In this retrospective study, we examined the diagnostic efficacy of three puncture needles (PC20, PC22, and AC22) for the first EUS-FNA/B in patients with pancreatic cancer. The PC20 and PC22 employ a core trap system, which contributes to increased tissue acquisition and preservation of the histological structure [22]. A core trap can reportedly ensure effective cellularity for diagnosis [20, 23]. Acquiring tissue specimens with minimal structural damage increases the histopathological quality of the samples. High-quality specimens are crucial for assessment of the subtypes of malignancy (e.g., adenocarcinoma, adenosquamous carcinoma, anaplastic carcinoma, neuroendocrine carcinoma, and metastatic carcinoma) and performance of immunohistochemistry [24, 25]. Therefore, a core trap is an advantage in histopathological diagnosis of pancreatic tumors by EUS-FNA/B. However, although both the PC20 and PC22 have a core trap, the present study demonstrated that diagnostic accuracy was significantly higher with the PC20 than PC22. Two possible factors led to the difference in diagnostic accuracy between the PC20 and PC22: the difference in the needle gauge and the difference in the design of the core trap. The PC20 is a 20G needle, while the PC22 is a 22G needle. A thicker needle is reportedly more useful for collection of core tissue [26] and preservation of the histological structure of specimens [18] than is a thinner needle. We must also consider the difference in the shape and size of the core trap between the PC20 and PC22. Armellini et al. [27] reported that the forward-bevel core trap enables the acquisition of more tissue than the reverse-bevel core trap. This is thought to be because the forward-bevel core trap catches the tissue while the needle moves forward [27]. Additionally, the core trap is larger in the PC20 than PC22 (2.9 vs. 2.0 mm, respectively). However, there is little difference in the distance from the tip to the core trap between the PC20 and PC22 (3.8 vs. 3.9 mm, respectively). This suggests that the stroke length measured by EUS imaging does not drive the needle choice between the PC20 and PC22. Our findings suggest that the PC20 is preferable for the histological diagnosis. The total accuracy (combined histology and cytology) with the PC20 (96.4%) was equal to the histological result; i.e., there was no additional diagnostic impact of cytology. Additionally, even in samples defined as clotted specimens by MOSE, the histological accuracy of clotted specimens was 92.5% with the PC20. In one study, the histological sensitivity was 92.4% in whitish tissue and 40.8% in blood clots [28]. Although a sample obtained by a thicker needle is contaminated by more blood [29], our results suggest that clotted specimens obtained by the PC20 can still be used for histological assessment.
We also assessed the diagnostic accuracy of the AC22 in this study. The AC22 is a relatively new needle with a diagnostic accuracy of 94.0 to 96.7% [30–32]. Iwashita et al. [33] reported that the AC22 was better than a conventional needle in acquiring cells. The AC22 is designed to hold and cut the tissue with multiple sharp structures on the tip (fork tip). This characteristic structure contributes to better sample collection even in a lesion with dense fibrotic change, different from most conventional EUS-FNA needles [31, 34]. In an animal experiment, there was no significant difference in the core tissue acquisition ability between the PC20 and AC22 [21]. Considering that the AC22 is thinner (22G) than the PC20 (20G), the fork-tip structure of the AC22 could be more advantageous for tissue collection than the core trap of the PC20. However, a recent study proposed that the side bevel, which moves to and fro in the tumor during the sampling procedure, can cut the tumor surface more effectively than the needle tip within the tumor [35]. Thus, considering the tip structure of the AC22, this needle might be more effective than other types particularly for cases in which a long puncture stroke length cannot be secured. Fujita et al. [36] compared the diagnostic accuracy between the AC22 and a conventional needle and found no difference between the two needles for > 20-mm tumors; however, there was a significant difference for < 19-mm tumors. There was no significant difference in the histological accuracy between the PC20 and AC22 in the present study (p = 0.02, p-adj = 0.06). A sufficient stroke length was secured for all patients; however, given that the puncture stroke length was significantly shorter with the AC22 than PC20, it is difficult to draw a strict conclusion regarding the efficacy of the PC20 versus AC22 from this study. However, our findings suggest that as long as a sufficient stroke length can be secured, the diagnostic accuracy of both the PC20 and AC22 is sufficiently high. Given our results that the histological accuracy tended to be higher with the PC20 than AC22 as well as the small number of patients in whom the AC22 was used, we speculate that the PC20 is suitable for cases in which histological assessment including immunohistochemistry can be critical for diagnosis, while the AC22 is suitable for small tumors in which it is difficult to secure a long stroke length.
In EUS-FNA/B, transduodenal puncture is challenging in some cases because of the steep up-angle of the scope [22, 37]. A thicker needle with less flexibility has a disadvantage in this procedure, and endoscopists need to use additional needles in some cases. In this study, however, the operability of the PC20 was not different from that of the PC22 and AC22. This finding suggests that the PC20 can be used for transduodenal puncture similarly to 22G needles. In terms of adverse events, one case of mild pancreatitis was observed in the PC20 group (0.9%), and no adverse events occurred in the PC22 and AC22 groups. No serious complications such as active bleeding or perforation occurred with any needles. A systematic review on the safety of EUS-FNA revealed a complication rate of 0.98% and mortality rate of 0.02% [38]. The needle diameter is not thought to be associated with the adverse event rate [39, 40]. Considering the results of both our study and previous studies, it seems that there is no considerable difference in safety among the PC20, PC22, and AC22.
This study had several limitations. First, because it was a single-center, retrospective study, the EUS-FNA/B skills might have varied over time and/or among endoscopists, and the needle selection by each operator might have been biased. Second, the AC22 was approved for EUS-FNA/B for pancreatic solid lesions in October 2016 in Japan, and in our facility, the AC22 has been used in much fewer patients than the PC20 and PC22. Further cases using the AC22 are needed to evaluate the efficacy of this needle. Additionally, in this study, the stroke length of puncture under EUS images was not controlled between the PC20 and AC22. A study in which the stroke length is identical between the needles is necessary for direct comparison of the efficacy of these needles.