Assessment of Clinical Plan Quality Metrics for IROC Houston SRS Phantom Performance Evaluation

Derek Garcia (Advisor: Mallory Glenn, PhD)

The gamma index is a widely used tool for radiotherapy quality assurance (QA). Despite this, it provides limited insight into the clinically relevant aspects of dose distributions including target coverage, conformity, and dose falloff. This study is intended to evaluate the feasibility of implementing clinical plan quality metrics into the IROC Houston SRS phantom evaluation framework in order to provide a more comprehensive phantom assessment. Additionally, this work aims to look into the relationship of these metrics with gamma-based QA and the impact of film digitization on these metrics.

A dataset of 110 brain SRS phantom irradiations from various institutions were analyzed using both treatment planning system (TPS) and film-measured dose distributions. Clinical quality metrics including RTOG conformity index (CI), ICRU homogeneity index (HI), Quality of Coverage (QoC), Paddick’s conformity index (PCI), gradient index (GI), and R50, were calculated as two-dimensional analogs based on film planes. Each metric was compared against thermoluminescent dosimeter (TLD) measurements and gamma pass rates to investigate possible relationships with the delivered dose agreement. A subset of cases performed between March 2025-February 2026 were used to form paired comparisons between CCD camera-based and flatbed scanner-based digitization workflows using statistical analysis and distributional comparisons.

Clinical plan quality metrics that were derived from film demonstrated stronger and more consistent correlation with TLD measurements than gamma pass rate, which suggests improved sensitivity to delivered dose characteristics. In contrast, there were minimal correlations between gamma pass rate and clinical metrics, indicating that these approaches capture different aspects of plan quality. Comparisons between digitization methods showed strong overall agreement in gamma pass rates, although with a small but statistically significant systematic bias (mean difference (0.76%). Metric-dependent differences showed a similar trend, with coverage and conformity metrics showing high robustness while gradient-based metrics showed increased sensitivity to digitization differences.

Clinical quality metrics provide complementary information to gamma analysis and could offer improved sensitivity to aspects of plan quality that are relevant to SRS treatments. However, one must be aware that the interpretation of these metrics is influenced by both metric definition and measurement workflow.

Advisory Committee:

• Mallory Glenn, PhD, Chair
• Stephen Kry, PhD
• Ruitao Lin, PhD
• Dennis Mackin, PhD
• Tucker Netherton, PhD

Join via Zoom (Please contact Mr. Derek for his Zoom meeting info.)