Abstract
Background: Pancreatic mass sampling has historically been performed by endoscopic ultrasound-guided fine needle aspiration (EUS-FNA). However, its sensitivity has been reported to be within a wide range, which limits its reliability. Fine needle biopsy (FNB) has been shown to have superior diagnostic performance and is increasingly replacing fine needle aspiration. In FNA, 25 gauge (G) needles appear to outperform 22G. Data comparing these sizes in FNB platforms is limited. We aimed to prospectively compare the performance of 22G and 25G Franseen-tip core biopsy needles in the sampling of solid pancreatic lesions. Patients and Methods: Patients who underwent EUS-FNB of pancreatic lesions at the Indiana University Hospital using 2 needle sizes: 25G (Study group) and 22G (Control group) using the Acquire needle (Boston Scientific Co., Natick, MA, USA) were enrolled. Needle choice was left to the discretion of the endosonographer. Tissue specimens were evaluated onsite, and underwent touch and smear and cellblock preparation. Specimens were independently evaluated by 2 expert cytopathologists blinded to diagnosis. Cytopathologists assessed cytological yield (on smears) and histological yield (on cellblock) using a validated scoring system reached by a consensus among our cytopathologists as we have previously published. Results: A total of 75 patients (42 males, median=65 years) underwent EUS-FNB during the study period (2017-2018): 50 using 25G and 25 using 22G needle. Diagnostic yield was numerically higher in 25G (98% vs. 88%, p=0.105). Number of passes for smears were similar, however the 25G group required additional passes for cell-block (1.6 vs. 0.4, p=0.001). 25G was used more frequently for pancreatic head and uncinate process sampling (70% vs. 52%, p=0.126). Four patients had self-limited adverse events in the 22G group, but none in the 25G group. Conclusion: We report no difference in the diagnostic yield between 25G FNB vs. 22G sampling device with Franseen style tip, however, the 25G needle use was associated with the need of additional passes to collect a sufficient cell block.
Endoscopic-ultrasound (EUS)-guided tissue acquisition of pancreatic lesions has primarily been conducted by fine needle aspiration (FNA). This method carries minimal risk of complications and has largely been successful; however, a major drawback is a wide reported range of sensitivity between 65%-92% (1, 2). This implies that the false-negative rate can be unpredictably high, and therefore a negative FNA may not rule-out pancreatic malignancy.
Fine-needle biopsy (FNB) is rapidly gaining popularity for sampling of pancreatic lesions. FNB allows for better preservation of tissue structure and hence histological analysis. This is beneficial in difficult cases that include significant inflammation or fibrosis from chronic pancreatitis that can obscure the diagnosis of malignancy. FNB may also provide an adequate sample for immune-histochemical staining, which is necessary for confirming neuroendocrine tumors and for the diagnosis of autoimmune pancreatitis (3-5). Lastly, a recent meta-analysis found that FNB requires fewer passes over FNA and maintains a similar diagnostic yield (6). This could result in overall shorter procedure duration while maintaining the same diagnostic utility (7).
The success of FNB is due to incremental improvements in the design of the needles over several generations. Earlier FNB platforms, such as 19 gauge Trucut® FNB needle (Cook Endoscopy, Winston Salem, NC, USA), offered improved tissue procurement over FNA but were also associated with high blood and cellular contamination. Further issues with the Trucut® needle include needle misfiring due to a spring-loaded mechanism and malfunction when the endoscope was in a torqued position, which is necessary for sampling pancreatic head and uncinate process masses (8). Thus, it was clear that both smaller and more flexible needles were needed. The 22 and 25 gauge Procore® (Cook Endoscopy, Winston Salem, NC, USA) were later introduced. The Procore® needle was an improvement over the Trucut®, but failed to show definite superiority over FNA in diagnostic adequacy according to several studies (9). One of the more recent additions to the FNB line of needles was the Acquire® (Boston Scientific, Natick, MA, USA), which features a Franseen tip equipped with three symmetrical cutting edges. Previous studies have demonstrated superior diagnostic adequacy of the Acquire® 22G FNB over FNA (10, 11).
Multiple studies have demonstrated the superiority of FNA using 25G compared to 22G needles (12, 13). In those studies, it was believed that smaller size needles lent more flexibility, thus allowed technically easier sampling of masses in the pancreatic head and uncinate process. Additionally, it is believed that a smaller caliber needle could reduce contamination from blood, non-pancreatic tissue, and tissue debris. It is unclear if the trend of improvement with smaller needles applies to FNB platforms as well. The aim of our study was to assess the diagnostic performance and safety of the 22 and 25G Franseen tip core biopsy needles in sampling solid pancreatic lesions.
Patients and Methods
Study design and population. This was a single-center non randomized prospective cohort study approved by the Indiana University School of Medicine Institutional Review Board. Consecutive patients referred for EUS-guided sampling of solid pancreatic lesions at the Indiana University Hospital were prospectively enrolled between January 2017 and December 2018. We used 2 needle sizes: 25G (Study group) and 22G (Control group) using the Acquire® needle (Boston Scientific Co., Natick, MA, USA) (Figures 1 and 2). The choice of needle size was left to the discretion of the endosonographer based on lesion location and characteristics. Only patients with available cell block material were included in the study.
Patient demographics and lesion-related characteristics, including cytological and pathological findings, the size and location of target lesions, and clinical outcomes, were collected. FNB-related data included: cell block histology, smear cytology score, number of passes for smears, number of additional passes for cell block, and total number of passes.
All specimens were evaluated by two expert cytopathologists to assess the cytological yield (on smears) and histological yield (on cell block preparations) using a validated standard scoring system we have previously published (Table I) (11). Both cytopathologists were blinded to the formal diagnosis rendered in each case.
Study definitions. A clear, non-equivocal diagnosis of malignancy was based on neoplastic cyto-histopathology by endoscopic sampling (FNB) and/or surgical resection. When cyto-histopathology was non-diagnostic, the diagnosis of malignancy was determined based on neoplastic EUS impression, in addition to clinical course, and imaging findings consistent with malignancy at ≥6 months of follow-up. Diagnostic yield for each needle was calculated based on the proportion of patients that underwent FNB that received a definitive diagnosis out of the total number of patients that underwent the procedure. Patients who had cyto-pathology results negative for malignancy but were ultimately found to have malignancy based off the aforementioned criteria, were determined to be non-diagnostic.
EUS-FNB technique. All procedures were performed by the same experienced endosonographer. EUS-FNB was performed using a linear array echoendoscope (Olympus America Inc., Center Valley, PA, USA) and a 22G or 25G Acquire® FNB needle (Boston Scientific Co., Natick, MA, USA). Monitored anesthesia care sedation was provided by anesthesiologists for all procedures. The slow-pull technique was implemented in removing the stylet in the first pass. For subsequent passes, application of 10 ml of negative pressure for 20 s, during which 4-6 actuations were performed. The fanning technique was employed whenever feasible. Following the needle withdrawal, obtained tissue was processed on site, and touch preparations or crush smears were prepared. Slides were air-dried and stained with modified Wright Giemsa staining for ROSE. If the first smear was inadequate on ROSE, the endosonographer performed additional smears until it turned adequate, or it was felt that further sampling would not increase the smears yield. All additional visible cores were placed in a cellular preservative (CytoRich Red; Becton Dickinson, Franklin Lakes, NJ, USA) for cell block preparation (Figures 3 and 4). Cell block was visually inspected to ensure sufficient material. This was determined by cloudiness of fluid rather than relying on the presence of whitish cores that can be missed on visual exam, especially mini-cores which may only affect fluid turbidity.
Follow-up. Procedure-related outcomes were collected prospectively. Our endoscopy Unit contacted all patients within 48-72 h post-procedure to assess for adverse events. Long-term adverse events were assessed through medical chart review. Adverse events were recorded according to the published American Society of Gastrointestinal Endoscopy criteria (14). Additional follow-up information after EUS was performed by review of medical charts and contact with the referring physicians.
