Extent of Lymphadenectomy

The two major lymphadenectomy (LND) trials from the Netherlands and the United Kingdom highlighted in the prior review showed no difference in survival and an increase in complications in patients undergoing D2 compared with D1 dissection. Although well designed and executed, both studies were criticized for methodological flaws with respect to actual extent of LND, which left this crucial question seemingly unanswered. Hartgrink and colleagues reported the final results of the Dutch Gastric Cancer Trial with a median follow up of 11 years. The authors defined D1 LND as removal of perigastric lymph nodes along the greater and lesser curvature. A dissection qualified as D2 when the nodal stations of the left gastric, common hepatic, celiac, and splenic arteries were removed. As in the earlier analysis, there was no difference in overall or disease-specific survival between D1 and D2 LND groups. In-hospital mortality, however, was significantly higher in patients randomized to D2 LND. Subset analyses suggested a possible benefit for patients with advanced nodal disease; however, this was not a planned analysis of the trial and therefore the possibility of random significance does exist.

The only study to demonstrate a difference in survival for D1 versus D2 LND was published in 2006. From a single institution in Japan, 221 patients were randomized to D1 or D2 dissection to be performed by surgeons well trained in the technique of extended LND; all specimens were examined by one pathologist. Five-year survival for D2 patients was significantly higher than for D1 patients; however, the absolute increase in survival was small (59.5% versus. 53.6% 5-year survival, P = .04). Additionally, there was no statistically significant difference in recurrence between treatment groups among patients who underwent an R0 resection. Perioperative morbidity and mortality were not reported in this study.

Two RCTs have been performed to evaluate the value of para-aortic nodal dissection (PAND), also known as a D4 dissection, in addition to D2 LND. The Japan Clinical Oncology Group (JCOG) randomized 523 patients from 24 centers to D2 LND with or without PAND. There was a trend toward an increase in operative morbidity in patients undergoing PAND (28.1%) compared with D2 LND alone (20.9%; P = .07), but no significant difference in overall or recurrence-free survival. Yonemura and colleagues evaluated 269 eligible patients randomized to D2 or D4 dissection and similarly found no difference in overall survival between treatment groups.

To evaluate the extent of LND for proximal lesions, the Japanese Cancer Oncology Group (JCOG) initiated an RCT comparing a left thoracoabdominal (LTA) approach to transhiatal (TH) approach for treatment of Siewert type 2 and 3 lesions. This study was halted at the first interim analysis owing to the unlikely probability of LTA having a survival advantage over TH. There was no statistically significant difference in 5-year overall survival between groups, however there were more complications observed in patients who had LTA.

Splenectomy and pancreatectomy are notable risk factors for complications associated with extensive LND. Two studies have focused on whether or not splenic preservation compromises oncologic outcome in patients undergoing at least a D2 LND. Csendes and colleagues randomized 187 patients to D2 LND with or without splenectomy. There was no difference in overall survival but a significant increase in the rate of subphrenic abscess, postoperative fever, or pulmonary complications in patients undergoing splenectomy. A similar and well-designed study by Yu and colleagues was unable to demonstrate a survival benefit from splenectomy and there was no difference in perioperative complications.

In summary, a D2 LND has not been consistently shown to improve overall survival over D1 LND and is associated with an increase in perioperative complications and mortality. Trials to date have not been designed to assess the impact of extended LND on specific subsets of gastric cancer patients. One single-institution study with very high quality control for surgeons and pathologists has shown a small survival benefit to extended LND, but these findings have not been replicated. Furthermore, data do not support a more extensive (D4) LND, as it is associated with more complications without conferring a survival benefit. Preservation of the spleen during D2 LND does not compromise overall survival and may reduce operative complications.

