Predicting Hospital Admissions for Genitourinary Toxicity After Prostate Radiotherapy
Overview
This study developed and validated a predictive model for hospital admissions due to genitourinary toxicity following external beam radiotherapy (EBRT) in localized prostate cancer patients. Key clinical factors including diabetes, smoking, and bladder outlet obstruction (BOO) status were identified as strong independent predictors of toxicity-related hospitalization.
Background
Prostate cancer is the second most common malignancy in men globally, with increasing numbers of long-term survivors. Radiotherapy and surgery offer similar local control but differ in toxicity profiles, particularly genitourinary side effects. Existing predictive models focus mainly on dosimetric parameters, while clinical factors influencing genitourinary toxicity remain underexplored. This study aimed to use pre-treatment clinical variables to forecast severe genitourinary toxicity requiring hospital admission after EBRT.
Data Highlights
Predictor
Hazard Ratio (HR)
95% Confidence Interval
p-value
Diabetes
1.35
1.13–1.60
<0.001
Smoking
1.78
1.40–2.12
<0.001
BOO without TURP
7.49
6.18–9.08
<0.001
BOO with TURP
4.96
4.10–5.99
<0.001
Baseline stress urinary incontinence
3.95
3.28–4.75
<0.001 (excluded from final model due to multicollinearity)
Key Findings
Bladder outlet obstruction without prior TURP was the strongest predictor of genitourinary toxicity-related hospital admission (HR 7.49).
Diabetes and smoking independently increased the risk of hospitalisation for genitourinary toxicity post-EBRT.
Patients with BOO who had undergone TURP also had significantly elevated risk (HR 4.96).
Baseline stress urinary incontinence was associated with increased risk but excluded from the final model due to multicollinearity.
The model demonstrated good discrimination and calibration using internal validation techniques.
Decision curve analysis supported the clinical utility of the predictive model for guiding patient management.
Clinical Implications
Clinicians should consider baseline clinical factors such as diabetes, smoking status, and bladder outlet obstruction when assessing risk for genitourinary toxicity after prostate radiotherapy. Identifying high-risk patients may guide closer monitoring and tailored interventions to mitigate severe toxicity requiring hospital admission. The predictive model and nomogram developed can aid in personalized risk stratification and shared decision-making.
Conclusion
This study presents a validated clinical model that effectively predicts hospital admissions for genitourinary toxicity following EBRT in localized prostate cancer patients. Incorporating key clinical variables enhances risk assessment beyond dosimetric factors alone.
References
SA-PCCOC Registry and Related Studies
Steyerberg et al. 2010 -- ABCD Approach for Model Validation
TRIPOD Statement 2015 -- Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis
