Background To evaluate the feasibility and accuracy of diagnosis using sentinel node (SN) biopsy in T1 gastric cancer, a multicenter trial was conducted by the Japan Clinical Oncology Group (JCOG). Methods Sentinel node biopsy with indocyanine green (ICG) was performed in patients with T1 gastric cancer. Green-stained nodes (GNs), representing SNs, were removed first, and gastrectomy with lymphadenectomy was then performed. GNs in one plane (with the largest dimension) were histologically examined intraoperatively by frozen section with hematoxylin and eosin (H&E) stain. All harvested lymph nodes (GNs and non-GNs) were histologically examined by paraffin section after surgery. The primary endpoint was to determine the proportion of false negatives, which was defined as the number of patients with negative GNs by frozen section divided by those with positive GNs and/or positive non-GNs by paraffin section. The sample size was set at 1,550, based on the expected and threshold value as 5 and 10 % in the proportion of false negatives. Results Accrual was suspended when 440 patients were enrolled because the proportion of false negatives was high. In the primary analysis, the proportion of false negatives was 46 % (13/28) after a learning period with 5 patients for each institution. Seven of 13 patients had nodal metastases outside the lymphatic basin. False negatives remained at 14 % (4/28) even by examining additional sections of GNs by paraffin section. Conclusions The proportion of false negatives was much higher than expected. Intraoperative histological
123Gastric Cancer (2014( ) 17:316-323 DOI 10.1007 examination using only one plane is not an appropriate method for clinical application of SN biopsy in gastric cancer surgery.