Gastric cancer is the fifth most common cancer and the second leading cause of cancer-related deaths worldwide.¹ Radical gastrectomy with lymph node dissection is the primary modality to treat gastric cancer in most patients with those malignancies.^2,3 Recently, minimally invasive surgery has been proven to show better short-term outcomes for gastric cancer than those by the open approach, including intraoperative blood loss, overall complications, wound problems, and shorter hospital stays.^4,5 Moreover, long-term outcomes from randomized controlled trials (RCTs) conducted in East Asia have also been comparable.^6,7 Although laparoscopic gastrectomy (LG), the most popular minimally invasive gastrectomy, is superior regarding postoperative recovery and cosmetic effects, and non-inferior in terms of oncological safety to open gastrectomy (OG),⁷ it still has some limitations in terms of technical dexterities, such as limited range of movement, a two-dimensional surgical field of view, amplification of operator hand tremors, and inconvenient surgical positioning.^8–11 To overcome these limitations, some surgeons have introduced advanced technologies, including articulated wrist-like instruments or three-dimensional (3D) cameras, during laparoscopic surgery; however, there is still an unmet need to solve these limitations.

Robotic surgery is widely used as an alternative to laparoscopic surgery for various malignancies, and its safety and validity have been proven in surgical oncology.^12,13 It provides magnified 3D high-definition vision under a stable camera platform, enhanced manual dexterity, better ergonomics, and endo-wristed instrumentation under control with tremor filtration.^14,15 Based on these benefits, robotic gastrectomy (RG) has been considered a potentially more effective surgical approach for gastric cancer. However, the evidence of the merits of RG over LG remains unclear because most studies are based on retrospective or non-randomized prospective designs or data from single-institutional studies.^11,16–22 Although there is an increasing number of prospective studies, only two RCTs analyzed the short-term outcomes of RG for gastric cancer.^23,24 Thus, we reviewed these RCTs and prospective studies (Table 1) to compare RG and LG to investigate RG’s current status and discuss future research regarding the technique. All reviewed articles were written in English and published until September 2021. This review assessed several postoperative outcomes (including operation time, blood loss, number of retrieved lymph nodes, postoperative laboratory findings, morbidity, and mortality), postoperative recovery, compliance with adjuvant chemotherapy, cost, and learning curve (Table 2).

Table 1 . Summary of Prospective Studies Reviewed in This Article.

Study	Year	Country	Study design	Group	Participant	Objective
Park et al20	2016	Korea	Multicenter POS	RG LG	223 211	Comparing subgroups who will benefit from robotic approach
Kim et al22	2016	Korea	Multicenter POS	RG LG	223 211	Postoperative morbidity and mortality
Lu et al23	2021	China	Single center RCT	RDG LDG	141 142	3-year disease-free survival Short-term postoperative outcomes
Ojima et al24	2021	Japan	Multicenter RCT	RG LG	113 117	Intra-abdominal infectious complications Short-term postoperative outcomes
Kim et al44	2021	Korea	Multicenter POS	RG	124	Learning curve

POS, prospective observational study; RCT, randomized controlled trial; RG, robotic gastrectomy; LG, laparoscopic gastrectomy; RDG, robotic distal gastrectomy; LDG, laparoscopic distal gastrectomy..

Table 2 . Comparison of Postoperative Outcomes Between Robotic Gastrectomy and Laparoscopic Gastrectomy Based on Two Recent Randomized Controlled Trials.

