Prognostic Impact of Frozen Section Investigation and Extent of Proximal Safety Margin in Gastric Cancer Resection

Background and Aims: Guidelines propose different extents of macroscopic proximal margin for gastric cancer and frozen margin investigation in selected cases, but data is lacking. This study was to evaluate the necessary extent of macroscopic proximal margin, accuracy of frozen margin investigation, and prognostic impact of tumor-free proximal margin length in pT2-pT4 gastric cancer. Study Design: Proximal and distal frozen margins were routinely investigated intraoperatively in all pT2-pT4 gastric cancers resected between 2011 and 2017. Macroscopic and microscopic proximal margin lengths were correlated. For R0-resections, survival analysis was performed for distal gastrectomy (DG) with microscopic proximal margin length <= 3 cm versus >3 cm. Results: Overall, 1484 patients were included. Microscopic proximal margin lengths were macroscopically more often misestimated in diffuse histology (P = 0.0004), but extent of underestimation in centimeter was similar to intestinal and mixed/undetermined type (P = 0.134). Fifteen cases (1.0%) resulted in R1-resection, 10 at distal, and 5 at proximal margin but none with macroscopic proximal margin >= 3 cm and negative frozen section. Overall agreement of frozen margin and final pathology was 2951/2968 (99.4%). Proximal margin length in DG did not correlate with survival or recurrence in R0-resected patients. Discussion: Diffuse histology is at higher risk for underestimation of proximal margin length, but extent of underestimation is similar in other Laure ' n subtypes. If >= 3 cm macroscopic proximal margin length is applied with intraoperative frozen margin confirmation, R1-resection can be avoided. Conclusion: In pT2-T4a gastric cancer, proximal margin of >= 3 cm plus frozen margin confirmation provides high oncological safety. In DG patients with R0-resection, proximal margin length does not correlate with survival or recurrence.