Some scholars have advocated TG for Siewert II AEG. Reason 1: Intolerable gastroesophageal reflux has a high probability of occurring with the use of oesophageal gastric-remnant anastomosis after PG. In a study by Nakamura M, the incidence of reflux oesophagitis above grade B in the traditional oesophageal gastric-remnant anastomosis group was 21.8% (12/55) , and Chen S reported an incidence of 35.3% (12/34) . Some patients needed to take acid suppressant drugs for a long time, which seriously affected their quality of life. Bile and food repeatedly stimulated the anastomosis and led to the development of anastomotic inflammation, further causing gastric stump cancer. Reason 2: Most AEG tumours are more malignant than other gastric cancer tumours in terms of biological behaviour . To achieve better radical treatment of tumours and reduce recurrence, TG is needed.
Currently, a number of studies have shown that there is no advantage in terms of the postoperative survival rate for Siewert II AEG patients treated with active TG and lymph node dissection. Recently, Yura  studied 202 cases and revealed a metastasis rate of 0.99% for lymph node 4d, 0.006% for lymph node 12a, and 0% for lymph nodes 5 and 6 in T2-3 AEG. Kaixuan Zhu  et al. analysed the Surveillance, Epidemiology and End Results database, and among 1584 patients treated with PG and 633 patients treated with TG, they found no difference in the 5-year overall survival rate, but PG resulted in better long-term survival in patients over 70 years of age. In this study, most of the patients had disease in a middle stage, including 14 patients with stage IIIA disease; these patients were followed up for 24 months, and 3 died from tumour recurrence. We believe that PG should be cautiously considered in the treatment of patients with stage III disease.
A disadvantage of TG is the possibility of a poor nutritional status, anaemia and dumping syndrome after the operation [17-18]. Due to the complete removal of gastric parietal cells, the secretion of intrinsic factors can no longer occur, and vitamin B12 absorption is poor, which results in megaloblastic anaemia. With the removal of principal cells, digestive enzymes of the stomach are lost, and nutrition-related complications are further aggravated. A study by Nishigori  showed that the weight loss after PG was significantly less than that after TG, and the gastrointestinal symptoms after PG were also less than those after TG. A study by Lee  showed significantly better nutritional serological indexes (albumin, total protein, haemoglobin) two years after PG than two years after TG.
Physiological anti-reflux mechanism: 1. The lower oesophageal sphincter is a high-pressure zone under normal conditions that forms a pressure barrier and blocks reflux. 2. The traction effect of the phrenic oesophageal ligament prevents reflux. 3. The angle formed by the right posterior wall of the gastric fundus and the left side of the abdominal segment of the oesophagus (His angle) is an acute angle; when the pressure in the gastric cavity increases, the expanded gastric fundus compresses the left side of the lower segment of the oesophagus and closes it [20-21]. PG can destroy the above structure; additionally, due to removal of the vagus nerve, the remnant stomach loses control of the vagus nerve, and gastric peristalsis weakens. The pylorus continues to contract, which inhibits gastric emptying and increases the probability of gastroesophageal reflux.
We were inspired by the mitral valve and tricuspid valve of the heart. When the arterial pressure is high, the blood compresses the valve root to close the opening and prevent the blood from flowing back. Why cannot a valve be made manually at the junction of the oesophageal and remnant stomach by surgery?
In this operation, a 45° oblique cut was first made with a linear stapler; then, the sloped edge of the row of staples formed the initial valve. The reverse-puncture opening was left in the lower right corner. When performing anastomosis with the anastomat rod and the nail base, the nail base was aligned vertically, and the anastomosis was made obliquely on the right side of the simulated valve. The operator grasped the remnant stomach and pushed it to the upper left, which made the deeply embedded the starting section of the valve in the lower left segment of the oesophagus, constructing a new His angle between the left posterior wall of the remnant stomach and the oesophagus. When food is ingested through the mouth, the simulated valve opens. When the pressure in the remnant stomach cavity increases and reflux occurs, the valve root is squeezed on the left side, and the valve closes. The difference in the GerdQ score between before and after the operation was not statistically significant in 28 patients, which preliminarily confirms the anti-reflux effect of the new anastomotic method; however, one patient scored 12 points. After the operation, only one case of reflux oesophagitis above grade B was confirmed by gastrofibroscopy, resulting in a low incidence of 3.57% (1/28).
