Multivisceral Resection for Locally Advanced Gastric Cancer: A Systematic Review and Evidence Quality Assessment

Patients with locally advanced gastric cancer (LAGC) often require multivisceral resection (MVR) of the involved organs to achieve R0 resection and local disease control. The aim of the present study was to systematically review all available literature on the postoperative and long-term outcomes of MVR for gastric cancer. The PubMed database was systematically searched by two independent investigators for studies concerning MVR for LAGC. In total, 30 original studies with 3362 patients met our inclusion criteria. R0 resection was achieved in 67.77% (95% CI, 65.75–69.73%) of patients. The spleen, colon and pancreas comprised the most frequently resected organs in the context of MVR. Pancreatic fistulae (10.08%, 95% CI, 7.99–12.63%), intraabdominal abscesses (9.92%, 95% CI, 7.85–12.46%) and anastomotic leaks (8.09%, 95% CI, 6.23–10.45%) constituted the most common postoperative complications. Using the available data, we estimated the mean 1-year survival at 62.2%, 3-year survival at 33.05%, and 5-year survival at 30.21% for the entire cohort. The survival rates were mainly correlated with lymphatic invasion, tumor size and patient age. Therefore, gastrectomy, together with MVR, is feasible and may offer a survival advantage compared to gastrectomy alone or no other surgical treatment in a selected group of patients. Consequently, both patient and tumor characteristics should be carefully assessed to optimize candidate selection.


Introduction
Gastric cancer (GC) represents the fifth-most common malignancy worldwide and remains a major cause of cancer-related mortality [1].Oncologic resection with negative macroscopic and microscopic margins (R0 resection) along with adequate lymphadenectomy (D2 lymphadenectomy) is considered the gold standard and mainstay of GC treatment [2,3].Unfortunately, GC typically presents at an advanced stage, and if the tumor invades adjacent structures, extensive surgical resections may be needed to achieve clear resection margins [2,4].
Locally advanced GC (LAGC) accounts for approximately 30-35% of locally advanced gastrointestinal malignancies [5].Contemporary literature defines locally advanced gastric cancer (LAGC) as any entity larger than early gastric cancer or T2-T4 lesions.Others define LAGC as T3-T4 gastric malignancies requiring neoadjuvant treatment [6,7].These patients usually have a poor prognosis compared to patients with early-stage GC [8].Patients with radiologic evidence of T4b disease are potential candidates for multivisceral resection (MVR) [9].The distal pancreas, spleen, transverse colon and left liver lobe constitute the most frequently resected organs in combination with the gastrectomy specimen during MVR procedures [6].Unsurprisingly, multivisceral resection carries a high risk of perioperative morbidity and mortality [10,11].In this context, identifying the patient population that would benefit the most from such radical resections is crucial [12].Nevertheless, this can be a challenging endeavor, since delineating invasion patterns with the current preoperative staging modalities is not always straightforward.Furthermore, there is a lack of consensus concerning patient characteristics and preoperative factors that can guide the decision-making process with regards to MVR.
The aim of the present study was to systematically review all available literature concerning patient characteristics and short-and long-term survival, as well as postoperative complications, of patients with T4b LAGC undergoing MVR.

Literature Search and Inclusion Criteria
This systematic review was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis) guidelines and in line with a protocol agreed on by all authors [13].Two independent investigators (IG and AP) searched the PubMed/Medline and Embase databases for articles reporting on T4b LAGC (last search: 11 October 2021).Boolean operators (AND and OR) in combination with the following keywords were utilized: "gastric cancer", "gastric neoplasm", "locally advanced", "multiorgan resection" and "multivisceral resection".
Eligible prospective or retrospective studies on MVR for T4b LAGC written in English were retrieved.The references of all the included papers were reviewed to identify additional (potentially eligible) manuscripts using snowball methodology.Any controversies were resolved by a third investigator (KSM).Articles reporting on gastric cancer with distant hematogenous or lymph nodal metastases, studies not written in English, reviews of the literature, case reports, letters to the editor not including original data, in vitro studies, animal studies and abstracts were excluded from this systematic review.

