Developing and validating a nomogram to forecast clinically relevant postoperative pancreatic fistula following distal pancreatectomy.

BACKGROUND

After distal pancreatectomy (DP), clinically relevant postoperative pancreatic fistula (CR-POPF) is a critical complication that adversely affects the prognosis of patients. The present study identified the risk factors for CR-POPF occurrences, as well as developing a nomogram to predict their risk after DP.

METHODS

We retrospectively examined 300 medical records, obtaining the patients' preoperative clinical baseline data, laboratory indicators, preoperative computed tomography (CT) data, and intraoperative clinical information. We determined CR-POPF independent risk factors using univariate as well as multivariate logistic regression analyses. We created a risk nomogram based on these variables and used the bootstrap method for internal validation. Area under the curve (AUC) assessed the nomogram's predictive power. The nomogram's clinical value and viability were evaluated using decision curve analysis (DCA) as well as clinical impact curve (CIC).

RESULTS

CR-POPF developed in 84 of the 300 patients (28.0% incidence). CR-POPF was found to be independently risked by operation time (P=0.002), preoperative C-reactive protein (CRP) levels (P<0.001), CT (pancreas)-to-CT (psoas major) ratio (P<0.001), and pancreatic thickness (PT) at transection site (P<0.001). The nomogram's AUC was 0.901, which, along with the DCA, highlighted the nomogram's excellent performance, surpassing those of four alternative CR-POPF prediction models. The nomogram has an immense net clinical advantage, according to the CIC.

CONCLUSIONS

The developed nomogram can be useful in identifying high-risk patients and formulating individualized perioperative plans to prevent the risk of CR-POPF formation in patients undergoing DP.