PG is regarded as a function preserving procedure for EGC located at the upper-third stomach and EGJ. The retrospective study by Yamashita et al. demonstrated that even for advanced adenocarcinoma of EGJ less than 4 cm in diameter, dissection of the distal LNs of the stomach is not necessary, such as No. 4d, 5, 6 and 12a LNs, arousing surgeons’ great interest for PG [
4]. However, the main obstacle for PG is the high incidence of postoperative reflux. Till now, there is no standard procedure for reconstruction after PG [
7]. DFT has attracted great attention since its introduction by Kamikawa et al. in 2001 for excellent anti-reflux effect [
7,
14]. However, DFT is extremely complicated and time-consuming. SOFY is a relatively simple anti-reflux procedure. But its anti-reflux effect varies among surgeons and patients, with a certain percentage of complications [
7,
15]. To overcome the shortcomings of the previous techniques, we designed esophagogastrostomy with ROSF following LPG.
Efficacy of anti-reflux
For esophagogastrostomy, anti-reflux effect is of great concern. Historically, simple esophagogastrostomy has a high rate of reflux esophagitis (20–65.2%) which prevented surgeons from performing PG. To overcome this, various techniques have been designed, among which DFT had the lowest incidence of reflux esophagitis (0–10.6%) [
7]. SOFY was also reported to have lower incidence of reflux esophagitis (7.1–17.8%), but higher than DFT [
15]. Table
5 shows the incidence of postoperative complications in patients who underwent ROSF and other reconstruction methods [
7,
15‐
19]. Compared with the these methods, it is encouraging that only one (5%) of our initial 20 patients had mild symptoms of reflux (Visick score = II) and endoscopic reflux esophagitis of LA Grade A was observed in only one patient (5%). However, the incidence of postoperative reflux esophagitis was not significantly different between ROSF and our previous experience of DFT in 23 patients (Table
6). The main mechanisms of anti-reflux of ROSF procedure may fall into two categories. The first is the compression of the anastomotic orifice and lower esophagus by the seromuscular flap. We adopted side overlap with the overlapping length of 3.0 cm, leaving a 5 mm long high-pressure zone in the esophagus above the highest point of the anastomosis. The second mechanism is assumed to be the twist of anastomosis similar to SOFY method. Although our series had low rate of symptomatic and endoscopic reflux, due to limited cases and time of follow-up, the actual performance of ROSF needs further investigation.
Table 5
Incidence of complications in reconstruction methods after proximal gastrectomy
EG (ROSF) | 0.0 | 0.0 | 5.0 |
EG (DFT) | 4.7–29.1 | 0.0–7.7 | 0.0–10.6 |
EG (SOFY) | 0.0–2.8 | 0.0 | 7.1–17.8 |
EG (gastric tube) | 7.1–28.6 | 0.0 | 4.3–30.8 |
EG (conventional) | 0.0–52.2 | 0.0–18.2 | 20.0–65.2 |
JI | 3.1–64.3 | 0.0–13.0 | 0.0–33.3 |
JPI | 0.0–27.8 | 0.0–17.2 | 4.0–27.8 |
DTR | 0.0–20.0 | 0.0–10.0 | 4.7–20.0 |
Table 6
Comparison between Right-Sided Overlap and Single-Flap Valvuloplasty and Double-Flap Technique performed at Second Affiliated Hospital of Soochow University
Surgery time (min), mean ± sd | 285.3 ± 71.3 | 336.5 ± 81.7 | 0.036 |
Anastomosis time (min), mean ± sd | 61.3 ± 11.2 | 67.9 ± 9.8 | 0.046 |
Anastomotic stenosis, n (%) | 0 | 5 (21.7%) | 0.027 |
Anastomotic leakage, n (%) | 0 | 0 | - |
Reflux esophagitis, n (%) | 1 (5.0%) | 4 (17.3%) | 0.206 |
Nutritional benefit
The advantage of PG in maintaining postoperative nutritional status has been confirmed by many studies [
20]. In our series, the similar benefit was observed (Table
4). At 1-year follow-up, the level of blood nutrition-related parameters did not change significantly, compared to preoperative levels. Similarly, the total body weight 1 year after surgery was not significantly different from the preoperative baseline. The value of ROSF in maintaining postoperative nutrition status of longer follow-up is still to be investigated.
Advantages in reducing anastomotic complications
Esophagogastrostomy with DFT was reported to have certain incidence of anastomotic stenosis, ranging from 4.7–29.1%, which required balloon dilation [
5‐
7]. The stenosis may develop after several months. On the other hand, SOFY was reported to have lower rate of anastomotic stenosis (0.0–2.8%) [
15]. Interestingly, none of our patients had anastomotic stenosis within follow-up period according to Nishikawa’s criterion [
12], showing a better effect in preventing anastomotic stenosis than existing methods (Table
5). Compared with our own experience of DFT, anastomotic stenosis decreased significantly in ROSF (DFT: 5, 21.7% vs. ROSF: 0,
p = 0.027, Table
6). Although four patient (20%) had mild dysphagia (Visick score II), their symptoms relieved simply by adjusting eating habit without any medication or balloon dilation.