Statistical analysis. Descriptive analysis was used according to the type of the variables used. A 2-tailed distribution was used and a p-value <0.05 was considered statistically significant. The categorical variables were measured as the count and percentage using the Chi-square test. Continuous variables were measured as the mean and standard deviation using Student t-tests. The sample size was not estimated a priori as this is the first study conducted with this novel needle in pancreatic solid masses. We chose 50 samples in the test group and 25 samples of controls as a convenience sample (2:1 inclusion ratio). Statistical analysis was carried out using IBM SPSS Statistics, version 23.0 (SPSS, Chicago, IL, USA).
Results
Between October 2017 and Dec 2018, 75 patients (42 males, median age=65 years) with solid pancreatic lesions underwent EUS-FNB: 50 using a 25G and 25 using 22G needle.
Baseline characteristics are reported in Table II. There was no significant difference in the age of patients between both groups (64±8 vs. 67±11, p=0.162). The final diagnosis was adenocarcinoma in the majority of masses (84% for 22G vs. 78% for 25G, p=0.327). No difference in pancreatic lesion size was found in the lesions in either group (3.2±1.8 cm vs. 3.3±1.3 cm, p=0.724). The number of passes for cytological smears were similar between both groups; however, the 25G group required more additional passes to obtain an adequate cell-block (1.6±0.6 vs. 0.4±0.7, p=0.001). The additional passes were necessary to build up adequate cellularity on visual exam.
Needle performance is reported in Table III. Cytological scores of smears were comparable between the two groups. Due to the 25G needle's increased flexibility in locations that involved scope angulation, it was utilized more often to sample lesions in the pancreatic head and uncinate process (70% vs. 52%, p=0.126). Diagnostic yield was higher among 25G group compared to 22G group with no statistical significance (98% vs. 88%. p=0.105).
Adverse events. All adverse events were classified as mild, and were recognized during or immediately after the procedure. Two patients developed post-procedural abdominal pain after pancreatic mass FNB using the 22G needle (confirmed adenocarcinoma on cytology). This required observation in the recovery room for 3 h in one patient and 23-h admission to the hospital for observation and pain management in another patient. In two cases who underwent FNB using the 22G needle, a hematoma developed outside the gastric and duodenal walls during FNB of the pancreatic body and head masses, respectively. Both patients underwent 4 passes each. The hematomas were monitored for 10 min and appeared stable on EUS. This did not result in clinically apparent adverse outcomes and both patients were discharged home after 90 min of observation without the need for further interventions. No patient developed any adverse events in the 25G group. No other adverse event was reported during or immediately after the procedure and up to 72 h of follow-up.
Discussion
Our primary aim was to compare the diagnostic performance of two FNB needles based on size. Previous FNA studies have shown a trend towards superior diagnostic performance in smaller caliber needles. Based on this, we hypothesized a similar result in our FNB study. We found a numerically higher diagnostic yield in the 25G FNB, however our results did not reach statistical significance. We believe the 25G needle's reduction in tissue debris and increased flexibility in sampling masses that required a trans-duodenal approach balanced the effect of the comparatively smaller tissue acquisition. The similarity in performance between both needles is supported further with no difference noted between smearing success, and cellularity of smears; an important feature in practice settings where ROSE is available, where the ability to render an onsite judgement on adequacy is important in the context of needles that provide mainly core-quality samples. Additionally, smear cytology and cell block histology scores devised by our own cytopathologists did not differ between both groups.
We report a significant increase in the number of passes required for cellblock in the 25G group. We believe this was related to the smaller diameter of the 25G needle and hence more passes were required to build up adequate cellularity for cellblock. In contrast, the 22G needle did not require additional passes as frequently due to producing an excess amount of tissue sample that was sufficient for both smears and cellblock preparation. We acknowledge that the diagnostic performance of 25G FNB could have been improved with multiple passes, however the magnitude is not measurable based off our study's design; we allowed unlimited passes to reach adequate cellularity and thus are not able to calculate the diagnostic yield per pass.
