Clinical outcomes of pancreas cancer patients with peritoneal spread: Results from the chord consortium.

Abstract

e19268 Background: Prior studies have demonstrated that clonal cells that give rise to pancreatic peritoneal metastases (PM) are geographically and genetically distinct from clonal cells, giving rise to lung and liver metastases. The objective of this study was to assess if there is a distinct difference in prognosis and therapeutic response among patients with pancreatic cancer with (PM compared to the lung/liver. Methods: Using a retrospective cohort design, medical records from adult patients diagnosed with metastatic adenocarcinoma of the pancreas at five Canadian academic cancer centers (2014 - 2019) were reviewed. Prognostic variables including age, Charlson comorbidity index, ECOG, cigarette smoking, nodal status, sites of metastases, and first line chemotherapy were collected. Cox proportional hazards model (MVA) was used to examine the association between peritoneal involvement and survival, adjusted for measured confounders. Analyses were completed using SAS, where alpha of 0.05 was defined as the level of significance. Results: A total of 1161 patients were included. Metastatic sites included peritoneum (n = 170, 14.6%), lung (n = 145, 12.5%) and liver (n = 563, 48.5%). Patients with PM received first-line FOLFIRINOX (FFX, n = 31), Gemcitabine + nab-paclitaxel (G/N, n = 20), Gemcitabine (G, n = 18), and no treatment (n = 97). In univariate analyses, worse ECOG PS was associated with PM (p = 0.002). The majority of patients died (89%), with a median overall survival (OS) of 3 vs 7 months for patients with PM and those without PM (p < 0.001), respectively. The median OS in patient whom receive first-line chemotherapy was 7 months in FFX group (95% CI 1.66-12.33), 6 months in G/N (95% CI 4.54-7.45) and 2 months in G group (95% CI 1.42-2.57). Patients had significantly better OS when treated with FFX or G/N compared to G alone (p = 0.002). Time to treatment failure was significantly shorter among patient treated with G alone compare to patients treated with FFX and G/N (P < 0.005). Conclusions: In the setting of combination chemotherapy for advanced pancreatic cancer, patients with PM continue to have a poor prognosis. This may be due to the impact of PM on PS and the inability to administer palliative chemotherapy. For eligible patients, FFX or G/N results in a higher OS than G monotherapy.