Data Extraction
Data extraction was performed by two independent researchers (IG and AP) using a pre-piloted template.The following variables were collected and tabulated: study characteristics (sample size, type of study and country of origin); patient age; tumor location; tumor size; macroscopic and histological findings; TNM classification; surgical treatment strategies; number of resected or invaded organs; lymphadenectomy and survival rates.Predictive factors were also analyzed.

Statistical Analyses
Continuous variables were summarized as the mean ± SD (standard deviation), while categorical variables were summarized using frequencies and percentages.Relative rates, along with the corresponding 95% confidence intervals (95% CI), were estimated based on the available data for each variable of interest.Statistical analysis was carried out using IBM SPSS Statistics for Windows, Version 24.0, Armonk, NY, USA: IBM Corp.

Assessment of Study Quality
The quality of the included case series was assessed using the tool developed by the National Heart, Lung, and Blood Institute (NHLBI) based on work from the Agency for Healthcare Research and Quality, the Cochrane Collaboration, the United States Preventive Services Task Force, the Scottish Intercollegiate Guidelines Network and the National Health Service Centre for Reviews and Dissemination [14].The NHLBI scale ranges from 1 to 9, with a score of 1-3 demonstrating poor quality, 4-6 fair quality and 7-9 showing good quality.The mean and SD values for the NHLBI score of this systematic review were calculated.Three independent reviewers (IG, AP and KSM) rated the quality of the included studies, and a synthesis of their reports was performed.

Article Selection and Patient Demographics
Our literature search yielded 171 unique articles following the removal of duplicate publications.Overall, 30 articles met the inclusion criteria (Figure 1).The eligible studies were published between 1988 and 2020.In total, 3362 patients were analyzed.Concerning the geographical distribution of the included studies, 56.7% (17 studies) originated from Asian countries, 20% (6 studies) from North America countries, 13.3% (4 studies) from European countries, 3.3% (one study) from an African country, 3.3% (one study) from a South America country and 3.3% (one study) from Australia.Out of the available data, 1612 patients were male and 790 were female, with a 2:1 (male:female) sex ratio.The median patient age ranged from 55.76 to 69.70 years among the studies.However, it should be mentioned that demographic data were not available for 900 patients.
Health Service Centre for Reviews and Dissemination [14].The NHLBI scale ranges from 1 to 9, with a score of 1-3 demonstrating poor quality, 4-6 fair quality and 7-9 showing good quality.The mean and SD values for the NHLBI score of this systematic review were calculated.Three independent reviewers (IG, AP and KSM) rated the quality of the included studies, and a synthesis of their reports was performed.

Article Selection and Patient Demographics
Our literature search yielded 171 unique articles following the removal of duplicate publications.Overall, 30 articles met the inclusion criteria (Figure 1).The eligible studies were published between 1988 and 2020.In total, 3362 patients were analyzed.Concerning the geographical distribution of the included studies, 56.7% (17 studies) originated from Asian countries, 20% (6 studies) from North America countries, 13.3% (4 studies) from European countries, 3.3% (one study) from an African country, 3.3% (one study) from a South America country and 3.3% (one study) from Australia.Out of the available data, 1612 patients were male and 790 were female, with a 2:1 (male:female) sex ratio.The median patient age ranged from 55.76 to 69.70 years among the studies.However, it should be mentioned that demographic data were not available for 900 patients.

Quality of Evidence Assessment
The mean NHLBI score for the 30 eligible studies that were included in this systematic review was 8.30 (SD: 1.58) (good quality: 28 studies; fair quality: 2 studies).Detailed NHLBI and JBI quality assessments for the included studies are provided in Supplementary Table S1.

