Author: James M. Brindle, Gene A. Cardarelli, Alexis Leuci, Mark J. Rivard 👨🔬
Affiliation: Warren Alpert Medical School of Brown University 🌍
Purpose: Leskell GammaPlan (LGP) is equipped with the Collapsed Cone Convolution (CCC) and TMR10 calculation algorithms. Use of the CCC algorithm accounts for tissue heterogeneities not considered in TMR10 calculations. This can reduce uncertainties in predicting the delivered physical dose distribution. We investigate the use of the CCC algorithm for planning Gamma Knife pituitary tumor treatments and complete an analysis or positional offset in the x/y/z directions.
Methods: The CCC algorithm was commissioned using an electron density phantom, homogenous phantom and an anthropomorphic head phantom. Positional offsets were investigated by shifting target isocenters along the x,y,z planes and comparing TMR10 and CCC calculations for each treatment plan. The application of this algorithm for pituitary treatments was analyzed using a cohort of 12 patients from 2020-2024. All treatment plans were recalculated using the CCC algorithm. Dosimetric differences to the target, optic nerves, and optic chiasm were evaluated.
Results: For pituitary tumors, TMR10 calculations showed a minimum dose to the target with a smaller deviation from the prescription dose than the CCC calculations. The minimum and maximum dose to the optic nerves was lower with the CCC calculation, but dose to the optic chiasm was higher. The positional analysis of x,y,z directions showed a smaller deviation between dose calculations in central locations.
Conclusion: The greatest variation between the CCC and TMR10 calculations occurs near the peripheries of the skull, closer to heterogeneities. For pituitary patients, the tumor volume had cold spots with a greater deviation from the prescription dose than observed from TMR10 calculations. The patient also received a lower dose to the optic nerves but a higher dose to the optic chiasm than identified by TMR10 calculations. Clinical planning goals may need to be modified prior to implementation of this algorithm to reflect the values identified with the CCC algorithm.