CR-Possum-Can It Be Used to Predict Morbidity? A Single-Centre Retrospective Study

Area of Science:

• Oncology
• Surgical Outcomes
• Medical Statistics

Background:

• Preoperative prediction of morbidity is crucial for optimizing surgical outcomes in colorectal cancer (CRC).
• The Colorectal Possum (CR-POSSUM) scoring system is established for mortality risk assessment.
• Validating and adapting existing scoring systems for morbidity prediction in CRC is essential.

Purpose of the Study:

• To develop a dedicated equation for predicting operative morbidity in colorectal cancer surgery.
• To validate the predictive accuracy of the CR-POSSUM scoring system for actual complications.
• To assess the utility of CR-POSSUM beyond mortality prediction.

Main Methods:

• Retrospective analysis of 322 patients undergoing colorectal cancer surgery (2004-2016).
• Collected CR-POSSUM mortality and morbidity risk factors.
• Logistic regression analysis used to derive a predictive equation for morbidity.
• Statistical analysis included Hosmer-Lemeshow test, Nagelkerke R², and ROC curve analysis.

Main Results:

• 103 out of 322 patients (32%) experienced complications; 10 (3%) died.
• The derived regression equation achieved 70% overall correct classification (P < 0.01).
• Positive predictive value was 60%, and negative predictive value was 73%.
• Model discrimination (Area under ROC) was 71.6%, with good model fit (Hosmer-Lemeshow P = 0.829).

Conclusions:

• The developed regression equation effectively predicts operative morbidity in colorectal cancer surgery.
• The CR-POSSUM scoring system, originally for mortality, can be extrapolated to predict surgical morbidity.
• This predictive capability can aid in optimizing surgical outcomes and patient management.