Location and Characteristics of Gastric Tumors
The majority of GCs that were treated in the context of MVR were found in the distal part of the stomach (41.3%), while 29.4% were found in the middle part, 23% in the upper stomach and only 6.3% were diffuse.The locations of gastric tumors in patients subjected to multivisceral resection are shown in Table 1.The mean tumor size ranged from 6.84 cm to 9.60 cm.Twenty articles (1200 patients) evaluated the grade of differentiation of the involved GC tumors, with 169 tumors well differentiated (169/1200; 14.08%, 95% CI, 12. 23 S2. Out of the available data, gastrectomy along with the resection of one organ was performed on 1500 patients (1500/3362; 44.62%, 95% CI, 42.94-46.30%).Furthermore, gastrectomy combined with the resection of two organs was performed on 585 patients (585/3362; 17.4% 95% CI, 16.16-18.72%),while gastrectomy along with the resection of three adjacent organs was performed on 73 patients (73/3362; 2.17%, 95% CI, 1.73-2.72%).The spleen, colon and pancreas were the most frequently resected organs in the context of MVR.Other excised organs included the small bowels, gallbladder, kidney, adrenal glands and the ovaries.The exact number of resected organs in the reported studies is presented in Table 2.

Discussion
The survival rates of patients undergoing radical procedures for advanced GC remain poor.Undoubtedly, T4b locally advanced gastric lesions that require multiorgan resection are associated with increased morbidity and mortality [12].That said, the refinement of surgical techniques and optimization of the pre-and postoperative management of GC patients have led to a considerable reduction in postoperative complications and improved patient survival [32].An increasing body of literature has shown comparable outcomes and survival rates between gastric surgery with MVR and gastrectomy alone [25,27].In accordance with the published medical literature, this study demonstrated that the prognosis of these patients mainly depends on the number and type of resected organs, size of the primary tumor and lymph nodal involvement.Within this framework, the invasion of gastric tumors into adjacent organs should not constitute a priori a contraindication for potentially curative resections.Our systematic analysis also demonstrated that the 1-year overall survival rates of patients ranged from 53.3% to 88%, 3-year survival ranged from 26.9% to 54.2% and 5-year survival ranged from 17.3% to 40%, suggesting that multiorgan resection in patients with LAGC could be performed safely in experienced hands with acceptable morbidity and mortality rates.
The available data are contradictory with regards to the impact that the type of resected organs exerts on the long-term outcomes of MVR [8].In this systematic review, the spleen, colon and pancreas constituted the most commonly resected organs during radical surgeries.Colonic resections during MVR for GC are historically associated with a potential survival benefit [32,33].In a retrospective study of advanced gastric cancer extended to the colon, the 1-, 3-and 5-year overall survival rates were 75.0%, 49.2% and 36.9%,respectively, with a median survival time of 24 months [20].Building on this knowledge, a recent multicenter trial reported no significant difference between gastrectomy combined with colonic resection alone and gastrectomy along with the resection of other involved organs [8].
For nearly two decades, splenectomy was routinely performed in patients undergoing gastrectomy for malignancies of the fundus, even in the absence of splenic infiltration.According to recent data, this approach offers no major survival benefits and has thus been abandoned.On the other hand, splenectomy in the setting of MVR for T4b lesions has been associated with over 80% and 45% 1-year and 2-year survival rates, respectively [8,22].Therefore, combining splenectomy with gastrectomy for T4 LAGC appears to significantly improve patient prognoses [8,34,35].
Encouraging outcomes have been reported from series performing pancreatectomies in the context of MVR.In this context, the 1-year survival has been shown to range from 61.5% to 83.8%.The two-year survival has been found to exceed 72%.Although scarce data exist, the 5-year survival rates may approximate 33% [32].Not surprisingly, pancreatectomyincluding MVR increases the incidence of postoperative complications and prolongs the length of the hospital stay substantially [11].In a study conducted by Piso at al., the longterm oncologic outcomes of patients with combined gastrectomy and pancreatic resection concerned a 5-year survival rate of 19% and a median survival at 13 months [36].The contradictory findings among published studies may be attributed to the limitations of retrospective studies, including selection bias, the coexistence of potential confounding factors and heterogeneity in practices among different surgeons [31].
