Oncological impact of unexpected horizontal tumour spread in oesophagogastric junction cancer.

Abstract

Background: Unexpected horizontal tumour spread towards the proximal and distal margins complicates the assessment of surgical margins in oesophagogastric junction (OGJ) cancer. Its impact on oncological outcomes remains unclear.

Methods: This study retrospectively analysed patients with OGJ adenocarcinoma undergoing proximal or total gastrectomy. Unexpected horizontal tumour spread was measured as the discrepancy between gross and pathological margins proximally (ΔPM) and distally (ΔDM). Clinicopathological features, recurrence-free survival (RFS), and overall survival (OS) were evaluated based on ΔPM and ΔDM.

Results: Based on cut-off values identified by time-dependent receiver operating characteristic curve analysis (ΔPM, 8 mm; ΔDM, 3 mm) in 197 patients, patients were classified into four groups: short; long ΔPM; long ΔDM; and both long ΔPM and ΔDM (both-long). RFS was best in the short group and worst in the both-long group. The long ΔPM and long ΔDM groups had intermediate and comparable RFS. Subsequently, patients were categorized into two groups: a short group and a long group, which included patients in the long ΔPM, long ΔDM, and both-long groups. The type of infiltrative growth and postoperative recurrence were significantly associated with the long group. Moreover, the long group had significantly worse RFS and OS than the short group. Multivariate Cox regression analyses identified the long group as an independent risk factor for both RFS and OS. Patients in the long group with clinical lymph node metastasis or tumours located in the proximal 2-cm segment of the OGJ, predominantly in the proximal rather than distal 2-cm segment of the OGJ, or equal involvement in both areas had markedly worse survival outcomes.

Conclusion: Unexpected horizontal tumour spread, represented by ΔPM and ΔDM, is a strong predictor of poor survival and recurrence in OGJ cancer. Intraoperative assessment of ΔPM and ΔDM using frozen section analysis may be useful in guiding additional resections, particularly when combined with other predictive factors.