Postoperative outcome	The FUGES-011 trial				The Japanese RCT
	RDG (n = 141)	LDG (n = 142)	P-value		RG (n = 113)	LG (n = 117)	P-value
Operation time (min)							0.001
Total operation time	201.2 ± 32.0	181.6 ± 44.4	< 0.001		297 (179–654)	245 (131–534)
Actual operation time	187.0 ± 32.4	181.6 ± 44.4	0.243		250
Estimated blood loss (mL)	41.2 ± 45.7	55.7 ± 70.5	0.045		25 (5–475)	25 (5–1,405)	0.18
Lymphadenectomy
Total LN	40.9 ± 11.2	39.9 ± 12.2	0.452		30 (10–103)	35 (5–92)	0.09
Perigastric	23.3 ± 8.6	24.0 ± 9.0	0.500
Extragastric	17.6 ± 5.8	15.8 ± 6.6	0.018
Metastatic LN					1.57 (0–21)	1.97 (0–28)	0.45
Postoperative recovery
First flatus (day)	3.2 ± 0.6	3.5 ± 0.9	< 0.001		2 (1–4)	2 (1–6)	0.01
First liquid diet (day)	3.5 ± 0.6	3.9 ± 1.3	0.001		1 (1–4)	1 (1–26)	0.49
Postoperative hospital stay (day)	7.9 ± 3.4	8.2 ± 2.5	0.062		12 (7–43)	13 (6–45)	0.93
Analgesic drug					1 (0–12)	2 (0–21)	0.001
Postoperative morbidity
Overall morbidity	13 (9.2)	25 (17.6)	0.039		10 (8.8)	23 (19.7)	0.02
Overall morbidity, ≥ CDC III					6 (5.3)	19 (16.2)	0.01
Intra-abdominal abscess, ≥ CDC II	3 (2.1)	2 (1.4)	0.684		7 (6.2)	10 (8.5)	0.50
Intra-abdominal abscess, ≥ CDC III					5 (4.4)	9 (7.7)	0.30
Anastomosis leakage	0 (0.0)	1 (0.7)	> 0.999		4 (3.5)	5 (4.3)	0.99
Pancreatic fistula					0 (0.0)	2 (1.7)	0.72
Intra-abdominal bleeding	1 (0.7)	3 (2.1)	0.622		0 (0.0)	0 (0.0)	0.99
Ileus	1 (0.7)	1 (0.7)	> 0.999		1 (0.9)	2 (1.7)	0.99
Cholecystitis					0 (0.0)	3 (2.6)	0.25
Wound infection	1 (0.7)	1 (0.7)	> 0.999		1 (0.9)	1 (0.9)	0.99
Pneumonia	8 (5.7)	16 (11.3)	0.091		1 (0.9)	5 (4.2)	0.21
Cardiovascular system	1 (0.7)	1 (0.7)	> 0.999		0 (0.0)	0 (0.0)	0.99
Liver system	2 (1.4)	1 (0.7)	0.622		0 (0.0)	0 (0.0)	0.99
Urinary system	1 (0.7)	2 (1.4)	> 0.999		0 (0.0)	1 (0.9)	0.99
Thrombosis	0 (0.0)	1 (0.7)	> 0.999		0 (0.0)	0 (0.0)	0.99
Postoperative mortality	0 (0.0)	0 (0.0)			0 (0.0)	0 (0.0)	0.99

Values are presented as mean ± standard deviation, number (%), or median (range)..

RDG, robotic distal gastrectomy; LDG, laparoscopic distal gastrectomy; RG, robotic gastrectomy; LG, laparoscopic gastrectomy; CDC, Clavien-Dindo classification; LN, lymph nodes..

Most studies demonstrated that the operation time was significantly longer in RG than in LG.^{19,22,25–27} A Korean prospective multicenter comparative study demonstrated 221 min of surgical time in RG, which is longer than 178 min in LG.²² Similarly, a recent Chinese RCT (FUGES-011) showed a longer operation time in RG (approximately 20 minutes).²³ This difference in operation time might be attributed to the setting and docking of the robotic arms in the initial phase or the change in the equipment of additional devices during surgery (Fig. 1). In the FUGES-011 trial, the docking time was approximately 9.1 minutes, and 5.1 minutes was required for the undocking process. If these times were excluded, the actual surgery time would not differ between the two surgical approaches. This overall time gap but similar actual surgical times after excluding the setting and docking-related time were also observed in a Japanese RCT conducted by Ojima et al.²⁴ Based on these findings, the surgical resection time may be similar between RG and LG; however, the robotic approach requires additional time to mount and detach the robotic system in the operation field.

Figure 1. Placement of trocars and docking of the robotic arms using the DaVinci Xi® Model. (A) Trocar placement before setting the robotic system. (B) The completed status after setting and docking the robotic arms.

Although there was no significant difference in the intraoperative transfusion rate, the median amount of blood loss was significantly lower in RG in the FUGES-011 trial,²³ which has been commonly reported in some retrospective studies as well.^25,28,29 However, the Japanese RCT reported an equivocal blood loss of 25 mL during both approaches.²⁴ This feature was also found in a multicenter, prospective trial, with about 50 mL of median blood loss during both RG and LG without statistical difference.²² In the Japanese RCT, surgeons used a 3D or 4K laparoscope during LG, which helped control bleeding in small blood vessels through a clear surgical view.

No significant difference was observed in the number of retrieved or metastatic lymph nodes between the RG and LG groups in the two RCTs,^23,24 and the rate of curability was also similar between the two groups in the Japanese RCT. Previous studies have reported that RG potentially enables more accurate lymph node dissection than LG.^29,30 Lymph node dissection may be sophisticated using endo-wristed instruments and a clear surgical view (Fig. 2).^25,31 However, despite these benefits, skilled surgeons can perform excellent lymphadenectomy with the help of a 3D or 4K monitor system, even during laparoscopy.

