SETs can be identified via endoscopy or EUS. The size and the pillow sign can be evaluated with endoscopy. EUS can be used to distinguish intramural from extramural lesions as well as the origin of the layer and echo patterns such as hyperechoic, hypoechoic, or anechoic lesion. However, the diagnostic accuracy of EUS is relatively low without histologic confirmation [7, 11]. The concordance of EUS with the histology is only about 43–66.7% [7, 11]. In our study, the agreement between EUS-FNTA and EUS (48.5%, 16/33) was similar to that of a previous study.
SETs can be histologically diagnosed via both endoscopic and surgical resection [12, 13]. Recently, new techniques such as endoscopic full-thickness resection and endoscopic submucosal tunnel resection were developed to remove SETs in deeper layers where it was difficult to enucleate SETs using conventional methods [14, 15]. However, it is not always easy to perform histological diagnosis via endoscopic or surgical resection for all benign looking lesions because of their invasiveness and considerable complication rates. Therefore, it is reasonable to perform EUS-FNTA for histologic confirmation of SETs [16, 17].
Several studies confirmed the usefulness of EUS-FNTA.[18–21] EUS-guided fine-needle biopsy for suspected GI stromal tumors was technically similar and the safety was equivalent to that of fine-needle aspiration, with better tissue acquisition [18]. Needles of various sizes ranging from 19 to 25 gauge were used to perform EUS-FNTA [10, 19, 22, 23]. It was assumed that the19-gauge large-bore needles might be increase the diagnostic yield of SETs compared with small-bore needles. However, the 19-gauge needle showed no superiority compared with 22-gauge needles [10]. Various needle passage frequencies (1 to 7) were used in other studies [10, 22]. To date, however, no clear consensus is available regarding the optimal frequency of needles. In the present study, the mean frequency of needle passage was 2.78. Our method using EUS-FNTA was similar to that of other studies.
In the present study, the histopathological diagnosis of gastric SETs based on FNTA in 33 cases was analyzed. GIST was the most common cause of SETs. The diagnostic yield was 69.7% (23/33). The reliability of diagnostic yield with EUS-FNTA was 100% (13/13). EUS-FNTA was more accurate compared with EUS only (82.4%, 14/17). The concordance between EUS and EUS-FNTA was relatively low (48.5%, 16/33).
In our study, the needle passage was more frequent in the lower third than in the upper to mid third, which was 3.6 times versus 2.79 times. However, the diagnostic yield of SETs obtained via EUS-FNTA in the lower third of the gastric region was very low (0/5). On the other hand, a considerable diagnostic yield of SETs was obtained with EUS-FNTA in the upper to middle third of the stomach (23/28, 82.1%). According to published studies in English language, three studies reported the diagnostic yields of EUS-FNTA according to the gastric location [10, 24, 25]. Eckardt et al. [10] showed lower diagnostic yield in the lower third area (36.4%) compared with the upper to middle third of the stomach (75%, p = 0.008). Lee et al. [24] (25% vs. 59%, p = 0.183) and Suzuki et al. [25] (33.3% vs. 80.5%, p = 0.013) also showed a lower diagnostic yield in the lower third compared with the upper to middle third of the stomach. The diagnostic yield of overall cases, including our study, were significantly lower in the lower third of the stomach compared with the upper to middle third (29.7% vs. 71.4%, p < 0.001, n = 191) (Table 3).
Table 3
Review of four original studies published in English showing diagnostic rates of EUS-FNTA by gastric location
Case (n) | Biopsy needle (gauge) | Size of SET (median) (mm) | Diagnostic rate depending on location of SET (%) | Upper + middle vs Lower | References |
| | | Upper third | Middle third | Lower third | P value | |
46 | 19G | 24 | 10/12 (83.3) | 8/12 (66.7) | 8/22 (36.4) | 0.008 | Eckardt [10] |
65 | TCB | 37 | 15/26 (57.7) | 21/35 (60) | 1/4 (25) | 0.183 | Lee [24] |
47 | 22G or Echo Tip | N/A | 22/26 (84.6) | 11/15 (73.3) | 2/6 (33.3) | 0.013 | Suzuki [25] |
33 | 20G (90%) | 26 | 8/10 (80) | 15/18 (83.3) | 0/5 (0) | < 0.001 | Our case |
191 | | | 55/74 (74.3) | 55/80 (68.8) | 11/37 (29.7) | < 0.001 | |
EUS-FNTA: endoscopic ultrasound-guided fine-needle tissue acquisition; G: gauze; TCB: true cut biopsy, N/A: not available, SET: sub-epithelial tumor |
The two most common causes of non-diagnostic EUS-FNTA were insufficient tissue acquisition and puncture failure in the gastric wall. The antral wall of the stomach is known to be relatively thicker than that of the corpus or cardia. Lee et al. [24] reported that the puncture failure rate was relatively higher in the lower third (50%) compared with upper to middle third (11.4%) of the stomach. These results suggested that thickening of the gastric wall influences the diagnostic yields of EUS-FNTA. Adequate needle gauge and frequency of needling vary with the gastric location of SETs. Suzuki et al. [25] suggested that it was more difficult to maintain a scope stably in the lower third area of the stomach. In addition, it is known that ectopic pancreas is more frequently detected in the antrum compared with corpus or cardia [26–29]. Ectopic pancreas is a heterogeneous lesion associated with muscular wall thickening, so it might be difficult to obtain adequate tissue from ectopic pancreas using EUS-FNTA [30].
Recently, an electrocautery-enhanced delivery system was applied to facilitate self-expandable metal stent insertion under combined endoscopic and EUS guidance at once [31–33]. EUS-FNTA needle with an electrocautery function may enhance the penetration ability for thick gastric antral wall and reduce the mechanical force to the wall. The electrocautery function can overcome difficulty of the antral FNTA procedure to obtain a better sample rate.
Our study has several limitations. First, it was a single-center study with a retrospective design based on observational data. Therefore, the possibility of selection bias exists. Second, in the absence of on-site cytopathologists during the EUS-FNTA procedure, the specimen adequacy was only assessed macroscopically by endosonographers.