OGC O02 The prognostic effect of pathological lymph node regression after neoadjuvant chemotherapy for oesophageal adenocarcinoma – a multicentre study

Abstract Background The prognostic benefits of primary tumour and lymph node (LN) downstaging after neoadjuvant chemotherapy for oesophageal adenocarcinoma are well described. Pathological primary tumour regression grading (TRG) is widely used in the assessment of response to chemotherapy and has been shown to have prognostic value in this patient group. However, there is a lack of robust evidence regarding the prognostic effect of pathological response in LN despite emerging evidence of a discrepancy, up to 25% of patients in some studies, between TRG in the primary tumour and response in regional LNs. Although primary tumour regression is routinely documented as part of the standard pathological reporting of oesophagectomy specimens, LN regression is generally overlooked despite it's potential prognostic value. The aim of this study was to investigate the relationship between pathological response in LN, tumour recurrence and survival. Methods Multicentre cohort study including 763 patients with oesophageal adenocarcinoma treated with neoadjuvant chemotherapy followed by surgery at Guy's and St Thomas’ NHS Foundation Trust (NHS-FT), The Royal Marsden NHS-FT, University Hospitals Birmingham NHS-FT, University Hospital Southampton NHS-FT and Belfast Health and Social Care Trust. Tumour regression was assessed in the primary tumour (as described by Mandard) and, retrospectively, in LNs retrieved from oesophagectomy specimens. LN were graded according to the proportion of fibrosis and residual tumour providing a LN regression score (LNRS). LNRS 1, complete response; LNRS 2, < 10% remaining tumour; LNRS 3, 10–50% remaining tumour; LNRS 4, > 50% viable tumour; LNRS 5, no evidence of response. Regression was defined as a LNRS of 1–3. Patients were classified as LN negative (no evidence of tumour or regression in any LN), complete LN-responders (evidence of regression ≥1 LN, no residual tumour in any LN), partial LN-responders (evidence of regression ≥1 LN with residual tumour ≥1 LN) and LN non-responders (no or minimal regression in any LN). Survival analysis was performed using multivariable Cox regression providing hazard ratios (HR) with 95% confidence intervals (CI) adjusting for age, gender, chemotherapy regimen, clinical stage, tumour grade, lympho-vascular invasion and primary tumour response. Results Mean age was 63 years with the majority male (86.2%). In total, 17,930 LN from 763 patients were analysed for evidence of response to chemotherapy. Overall, 243 (31.8%) patients were classified as LN negative, 62 (8.1%) as complete LN-responders, 155 (20.3%) as partial LN-responders and 303 (39.7%) as LN non-responders. Less than half (322/763, 42.2%) of patients demonstrated a pathological response in the primary tumour (Mandard score 1–3). Some patients had a LN response in the absence of a response in the primary tumour (97/431, 22.5%). Multivariable Cox regression survival analysis demonstrated improved overall survival in complete LN-responders (HR 0.37 95% CI 0.24–0.58), partial LN-responders (HR 0.70 95% CI 0.55–0.89) and LN negative patients (HR 0.34 95% CI 0.26–0.44) compared to LN non-responders. Similar results were observed for disease-free survival (complete LN-responders, HR 0.34 95% CI 0.22–0.53; partial LN-responders, HR 0.74 95% CI 0.58–0.93; LN negative, HR 0.33 95% CI 0.25–0.42). Rates of tumour recurrence were lower in patients who demonstrated a LN response or had negative LN (LN negative 23.0% vs complete LN-responders 19.4% vs partial LN-responders 50.3% vs LN non-responders 66.7%, p<0.001). Conclusions In this cohort of patients with oeosphageal adenocarcinoma treated with neoadjuvant chemotherapy prior to surgical resection, LN regression was a strong predictive factor for better survival. This relationship was independent of primary tumour response, which was discordant in a significant number of patients. Complete LN-responders had equivalent survival to those with negative LN. Complete and partial LN-responders had better survival than LN non-responders. Evaluation and documentation of LN regression should be considered during the standard pathological reporting of oesophagectomy specimens.