In the present study, we analyzed preoperative diagnosis of 293 patients with pancreatic cancer. The overall diagnostic sensitivity of EUS-FNA for pancreatic cancer was 94.4% which was significantly higher than that of ERCP (45.5%) (p<0.001). Adverse event rates associated with endoscopic procedures were 1.3% (2/160) with EUS-FNA, whereas 10.2% (9/88) with ERCP related procedures, and the adverse event rate was significantly lower with EUS-FNA (p=0.001). In the comparison between patients underwent EUS-FNA (FNA group) and those who did not undergo FNA (non-FNA group), there were no significant difference in the surgical outcomes, surgical curability, and recurrence rate. Kaplan–Meier analysis also confirmed the absence of a significant difference in overall survival between the two groups (p=0.735). In the Cox proportional-hazards model for the overall survival, EUS-FNA was not an independent risk factor.
There are two methods to obtain pathological specimens from pancreatic tumors: the transpapillary approach (ERCP) and the transintestinal approach (EUS-FNA). The sensitivity of malignant diagnosis of the transpapillary approach is not sufficient, and endoscopic-related adverse events such as post-ERCP pancreatitis are also a concern (17). On the other hand, EUS-FNA has higher preoperative diagnostic capability than other modalities, with a diagnostic accuracy of 75%-95%(18-23). It is noteworthy that the reported specificity of EUS-FNA for pancreatic solid neoplasms is almost 100% (5) and that the associated complication rate is < 1% (24). Previous study by Wakatsuki et al. compared diagnostic ability of EUS-FNA (53 patients) and ERCP (30 patients) in pancreatic mass. They reported that in the FNA group, the overall results for the available samples were sensitivity 92.9% and accuracy 94.3%. In contrast, in the ERCP group the overall results were sensitivity 33.3% and accuracy 46.7%. The result suggested that, EUS-FNA is more accurate for the cytopathological diagnosis of suspected pancreatic masses as compared with cytology during ERCP(17). In this study, the diagnostic ability of EUS-FNA revealed a sensitivity of 94.4% (151/160), it was significantly higher than that of ERCP 45.5% (40/88). The major adverse events reported with EUS-FNA are infections, bleeding, pancreatitis, and duodenal perforation, with an overall rate between 1 and 2 %(25). In our study, the adverse event rate of EUS-FNA was only 1.3% (2/160), all were relieved with conservative treatment. It was significantly lower than that of ERCP. Considering diagnostic ability and short-term adverse events, EUS-FNA is considered to be an adequate and much-needed procedure in preoperative diagnosis of patients with suspected pancreatic cancer.
There are a few articles reported about long-term outcomes of preoperative EUS-FNA in patients underwent surgical resection of pancreatic cancer. A retrospective study by Ngamruengphong et al. evaluated 256 patients underwent pancreatectomy, including 208 patients had EUS-FNA for pancreatic tumor (FNA group) and 48 patients who did not have FNA (non-FNA group) with the median length of follow-up period of 23 months (range 0 - 111 months) and showed a gastric or peritoneal recurrence in a total of 19 patients: 13 patients (7.7%) in the FNA group versus 6 patients (15.4%) in the non-EUS-FNA group (p=0.21). In this study, three patients had recurrence in the gastric wall: one (2.6%) patient in the non-EUS-FNA group versus two patients (1.2%) in EUS-FNA group (p=0.46) (8). Another retrospective study by Kudo et al. evaluated 82 patients with resectable pancreatic cancer. Of these, 54 underwent EUS-FNA before surgery (FNA group) and 28 underwent surgery without preoperative EUS-FNA (non-FNA group). The study reported thatthe median relapse-free survival (RFS) of FNA group and non-FNA group was 742 and 265 days, respectively (p=0.009), and the median overall survival (OS) was 1042 and 557 days, respectively (p=0.007). FNA group had better RFS and OS than non-FNA group, because more patients in the FNA group underwent adjuvant chemotherapy. Multivariate analysis revealed that tumor size and adjuvant chemotherapy were both prognostic factors in this study. It was possible that patients in the FNA group benefited from the chemotherapy administered immediately after surgery (9). Tsutsumi et al. also carried out a retrospective study evaluating the impact of preoperative EUS-FNA. They divided 209 patients with pancreatic cancer into two groups: 126 patients who underwent preoperative EUS-FNA (FNA group) and 83 patients who did not undergo preoperative EUS-FNA (non-FNA group). They evaluated long-term outcomes of preoperative EUS-FNA, especially disease-free survival, needle tract seeding and recurrence. Kaplan-Meier analysis indicated no significant difference in disease-free survival between the FNA and non-FNA groups. The site of recurrence was not significantly different between the two groups, and needle tract seeding was not observed (10). In our study, the surgical curability (R0) was not significantly different between the FNA and non-FNA group. And there was no significant difference among two groups in the recurrence rate, peritoneal dissemination incidence, and survival time after surgery. Moreover, in multivariate analysis of factors related to prognosis, staging, curability and adjuvant chemotherapy were identified as dominant factors, but EUS-FNA itself was not. These studies including our one indicate that preoperative EUS-FNA does not adversely affect surgery or prognosisnor does it increase the risk of gastric wall or peritoneal recurrence in patient with resectable pancreatic cancer.
