Sentinel Lymph Node Sampling for Early Gastric Cancer-Preliminary Results of A North American Prospective Study

Although endoscopic resection for early gastric cancer is well established, anatomical resection with regional lymphadenectomy is recommended for lesions at high risk for occult lymph node metastasis (e.g., lymphovascular invasion, poor grade, and deep submucosal invasion). However, 75-95% high-risk early gastric cancer (HR-EGC) patients ultimately have node-negative disease and could potentially have undergone organ-sparing resection. Due to the inadequacy of standard modalities to reliably rule out nodal metastases in HR-EGC patients, sentinel lymph node (SLN) sampling was developed in Asia with promising results. However, the applicability of this technique in the West has been brought into question due to potential differences in tumor histology and body habitus. This prospective study aimed to test SLN sampling for North American EGC patients.

All patients with biopsy-confirmed T0-2 N0-1 M0 gastric adenocarcinoma at the Montreal General Hospital-McGill University Health Centre were eligible for enrollment. Esophageal and GEJ cancers were excluded due to the high rate of intrathoracic lymph node involvement. Peritumoral submucosal injection with T⁹⁹ radiocolloid was performed endoscopically 24-30 h prior to surgery. Methylene blue dye injection was performed after induction of anesthesia. SLN basins were identified as those having > 10% of baseline tumor radiation signal or blue color, or both. After basins were individually removed, standard laparoscopic anatomical resection was then performed with D2 lymphadenectomy. ( ClinicalTrials.gov identifier: NCT03049345). Data are presented as median (interquartile range).

From July 2016-April 2018, 253 patients with esophagogastric adenocarcinoma were evaluated. Of these, 10 met inclusion criteria (90% male, age 66(30) years). Subtotal gastrectomy was performed in nine patients (90%) and length of stay was 4 (2) days. At least one SLN basin was identified in nine cases (90%). The median #SLN basins identified was 2(2) with a median of 5(5) total SLNs retrieved per patient. In the one case for which no SLN basins were identified, only blue dye injection was used, whereas SLNs were identified in all cases using the dual tracer method. Final T-stage was pT1b/T2 in four (40%), pT1a in two (20%), and Tx in four (40%). Two patients (20%) had lymph node metastases on final pathological analysis, both of which were identified by SLN sampling (accuracy 100%; false negative rate 0%). No adverse events related to SLN retrieval were identified.

This study represents the first prospective feasibility evaluation of sentinel lymph node sampling for early gastric cancer in North America with promising preliminary results. The dual tracer method was superior to single agent blue dye in identifying sentinel nodal basins. Considerable further study is necessary to verify the safety and utility of SLN mapping in North American patients with early gastric adenocarcinoma.