The appropriate anastomotic diameter and sufficient blood supply may contribute to the outstanding performance of ROSF in this series. Shoji et al. reported that the diameter of the esophagus < 18 mm was one of the independent risk factors for postoperative anastomotic stricture [
5]. The end-to-side esophagogastrostomy may be related to the possibility of stricture, especially in patients with a small lumen of the lower esophagus. Contrarily, overlap esophagogastrostomy is less likely to develop stenosis due to the more spacious anastomosis. In the present study, we utilized overlap anastomosis, ensuring an adequate diameter.
Insufficient blood supply of the seromuscular flap may also contribute to the development of stenosis. Since both flaps are supplied mainly by the right gastroepiploic vessels in DFT, interruption of blood flow by the incision between the flaps might theoretically impair the bloody supply of the right flap. In fact, sometimes we did observe the apparent color change of the right flap after dissection of the flaps when performing DFT. In our study, we speculate that a left-sided single flap may ensure sufficient blood supply to the flap from the left side, which possibly prevents ischemia and consequential stenosis.
Moreover, none of the patients in our study developed anastomotic leakage. Similar to the reasons described above, left-sided single flap may also prevent potential leakage as a consequence of flap ischemia and necrosis, though anastomotic leakage is rare in DFT patients [
21,
22] (Tables
5,
6). Theoretically, ROSF may reduce the possibility of leakage better than SOFY, since the anastomosis is covered by the flap, though anastomotic leakage was not reported in SOFY [
15]. Based on the considerations above, ROSF might have some advantages in preventing anastomotic stenosis and leakage.
Simplification of surgical manipulations
While the efficacy of DFT is gradually recognized, technical difficulty and laborious suturing hinders its popularity. In previous studies of DFT, the mean/median surgery and anastomosis time was 235.3–420 and 79.4–109 min respectively [
5,
6,
21,
23]. Compared to our previous experience of 23 cases who underwent DFT, the anastomosis time (DFT: 67.9 ± 9.8 min vs. ROSF: 61.3 ± 11.2 min,
p = 0.046) and overall surgery time (DFT: 336.5 ± 81.7 min vs. ROSF: 285.3 ± 71.3 min,
p = 0.036) were reduced significantly (Table
6).
In valvuloplasty (both double flap and single flap), closing the flap(s) to cover the esophagus and anastomosis is the most time-consuming procedure. In ROSF, fixing the three brims of the flap with two barbed sutures reduced the complexity and time of suturing. Additionally, in DFT, closing of the flaps is assumed difficult due to lack of anchoring. However, in ROSF, a suture made on the upper-right corner of the flap facilitated the subsequent continuous suturing. Furthermore, overlap anastomosis with linear stapler shortened the time for anastomosis. Though ROSF was shown relatively simple and time-saving in this study, there are still some technical challenges as a seromuscular flap anastomosis method.
Firstly, although cases were not subgrouped according to the length of esophageal invasion, we noticed that performing ROSF in patients with highly-located tumor or short esophagus took more efforts. It needs to be further studied whether ROSF can be easily performed in these patients and what the upper limit of the technique is. Secondly, operation around the esophageal hiatus is difficult due to limited space obstructed by left hepatic lobe. Thus, we applied the liver retraction method shown in Fig.
1A, which provided a satisfactory field of view. However, further studies will be needed to find the optimal liver retraction technique. Lastly, performing ROSF might be hard for surgeons without extensive experience of hand-sewn suturing under laparoscopy, despite that barbed sutures were adopted instead of interrupted sutures. In the future, standardization of the procedure and robotic surgery might contribute to lowering its threshold.
Although ROSF showed benefits in preventing reflux and stenosis and simplifying manipulations, there are some limits of this study. Firstly, this retrospective study had limited number of cases and no comparative analysis. Cohort studies and clinical trials with larger samples are needed to further prove ROSF’s advantages. As a newly introduced technique, follow-up period of ROSF was short. Disease-free survival, long-term complications (such as sliding hernia), and long-term QOL should be investigated in longer time follow-up to confirm its safety and effect. Moreover, 24-h pH monitoring and manometer were not introduced to our institution and thus not conducted in patients. These examinations should be conducted in the future to precisely evaluate postoperative reflux and dysphagia, especially for patients without positive endoscopic results. Lastly, whether ROSF is feasible for all types of adenocarcinoma of EGJ and upper-third stomach needs further investigation.