Poster - Thur Eve - 21: ROC analysis in patient specific quality assurance.

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiotherapy Quality Assurance

Background:

• Institutions commonly use patient-specific measurements for Intensity-Modulated Radiation Therapy (IMRT) and Volumetric Modulated Arc Therapy (VMAT) quality assurance (QA).
• Receiver Operator Curves (ROC) are utilized in diagnostic imaging to evaluate the performance of tests.
• The application of ROC methodology to patient-specific IMRT QA is explored.

Purpose of the Study:

• To investigate the utility and value of Receiver Operator Curves (ROC) methodology for patient-specific IMRT QA.
• To determine if ROC analysis can provide optimal cut-off values for gamma analysis in IMRT QA.

Main Methods:

• Analyzed beam fluences from 34 prostate IMRT patients using gamma analysis.
• Introduced perturbations to Multi-Leaf Collimator (MLC) positions in half of the patients.
• Calculated true positive and false negative distributions assuming unperturbed fluences as positive measurements.

Main Results:

• The Area Under the Curve (AUC) for ROC analysis increased by approximately 10% for high-performance beam delivery systems.
• At a 3%/3mm gamma-DTA (dose-to-agreement) criterion, optimal cut-off values were largely independent of MLC performance.
• At a stricter 2%/2mm gamma-DTA criterion, test sensitivity and specificity became more dependent on MLC performance.

Conclusions:

• Selecting optimal cut-off values is crucial for maximizing sensitivity and specificity in pass-fail tests like gamma-DTA maps.
• ROC methodology offers a prescriptive approach to determine ideal cut-off values for gamma analysis.
• ROC analysis can effectively assess improvements in sensitivity and specificity associated with higher-performing beam delivery systems.