A Nomogram to Predict Major Postoperative Complications After Cytoreductive Surgery and HIPEC Based on Pre and Peroperative Criteria: Which Patient Require Intensive Monitoring?

This is a retrospective study from a single center of a prospectively maintained database of patients who underwent cytoreductive surgery with hyperthermic intraperitoneal chemotherapy for peritoneal carcinomatosis of various primitives. The postoperative monitoring was carried out by a daily clinical examination and laboratory tests were drawn every two days during the first postoperative week. All patients had a CT-scan when a postoperative complication was suspected or at postoperative day 7 before hospital discharge. All patients were followed up to 90 postoperative days. Collected variables included age, gender and body mass index (BMI), smoking, diabetes mellitus, American Society of Anesthesiology classification of physical health (ASA), nutritional status (albuminemia ≤ 35mg/L), preoperative hemoglobin level, tumor origin, synchronous/metachronous peritoneal carcinomatosis, preoperative chemotherapy and/or biotherapy, presence of ascites at surgery, Peritoneal Cancer Index, type and number of resected organs, type and number of anastomoses, and number of packed red blood cells transfused. Wedge resection, cauterization, or superficial excisions of the bowel wall were not recorded. Minor hepatectomy concerned wedge resections or resection of less than 3 hepatic. Postoperative complications, recorded up to postoperative day 90, were graded according to Clavien-Dindo classification.

The endpoint of this study was the occurrence of a major postoperative complication that was defined as a Clavien-Dindo ≥ 3 grade complication.

To build the multivariate model the investigators used a backward stepwise approach with removal of individual variables one by one starting with the highest p-value until all variables had a p-value lower than 0.05. Results were presented as Odds Ratio and 95% confidence interval. Based on the multivariate model with identified risk factors, a nomogram was constructed to predict major postoperative complication probability after cytoreductive surgery with hyperthermic intraperitoneal chemotherapy. The coefficients derived from the multivariate analysis were used as weights. The nomogram assessed the probability of major postoperative complication after cytoreductive surgery with hyperthermic intraperitoneal chemotherapy by adding up the scores identified on the "Points" scale for each factor. The total score projected from the "Total points" axis to the "Risk of major postoperative complication" axis, indicated the probability of major postoperative complication occurrence after cytoreductive surgery with hyperthermic intraperitoneal chemotherapy. The calibration curve of the prediction model was plotted to compare the predicted and actual probability of major postoperative complication.

The concordance index (C-index) was reported as a measure of internal validation using both 10-fold cross-validations repeated for 20 times, and bootstrap validation of 200 resamples. The area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate the performance of the model. The AUC is an effective method for quantifying the discriminatory capacity of a nomogram to correctly predict patients at low or high risk of major postoperative complication.

The Youden Index, which maximizes the sum of sensitivity and specificity, was determined according to the ROC curve. At this threshold, the sensitivity, the specificity, the predictive positive value (PPV), the negative predictive value (NPV) and the overall diagnostic accuracy (i.e., probability for a patient to be classified correctly by the nomogram) were calculated. A calibration curve was used to show consistency between the observation frequency and prediction probability.