Clinical Scorecard: A Radiomics Model Utilizing CT Imaging to Forecast Pain Relief Following Radiotherapy in Bone Metastasis Patients: Findings from a Dual-Center Investigation

At a Glance

Key Highlights

• Developed and validated a CT radiomics model using 7 selected features to predict pain relief after radiotherapy in bone metastasis patients.
• Compared 11 machine learning algorithms; k-nearest neighbors (KNN) model showed best predictive performance with AUC > 0.8 across training, internal validation, and external test sets.
• Model demonstrated potential clinical utility to identify patients likely to benefit from radiotherapy-induced pain relief, aiding personalized treatment planning.

Guideline-Based Recommendations

Diagnosis

• Use CT imaging with bone-window settings to delineate tumor regions of interest for radiomic feature extraction.
• Assess pain response following International Consensus on Endpoints for Palliative Radiotherapy in Bone Metastases criteria.

Management

• Administer palliative radiotherapy with standard fractionation regimens (e.g., 40 Gy in 20 fractions or 30 Gy in 10 fractions) for bone metastases.
• Consider radiomics-based predictive models to guide individualized radiotherapy decisions and optimize pain relief outcomes.

Monitoring & Follow-up

• Evaluate pain response post-radiotherapy using standardized criteria to classify complete/partial response versus stable/progressive disease.
• Monitor clinical variables including primary tumor type, bone destruction pattern, and metastatic site alongside imaging features.

Risks

• Recognize that approximately 40% of patients may not achieve satisfactory pain relief despite radiotherapy.
• Be aware of heterogeneity in pain response rates and limitations of current clinical predictors without radiomics integration.

Patient & Prescribing Data

134 patients with bone metastases treated with palliative radiotherapy from two centers

Pain relief achieved in 53 patients (complete or partial response); KNN radiomics model predicted pain relief with AUC ~0.82, supporting its use for treatment stratification.

Clinical Best Practices

• Incorporate CT radiomics analysis into routine assessment to enhance prediction of radiotherapy-induced pain relief in bone metastasis patients.
• Use standardized pain response criteria (ICPRE) for consistent evaluation of treatment outcomes.
• Apply machine learning models validated with external datasets to ensure generalizability and clinical applicability.

References

• International Consensus on Endpoints for Palliative Radiotherapy in Bone Metastases
• Meta-analysis of pain response rates in bone metastases radiotherapy
• Radiomics in predicting radiotherapy response