In this manuscript, we investigated the efficacy of EUS-FNB using a Franseen needle for UR SPL patients. As a result, more histological specimens were obtained by EUS-FNB than by EUS-FNA. Furthermore, EUS-FNB required fewer punctures than EUS-FNA for diagnosing UR SPLs, and the MSI evaluation was achieved more often by EUS-FNB than EUS-FNA.
Several factors that may affect the diagnostic value of EUS-FNA have been reported. Regarding the relationship between EUS-FNA and the tumor size, there are multiple points of view [5, 21]. Haba et al. [21] reported that in cases of smaller lesions, the diagnostic value of EUS-FNA is reduced. On the other hand, Uehara et al. [5] found that the size of the tumor does not influence the diagnostic value of EUS-FNA. It is true that EUS-FNA is difficult for small lesions; however, it is not true that the diagnostic value of EUS-FNA increases with increasing tumor size because the presence of necrosis could affect the accuracy of EUS-FNA. Therefore, contrast-enhanced EUS-FNA was reported to evade the puncture of necrotic areas [22–24]. Necrotic areas can be observed more clearly in larger SPLs than small SPLs, and it is difficult to identify necrotic areas using only B-mode imaging. In this study, the UR SPLs were larger than the R SPLs (median (range): 30 (10–82) mm vs 20 (7–40) mm, P value < 0.01). Although the evaluation of necrotic areas on EUS is difficult, the superiority of EUS-FNB using a Franseen needle over EUS-FNA was proven in not only SPLs but also UR SPLs.
In past reports, the frequency of MSI-high tumors was 0–29% in pancreatic cancer [25–30]. However, frequencies of MSI-high tumors of more than 10% were reported in small studies. In a study with a large number of patients (n = 3954) [27], MSI-high pancreatic cancer was found in very few patients (0.5%). In a study that targeted 12,019 patients with 32 types of cancer, the rate of MSI-high tumors was less than 2% in pancreatic cancer [31]. Therefore, it is reasonable that MSI-high tumors were not observed in this study.
What about the difference in the measurement method? In recent reports using next-generation sequencing, the frequency of MSI-high tumors in pancreatic cancer was 0.5%, or less than 2% [27, 31]. On the other hand, Yamamoto et al. [29] reported that 13% of pancreatic cancer patients showed MSI-high tumors according to PCR, and the results were satisfactory. In the report written by Yamamoto et al. [29], the Bethesda panel was used. In this study, the Promega panel was used. In the report written by Murphy et al. [32], the Promega panel was superior to the Bethesda panel. Therefore, there should be no problem with the MSI measurement method in this study.
As described above, the frequency of MSI-high tumors is low in pancreatic cancer. However, dramatic efficacy in MSI-high pancreatic cancer has been reported for pembrolizumab [30]. Therefore, EUS-FNB using a Franseen needle should be performed to collect sufficient specimens for MSI evaluation.
There are several limitations to this study. First, this study was a small retrospective study performed at a single institution. In the future, it is desirable for a multicenter prospective study to prove the argument of this study. Second, the types of SPLs were not unified. However, only UR pancreatic cancer patients were involved in the MSI comparison. Therefore, there was no influence of the SPL type in the MSI comparison between FNB and FNA. Third, the number of cells was not mechanically counted. Instead, the suitability of histological specimen sampling was compared between the two groups. The histological evaluation was performed by experienced pathologists.