Minimally Invasive Resection

Minimally invasive approaches in surgical oncology are gaining popularity and several trials were performed to evaluate laparoscopic distal gastrectomy; no RCT to date has applied this technique to total gastrectomy. Despite the efforts to study laparoscopic resection, none of the studies met level 1a evidence criteria because of the small numbers of patients and relatively short follow-up. The largest study performed with long-term follow-up is reported by Huscher and colleagues in which patients were randomized to open (N = 29) or laparoscopic (N = 30) subtotal gastrectomy performed by a single surgeon. Perioperative morbidity and mortality were equivalent, although patients who had laparoscopic resection had less blood loss, were started on oral intake earlier, and had a shorter hospital stay than patients treated with open gastrectomy. There was no difference in the number of lymph nodes harvested between the open (33.4 ± 17.4) and laparoscopic (30.0 ± 14.9) groups. Overall 5-year and disease-specific survival were equivalent. The short-term benefits of laparoscopic resection (blood loss, analgesic use, hospital stay) were also reported in other studies with small numbers of patients. As expected, the laparoscopic approach was associated with longer operative times in each of the studies performed.

Kim and colleagues initiated the largest RCT to date of 164 patients with early gastric cancer, and recently reported their interim analysis. This is a noninferiority trial with a primary end point of disease-free survival at 5 years. Consistent with earlier studies, patients who underwent laparoscopic resection had less blood loss and required less postoperative analgesia. Pathologic analysis revealed that patients who underwent laparoscopic resection had significantly fewer lymph nodes harvested than did those who had open gastrectomy (39 ± 11.0 versus 45 ± 13.8, P < .05). The proximal resection margin was also an average of 1 cm shorter in the laparoscopic group, but this did not reach statistical significance. The authors also studied several quality of life (QOL) measures and patients randomized to laparoscopic resection had significantly better QOL scores up to 3 months after resection. Similar findings came out of a trial from Korea in which 47 patients were randomized to open or laparoscopic distal gastrectomy. In summary, despite the obvious short-term benefits of the laparoscopic approach to gastric resection, the long-term oncologic outcomes remain unknown.

Surgical Site Infection Prophylaxis

There is one large RCT from Japan addressing antibiotic prophylaxis in gastric cancer surgery. Ten centers randomized 501 patients to single-dose cefazolin or ampicillin-sulbactam 30 minutes before surgery or to a multiple dose regimen of the same antibiotic for 3 days. The surgical site infection rate was equivalent between groups and there was no difference in complications.

Reconstruction After Gastrectomy

Several RCTs evaluated reconstruction techniques after gastrectomy. One study compared Billroth I and Roux-en-Y reconstruction after distal gastrectomy in 50 randomized patients. Patients with Roux-en-Y reconstruction had significantly less bile reflux and inflammatory changes in the remnant stomach on endoscopic examination 5 months postoperatively, but equivalent rates of esophagitis. Because patients in the Roux-en-Y group had a significant increase in hospital stay attributed to gastrojejunal stasis, the authors conclude that Roux-en-Y reconstruction is of limited value. Two RCTs compared Roux-en-Y with or without jejunal pouch reconstruction after total gastrectomy. Iivonen and colleagues randomized patients 49 patients and found a decrease in dumping syndrome and early satiety but showed equivalent weight gain and nutritional status to those without pouches 15 months after resection. A larger RCT with long-term follow-up showed that quality of life was similar at 1 year, but was significantly improved at 3, 4, and 5 years after surgery in patients with pouch reconstruction. It appears that there may be some long-term benefits to pouch reconstruction, although larger studies with long-term follow-up are necessary before this can be recommended as a standard reconstruction option.

Intraperitoneal Drainage After Gastrectomy

This highly debated topic has been studied in numerous tumor types across many different centers. Two RCTs address the question of the value of intraperitoneal drainage following gastrectomy for cancer. One single-surgeon study randomized 170 patients undergoing gastrectomy with at least a D2 dissection and evaluated the impact of drain placement based on the operation performed (subtotal gastrectomy N = 118, total gastrectomy N = 52). There was no difference in postoperative complications between groups, although the incidence of postoperative abscess in this study was only 2% (3/170). A study from Chile randomized 60 patients after total gastrectomy to no drain or placement of two drains surrounding the gastrojejunostomy. There was a statistically significant increase in hospital stay, morbidity, and need for reexploration in patients with intraperitoneal drains. These studies suggest that there is minimal value to and may be harmful consequences of routine intraperitoneal drainage following gastrectomy.