According to the principle, we divided anti-reflux operational methods after PG into two categories:
- Method of changing the physiological digestive tract pathway: single-lumen jejunal interposition, pouch jejunal interposition, and double-tract reconstruction.
- Method of improving the traditional oesophageal gastric-remnant anastomosis: oesophagogastric anterior wall anastomosis, double-flap (Kamikawa) anastomosis, anastomosis of gastroesophageal side overlap, and tubular gastric anastomosis.
Jejunal interposition involves inserting a segment of the pedicled jejunum between the oesophagus and the remnant stomach. The rhythmic peristalsis of the jejunum and the alkaline solution neutralize the residual gastric acid and therefore plays a role in anti-reflux. We think jejunal interposition has some shortcomings. First, the operation is complex, requires a long time to perform, forms three anastomoses, and carries the risk of anastomotic complications. The principle of Kamikawa anastomosis is similar to that of ours, with the aim of forming an embedded valve and constructing a His angle. In 33 cases of Kamikawa anastomosis reported by Kuroda , the number of reflux oesophagitis cases above grade B was 0, while the average time required for anastomosis reconstruction was 109 min at the one-year follow-up, without a grade B record. We believe that Kamikawa anastomosis requires cutting of the sarcoplasmic layer of the "H" type, inserting the anastomosis into the submucous layer by hand, and then folding and covering the two layers of the gastric muscle flap. This procedure is also complicated. Excessive overlapping of the valve is likely to cause ischaemic necrosis of the valve or anastomotic stenosis. Yoshito Yamashita  et al. reported an anastomotic anti-reflux method of oesophageal-remnant stomach side overlap via a linear cutter stapler under complete laparoscopy. We believe that in this method, a long segment of the lower oesophagus needs to be reserved, which cannot guarantee upper margin R0 resection of Siewert II AEG tumours with dentate line invasion.
There is only one anastomosis in this anastomotic method. The final anastomosis is a circular anastomosis performed under direct vision, which requires less of the reserved lower oesophageal segment and is safer. There were no cases of anastomotic leakage in the 28 treated patients. The principle of our anastomotic method is to construct a one-way valve at the junction of the oesophagus and remnant stomach, which does not affect the circular anastomosis; therefore, we can also combine this method with other anti-reflux surgical methods freely, such as oesophageal and gastric forearm anastomosis and tubular gastric anastomosis, to achieve a "double-guaranteed" anti-reflux effect.
We invented three united laparoscopic surgeries. Using a patented instrument, the operator could freely switch between using single-port laparoscopy, complete laparoscopy and hand-assisted laparoscopy during the operation . In the initial stage, complete laparoscopic surgery was used to free the stomach and dissect the lymph nodes. Then, the operator switched to hand-assisted laparoscopic surgery for anastomosis. The exposure and traction effects of the hand assistance made it more convenient to place the nail base. In this study, the average time required to construct the valve and perform anastomotic reconstruction was only 35.4 min, indicating that this procedure greatly shortened the total operative duration. Sample collection, anastomosis and hand assistance were all performed through the incision made for placement of the patented instrument, making reasonable use of the auxiliary incision.
There are some shortcomings in this method. If the valve is too large, anastomotic stenosis may be caused by the circular anastomat. There may be subjective bias in the GerdQ score of patients in this retrospective analysis, and more objective 24-h pH impedance tests and other experiments are required. Furthermore, there were few cases in this study, which may affect the statistical results.