Japanese gastric cancer treatment guidelines 2018 (5th edition)

The algorithm is shown in Fig. 1. Description of the tumor status (T/N/M and stage) in this edition is based on the 15th edition of the Japanese Classification of Gastric Carcinoma [1], which is identical to the 8th edition of the International Union Against Cancer (UICC)/TNM Classification [2].

Non-curative surgery is offered to the patients who are considered to be incurable. It can be semi-classified into either palliative surgery or reduction surgery depending on the aim of surgery.

Surgery for gastric cancer is defined as follows in the order of the stomach volume to be resected.

In addition, surgery for cancer of the gastric remnant is defined as follows.

The extent of lymphadenectomy is classified by the D-level criteria into D1, D1+ or D2, and is defined as follows according to the type of gastrectomy conducted. The indications for each of the D levels are described in the subsequent section. See descriptions under the title “Junctional cancer” for the current recommendations on the extent of lymph node dissection for the esophagogastric junctional carcinoma.

The extent of systematic lymphadenectomy is defined as follows, according to the type of gastrectomy conducted. When the extent of lymphadenectomy performed does not fully comply with the D-level criteria, the lymph node station that has been additionally resected or left in situ could be recorded as in the following examples: D1 (+ No. 8a), D2 (− No. 12a). However, when registering data to the nationwide database, the D levels need to be strictly determined and should be downgraded in case resection of any of the lymph node stations that should have been resected to fulfill the criteria of a D level was omitted. (e.g., D2(− No. 12a) should be registered as D1+).

In principle, D2 lymphadenectomy is indicated for cN+ or ≥ cT2 tumors and a D1 or D1+ for cT1N0 tumors. Since pre- and intraoperative diagnoses regarding the depth of tumor invasion and nodal involvement remain unreliable, D2 lymphadenectomy should be performed whenever the possibility of nodal involvement cannot be dismissed.

The current edition of the Japanese Gastric Cancer Treatment Guidelines defines the extent of lymphadenectomy according to the type of gastrectomy regardless the tumor location. However, only for esophagogastric junctional cancer (adenocarcinoma or squamous cell carcinoma), of which center locates within 2 cm of the esophagogastric junction, there is no consensus over the type of resection and the extent of lymphadenectomy that could be a standard of care for this category. In 2012–2013, the Japanese Gastric Cancer Association and Japan Esophageal Society joined forces to conduct a nationwide surveillance of junctional cancer of less than 4-cm diameter, and retrospective data of 3177 patients operated on between 2001 and 2010 were collected from 273 institutions [7]. An algorithm showing the tentative standard in the extent of lymphadenectomy based on the tumor location, histology and T-categories was constructed based on this surveillance (Fig. 6 ). A prospective phase II study by the same joint force to further investigate this issue is on-going.

Either of the following procedures is selected: proximal gastrectomy with or without lower esophageal resection, total gastrectomy with or without lower esophageal resection, esophageal resection and upper gastric resection.

Lymphadenectomy as shown in the algorithm is the tentative recommendation, although survival benefit of more extensive lymphadenectomy cannot be denied at this time. D levels and resected lymph nodes should be recorded according to the current classification for total gastrectomy or proximal gastrectomy as shown in Fig. 2 and 5. Refer also to the section “Lymph node dissection”. For example, in case of total gastrectomy, when lower esophageal resection and lymphadenectomy as indicated in the algorism (dissection of Nos. 1, 2, 3, 7, 8a, 9, 11p, 11d, 19, 20) are performed for cT3 esophagogastric junctional carcinoma of 3.0-cm diameter, the D level should be recorded as D2 (− No. 110, 111).

It is reported that preservation of the hepatic branch of the anterior vagus and/or the celiac branch of the posterior vagus contributes to improving postoperative quality of life through reducing post-gastrectomy gallstone formation, diarrhea and/or weight loss. In case of PPG, the hepatic branch should be preserved to maintain the pyloric function (CQ2).

Removal of the greater omentum is usually integrated in the standard gastrectomy for T3 or deeper tumors. For T1/T2 tumors, the omentum more than 3 cm away from the gastroepiploic artery may be preserved.

