TL;DR
This study developed and validated a predictive model for stoma prolapse in colorectal cancer patients, identifying age, elevated intra-abdominal pressure, stoma type, and hypoproteinemia as key risk factors. The model, presented as a nomogram, offers clinicians a quantitative tool for early identification of high-risk patients, enabling targeted preventive measures and improving patient outcomes.
What Problem Does This Paper Address?
Colorectal cancer is a significant global health concern, and while ostomy surgery is a crucial part of treatment, stoma prolapse is a common and distressing complication. Currently, it's challenging for clinicians to identify patients at high risk of developing stoma prolapse during the perioperative and early postoperative periods. This lack of early identification hinders the implementation of timely preventive strategies, leading to potential patient discomfort, reduced quality of life, and increased healthcare resource utilization. Existing predictive tools are often based on limited clinical indicators and lack comprehensive validation, making them less effective in real-world clinical practice.
Methods
The researchers employed a retrospective case-control study design. They collected data from 270 colorectal cancer patients who underwent stoma surgery. This cohort was randomly split into a training set (70%) and a validation set (30%).
The core of their methodology involved:
- Univariate Logistic Regression: Initially, various clinical characteristics were analyzed to identify potential risk factors for stoma prolapse. Factors with a p-value less than 0.2 were considered for further analysis.
- Multifactorial Logistic Regression: The identified potential risk factors were then subjected to a multifactorial analysis to determine independent predictors of stoma prolapse.
- Nomogram Construction: Based on the independent risk factors identified through multifactorial analysis, a nomogram was constructed. A nomogram is a graphical tool that visually represents a statistical prediction model, allowing for easy calculation of an individual's risk score.
- Internal Validation: The nomogram's predictive performance was internally validated using the Bootstrap method (1000 repetitions) in R software. This process assesses the model's calibration (how well predicted probabilities match actual outcomes) and discrimination (how well it distinguishes between those who will and will not develop prolapse) using ROC curves.
- Decision Curve Analysis (DCA): DCA was used to evaluate the clinical utility of the nomogram by comparing the net benefit of using the model versus treating all patients or no patients.
Datasets
The study utilized a dataset of 270 colorectal cancer patients who underwent stoma surgery at the Third Hospital of Hebei Medical University between January 2021 and June 2024.
- Training Set: 189 patients
- Validation Set: 81 patients
This dataset is a newly collected cohort for this specific study and is not presented as a publicly released benchmark.
Key Insights and Ideas
- Multifaceted Risk: Stoma prolapse isn't due to a single factor but a complex interplay of patient characteristics and physiological states. The model's success lies in integrating these diverse factors.
- Aging as a Silent Aggravator: The study highlights that aging significantly impacts tissue elasticity and repair mechanisms around the stoma, making older individuals more vulnerable.
- The Unseen Pressure: Elevated intra-abdominal pressure, often stemming from common conditions like chronic cough or constipation, creates mechanical stress that can lead to stoma prolapse. The study emphasizes the importance of quantifying and managing this pressure.
- Nutritional Status Matters: Hypoproteinemia, a state of low protein levels, was identified as a crucial independent risk factor. This suggests that compromised tissue integrity due to poor nutrition directly contributes to prolapse.
- Stoma Type as a Predictor: Different types of stomas have inherent structural differences that influence their susceptibility to prolapse, a factor the model effectively incorporates.
- Clinical Utility of Nomograms: The nomogram provides a practical, visual, and quantitative tool that translates complex statistical findings into actionable clinical insights, bridging the gap between research and practice.
Key Results
The study successfully identified four independent risk factors for stoma prolapse within six months post-surgery:
- Age
- Elevated Intra-abdominal Pressure
- Type of Stoma
- Hypoproteinemia
The developed nomogram demonstrated strong predictive performance:
- Training Set ROC AUC: 0.852 (95% CI = 0.797–0.907)
- Validation Set ROC AUC: 0.830 (95% CI = 0.734–0.926)
These AUC values indicate a good ability of the model to discriminate between patients at high and low risk of stoma prolapse. The decision curve analysis suggested that the model provides more net benefit than "all intervene" or "none intervene" strategies when the threshold probability of an individual is >0.05, highlighting its clinical utility.
Limitations and Future Work
The authors acknowledge several limitations:
- Retrospective Design: The retrospective nature of the study may introduce biases, and future prospective cohort studies are recommended to mitigate this.
- Data Collection Gaps: Incomplete recording of some clinical information, particularly in primary hospitals, may have affected the assessment of certain risk factors.
- Sample Size: While 270 patients were included, the authors suggest that a larger sample size could improve the model's generalizability.
- Geographical and Racial Diversity: The study was conducted at a single institution, and future research should aim to include patients from diverse geographical regions and racial backgrounds to enhance the model's applicability.
- Measurement Errors: Potential biases in measuring intra-abdominal pressure due to variations in tools and procedures are acknowledged.
- Confounding Factors: Despite efforts to control for bias, the retrospective design makes it difficult to completely exclude the influence of confounding factors like variations in patient care and postoperative rehabilitation instructions.
Future work suggested includes:
- Expanding the sample size and geographical diversity.
- Including patients with a wider range of disease characteristics and treatment options.
- Developing specialized clinical software plug-ins to seamlessly integrate the nomogram into electronic medical record systems for easier clinical application.
- Conducting multicenter, large-sample prospective cohort studies.
Summary updated