Figure 2. Intraoperative findings after lymph node dissection during robotic total gastrectomy. (A) After proper lymph node dissection around pancreas and spleen, structures including distal portion of splenic artery (yellow arrow) and peri-splenic area (red arrow) are observed. (B) After excising lymph nodes in the hepatoduodenal ligament with the wrist-like robot instrument, proper hepatic artery (red arrow) and portal vein (yellow arrow) are exposed. (C) Left gastric artery (yellow arrow) is isolated after clearing suprapancreatic lymph nodes. (D) Proximal half of splenic artery (yellow arrow) and vein (red arrow) are clearly seen after removing lymph node around vessels.

Despite the grossly similar number of retrieved lymph nodes, the FUGES-011 trial demonstrated improved lymphadenectomy in patients receiving RG through subgroup analyses. When lymph nodes were dichotomized into perigastric and extra-perigastric nodes, more extra-perigastric nodes were surgically removed in the RG group, which was more prominent in patients with a body mass index < 25 kg/m². The number of harvested lymph nodes at station 12a (referring to nodes around the proper hepatic artery) was significantly higher in patients undergoing RG. Additionally, the authors tried to assess the absence rate of lymph nodes to be removed, which means that no lymph nodes were found in the pathologic examination despite surgical resection in individual lymph node stations that should have been excised. A significantly lower absence rate was observed in the RG group than in the LG group, especially at the extra-perigastric station.

In addition to the fidelity of lymph node dissection, whether lymphadenectomy is more effective in robotic surgery is essential to determine for gastric cancer surgery because lymph node metastasis is the most important prognostic factor. However, this issue has remained controversial in previous studies^32–34 and in these two RCTs. Although it is not clear whether a surgeon can perform more qualified lymphadenectomy using robotic instruments, a few previous studies revealed that RG could provide a greater number of lymph nodes from the extragastric, suprapancreatic, and splenic hilum regions.^23,29,30 Some investigators have assumed that the technical advantages, especially articulated wrist-like instruments, would enable improved lymph node dissection in the extra-perigastric area.

There were no differences between the LG and RG in terms of the maximum value and trend in postoperative body temperature and blood laboratory findings. In the Japanese RCT, amylase levels in drainage fluid on postoperative day (POD) 1 were significantly lower in the RG group. However, these values on POD 3 were not significantly different. The main reason for the lower amylase level on POD 1 could be explained by the robotic approach avoiding unnecessary pancreatic compression or injury;³⁵ however, it did not eventually lead to a significant reduction in pancreas-related complications.²⁴

Regarding blood cell counts, FUGES-011 demonstrated that the hemoglobin levels on POD 1 and 5 were significantly higher, and the white blood cell count on POD 1 was lower in the RG group, although it was not associated with clinical adverse events, including transfusions or complications.²³ Previous studies have shown that both C-reactive protein (CRP) and interleukin (IL)-6 are predictive factors of postoperative surgical stress and morbidity.^5,36 However, as these values were not routinely evaluated, this RCT could not determine the clinical significance of specific laboratory findings.

Although the comparison of detailed complications did not show any statistical differences, the overall incidence of postoperative morbidities was significantly higher in the LG group. The Japanese RCT demonstrated that the protective effect of the robotic approach on morbidity was remarkable, especially for clinically more severe complications of Clavien-Dindo classification (CDC) grade III or higher, although this was not proven in the FUGES-011 trial. When complications were classified as surgical or medical, both studies showed better outcomes in medical morbidity in the RG. In contrast, the FUGES-011 trial showed no difference in surgical morbidity according to the surgical approach. It is noteworthy that the FUGES-011 trial demonstrated that the robotic approach was an independent protective factor for overall postoperative complications compared with LG (odds ratio, 0.472; 95% confidence interval [CI], 0.225–0.993; P = 0.048). Pneumonia, the most common complication after both RG and LG, was significantly less likely to occur after RG.²³

The Japanese RCT attempted to compare complications related to intra-abdominal infection as a primary endpoint. Similar outcomes were observed between LG and RG regarding the incidence of CDC grade II or higher complications, including anastomotic leakage, pancreatic fistula, and intra-abdominal abscess. Additional analyses confined to grade III or higher or revealing the proportion of complications classified as anastomosis leakage, pancreatic fistula, and intra-abdominal abscess also showed a similar overall incidence.²⁴