Bursectomy for tumors penetrating the serosa of the posterior gastric wall had been performed with the aim of removing microscopic tumor deposits within the omental cavity. However, survival benefit of this procedure has been denied by a large-scale randomized trial (JCOG1001), not only for all patients registered but also for subsets with T4a tumors and tumors located in the posterior wall [8].

For tumors in which the primary or metastatic lesion directly invades adjacent organs, combined resection of the involved organ may be performed to obtain an R0 resection.

For gastric cancers invading less than 3 cm of the distal esophagus, a transhiatal abdominal approach is recommended (JCOG9502) [9]. Where a greater length of esophagus is involved, a transthoracic approach should be considered if the surgery is potentially curative.

Laparoscopic surgery can be considered as an option to treat cStage I cancer that is resectable by distal gastrectomy. In the 2014 version of the guidelines by the Japan Society for Endoscopic Surgery, distal gastrectomy by the laparoscopic approach was recommended for gastric cancer with the diagnosis of cStage I, according to the Japanese Classification of Gastric Carcinoma 14th edition (rated recommendation B). These decisions reflect the fact that superiority of the laparoscopic approach in terms of short-term outcome has been reported through small-scale randomized trials and meta-analyses, while safety was proven in a prospective phase II study (JCOG0703) conducted by certified surgeons with sufficient skill and knowledge in laparoscopic surgery [10]. However, surgeons will have to be aware that the learning curve exists, and the indication for this approach should, therefore, be decided discreetly in each institution based on the expertise of the surgical team.

Data regarding the long-term outcome are yet to be available, and results of pivotal phase III studies conducted in Japan (JCOG0912 [11]) and Korea (KLASS01 [12]) are awaited*. As for more advanced cancer, there is currently no evidence to recommend a laparoscopic approach, since randomized trials to look at safety and long-term outcome are currently ongoing (JLSSG0901) [13].

No prospective trial has been reported regarding total gastrectomy for early gastric cancer by the laparoscopic approach**. Thus, laparoscopic total gastrectomy has been rated by the guidelines of the Japan Society for Endoscopic Surgery (2014) as recommendation C1 (may be considered for a patient in need of total gastrectomy, but no scientific evidence in support of the procedure is currently available) (CQ7). Those who consider to try this procedure in their institution should plan to do so with sufficient caution, since postoperative complications were reported to be significantly more frequent in the first year of its introduction.

When conducting gastrectomy by the laparoscopic approach, informed consent should be obtained from all patients after providing sufficient information, including the lack of data regarding long-term consequences.

*Survival data from the KLASS01 trial are now available [14]. Intention-to-treat analysis confirmed non-inferiority of the laparoscopic approach compared with open approach, the 5-year overall survival being 94.2% in the laparoscopic group and 93.3% in the open surgery group.

**Regarding the safety issue of laparoscopy-assisted total or proximal gastrectomy, evidence from a single-arm (JCOG1401) is now available. In this trial, incidence of the Grade 2–4 esophagojejunal anastomotic leakage was 2.5% (6/244, 95% CI 0.9–5.3) and met the required level of safety [15].

The lesion, together with the surrounding mucosa, is lifted by submucosal injection of saline (normo- or hypertonic) and removed using a high-frequency steel snare [16, 17].

The mucosa surrounding the lesion is circumferentially incised using a high-frequency electric knife (usually insulation-tipped), and the submucosal layer is dissected from the proper muscle layer [18‐20].

The resected specimens should be handled according to the rules described in the Japanese Classification of Gastric Carcinoma 15th edition [1].

The tumor biopsy specimens and endoscopically resected tumors are histologically classified into either the differentiated or undifferentiated type. The former includes malignant epithelial tumor, general type, of papillary adenocarcinoma (pap) and tubular adenocarcinoma (tub1, tub2), and the latter includes that of poorly differentiated adenocarcinoma (por1, por2) and signet ring cell carcinoma (sig) according to the Japanese Classification of Gastric Carcinoma 15th edition. In case mucinous adenocarcinoma (muc) was found at the submucosal layer, resected specimen is handled as undifferentiated type, regardless of whether it is considered to derive from the differentiated or undifferentiated type.