Nasojejunal Decompression After Total Gastrectomy

The largest RCT was from the Italian Total Gastrectomy Study group who evaluated routine nasojejunal tube placement after total gastrectomy with Roux-en-Y reconstruction. A total of 237 patients were randomized to NJT placement or not. All patients underwent radiographic examination with water-soluble contrast on postoperative day 7, and if no leak was detected, a diet was initiated. There was no difference in postoperative morbidity, anastomotic leak, hospital stay, or time to initiate diet between groups. The results of this study and others with similar results, which have been reported in a recent meta-analysis, obviate the need for routine placement of nasojejunal tubes following total gastrectomy.

Neoadjuvant Therapy

Four studies investigated the role for neoadjuvant chemotherapy in resectable gastric cancer. The most notable is the Medical Research Council Adjuvant Gastric Infusional Chemotherapy (MAGIC) Trial published in 2006. This study has had a significant impact on the management of resectable gastric cancer. Patients were randomized to surgery alone or to three preoperative plus three postoperative cycles of epirubicin, cisplatin, and 5-fluorouracil (5-FU). The perioperative chemotherapy group seemed to have been down-staged by neoadjuvant chemotherapy, manifest by smaller tumor size and fewer positive nodes than the surgery-only group. At a median follow-up of 4 years, there was a significant improvement in 5-year overall and progression-free survival in patients treated with perioperative chemotherapy, although only 42% of patients in that group completed protocol therapy. Two smaller studies using different regimens were unable to demonstrate a survival benefit of neoadjuvant chemotherapy over surgery alone.

One RCT investigated neoadjuvant chemotherapy with or without radiation in locally advanced gastroesophageal (GE) junction tumors (55% Type I, 45% Type II/III) across 15 centers in Germany. The study was closed early because of poor accrual. Patients (N = 126) were randomized to chemotherapy followed by surgery or chemotherapy followed by chemoradiotherapy before surgery. Patients who had chemoradiation had a significantly higher complete pathologic response rate (15.6% versus 2.0%) and more often had tumor-free lymph nodes compared with patients treated with chemotherapy alone. There was no difference in operative complications or survival at 3 years, although there was a trend toward better survival in patients treated with chemoradiation (47.4% versus 27.7%, log rank P = .07). In sum, neoadjuvant therapy has become a standard approach in many centers for patients with locally advanced gastric cancer (T3, N any) who are fit to tolerate treatment. Radiation therapy in the neoadjuvant setting should be reserved for appropriately staged distal esophageal and proximal GE junction tumors.

Adjuvant Chemoradiation

One study has been performed during the period of this review that meets level 1a criteria. The Southwest Oncology Group published an RCT (Intergroup 0116) of surgery alone compared with surgery plus adjuvant chemoradiotherapy (CRT). Following R0 resection of gastric or gastroesophageal junction tumors, 556 well-selected patients were randomized to observation or adjuvant CRT. Of note, 54% of patients underwent a D0 lymphadenectomy. Although there was significant toxicity, overall and disease-specific survival were significantly improved in patients treated with adjuvant CRT.

Adjuvant Chemotherapy

There were 16 randomized adjuvant trials of chemotherapy without radiation performed during the period of this review. Numerous attempts to demonstrate a survival advantage to adjuvant systemic chemotherapy have failed, summarized by the prior review and multiple meta-analyses. Critical problems associated with these trials include the lack of surgical standardization and the heterogeneity of patient populations being studied. One trial demonstrating a survival benefit from adjuvant therapy was from Japan where patients were randomized to surgery alone or adjuvant S-1, an oral fluoropyrimidine. Surgical quality control was superior to that of the Intergroup 0116 trial, as all patients received D2 or more LND and were staged with peritoneal cytology. The trial was stopped after the first interim analysis at 1 year because of significantly better survival in patients treated with adjuvant S-1.

The remaining trials of adjuvant chemotherapy have demonstrated no benefit over surgery alone. Di Costanzo and colleagues randomized 258 patients to surgery alone or adjuvant chemotherapy with four cycles of cisplatin, epirubicin, 5-FU, and leucovorin (LV) (PELF). Of note, 47% of the patients in the study had fewer than 15 lymph nodes examined on pathologic analysis. At a median follow-up of 73 months, the rate of recurrence and overall and disease-free survival were similar between groups. Similar findings were reported by De Vita and colleagues who randomized 228 patients to adjuvant 5-FU, LV, epirubicin, and etoposide or to surgery alone. In a study to evaluate the addition of epirubicin and cisplatin to 5-FU, Cascinu and colleagues randomized patients to adjuvant PELF versus adjuvant 5-FU/LV; there was no measurable survival benefit. These studies also demonstrated that chemotherapy, when administered in the adjuvant setting, can impose significant toxicity.