To assess patients’ recovery after surgery, researchers have evaluated the resumption of eating or drinking, postoperative hospital stays, days of drainage tube extraction, the median time to start ambulation, and the date of the first flatus. Kim et al²² demonstrated that the time to recovery of bowel function, diet build-up, and length of hospital stay were similar between the two surgical approaches in a prospective comparative multicenter study. Similarly, two RCTs showed that short-term recovery after RG was not significantly different from that of LG.^23,24 However, the Japanese RCT demonstrated that the median time to first flatus was significantly shorter after robotic surgery (95% CI, 1.84–2.17 days) than that after LG (95% CI, 2.15–2.49 days, difference: -0.31 days, P = 0.001). The stable and flexible movement of robotic instruments may prevent excessive tissue traction and enhance the early recovery of gastrointestinal peristalsis.^37,38 The dose of postoperative analgesics was also significantly lower in the RG group, which might be associated with the rapid recovery of intestinal peristalsis and the lower surgical stress of robotic instruments in the Japanese RCT.²⁴ The FUGES-011 trial reported a shorter time to the first flatus, ambulation, and liquid diet intake as well.²³

There were no significant differences in the proportion of patients who initiated adjuvant chemotherapy or the number of completed cycles. However, patients receiving RG were more likely to initiate adjuvant chemotherapy earlier than those receiving LG,²³ which could be explained by the higher incidence of postoperative morbidity in the LG group. Adjuvant chemotherapy could be delayed due to postoperative complications, leading to an increased risk of recurrence,³⁹ even though it is integral to improving the long-term survival of patients with stage 2 and 3 gastric cancer.^40,41 Consistently, the FUGES-011 trial demonstrated that the RG group started adjuvant chemotherapy earlier than the LG group; however, it failed to show any effect on long-term survival. Further research is needed to reveal the effects on long-term survival.

The total cost of surgical treatment was divided into direct and indirect costs. Direct costs refer to the cost of patient management, including the money spent on hospitalization, laboratory tests, radiology tests, or medication and injection. Indirect costs include the overhead cost of the building, amortization of capital equipment and supplies, and maintenance of services, utilities, and administrative staff.^42,43

The total cost of RG tends to be higher than that of LG, but the direct cost of RG is significantly lower.²³ The direct cost was reduced in the RG group because it was associated with better postoperative recovery. The higher total cost of RG was also demonstrated in another prospective study;²² however, they assumed that the cost difference would be solved in the future because of the high maintenance and premium costs.³⁸

Kim et al⁴⁴ attempted to determine the learning phases and learning-associated morbidity in RG. Operative time, which gradually increased in the first 25 cases, subsequently decreased till the 65th case to reach a plateau. The complication learning curve was divided into four phases: the initial learning phase, proficiency phase, transitional phase, and mastery phase, which were defined by the 25th, 65th, 88th, and 125th cases, respectively. The overall complications, estimated blood loss, and procedure-related complications decreased as the phases progressed. The complications of CDC grade 2 or higher, however, were increased in the transitional phase compared to that in the other phases, which may result from the extension of indications or increased attempts of technically demanding procedures.

The reviewed RCTs showed non-inferiority of RG to LG. The FUGES-011 trial provides solid evidence for the application of RG in patients with gastric cancer, demonstrating its superiority in reducing intraoperative blood loss and improving lymphadenectomy. However, both RCTs still have some limitations, including a relatively small number of enrolled patients, confinement to cases receiving distal gastrectomy, and an insufficient number of participating institutions. Therefore, additional large-scale multicenter RCTs are needed to acquire strong evidence in the future.

Some surgeons agree on the benefits of robotic surgical systems. Compared with laparoscopic surgery, robotic surgery provides more ergonomic postures and a technically superior operative environment, thereby reducing the surgeon’s efforts to reach maximum proficiency.^30,45 Therefore, there has already been significant progress in the RG field. For example, the feasibility and safety of totally robotic distal gastrectomy and total/proximal gastrectomy using the reduced port system have been reported.^46–48 Nevertheless, robotic gastrectomy still requires more evidence to broaden its application in gastric cancer in clinical practice.

Surgical tissue damage is associated with an inflammatory response, which leads to systemic complications. However, based on the evidence discussed in this review, although there were no differences in laboratory findings, the RG group showed better results regarding postoperative complications. A future study with a larger sample size may yield more comprehensive information regarding the association of RG and postoperative complications.

Most importantly, long-term oncologic outcomes should be confirmed for safe and successful introduction of RG in gastric cancers. The efficiency and fidelity of lymph node dissection, earlier initiation of adjuvant chemotherapy, and application of RG in patients receiving neoadjuvant treatment should be further assessed in well-designed studies.

In conclusion, RG is a promising technology in the field of gastric cancer surgery that could provide equivalent short-term outcomes to LG. Potential benefits from recent RCTs, including less blood loss, fidelity of removing lymph nodes in extragastric area, fewer postoperative complications, and faster recovery or initiation of adjuvant chemotherapy, are expected to lead to better oncologic outcomes. However, more consistent results are needed and long-term outcomes should be verified in future studies.

None.

No potential conflict of interest relevant to this article was reported.