A tumor consisting of components of both differentiated- and undifferentiated-type carcinoma is, nevertheless, classified into one of the two types according to the quantitative predominance. In addition, when more than one histological type is found in a tumor, all histological types are to be recorded in the order of quantitative predominance, e.g., tub2 > tub1. Diagnosis of UL1 is principally made based on the histological evidence of ulcerative findings. However, the histological diagnosis of UL is sometimes difficult because of a biopsy-derived scar. Thus, endoscopic and/or radiological evidence should also be taken into consideration when making a conclusive diagnosis. A biopsy-derived scar is usually observed histologically as fibrosis restricted to small areas just beneath the muscularis mucosae [21]. If it cannot be discriminated from the ulcer scar, it should be classified as UL1.

Endoscopic resection is considered for tumors that have a very low possibility of lymph node metastasis and are suitable for en bloc resection [22].

Absolute indication of EMR or ESD [23, 24]

A differentiated-type adenocarcinoma without ulcerative findings (UL0), in which the depth of invasion is clinically diagnosed as T1a and the diameter is ≤ 2 cm.

Absolute indication of ESD

Expanded indication [25]

A standard therapy is surgical resection for tumors that do not fulfill the absolute or expanded indications. However, endoscopic resection could be an option for the elderly and high-operative-risk patients with severe comorbidities. Such case is considered as a relative indication, and endoscopic resection could be performed, provided a consent was obtained from the patient after explaining the risk of residual disease, possibly in the form of lymph node metastasis.

Indication of endoscopic resection for local recurrence after EMR/ESD [26]

Local recurrence within the mucosa after initial EMR/ESD for tumors that had fulfilled the absolute indication could be considered as expanded indication for repeat endoscopic resection. However, given paucity of hard evidence in support of the repeat endoscopic resection, a large-scale observational study looking at the long-term outcome of this procedure is warranted.

Two factors should be considered for curability assessment: completeness of the primary tumor removal and possibility of lymph node metastasis.

The resection is classified as endoscopic curability A (eCuraA) when all of the following conditions are fulfilled, provided cancer is without ulcerative findings (UL0): en bloc resection, any tumor size, histologically differentiated type-dominant, pT1a, negative horizontal margin (HM0), negative vertical margin (VM0) and no lymphovascular infiltration (Ly0, V0). However, if the undifferentiated component of the lesion exceeds 2 cm in length, the endoscopic curability is classified as C-2 (eCuraC-2).

When cancer is with ulcerative findings (UL1), the resection is still classified as eCuraA when all of the following conditions are fulfilled: en bloc resection, tumor size ≤ 3 cm, histologically differentiated type-dominant, pT1a, HM0, VM0, Ly0, V0.

The resection is classified as endoscopic curability B (eCuraB) for histologically undifferentiated type-dominant when all of the following conditions are fulfilled: UL0, en bloc resection, pT1a, HM0, VM0, Ly0, V0, tumor size ≤ 2 cm.

The resection is also classified as endoscopic curability B (eCuraB) for pT1b cancer when all of the following conditions are fulfilled: en bloc resection, histologically of differentiated type-dominant, pT1b1 (SM1) (< 500 μm from the muscularis mucosae), HM0, VM0, Ly0, V0, tumor size ≤ 3 cm. However, if the undifferentiated component is included in the portion of submucosal invasion, the endoscopic curability is classified as C-2 (eCuraC-2) [27].

The resection is classified as endoscopic curability C (eCuraC) when it does not fulfill the conditions described above to be classified as either eCuraA or eCuraB.

The resection is classified as endoscopic curability C-1 (eCuraC-1) when it is histologically differentiated type and fulfills other criteria to be classified into either eCuraA or eCuraB but was either not resected en bloc or had positive horizontal margin. All other eCuraC resections are subclassified as endoscopic curability C-2 (eCuraC-2).

