BackgroundPancreatic tail cancer (Pt-PC) is generally considered resectable when metastasis is absent, but doubts persist in clinical practice due to the variability in local tumor extent. We conducted a multicenter retrospective study to comprehensively identify prognostic factors associated with Pt-PC after resection. MethodsWe enrolled 100 patients that underwent distal pancreatectomy. The optimal combination of factors influencing relapse-free survival (RFS) was determined using the maximum likelihood method (MLM) and corrected Akaike and Bayesian information criteria (AICc and BIC). Prognostic elements were then validated to predict oncological outcomes. ResultsTherapeutic interventions included neoadjuvant treatment in 16 patients and concomitant visceral resection (CVR) in 37 patients; 89 patients achieved R0. Median RFS and OS after surgery were 23.1 and 37.1 months, respectively. AICc/BIC were minimized in the model with ASA-PS (≥2), CA19-9 (≥112 U/mL at baseline, non-normalized postoperatively), need for CVR, 6 pathological items (tumor diameter ≥19.5 mm, histology G1, invasion of the anterior pancreatic border, splenic vein invasion, splenic artery invasion, lymph node metastasis), and completed adjuvant treatment (cAT) for RFS. Regarding the predictive value of these 11 factors, area under the curve was 0.842 for 5-year RFS. Multivariate analysis of these 11 factors showed that predictors of RFS include CVR (hazard ratio, 2.13; 95 % confidence interval, 1.08–4.19; p = 0.028) and cAT (0.38, 0.19–0.76; p = 0.006). ConclusionsThe MLM identified certain Pt-PC cases warranting consideration beyond resectable during clinical management. Particular attention should be paid to conditions requiring CVR, even though immortal time bias remains unresolved with adjuvant treatment.
Read full abstract