How should tracers be injected to detect for sentinel nodes in gastric cancer - submucosally from inside or subserosally from outside of the stomach?

Background: In sentinel node (SN) detection for cases of early gastric cancer, the submucosal dye injection method appears to be more reasonable than the subserosal injection. To compare the two injection methods, we have focused on the rate of concordance between hot nodes (HNs) obtained from the radioisotope (RI) method and green nodes (GNs) obtained from the dye-guided method in addition to the number and distribution of GNs detected, and the sensitivity of metastatic detection. Methods: The subjects of this study were 63 consecutive patients with gastric cancer (sT1-T2, sN0, tumor diameter % 4 cm) in whom we attempted SN detection using a combination of RI and dye methods. Tc-99m-tin colloid was injected a day before the surgery, and indocyanine green was injected either submucosally (n=43) with endoscopes or subserosally (n=20) by direct vision. Results: An average of hot and green nodes (H&G: 4 +/- 3 vs. 4 +/- 3), hot and non-green nodes (H&NG: 2 +/- 3 vs. 1 +/- 2), cold and green nodes (C&G: 2 +/- 2 vs. 3 +/- 4), and the rate of concordance (H&G/H&G + H&NG + C&G: 45 + 27% vs. 48 +/- 30%) were not significantly different between the submucosal and subserosal injection methods. The spread of GNs to tier 2 stations (24% vs. 30%) and metastatic detection sensitivity (86% vs. 100%) were also not different between the submucosal and subserosal injection methods. Conclusion: The tracer injection sites do not have to be limited to the submucosa.