Follow-up with annual or biannual endoscopy is recommended after the eCuraA resection [28]. In addition, follow-up with abdominal ultrasonography or computed tomography (CT) for surveillance of metastases is recommended after the eCuraB resection [29, 30]. For both eCuraA and eCuraB resection, it has been recommended that Helicobacter pylori be examined and, if positive, be eradicated (CQ12). However, some studies showed that the Helicobacter eradication after endoscopic resection had no impact on the occurrence of metachronous cancer. Further investigations regarding this issue are warranted [31, 32].

Since the risk for harboring lymph node metastasis is low, one of the following alternatives could be selected according to the institutional policy after obtaining patient’s consent: repeat ESD, surgical resection, close observation expecting a burn effect of the initial ESD, and endoscopic coagulation using a laser or argon-plasma coagulator [33].

When the lesion is differentiated type of ≤ 3 cm and either UL1or pT1b1 (SM1), size of the residual mucosal lesion should be reassessed by endoscopy. When the sum of the lengths of the resected and residual lesions exceeds 3 cm, gastrectomy with lymphadenectomy should be considered the standard of care. In addition, patients with positive horizontal margin within the portion of submucosal invasion and those who underwent piecemeal resection in which the resection line involved the portion of submucosal invasion should be recommended to undergo gastrectomy with lymphadenectomy, since the histological diagnosis under these circumstances is destined to be uncertain.

Gastrectomy with lymphadenectomy should be considered as the standard of care. When surgery cannot be recommended because of old age or severe comorbidities, the risk of residual disease in the form of lymph node metastasis (Tables 2 [34] and 3 [35]) and possibility of the subsequent local recurrence and/or distant metastasis should be assessed and explained sufficiently to the patients, along with the information that the recurrent disease is usually incurable with dismal prognosis.

The following chemotherapeutic or molecular targeted agents are administered in chemotherapy for AGC: fluorouracil (5-FU), tegafur/ 5-chloro-2,4-dihydroxypyridine/potassium oxonate (S-1), levofolinate calcium, capecitabine, cisplatin, oxaliplatin, irinotecan, docetaxel, paclitaxel, nab-paclitaxel, trastuzumab, ramucirumab, and nivolumab. These agents are used either as monotherapy or combination therapy based on the evidence obtained through clinical trials.

Recommended regimens are defined in this guideline as those that fulfill either of the following requirements for patients who are in sufficient general condition to meet inclusion criteria of clinical trials.

Conditionally recommended regimens are defined as those that fulfill either of the following requirements and could substitute for the “Recommended regimens” when deemed more appropriate after considering the factors such as i) general condition of the patient including disease status, age, organ functions and comorbidities, ii) social factors such as the necessity for hospital admission, cost of the treatment, distance to the hospital that limits the frequency of visit, and iii) personal preference that derives from the type of adverse events.

The “Recommended regimens” and “Conditionally recommended regimens” are listed in Figs. 8 and 9 based on the voting from six medical oncologists who were members of the JGCA Guidelines Committee (the decision supported by at least 70% [5 out of 6] of the medical oncologists). However, the readers are not necessarily discouraged from using regimens which are not listed in these figures. The selection was stringent, and even the regimens which were supported by 50–69% [4 out of 6] of the medical oncologists are not listed. Given the complexity of daily clinical practice, there could be situations where regimens that are not listed could, nevertheless, serve as a useful option.

Due to paucity of clinical trial results specific for elderly patients and for patients with impaired organ function or comorbidities, it is not possible to indicate with sufficient evidence whether a “conditionally recommended regimen” is superior to or safer than a “recommended regimen” delivered at a reduced dosage or modified treatment interval. Therefore, the optimal therapeutic regimen for these patients should be selected on a case-by-case basis, with the guidelines serving only as a reference.

The evidence level according to the criteria of the MINDS clinical guideline manual version 2.0 (Table 4) was provided only for the “Recommended regimens”. For the “Conditionally recommended regimens”, the evidence level is not described because no evidence is available for the specific clinical condition where the “Conditionally recommended regimens” could be more appropriate than the “recommended regimens”.

