Abstract: Objective To construct a nomogram model for predicting the early recurrence of pancreatic head cancer after radical pancreaticoduodenectomy and to evaluate its application value. Methods This retrospective cohort study selected patients who underwent radical pancreaticoduodenectomy for pancreatic head cancer in our hospital from June 2017 to July 2019 as the research objects. The outcome of the study was early postoperative recurrence. The influencing factors of early recurrence were analyzed by univariable and multivariable logistic regression, and a nomogram model was constructed based on the influencing factors. The discrimination of the nomogram model was evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC)  curve, the calibration was evaluated by the calibration curve and Hosmer-Lemeshow test, and the value of clinical application was evaluated by the decision curve. Results A total of 137 patients met the criteria were included in the final analysis. 58 patients (42.3%) had early recurrence after surgery. Multivariable logistic regression showed that tumor size ≥3 cm, preoperative CA19-9 level >37 U/mL, poorly differentiated tumor and the number of lymph node metastases >3 were risk factors affecting the early postoperative recurrence (all P<0.05). Based on these factors, a nomogram model was successfully constructed. The AUC was 0.807 (95%CI: 0.729-0.885). The Hosmer-Lemeshow test of the calibration curve showed that the model had a good calibration degree (P=0.569). The decision curve showed that the nomogram had good clinical application value, and intervention could be taken when the early recurrence probability reached 22%. Conclusions A nomogram model is successfully constructed for predicting the early recurrence of pancreatic head cancer after radical pancreaticoduodenectomy, and it is helpful in the early clinical screening and identification of patients at risk.

Key words: Pancreatic head cancer, Radical pancreaticoduodenectomy, Early recurrence, Nomogram