A secondary aim of our study was to assess the safety of both needles. In the 22G group, we noted a mild hematoma in two patients and abdominal pain requiring hospital admission in another two patients. There were no adverse events in the 25G group despite having twice the number of patients. We hypothesize that a larger caliber needle has an increased propensity to cause tissue damage and result in bleeding. Our study was not powered to detect subtle differences in adverse events. Larger studies are needed to confirm our findings.
It should be noted that the 25G needle was used more often than 22G in sampling pancreatic head and uncinate masses (70% vs. 52%, p=0.105). Needle selection was at the operator's discretion in this non-randomized study. Throughout the trial, our operator selected the 25G for head and uncinate masses due to it being technically less challenging when scope-torque was necessary with a trans-duodenal sampling approach.
There is a paucity of studies comparing the outcomes of various FNB platforms. To our knowledge there is only one study conducted by Park et al. comparing 22G to 25G FNB. Our results are consistent with that study in which 25G FNB was associated with numerically higher histological procurement and diagnostic accuracy compared to 22G, but results also did not reach statistical significance (15). One caveat is that they used a different needle platform (reversed bevel Procore®) which was an earlier core needle design that failed to demonstrate superiority over FNA in previous studies (9). Our results differed in that we found a lower histological procurement in 25G, as evidenced by the number of additional passes needed for cell-block.
Our study has several limitations. We acknowledge that because we used a non-randomized design there is a degree of selection bias. Our study did not sample the same lesion with both needle types to allow a direct, head-to head comparison. This may have increased the validity of our findings, however, it is theoretically possible that multiple biopsy attempts on a lesion can increase blood contamination in subsequent samples, thereby decreasing diagnostic performance and distorting results. To our knowledge, there are no studies that have assessed the impact of repeated passes on pathology specimen integrity on pancreatic masses in particular. Another limitation of our study was a low sample size of 75 total cases with a 2:1 inclusion of 25G vs. 22G cases. The 2:1 ratio was due to the availability of needles at our institution where 22G needles became available first followed by a later release of the 25G line of needles. We acknowledge that the lack of blinding introduced bias as evident by the increased usage of 25G to sample head/uncinate masses. In addition, we acknowledge that because we used a non-randomized design and convenience sample, there may be an increased sampling error. In other regards, our study did well in eliminating factors that can affect the diagnostic yield of any EUS-guided procedure. This includes sampling technique (which was standardized), number of passes, lesion-specific characteristics, location, and size, and endosonographer style and experience (1, 16, 17). We believe that our study did well in eliminating external variables by using the same endoscopist to perform all tissue acquisition and limiting cytological analysis to two blinded cytopathologists. The endosonographer on the study has performed over 1500 pancreatic mass samplings using various needle platforms. We acknowledge that there may still be a learning curve associated with using a new platform, but we expect that to be mastered during the course of this study.
Conclusion
FNB needles are rapidly gaining popularity in EUS sampling of solid lesions. Both the 22G and 25G Franseen tip FNB needles offer high sensitivity and specificity for the sampling of solid pancreatic lesions. We found that diagnostic yield was similar between both needles, however more passes were required in 25G needles to acquire an adequate cell block. Although 25G needles may have been easier to sample masses in technically challenging areas. The safety of the 22G Franseen-tip needle should be carefully assessed in prospective studies. Randomized-controlled studies are needed in the future to confirm our findings.
Footnotes
Authors' Contributions
D.A.S. performed the literature search. D.A.S. and M.A.H. conceived and designed the study. D.A.S., M.A.R., and M.A.H. interpreted the data. D.A.S., M.A.R., M.A.H., K.T. and H.H.W. collected the data, prepared and the edited manuscript. All Authors read and approved the final manuscript.
Conflicts of Interest
Dr. Al-Haddad has received research and teaching support from Boston Scientific company. Drs. Siddique, Rahal, Trevino, and Wu do not have any conflicts of interest or financial ties to disclose.
- Received June 29, 2020.
- Revision received July 18, 2020.
- Accepted July 21, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved