Clinical Scorecard: Hybrid Artificial Intelligence for Personalized Dose Recommendations and Outcome Forecasting in Gamma Knife Radiosurgery for Recurrent Glioblastoma: A Proof-of-Concept Study

At a Glance

Key Highlights

• Strong internal validation performance of local control survival model (C-index 0.80; IBS 0.14).
• CGH-AI provides individualized prescription dose recommendations.
• Model predicts local control probabilities and expected local control duration.
• Addresses the challenge of dose selection in previously irradiated brain.
• Supports personalized decision-making and follow-up planning.
• Sensitivity analyses confirm robustness of model performance.

Guideline-Based Recommendations

Diagnosis

• Utilize imaging and clinical assessment to confirm recurrent GBM, including MRI and CT scans.

Management

Monitoring & Follow-up

Risks

Patient & Prescribing Data

Patients with recurrent GBM treated with GKRS from 2014 to 2024.

Incorporates clinical variables, tumor characteristics, and dosimetric parameters for dose selection.

Clinical Best Practices

• Implement data-driven approaches for individualized GKRS dose recommendations.
• Utilize AI-driven models to enhance decision support in treatment planning.
• Regularly assess and validate treatment outcomes to refine predictive models and ensure accuracy.

Related Resources & Content

• University of Pittsburgh Medical Center Study