Oncological recurrence following pathological complete response after neoadjuvant treatment in patients with esophageal cancer — a retrospective cohort study

BACKGROUND: To evaluate recurrence in patients with post-neoadjuvant pathological complete response (pCR) and in patients with complete response of primary tumor but persisting lymphatic spread of disease (non-pCR, ypT0ypN +) of esophageal cancer. METHODS: Seventy-five patients (63 pCR, 12 non-pCR) were analyzed retrospectively. Pattern and incidence of local and distant recurrence as well as the impact on overall (OS) and disease-free survival (DFS) were evaluated. The efficacy of neoadjuvant chemotherapy according to FLOT protocol was compared to neoadjuvant chemoradiation according to CROSS protocol. RESULTS: In the pCR group, isolated local recurrence was diagnosed in 3%, while no isolated local recurrence was observed in the non-pCR group due to the high incidence of distant recurrence. Distant recurrence was most common in both cohorts (isolated distant recurrence: pCR group 10% to non-pCR group 55%; simultaneous distant and local recurrence: pCR group 3% to non-pCR group 18%). Median time to distant recurrence was 5.5 months, and median time to local recurrence was 8.0 months. Cumulative incidence of distant recurrence (with and without simultaneous local recurrence) was 16% (± 6%) in pCR patients and 79% (± 13%) in non-pCR patients (hazard ratio (HR) 0.123) estimated by Kaplan–Meier method. OS (HR 0.231) and DFS (HR 0.226) were significantly improved in patients with pCR compared to patients with non-pCR. Advantages for FLOT protocol compared to CROSS protocol, especially with regard to distant control of disease (HR 0.278), were observed (OS (HR 0.361), DFS (HR 0.226)). CONCLUSION: Distant recurrence is the predominant site of treatment failure in patients with pCR and non-pCR grade 1a regression, whereby recurrence rates are much higher in patients with non-pCR. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s00423-023-03100-2.