Kimura or Warshaw? Development and Validation of a Preoperative Prediction Model Based on 3D Reconstruction Technology.

Abstract

Background: The purpose of this study was to investigate the risk factors for splenic vessel preservation in laparoscopic Kimura procedure and to establish and validate a nomogram model for predicting the failure of splenic vessel preservation.

Patients and methods: Clinical data of 169 patients who underwent spleen-preserving distal pancreatectomy for benign or low-grade malignant tumors of the distal pancreas at the Affiliated Hospital of Qingdao University from July 2018 to April 2024 were retrospectively collected. Among them, there were 55 male and 114 female individuals, with an average age of 48.24 ± 14.62 years. All patients were planned to undergo laparoscopic Kimura procedure on the basis of preoperative discussions. A total of 118 patients (69.82%) were randomly assigned to the modeling group, and 51 patients (30.18%) to the validation group. No significant differences were observed in baseline characteristics between the two groups (all P > 0.05). On the basis of data from the modeling group, independent risk factors for splenic vessel preservation were identified using rank sum test and multivariate logistic regression analysis. R software was used to establish a nomogram model for predicting the failure of splenic vessel preservation, and external validation was performed using data from the validation group.

Results: On the basis of data from the modeling group, the cutoff values of tumor volume, ratio of splenic vein (SV) circumference embedded in pancreas, and pancreas-SV contact area were determined using receiver operating characteristic curves as follows: 9828.8 mm³ [area under the curve (AUC) = 0.745, 95% confidence interval (CI) 0.654-0.835], 0.45 (AUC = 0.751, 95% CI 0.610-0.893), and 501 mm² (AUC = 0.715, 95% CI 0.588-0.903), respectively. The AUC for tumor location was 0.733 (95% CI 0.622-0.874). Univariate and multivariate logistic regression analyses showed that tumor volume, tumor location, ratio of SV circumference embedded in pancreas, and pancreas-SV contact area were valuable preoperative indicators for predicting the failure of splenic vessel preservation (P < 0.05). The nomogram prediction model established using R software on the basis of these four indicators exhibited good predictive performance, with C-indices of 0.747 in the modeling group and 0.756 in the validation group. Decision curve analysis indicated that the nomogram model had favorable clinical benefits.

Conclusions: Tumor volume, tumor location, ratio of SV circumference embedded in pancreas, and pancreas-SV contact area are valuable preoperative indicators for predicting the risk of splenic vessel preservation failure. The nomogram model established on the basis of these indicators has certain clinical practical value. As a representative of the integration of medicine and engineering, three-dimensional medical reconstruction can provide an important reference for the selection of surgical procedures in laparoscopic spleen-preserving distal pancreatectomy.