A combination of S-1 and cisplatin (SP) is the standard of care (the recommended regimen) based on the results of two phase III trials conducted in Japan (JCOG 9912 trial [41] and SPIRITS trial [42]) (evidence level A). After its non-inferiority to the 5-FU/cisplatin combination (FP) had been proven, a combination of capecitabine and cisplatin (XP) became one of standard treatments overseas and was employed as a control group in several global phase III studies including the ToGA trial [37] and AVAGAST trial [43]. Since safety and efficacy of XP have been recognized in subset analyses of the Japanese participants in these trials, this combination was added to the list of “Recommended regimens” (Evidence level A). CapeOX, a combination of capecitabine and oxaliplatin which was approved in 2014 in Japan, has shown efficacy that is at least equivalent to the FP in the subset analysis of a phase III study (no Japanese patients included) which evaluated triplets that combined these regimens with epirubicin (Evidence level B) [44]. A combination of S-1 and oxaliplatin (SOX) also demonstrated efficacy similar to SP in the G-SOX study (Evidence level B) [45]. These oxaliplatin-containing regimens could be delivered more easily than SP or XP because hydration is not required. Furthermore, a combination of 5-FU/levofolinate calcium (LV) with oxaliplatin (FOLFOX) has been employed as a control regimen in the recent comparative studies (Evidence level B) [46, 47] and has now been approved in Japan. FOLFOX can be particularly useful among patients who have difficulty in oral intake, such as those with bowel obstructions. To summarize, the list of “Recommended regimens” for the first-line treatment of unresectable advanced or recurrent gastric cancer includes various combinations of fluoropyrimidine and platinum except for the original FP regimen (Fig. 8), and selecting the most suitable regimen for each patient after taking into consideration various factors is a challenge for the physicians (CQ13).

A combination of S-1 and docetaxel failed to show superiority to S-1 monotherapy in the primary survival analysis of the START trial, but superiority in overall survival was observed in a reanalysis [48]. This regimen could be recommended for a limited population such as those who are not suitable to receive a platinum-containing regimen (“Conditionally recommended regimen”) (CQ14).

The combinations of irinotecan with cisplatin or S-1 are not recommended in the first-line treatment because they did not show significant superiority over S-1 alone in randomized trials conducted in Japan [41, 49].

Regarding the triplet regimens, superiority of adding docetaxel to a combination of infusional 5FU and cisplatin was proven in the V325 study [50] conducted in the Western countries. This triplet was avoided in Japan at the time, however, due to the excessive toxicity that did not balance well with the benefit in efficacy. More recently, following a promising phase II evidence, a triplet regimen consisting of S-1, cisplatin and docetaxel (DCS) was compared with SP in a phase III trial, JCOG1013. Since no benefit in overall survival was shown in that trial [51], triplet regimen containing taxane is currently not recommended as a first-line therapy.

Evidence is lacking regarding chemotherapy for specific types of patients such as those with impaired oral intake, peritoneal carcinomatosis (patients with a moderate to high volume of ascites or bowel obstruction) and the elderly. For these patients, conditionally recommended regimens could substitute for the recommended regimens as can be seen in CQ21.

The definition of HER2 positive in the ToGA trial had been either IHC3+ or FISH positive [37]. In the subgroup analyses of the trial, survival benefit was more distinct among the IHC3+ or FISH positive/IHC2+ cohorts. Thus, trastuzumab-containing regimens are currently recommended for patients with IHC3+ or FISH positive/IHC2+ status in clinical practice. Since continuous infusion of 5-FU is rarely used nowadays, a combination of trastuzumab with XP which was employed in the ToGA study (Evidence level A), and combinations of trastuzumab with either triweekly or conventional SP where efficacy of both regimens were satisfactory in two successive phase II studies (Evidence level B) are the recommended regimens [52, 53].

In addition, results of phase II studies assessing combinations of trastuzumab with CapeOX [54] and SOX [55] have been reported. These regimens are rated as “Conditionally recommended regimens”, suitable for those who may not be able to tolerate cisplatin.