The current evidence to date demonstrates that adjuvant therapy is beneficial in most patients. In Western patients, the standard options supported by level 1a studies are perioperative chemotherapy or adjuvant chemoradiotherapy. No study to date has demonstrated a survival benefit to adjuvant chemotherapy alone in a Western population.

Intraperitoneal Therapy

The largest RCT to date investigating intraperitoneal (IP) chemotherapy was published by Yu and colleagues, who randomized 248 patients to surgery alone or early adjuvant IP mitomycin-C (MMC) and 5-FU. All patients underwent D2 LND and chemotherapy patients received a Tenckhoff catheter for delivery starting on the first postoperative day. Patients treated with chemotherapy had a significant increase in intra-abdominal hemorrhage and abscess formation compared with controls; 24% of patients treated with chemotherapy had mild to moderate abdominal pain during treatment. Eight patients (6.4%) in the chemotherapy group died in the perioperative period compared with two (1.6%) in the control group. Despite the significant increase in perioperative morbidity and mortality, treated patients experienced a significant increase in survival and a decrease in peritoneal recurrence. On subgroup analysis, patients with stage I and II disease did not benefit from IP chemotherapy advocating for a better selection algorithm in these patients.

Intraperitoneal chemotherapy has not gained acceptance as a standard adjuvant therapy for gastric cancer in many centers because of the significant potential for perioperative morbidity. Moreover, there exists the potential to over treat patients with early stage disease. With the development of newer systemic agents, IP chemotherapy is unlikely to become a standard modality in the treatment of gastric cancer.

Helicobacter Pylori Eradication

One trial investigated the role for Helicobacter pylori eradication following endoscopic mucosal resection (EMR) in patients with early gastric cancer. This was a multicenter RCT from Japan in which 544 patients with documented H pylori infection were randomized to treatment or not following EMR. Patients were evaluated for metachronous cancer up to 36 months following resection and it was found that those who underwent treatment for H pylori had a significant reduction in the incidence of metachronous cancer (9 new cancers in treated patients versus 24 controls; odds ratio = 0.35).

Neoadjuvant Therapy

Four studies investigated the role for neoadjuvant chemotherapy in resectable gastric cancer. The most notable is the Medical Research Council Adjuvant Gastric Infusional Chemotherapy (MAGIC) Trial published in 2006. This study has had a significant impact on the management of resectable gastric cancer. Patients were randomized to surgery alone or to three preoperative plus three postoperative cycles of epirubicin, cisplatin, and 5-fluorouracil (5-FU). The perioperative chemotherapy group seemed to have been down-staged by neoadjuvant chemotherapy, manifest by smaller tumor size and fewer positive nodes than the surgery-only group. At a median follow-up of 4 years, there was a significant improvement in 5-year overall and progression-free survival in patients treated with perioperative chemotherapy, although only 42% of patients in that group completed protocol therapy. Two smaller studies using different regimens were unable to demonstrate a survival benefit of neoadjuvant chemotherapy over surgery alone.

One RCT investigated neoadjuvant chemotherapy with or without radiation in locally advanced gastroesophageal (GE) junction tumors (55% Type I, 45% Type II/III) across 15 centers in Germany. The study was closed early because of poor accrual. Patients (N = 126) were randomized to chemotherapy followed by surgery or chemotherapy followed by chemoradiotherapy before surgery. Patients who had chemoradiation had a significantly higher complete pathologic response rate (15.6% versus 2.0%) and more often had tumor-free lymph nodes compared with patients treated with chemotherapy alone. There was no difference in operative complications or survival at 3 years, although there was a trend toward better survival in patients treated with chemoradiation (47.4% versus 27.7%, log rank P = .07). In sum, neoadjuvant therapy has become a standard approach in many centers for patients with locally advanced gastric cancer (T3, N any) who are fit to tolerate treatment. Radiation therapy in the neoadjuvant setting should be reserved for appropriately staged distal esophageal and proximal GE junction tumors.

