Adjuvant capecitabine: safe and tolerable option for adjuvant treatment in patients with resected pancreas and ampullary adenocarcinoma

Introduction: Since the ESPAC-1 and CONKO-001 studies were published, sixmonths of adjuvant chemotherapy with 5-FU or gemcitabine, respectively, are givenpost-operatively to improve disease-free survival (DFS) and overall survival (OS).Capecitabine (5-FU pro-drug) is widely employed as an oral alternative to infusional5-FU, reducing cannulations and hospital visits. However, no quality data supports its use in this setting.Methods: In this single-centre, retrospective study, eligible patients had resectedpancreatic or periampullary adenocarcinoma treated with adjuvant chemotherapy with availability of outcome and toxicity data (Aug’08-Aug’14; data cut-off: 25/Feb/2015). Kaplan-Meier, Cox regression and ROC analyses were employed, as appropriate. Uni-/multivariable analysis was performed (Stata v.12 Package). Ampullary malignancies were excluded from analysis for the identification of prognostic factors.Results: Sixty-two of 76 screened patients were eligible; median age at diagnosis 65.9 years (41.8-84.3); male (61%); none/mild comorbidities (89%); ECOG performance status (PS) 0 (34%), 1 (65%), 2 (1%); primary pancreas (n = 48 (77%); 85% head) and ampulla (n = 14 (23%)). Capecitabine or gemcitabine were given to 28 (45%) and 34 (55%) patients; 23(37%) within and 11(18%) outside a clinical trial, respectively. Pathological characteristics: differentiation [moderate (57%); poor (42%)]; lymphovascular (59%) and perineural (85%) invasion; resection [R0 (34%); R1 (66%)]; median involved/resected lymph node ratio (LNR) 0.1 (0-0.8); median tumour size (T) 30mm (4-80); stage IIa (11%) and stage IIb (77%); with no differences in baseline characteristics between capecitabine and gemcitabine cohorts. Median time between surgery and chemotherapy was 2.4 months (1.6-3.8). Median dose-intensity received was 87.6% (20.5-102.6); [capecitabine (79.3%; grade3/4 toxicity rate 32%); gemcitabine (98.5%; grade3/4 toxicity rate 50%)]. With a median follow-up of 19.6 months (range 6-69.8), 74% of patients have relapsed (45% distant metastases) and 60% died. Estimated median DFS and OS were 13.3 (95%CI 11.6-16) and 25.1 (95%CI 19.4-31.9) months, respectively. With ampullary malignancies excluded; N1 (vs. N0; p = 0.03), high LNR ( p = 0.03) and moderate comorbidities (vs. none; p = 0.025) were independent factors associated with shorter DFS when adjusted for other prognostic factors (ECOG PS, T, comorbidities, R0/R1, dose-intensity, baseline CA19-9, trial/non-trial and type of adjuvant treatment) in the multivariable analysis. A LNR cut-off of 0.2 provided a sensitivity of 61.9% and specificity of 80.8% for identification of patients with high risk of relapse within 12 months of surgery. Shorter OS was shown in patients with ECOG PS2 (vs. PS0; p = 0.016), N1 (vs. N0; p <0.001) and high CA19-9 ( p = 0.002) in the multivariable analysis (adjusted for comorbidities, R0/R1, T, number of resected lymph nodes, dose-intensity, trial/non-trial and type of adjuvant treatment). CA19-9 >17U/ml prior to starting adjuvant treatment identified patients with higher 1-year death rate (sensitivity 85.7%; specificity 70.6%). There were no significant differences between the cohort of gemcitabine and capecitabine for DFS(HR 3.5; 95% CI 0.9-14.3; p = 0.074) or OS (HR 0.6; 95% CI 0.1-3.2; p = 0.58).Conclusion: Within the limitations of this retrospective study and the small samplesize, adjuvant capecitabine may be an option for patients with resected pancreaticadenocarcinoma, with no differences in DFS or OS demonstrated when compared topatients receiving adjuvant gemcitabine. This would require validation in a prospective study