SO-11 Response evaluation after neoadjuvant chemotherapy for resectable gastric cancer: Downsizing vs downstaging

Abstract

Response evaluation following neoadjuvant chemotherapy (NAC) in resectable gastric cancer is debated. The aims of the present study were to investigate the value of strict radiologic methods, hence UICC-downstaging and RECIST, as mode of response evaluation following a MAGIC-style regimen of NAC applied to a large Western cohor. Further to explore a composite radiologic/pathologicdownstaging comparing radiologic stage at diagnosis (rTNM) to pathologic stage following chemotherapy (ypTNM). Population-based study (2007-2016) on 171 consecutive patients with gastric adenocarcinoma, receiving a MAGIC style neoadjuvant chemotehrapi (NAC). Three methods of response evaluation are investigated, two strict radiologic methodes; one comparing initialt rTNM-stage and radiologic stage following NAC (yrTNM) and one using RECIST (downsizing), and a composite radiologic/pathologic comparing initial rTNM-stage to pathologic ypTNM-stage (downstaging). Of 171 patients receiving NAC, 169 were available for response evaluation. For radiologic TNM response mode, 43% responded, 50% had stable disease and 7% progressed at CT. Crosstabulating yrTNM stage to ypTNM stage, 24% had concordant stages, with CT overstaging 38% and understaging 38% of the tumours, Cohen kappa ƙ=0,06 (95%CI 0,004-0,12). Similar patterns of discordance were found for T-stages and N-stages separately. For M-category, restaging CT detected 12 patients with carcinomatosis, with an additional 14 diagnosed with carcinomatosis only at operation. No patient developed parenchymal or extra abdominal metastases, and none developed locally non-resectable tumour during delivery of NAC. Restaging CT with response evaluation was not able to stratify patients into groups of different long-term survival rates based on response mode. RECIST failed to identify half of the patients progressing to metastatic disease, and were not able to assign patients to subsets with different long-term survival rates. The compositeTNM response mode, comparing initialt radiologic stage (rTNM) to the pathologic TNM stage (ypTNM) did achieve this objective. A totale of 78/164 (48%) were downstaged, 25/164 (15%) had stable disease and 61/164 (37%) were upstaged. Histopathologic complete response was found in 15/164 (9%). The 5-year overall survival was 65.3% (95% CI 54.7% - 75.9%) for responders, 40.0% (95 CI 20.8% - 59.2%) for stable disease and 14.8% (95% CI 6.0% - 23.6%) for patients with progression, p < 0.001. In a multivariable ordinal regression Lauren category and tumour location were the only significant determinants of response mode. Routine CT-scan following NAC is of limited value. Accuracy of CT staging compared to final pathologic stage is poor, and strict radiologic downstaging as measure of response evaluation is unreliable and unable to discriminate long-term survival rates based on response mode. RECIST as method of response evaluation following NAC in gastric cancer is also discouraged. Comparing CT-stage at diagnosis to pathologic stage following NAC appears as a useful method of response evaluation serving the every-day situation well.