Reappraising the Concept of Conditional Survival After Pancreatectomy for Ductal Adenocarcinoma: A Bi-institutional Analysis

OBJECTIVE: To reappraise the concept of conditional survival (CS) following pancreatectomy for pancreatic ductal adenocarcinoma (PDAC), accounting for the patient's present disease status relative to recurrence. BACKGROUND: CS, defined as the probability of surviving an additional time frame based on accrued lifespan, offers dynamic survival projections as compared with baseline overall survival. METHODS: Patients undergoing pancreatectomy for PDAC at 2 institutions from 2000 to 2013 were retrospectively analyzed. The 12-month CS was estimated separately for patients who were disease-free or with recurrence at the given time points. Next, the conditional probability of reaching 60-months of survival was examined in each conditioning set across strata of prognostic covariates, including American Joint Committee on Cancer stage, tumor grade, R-status, and adjuvant treatment. RESULTS: The study population consisted of 1005 patients. In disease-free patients, the 12-month CS increased as a function of time already survived, showing an opposite trend compared with overall survival. In patients who recurred, the 12-month CS was lower than the disease-free counterpart, especially within 24 months postoperatively. When stratifying by the levels of prognostic covariates, the 60-months CS estimates for disease-free patients tended to level off progressively, indicating that factors independently associated with survival at the time of pancreatectomy lost power over time. This concept did not apply to the conditioning set of patients with recurrence, where CS estimates across variables strata diverged with accrued lifespan. CONCLUSION: This paper provides new information on how prognosis following pancreatectomy for PDAC evolves over time, adjusting for the time the patient already survived, and for the patient's present disease status relative to recurrence