The incidence of gastric cancer (GC) has been decreasing in recent years. However, the increasing age of populations worldwide makes its detection still frequent. Even if GC treatment has been tailored according to disease stage, endoscopic or surgical resection must be always undertaken to guarantee a good outcome. Moreover, there is no hope of survival after incomplete surgery.

Optimal preoperative staging and a multimodal approach can lead to differentiated treatment, allowing also more specific results on complications, reducing morbidity and mortality. This therapeutic approach enables consideration of the hospital type and hospital-specific disease volume, considered as an important parameter for good results. Indeed, many studies have been done showing that morbidity and mortality improves in patients with esophageal cancer treated at high-volume hospitals, but few studies have described the same good results for GC, even if decreasing morbidity and mortality have been reported. A reason for this finding could be the lack of very-high-volume centers in western countries, which does not allow comparisons of different situations. However, we should probably also consider other variables, such as the availability of an experienced surgical team in medium-sized hospitals. In this regard, a study by Jensen et al. in 2010 showed results after the centralization of GC patients in a high-volume hospital in Denmark. The decrease from 37 to 5 departments after centralization resulted in an improvement in the mortality rate from 8.2% to 2.4% [1].

In recent years, the endoscopic approach for high-grade dysplasia and for a subset of early gastric cancer (EGC) not requiring lymphadenectomy has changed the quality of life of patients. Several studies and Japanese guidelines defined these subsets [2], i.e., mucosal differentiated cancer of diameter <2 cm without ulceration can be treated by endoscopic mucosal resection (EMR) or with endoscopic submucosal dissection (ESD). Not all mucosal cancer, differentiated tumors or small cancers, but only a subset of EGC with well-defined and associated characteristics did not present lymphatic spread; this subset should be considered for local endoscopic resection. If endoscopic treatment can cure patients considered to be N0, surgery must be proposed for all other patients suspected to have lymphatic diffusion.

The second-level lymphadenectomy (D2) proposed by Japanese surgeons can improve survival also in specialized western centers where morbidity and mortality is not significantly increased [3, 4]. A recent meta-analysis referring to old trials suggested that D1 is the better dissection, but revised studies also on the Dutch trial confirmed that D2 dissection improves survival if morbidity and perioperative mortality is low [5]. The McDonald study based in America added radiotherapy on perigastric areas that are submitted to insufficient lymphadenectomy, and revealed better long-term survival in relation to surgery alone [30]; radiotherapy could be considered in those patients as a sort of „radio-lymphadenectomy”. Limited lymphadenectomy is, in general, proposed by Japanese guidelines for EGC patients. Even if correct, considering the eastern preoperative accuracy in diagnosis, such limited indications are not always applicable in western centers. In our opinion, surgically treated patients should be submitted to D2 lymphatic dissection because of the following reasons:

In 2010, the Japanese Gastric Cancer Association presented a new D2 definition related not to the site of the cancer but to the type of gastrectomy, with the definition being different for subtotal or total gastrectomy [2]. Looking at lymphatic dissection, differences from previous definitions focused principally on gastric artery station 7 now being considered as D1 and mesenteric vein station (n14) being not dissected further in the new D2 definition and only being suggested in pylorus cancer. After subtotal gastrectomy, D2 lymphadenectomy requires dissection of 1, 3, 4sb, 4d, 5, 6, 7, 8a, 9, 11p and 12a node stations. Total gastrectomy requires dissection of 1, 2, 3, 4, 5, 6, 7, 8a, 9, 10, 11p, 11d, and 12a node stations [2].

More extended lymphadenectomy D3 is now not recommended as prophylactic treatment in Japan after the JCOG 9501 trial [6]. This trial did not find any improvement for patients submitted to para-aortic dissection (PAND). In western centers (especially in Italy), this option is not completely excluded because 5-year survival rates of 17% have been reported after surgical dissection for patients with positive para-aortic nodes. Even if these patients are considered to be M1, their survival is better than that for other metastatic sites. In Japan, even if prophylactic PAND is not recommended, if the involvement of the para-aortic nodes is suspected, these patients are submitted to neoadjuvant treatment, and then to D3 dissection. [6, 7]. D3 dissection involves removal of the posterior lymph nodes of the hepatic artery (8p), hepatoduodenal ligament (12p), retro-pancreatic segments (13), and peri-aortic segments (16). Morbidity is improved, but has been reported to be 28%; mortality (2.1%) is quite similar in selected centers [7]. Western centers are looking for selective criteria to identify patients with suspected involvement of the para-aortic lymph nodes [8].

