High expression of krüppel-like factor 10 or Smad4 predicts clinical benefit of adjuvant chemoradiotherapy in curatively resected pancreatic adenocarcinoma: from a randomized phase III trial

Abstract Purpose Our previous studies have demonstrated that Kruppel-like factor 10 (Klf10) modulated tumor radiation resistance and helps to predict clinical outcomes of pancreatic adenocarcinoma (PDAC). This study aimed to evaluate whether the expression levels of Klf10, Smad4 and Runx3 can help predict the benefits of adjuvant chemoradiotherapy (CRT) in resected PDAC. Methods and Materials Tissue specimens were collected from 111 patients with curatively resected PDAC who were enrolled into a randomized trial comparing adjuvant gemcitabine with or without CRT. Immunohistochemical expression of biomarkers was quantified by pathologists blinded to patient outcomes through a grading system based on the extent and intensity of staining. The predictive value of biomarkers was analyzed using SAS statistical software. Results In total, 56 and 55 patients received adjuvant gemcitabine alone and additional CRT, respectively. The expression levels of Klf10, Smad4 and postoperative CA19-9 were significantly correlated with overall survival (OS) (p = 0.013, 0.045, and 0.047, respectively). Multivariable analysis showed that the expression level of postoperative serum CA19-9 and tumor tissue Klf10 expression level were significant predictors for OS (p = 0.038, and 0.028, respectively). Patients with high Klf10 or Smad4 (n=55), had a significantly better local recurrence-free survival (∞ vs 19.8 months; p = 0.026) and a longer OS (33.0 vs 23.0 months; p = 0.12) if they received additional adjuvant CRT than gemcitabine only. The results were similar after adjusted by postoperative level of CA19-9. Conclusion Patients with curatively resected PDAC and a high expression of either Klf10 or Smad4 have high chances of benefiting from adjuvant CRT. Combining Klf10 and Smad4 to predict the benefits of adjuvant CRT in resected PDAC deserves further validation.