Recently, neoadjuvant chemotherapy for borderline resectable pancreatic cancer has been reported to have excellent effects on survival (26-30). Motoi et al. also evaluated neoadjuvant chemotherapy for resectable pancreatic cancer. They performed a randomized, controlled trial to compare neoadjuvant chemotherapy using gemcitabine and S-1 (NAC-GS) with upfront surgery (Up-S) for patients planned resection of pancreatic cancer. It found that the NAC-GS group showed a significant increase in overall survival (Prep-02/JSAP05) (31). Considering these results, importance of preoperative pathological diagnosis might increase even in cases with indications of surgical resection. EUS-FNA is recommended as the first choice for tissue sample collection method considering the high diagnostic capability and lower adverse event rate in comparison with transpapillary approach.
Although preoperative EUS-FNA for pancreatic cancer did not influence RFS, OS, or peritoneal dissemination as we discussed above, this is a matter of concern especially in patients with pancreatic body or tail cancer, since FNA was performed into the pancreatic tumors through the intestinal wall and peritoneal cavity and needle tract is not included surgically resected area. There have been several case reports regarding suspected tumor seeding related to preoperative EUS-FNA. In those reports, all recurrences occurred in the gastric wall (11-14). Minaga et al. summarized the clinical features and outcomes of 15 cases with needle tract seeding. In 13 (86.7%) of 15 cases, EUS-FNA was performed via the gastric body to preferentially diagnose pancreatic body or tail lesions. The remaining 2 cases performed EUS-FNA for perigastric and mediastinal lymph nodes(32). Yane et al. investigated the long-term outcomes, including the needle tract seeding ratio, of patients undergoing distal pancreatectomy for pancreatic body and tail cancer diagnosed preoperatively by EUS-FNA. Of the 301 total patients analyzed, 176 underwent preoperative EUS-FNA (FNA group) and 125 did not (non-FNA group). The median RFS or OS was not significantly different between the FNA group and the non-FNA group (23.7 vs 16.9 months: p=0.205; 48.0 vs 43.9 months: p=0.392). However, six patients (3.4%) in the FNA group were diagnosed as having needle tract seeding, the 5-year cumulative needle tract seeding rate estimated using Fine and Gray’s method was 3.8% (95% CI 1.6-7.8%)(33). These data suggest that needle tract seeding after EUS-FNA is observed to have a non-negligible rate.In particular, when performing EUS-FNA for a resectable tumorlocated in the pancreatic body or tail, there is a possibility of needle tract seeding, so it is necessary to pay attention to the size of the needle and the number of punctures.
The present study has several limitations. a retrospective study design might biased study outcomes. Since the study included only high-volume center for both EUS and ERCP related procedures, the external viridity might be low especially in the endoscopic procedure.
In conclusion, EUS-FNA was safe with high diagnostic ability as a preoperative examination of pancreatic cancer. It was considered to be the first choice without the influence of surgical curability, postoperative recurrence, peritoneal dissemination and prognosis. However, since needle tract seeding is observed to have certain probability, EUS-FNA for resectable tumors especially located in the pancreatic body or tail, requires careful consideration of the relationship between risk and benefit. A randomized controlled trial in a multicenter setting is needed to confirm the study results.