Neoadjuvant chemoradiation for patients with advanced oesophageal cancer – which response grading system best impacts prognostic discrimination?

AIMS: Neoadjuvant chemoradiation reduces tumour volume and improves the R0 resection rate, followed by extended survival for patients with advanced oesophageal cancer. The degree of tumour regression has high prognostic relevance. To date, there is still no generally accepted tumour regression grading system. The aim of this study was to compare the prognostic discrimination power of different histological regression grading systems: (i) the fibrosis/tumour ratio within the primary tumour (Mandard classification), (ii) the percentage of residual vital tumour cells (VTC) compared to the original primary tumour (Cologne Regression) and (iii) the ypT category, in patients with cT3 carcinoma of the oesophagus after neoadjuvant chemoradiation. METHODS AND RESULTS: This study included 216 patients with oesophageal cancer clinically staged as cT3NxM0 and treated from 2009 to 2012 with standardised chemoradiation followed by oesophagectomy [median age 62 years, 176 (81%) male and 138 (64%) adenocarcinoma patients]. The subgroup frequencies of the three classification systems were ypT category: ypT0 = 18%, ypT1 = 14%, ypT2 = 23%, ypT3 = 44%, ypT4 = 1%; Mandard classification: TRG1 = 18%, TRG2 = 26%, TRG3 = 24%, TRG4 = 30%, TRG5 = 2%; and Cologne Regression Scale: no tumour = 18%, 1-10% VTC = 27%, 10-50% VTC = 26% and >50% VTC = 29%. The Mandard and Cologne Regression classifications showed better prognostic differentiation for the subgroups than the ypT category. The four-tiered Cologne Regression system had a good prognostic relevance. Comparing results of the re-evaluated Cologne Regression classification with the classification by routine pathological report showed very good inter-rater agreement, with kappa value 0.891. CONCLUSION: Compared to the original primary tumour, the tumour regression grading system using the percentage of residual vital tumour has prognostic relevance.