Gastric Cancer: D2 All Over Again

Why do reports from large volume single institutions continue to demonstrate low operative mortalities, short hospital stays, and prolonged patient survivals for patients with gastric cancer treated by D2 lymphadenectomy when randomized trials show no survival advantage and increased operative mortality with this procedure? Should gastric cancer resections be added to the growing list of procedures with reduced mortality and improved survivorship when performed in highvolume institutions? Although not a randomized trial, the report by Dr. Sierra and colleagues from the University of Navarra in Paploma, Spain, adds credence to the concept that D2 resections can be performed safely and provide enhanced survival when performed in high-volume hospitals by a small group of experienced surgeons. 1,2 The D2 dissection that they performed included total gastrectomy and splenectomy but did not require resection of the tail of the pancreas as performed in the Dutch and British randomized trials. 3,4 Instead, lymph nodes along the splenic artery (station 11 in the Japanese classification) were usually dissected away from surrounding tissues to be included with the resected specimen. Pancreatic fistulas did occur in three of the eight D2 patients who required pancreatic resection. This suggests that, in the randomized trials, the pancreatic resection, rather than the splenectomy and the D2 lymphadenectomy, was the major cause of the excess operative mortality in the D2 group. 3,4 In this nonrandomized comparison, how comparable were the two groups of patients? Could patient characteristics explain the better survival in the D2 group? Patients in the D2 group were younger and had fewer comorbidities (ASA I index in 82% of D2 patients vs. 52% of the D1 group) and were treated by total gastrectomy (100% of D2 vs. 44% of D1 patients) and adjuvant radiotherapy (47% of D2 vs. 27% of D1 patients). On the other hand, patients treated by D1 resection were less likely to have lymph node metastases (45% of D1 patients were stage N0 vs. 37% of the D2 group) and M1 disease (13% of D1 vs. 20% of D2 patients). Thus, survival benefit of D2 dissection could only be explained from these characteristics if total gastrectomy vastly improved survival, which it does not. On average, more than twice as many lymph nodes were removed by D2 resections as compared with D1 (31 vs. 14 lymph nodes, respectively). The authors evaluated the fraction of lymph nodes containing metastases divided by the total number of lymph nodes removed (nodal index), noting that this fraction was 20% or less in 34% of patients in the D2 group but was only 17% among D1 patients. Patients with a nodal index of 20% or less had a significantly prolonged survival compared with the group of gastric cancer patients as a whole, whether treated by D1 or D2 resection. Although increased survival of more advanced stage patients in a series of D2 resections has been attributed to stage mi