Adjuvant Chemoradiation

One study has been performed during the period of this review that meets level 1a criteria. The Southwest Oncology Group published an RCT (Intergroup 0116) of surgery alone compared with surgery plus adjuvant chemoradiotherapy (CRT). Following R0 resection of gastric or gastroesophageal junction tumors, 556 well-selected patients were randomized to observation or adjuvant CRT. Of note, 54% of patients underwent a D0 lymphadenectomy. Although there was significant toxicity, overall and disease-specific survival were significantly improved in patients treated with adjuvant CRT.

Adjuvant Chemotherapy

There were 16 randomized adjuvant trials of chemotherapy without radiation performed during the period of this review. Numerous attempts to demonstrate a survival advantage to adjuvant systemic chemotherapy have failed, summarized by the prior review and multiple meta-analyses. Critical problems associated with these trials include the lack of surgical standardization and the heterogeneity of patient populations being studied. One trial demonstrating a survival benefit from adjuvant therapy was from Japan where patients were randomized to surgery alone or adjuvant S-1, an oral fluoropyrimidine. Surgical quality control was superior to that of the Intergroup 0116 trial, as all patients received D2 or more LND and were staged with peritoneal cytology. The trial was stopped after the first interim analysis at 1 year because of significantly better survival in patients treated with adjuvant S-1.

The remaining trials of adjuvant chemotherapy have demonstrated no benefit over surgery alone. Di Costanzo and colleagues randomized 258 patients to surgery alone or adjuvant chemotherapy with four cycles of cisplatin, epirubicin, 5-FU, and leucovorin (LV) (PELF). Of note, 47% of the patients in the study had fewer than 15 lymph nodes examined on pathologic analysis. At a median follow-up of 73 months, the rate of recurrence and overall and disease-free survival were similar between groups. Similar findings were reported by De Vita and colleagues who randomized 228 patients to adjuvant 5-FU, LV, epirubicin, and etoposide or to surgery alone. In a study to evaluate the addition of epirubicin and cisplatin to 5-FU, Cascinu and colleagues randomized patients to adjuvant PELF versus adjuvant 5-FU/LV; there was no measurable survival benefit. These studies also demonstrated that chemotherapy, when administered in the adjuvant setting, can impose significant toxicity.

The current evidence to date demonstrates that adjuvant therapy is beneficial in most patients. In Western patients, the standard options supported by level 1a studies are perioperative chemotherapy or adjuvant chemoradiotherapy. No study to date has demonstrated a survival benefit to adjuvant chemotherapy alone in a Western population.

Intraperitoneal Therapy

The largest RCT to date investigating intraperitoneal (IP) chemotherapy was published by Yu and colleagues, who randomized 248 patients to surgery alone or early adjuvant IP mitomycin-C (MMC) and 5-FU. All patients underwent D2 LND and chemotherapy patients received a Tenckhoff catheter for delivery starting on the first postoperative day. Patients treated with chemotherapy had a significant increase in intra-abdominal hemorrhage and abscess formation compared with controls; 24% of patients treated with chemotherapy had mild to moderate abdominal pain during treatment. Eight patients (6.4%) in the chemotherapy group died in the perioperative period compared with two (1.6%) in the control group. Despite the significant increase in perioperative morbidity and mortality, treated patients experienced a significant increase in survival and a decrease in peritoneal recurrence. On subgroup analysis, patients with stage I and II disease did not benefit from IP chemotherapy advocating for a better selection algorithm in these patients.

Intraperitoneal chemotherapy has not gained acceptance as a standard adjuvant therapy for gastric cancer in many centers because of the significant potential for perioperative morbidity. Moreover, there exists the potential to over treat patients with early stage disease. With the development of newer systemic agents, IP chemotherapy is unlikely to become a standard modality in the treatment of gastric cancer.

Helicobacter Pylori Eradication

One trial investigated the role for Helicobacter pylori eradication following endoscopic mucosal resection (EMR) in patients with early gastric cancer. This was a multicenter RCT from Japan in which 544 patients with documented H pylori infection were randomized to treatment or not following EMR. Patients were evaluated for metachronous cancer up to 36 months following resection and it was found that those who underwent treatment for H pylori had a significant reduction in the incidence of metachronous cancer (9 new cancers in treated patients versus 24 controls; odds ratio = 0.35).