inc, Chicago, IL, USA) was used for statistical analyses. A value of P <0.05 was considered statistically significant. Supplementary Table 1, https://links.lww.com/CM9/B853 shows the characteristics of the patients. A total of 79 patients fulfilled the inclusion criteria and were enrolled in the study. Fourteen patients had multiple masses around the gastrointestinal tract. EUS-FNA with a 22-G needle was attempted for 93 solid masses. The masses were mainly located in the pancreas (n = 52, 55.3%), and the median lesion maximum diameter was 25 (Q1, Q3: 18–40) mm. Supplementary Table 2, https://links.lww.com/CM9/B853 shows the EUS-FNA results, including the total needle passes, the median number of passes, the complication rate, and the number of different puncture methods. In this study, 93 lesions were punctured, and the EUS-FNA histological diagnosis suggested that 73 were malignant lesions, of which 13 lesions had pathological confirmation after surgery, while the other 60 patients who did not undergo surgery were proven clinically progressive during the follow-up. The remaining 20 lesions were diagnosed as benign by EUS-FNA, of which 8 were true negatives and 12 were false negatives. In this study, EUS-FNA yielded a sensitivity of 85.9% (73/85), specificity of 100.0% (8/8), positive predictive value of 100.0% (73/73), negative predictive value of 40.0% (8/20), and histological diagnostic yield of 87.1% (81/93). Meanwhile, the kappa statistic indicated a moderate agreement between EUS-FNA histological diagnoses and final diagnoses (kappa = 0.511, P <0.05). Supplementary Table 3, https://links.lww.com/CM9/B853 shows the subgroup analysis results of accuracy of different methods in the EUS-FNA histological diagnosis. There was no statistically significant difference between the diagnostic accuracy obtained by the different negative pressure and suction methods (P >0.05).There were 221 (88.8%) needle passes contained VTC, while 28 (11.2%) needle passes did not contain VTC. The histological diagnostic yield of needle passes contained VTC was significantly higher than that of without VTC (91.0%, 201/221 vs. 46.4%, 13/28, P <0.05). All samples containing VTC were evaluated by two pathologists to precisely assess the relationship between the VTC length and quality of EUS-FNA samples by quantifying the histological quality of the samples. Spearman rank correlation analysis revealed that the sample quality (the integrity of the tissue structure) was positively associated with the VTC length (rs = 0.427, P <0.05) [Supplementary Figure 4A, https://links.lww.com/CM9/B853]. The ROC curve of the VTC length showed a cut-off VTC length of 7.45 mm, with an area under the curve of 0.838 (95% confidence interval [CI], 0.759–0916, P <0.01) [Supplementary Figure 4B, https://links.lww.com/CM9/B853]. At this cut-off level, the sensitivity and specificity of EUS-FNA were 80.8% and 66.7%, respectively, with a highest Youden index of 0.475. Multivariable analysis was performed, and found that the maximum diameter of the lesion (odd ration [OR] 1.046, 95% CI: 1.008–1.086; P <0.05) and the VTC length (OR: 2.632, 95% CI: 1.157–5.989; P <0.05) were independent factors affecting the histological diagnosis rate of EUS-FNA [Supplementary Table 4, https://links.lww.com/CM9/B853]. However, the tissue diagnostic yield was not related to the location of the lesion and the puncture method. EUS-FNA is a minimally invasive biopsy method, providing a limited puncture opportunity to collect adequate specimens. To further improve its diagnostic accuracy, our group made a comprehensive improvement to the routine visual observation of FNA specimens and finally proposed a magnifier-based MOSE of tissue specimen quality. First, we referred to the observations of Iwashita et al[3]. Endoscopists could place the puncture sample directly from the needle lane into the petri dish, preventing the VTC from being crushed into smaller pieces on the slide, improving the detection rate of VTC. Second, the sample pretreatment method of Oh D et al[4] was considered, absorbing sample blood with filter paper before observation. In our study, the puncture samples were directly injected into the liquid-based cell preservation solution for rinsing, which reduced unwanted blood in the sample and could be sent for liquid-based cytology. This measure makes full use of the sample and provides a higher quality sample for pathological diagnosis. MOSE can provide the endoscopist with a clear endpoint to minimize unnecessary punctures by sample quality score. The median number of passes per lesion in this study was three, which was lower than the recommended number of EUS-FNA needle passes without ROSE.[5] The diagnostic accuracy of EUS-FNA in this study was 87.1%, indicating that MOSE can reduce the number of needle passes and avoid excessive puncture while increasing the diagnosis rate of EUS-FNA. Reducing the number of needle passes is particularly relevant for patients with pancreatic tail tumors. A multicenter retrospective cohort study found that tumors at this site carried an additional risk of needle tract metastasis through the gastric wall when undergoing EUS-FNA, with an incidence of 3.4%, which is much higher than the rate of tumors at other sites.[6] The incidence of diseases related to solid masses around the gastrointestinal tract is widespread, and their origin is complex and variable. Without a pathological diagnosis, it is not possible to formulate the next treatment plan. Meanwhile, as tumor-targeted therapy and immunotherapy are widely accepted, there is an increasing need to rely on tissue specimens for defining drug treatment options. Therefore, the aim of this study was to investigate the value of MOSE in judging the specimen quality during EUS-FNA based on magnifying glass; and found that the diagnostic efficacy of this method to assist EUS-FNA is satisfactory. This method is simple, cheap, and does not increase the burden on medical resources, patient, or operation difficulty, and it can be carried out in hospitals of all sizes. This study had several limitations. First, this is a single-center and single-arm exploratory study, with an insufficient sample size and uneven distribution of diseases. These factors may bias the study results, so the results should be confirmed in a multi-center study with a larger sample size. Second, even if a magnifying glass was used, accurate macroscopic evaluation of a sample for the presence of VTC remains difficult. This study developed simple visual assessment criteria for endoscopists, mainly relying on the length of VTC, which is based entirely on clinical judgment. Other parameters, which may provide a better assessment of sample adequacy such as quantifying the characteristics of VTC (intensity of hue, transparency, and hardness) using computer software, were not included in this study. In conclusion, with the help of MOSE, endosonographers can assess the specimen quality by VTC length to independently predict specimen adequacy. The histological diagnostic yield was significantly improved when the VTC lengths obtained by EUS-FNA were ≥7.45 mm. Performing EUS-FNA with this cut-off value as an index may be especially useful in institutions where ROSE is unavailable. Funding This study was supported by the Program for GUSU medicine talents (No. GSWS2020026), Program for Suzhou Science and Technology Project (No. SKY2021044), and the State Key Laboratory of Radiology Medicine and Radiation Protection (No. GZK1202214).