Background: Advanced pancreatic ductal adenocarcinoma (aPDAC) has a poor prognosis with median overall survival (OS) of about 12 months. It is therefore important to explore factors that predict the efficacy of third-line chemotherapy (L3) to identify patients who may benefit from this controversial treatment.
Methods: We conducted a multicenter retrospective cohort-based study of 202 French patients treated for aPDAC who received at least three treatment lines from January 2011 to March 2022. We used penalized Cox regressions to predict progression-free survival (PFS) and OS in patients on L3.
Results: Median age at the start of L3 was 64.3 years old and 63.5% had an Eastern Cooperative Oncology Group Performance Status (ECOG-PS) of 0 or 1. The most frequent regimens for L3 were FOLFIRI (25.2% of patients). Median PFS was 2.2 months, while median OS was 4.2 months. In multivariate models, we identified the following predictors of both PFS and OS: age, sex, surgery for the primary tumor, FOLFIRINOX as the first-line therapy, duration of first and second-line treatments, and for L3: ECOG-PS level, peritoneum, liver and/or lung metastasis and depletion of therapeutic resources. The model incorporating these factors provided acceptable discrimination between event and event-free patients at 6 months post-L3 (area under the ROC curve of 0.83 for PFS and 0.73 for OS).
Conclusion: The characteristics of patients and their aPDAC are readily available in clinical practice and were able to predict survival with L3. The online calculator we propose here could help physicians determine whether L3 chemotherapy would be beneficial.
Keywords: cohort; pancreatic cancer; predictive factors; third-line chemotherapy.
© The Author(s) 2025. Published by Oxford University Press.