Nomogram for predicting prolonged postoperative ileus after laparoscopic low anterior resection for rectal cancer

Personalized risk prediction for prolonged ileus after minimally invasive colorectal cancer surgery: in-depth risk factor analysis and model development

PURPOSE

Despite the increasing preference for minimally invasive surgery for colorectal cancer (CRC), the incidence of prolonged postoperative ileus (PPOI) remains high. Thus, this study aimed to identify risk factors for PPOI in patients with CRC who underwent minimally invasive surgery (MICRS) and to develop a practical nomogram for predicting individual PPOI risk.

METHODS

A consecutive series of 2368 patients who underwent MICRS between 2013 and 2023 at two tertiary academic centers were retrospectively studied. Using the data from 1895 patients in the training cohort, a multivariable logistic regression model was employed to select significant variables for the construction of a best-fit nomogram. The nomogram was internally and externally validated.

RESULTS

PPOI occurred in 9.5% of patients. Six independent risk factors were identified to construct a nomogram: advanced age (OR 1.055, P = 0.002), male sex (OR 2.914, P = 0.011), age-adjusted Charlson comorbidity index ≥ 6 (OR 2.643, P = 0.025), preoperative sarcopenia (OR 0.857, P = 0.02), preoperative prognostic nutritional index (OR 2.206, P = 0.047), and intraoperative fluid overload (OR 2.227, P = 0.045). The AUCs of the model for predicting PPOI in the training and external validation cohorts were 0.887 and 0.838, respectively. The calibration curves demonstrated excellent consistency between the nomogram-predicted and observed probabilities in both cohorts. Individuals with a total nomogram score of < 197 or ≥ 197 were considered to be at low or high risk for PPOI, respectively.

CONCLUSIONS

The integrated nomogram we developed could provide personalized risk prediction of PPOI after MICRS. This quantification enables surgeons to implement personalized prevention strategies, thereby improving patient outcomes.

Development and validation of a nomogram to predict the risk factors of major complications after radical rectal cancer surgery

Purpose

The aim of this study was to establish a validated nomogram to predict risk factors for major post-operative complications in patients with rectal cancer (RC) by analyzing the factors contributing to major post-operative complications in RC patients.

Methods

We retrospectively collected baseline and surgical information on patients who underwent RC surgery between December 2012 and December 2022 at a single-center teaching hospital. The entire cohort was randomly divided into two subsets (60% of the data for development, 40% for validation). Independent risk factors for major post-operative complications were identified using multivariate logistic regression analyses, and predictive models were developed. Area under the curve (AUC) was calculated using receiver operating characteristic curve (ROC) to assess predictive probability, calibration curves were plotted to compare the predicted probability of the nomogram with the actual probability, and the clinical efficacy of the nomogram was assessed using decision curve analysis (DCA).

Results

Our study included 3151 patients who underwent radical surgery for RC, including 1892 in the development set and 1259 in the validation set. Forty (2.1%) patients in the development set and 26 (2.1%) patients in the validation set experienced major post-operative complications. Through multivariate logistic regression analysis, age (p<0.01, OR=1.044, 95% CI=1.016-1.074), pre-operative albumin (p<0.01, OR=0.913, 95% CI=0.866-0.964), and open surgery (p<0.01, OR=2.461, 95% CI=1.284-4.761) were identified as independent risk factors for major post-operative complications in RC, and a nomogram prediction model was established. The AUC of the ROC plot for the development set was 0.7161 (95% Cl=0.6397-0.7924), and the AUC of the ROC plot for the validation set was 0.7191 (95% CI=0.6182-0.8199). The predicted probabilities in the calibration curves were highly consistent with the actual probabilities, which indicated that the prediction model had good predictive ability. The DCA also confirmed the good clinical performance of the nomogram.

Conclusion

In this study, a validated nomogram containing three predictors was created to identify risk factors for major complications after radical RC surgery. Due to its accuracy and convenience, it could contribute to personalized management of patients in the perioperative period.

Artificial intelligence algorithms for predicting post-operative ileus after laparoscopic surgery

Objective

By constructing a predictive model using machine learning and deep learning technologies, we aim to understand the risk factors for postoperative intestinal obstruction in laparoscopic colorectal cancer patients, and establish an effective artificial intelligence-based predictive model to guide individualized prevention and treatment, thus improving patient outcomes.

Methods

We constructed a model of the artificial intelligence algorithm in Python. Subjects were randomly assigned to either a training set for variable identification and model construction, or a test set for testing model performance, at a ratio of 7:3. The model was trained with ten algorithms. We used the AUC values of the ROC curves, as well as accuracy, precision, recall rate and F1 scores.

Results

The results of feature engineering composited with the GBDT algorithm showed that opioid use, anesthesia duration, and body weight were the top three factors in the development of POI. We used ten machine learning and deep learning algorithms to validate the model, and the results were as follows: the three algorithms with best accuracy were XGB (0.807), Decision Tree (0.807) and Neural DecisionTree (0.807); the two algorithms with best precision were XGB (0.500) and Decision Tree (0.500); the two algorithms with best recall rate were adab (0.243) and Decision Tree (0.135); the two algorithms with highest F1 score were adab (0.290) and Decision Tree (0.213); and the three algorithms with best AUC were Gradient Boosting (0.678), XGB (0.638) and LinearSVC (0.633).

Conclusion

This study shows that XGB and Decision Tree are the two best algorithms for predicting the risk of developing ileus after laparoscopic colon cancer surgery. It provides new insight and approaches to the field of postoperative intestinal obstruction in colorectal cancer through the application of machine learning techniques, thereby improving our understanding of the disease and offering strong support for clinical decision-making.