An interesting study based in Korea in 2011 confirmed that free cancer cells can be released from lymphovascular pedicles opened during surgery for GC, especially in advanced-stage disease. The authors found that this could be prevented by using an energy-based device [9]. The intraoperative use of devices makes en block dissection a secondary endpoint, whereas distant stations (e.g., 12a in the hepatoduodenal ligament or 9 at the splenic hilus) could be better dissected alone. An interesting question is who should dissect the lymphatic stations after surgery in the surgical specimen. Quarrels have arisen between pathologists and surgeons if a low number of lymph nodes are sampled. Lymphatic dissection on fresh specimens immediately after resection permits better collection of lymph nodes, and different stations are better recognized by surgeons (mainly if nearby stations have not been dissected intraoperatively). Dissection by surgeons in theatre immediately after resection must be suggested and stressed.

The type of patient who could be considered for radical resection is an open question. The old Japanese definition of R0 was: no residual disease with a high probability of cure. This condition was achieved by a distance of 10-cm margin from the cancer and resection lines in T1/T2 tumors with a D lymphadenectomy level more than (N) lymphatic stage [10]. A recent R0 definition from the International Union against Cancer/American Joint Committee on Cancer (UICC/AJCC) stated „complete resection of the primary cancer without macroscopic or microscopic residual” [11]. The high prevalence of relapse after suspected R0 resection makes this definition inadequate, and perhaps the concept of circumferential/ lateral resection margin must be added [12]. The seventh tumor/node/metastasis (TNM) classification now necessitates cytology upon peritoneal lavage to avoid peritoneal cancer cells causing peritoneal relapse; positive cytological lavage has been added as M1 in the new TNM classification [11] and must be performed to define a R0 resection.

The MRC Adjuvant Gastric Infusional Chemotherapy Trial (MAGIC) in England and FFCD 9703 trial (France) described significant survival improvement if perioperative chemotherapy was followed by surgery. In these clinical trials, several questions have been asked (for example: the number of patients recruited with cancer of the gastric cardia or early lesions, and surgical results) but a statistically robust solid improvement was observed. Unfortunately, a more selective trial with good preoperative staging, the European Organization for the Research and Treatment of Cancer (EORTC) trial, failed to find statistical power for the low number of patients recruited even if downstaging had been frequently observed. No information is available on survival in the Swiss Group for Clinical Cancer Research (SAKK) trial. Five trials on neoadjuvant treatment are ongoing: MAGIC B (England); Chemoradiotherapy after Induction chemotherapy in Cancer of the Stomach (CRITICS; the Netherlands); JCOG 0501 (Japan); Intergroup Trial of Adjuvant Chemotherapy in Adenocarcinoma of the Stomach (ITACA 2) and Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori (IRST 151) (both Italy). These trials using neoadjuvant treatment involve different drugs and, sometimes, combination with radiotherapy and biological therapy. The importance of preoperative staging must be stressed to avoid irrelevant treatment in early lesions or delayed specific treatment in metastatic disease. Moreover, collaboration with oncologists in this new approach should be improved.

Laparoscopic staging is accurate and allows collection of peritoneal fluid for cytology. Laparoscopic surgery for early lesions (T1 and T2) is proposed also in Japanese guidelines with modified limited lymphadenectomy. Doubt regarding the treatment of advanced forms has been wiped away due to a meta-analysis reporting less lymph-node dissection and difficult dissection of some stations with laparoscopic access [13]. In this regard, some trials are ongoing (JCOG 0912 and Korean Laparoscopic Gastrointestinal Surgery Study Group (KLASS)) and results are awaited but, at the moment, robotic surgery may offer a better and easier dissection compared with laparoscopy.