Interestingly, a desmoplastic reaction of the surrounding tissues due to widespread GC may be incorrectly recognized as an invasion of the primary tumor to the adjacent organs, leading to an en bloc MVR of the involved structures [11].This difficulty in recognizing cases with true malignant invasion of the adjacent structures persists due to imaging and staging limitations in delineating tumor anatomy [25].In this systematic review, 13.27% of the GCs were staged as T3, while the vast majority of patients [86.13%] presented with T4 lesions (42.89% had T4a and 43.58% had T4b).Of note, the majority of included patients that underwent MVR along with gastrectomy had gastric tumors staged as T3 or T4a and were incorrectly considered suitable for MVR.Furthermore, the patient selection for MVR should be more careful, since these patients may present with various postoperative complications.In our systematic analysis, pancreatic fistulae (10.08%), intraabdominal abscesses (9.92%), anastomotic leaks (8.09%) and surgical site infections (5.04%) comprised the most commonly recorded complications.Interestingly, rapid recognition of these complications could be based on various predictive markers, such as the neutrophile/lymphocyte ratio (NLR) for anastomotic leaks in gastrectomies [37].
In large cross-sectional studies, the overall survival of patients with LAGC was estimated at 13.5 months (ranging from 6.2 to 30.3 months), while no significant difference concerning the survival rates between gastrectomy with MVR and gastrectomy alone was reported [6,38].In our analysis, the estimated mean 1-year survival was 62.2%, 3-year survival was 33.05% and 5-year survival was 30.21%.Interestingly, the impact of the number of resected organs on the long-term survival of patients with MVR remains unclear.Historically, the resection of more than one organ has been associated with a poor prognosis and an increased incidence of intraoperative complications [12,22].Nevertheless, a recently published landmark study by Yang et al. demonstrated that the survival of patients with gastrectomy along with the resection of one organ had no significant difference compared to gastrectomy combined with the resection of more than one organ [21].In the present systematic review, the mean survival of patients undergoing gastrectomy in combination with a solitary additional organ was 62.2%, 40.9% and 24.9% at 1, 3 and 5 years from the time of surgery, respectively.Unsurprisingly, patients who underwent MVR of >1 organ appeared to have an inferior prognosis (1-year survival: 53.6%, 3-year survival: 24.9% and 5-year survival: 11%).Overall, our data synopsis suggests that the number of resected organs should not be considered a contraindication by definition for gastric cancer surgery [8,17,18].
Tumor boards should be cautious when evaluating patients as potential candidates for multivisceral resection.Importantly, peritoneal dissemination is a classical marker of stage IV disease and of a dismal prognosis [16,27].In this systematic review, lymph node involvement was found to be a poor prognostic factor in nine studies, increased tumor size in four studies and pancreatic involvement in two studies.
The current systematic review has certain limitations that should be acknowledged.The vast majority of the included studies were retrospective studies from single institutions that were published in different time periods with an evolving AJCC/TNM classification system.Secondly, although our intention was to analyze post-MVR outcomes strictly in the setting of T4b LAGC, our patient cohort included T3 and T4a lesions that were misclassified as T4b based on preoperative imaging.Thirdly, MVR is an unavoidably broad term.For instance, a small transverse colon resection is markedly different than a distal pancreatectomy and splenectomy.Unfortunately, limiting our analysis to major MVR was impossible, since granular subgroup data were not provided within the published studies.Fourthly, significant breakthroughs in surgical, anesthetic and medical management techniques have been introduced throughout the years.The heterogeneity of such robust changes could not be accounted for in the present work but, undoubtedly, should be taken into consideration by practicing surgeons looking for actionable information in our systematic review.More specifically, although neoadjuvant chemotherapy constitutes the initial therapeutic approach of these patients in recent years, few studies evaluated the role of adjuvant or neoadjuvant therapy in patients that underwent MVR resection.Finally, although we initially sought to meta-analyze our data, unfortunately, this could not be performed due to the high degree of heterogeneity in the reporting of most of our outcomes of interest (complications and survival rates).

Figure 1 .
Figure 1.Flowchart of this systematic review.

Figure 1 .
Figure 1.Flowchart of this systematic review.

Table 2 .
Type and number of organs resected during MVR.

Table 3 .
Survival of patients after multivisceral resection.

Table 4 .
Poor prognostic factors of the included studies.