Background: FOLFIRINOX (leucovorin, 5-fluorouracil, irinotecan, and oxaliplatin) is an option for fit patients with metastatic (MPC) and locally advanced unresectable (LAPC) pancreatic cancer. However, no criteria reliably identify patients with better outcomes.
Patients and methods: We investigated putative prognostic factors among 137 MPC/LAPC patients treated with triplet chemotherapy. Association with 6-month survival status (primary endpoint) was assessed by multivariate logistic regression models. A nomogram predicting the risk of death at 6 months was built by assigning a numeric score to each identified variable, weighted on its level of association with survival. External validation was performed in an independent data set of 206 patients. The study was registered at ClinicalTrials.gov (NCT03590275).
Results: Four variables (performance status, liver metastases, baseline carbohydrate antigen 19-9 level, and neutrophil-to-lymphocyte ratio) were found to be associated with 6-month survival by multivariate analysis or had sufficient clinical plausibility to be included in the nomogram. Accuracy was confirmed in the validation cohort (C index = 0.762; 95% confidence interval, 0.713-0.825). After grouping all cases, 4 subsets with different outcomes were identified by 0, 1, 2, or > 2 poor prognostic features (P < .0001).
Conclusion: The nomogram we constructed accurately predicts the risk of death in the first 6 months after initiation of FOLFIRINOX in MPC/LAPC patients. This tool could be useful to guide communication about prognosis, and to inform the design and interpretation of clinical trials.
Keywords: Clinical parameters; FOLFIRINOX; Laboratory parameters; Prognosis; Risk categories.