Conformity Distance Measure for Stereotactic Radiosurgery of Brain Metastases
Presentations
PO-GePV-T-400 (Sunday, 7/10/2022) [Eastern Time (GMT-4)]
ePoster Forums
Purpose: Conformity and gradient index are commonly used to evaluate the plan quality of stereotactic radiosurgery (SRS). Strong volume dependency of the metric hinder fair evaluation when target volume variation is > 10 thousand folds (0.001cc to 10cc) in Gamma Knife (GK) SRS plans for brain metastases. As a new quality measure, dosimetric margin of the GK plans were investigated using volume independent quality measures.
Methods: 58 brain metastasis treated in GK ICON were optimized to achieve CI < 2.0 and to lower GI as much as possible by experienced neuro-surgeons. Dose prescription was 24Gy, 18Gy and 15Gy for target maximum dimension of <2cm, <3cm and >3cm. New quality metric of conformity distance measure (CDM or dosimetric margin) and gradient distance measure (GDM) were used to quantify the average distance of prescription isodose volume (PIV) to target volume (TV) and the distance of ideal 50%PIV to real 50%PIV of the plan. Hypothetical CDM and GDM were calculated to quantify the distance of TV to 18Gy or 15Gy.
Results: The average CI and GI were 2.03 (1.23 to 3.7) and 4.53 (2.5-10.7) for the average TV of 1.1cc (0.01-15.8cc) in this study. Strong volume dependency was found in conventional CI and GI such that CI were > 2.0 and GI >3.0 for TV of < 1cc. Without strong volume dependency, CDM from TV to the isodose of 24Gy, 18Gy and 15Gy was 0.8 mm (0.33-1.8mm), 1.7 mm (0.6-3.5mm) and 2.4 mm (1.0-4.6mm), respectively. Although planning margin is not used in GK plans, optimization resulted in small CDM of 0.8-2.4 mm. Equivalently representing plans with prescription of 24Gy as those with 18Gy and corresponding CDM, the distribution of CDM becomes even more volume independent.
Conclusion: Conformity distance measure is useful tool to measure the quality of SRS plans for brain metastasis.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email