2384 - Evaluating the Impact of Workflow Modifications on Plan Quality in Gamma Knife Radiosurgery

S. Oh, K. Renick, A. Heermann, and T. Kim; Washington University School of Medicine in St. Louis, Department of Radiation Oncology, St. Louis, MO

Purpose/Objective(s): To evaluate the impact of simulation/planning workflow in Gamma Knife Radiosurgery (GKRS) by analyzing plan quality parameters and Gaussian Weighted Conformity Index (GWCI). Materials/

Methods: We have implemented an updated GKRS workflow in our department to mitigate the transmission of COVID-19 since the onset of the pandemic. This involved integrating the utilization of our departmental MR simulator and CBCT on GK ICON into our procedures. In the updated workflow (W2), a departure from the previous practice (W1) was made to enhance safety measures. Instead of conducting imaging and planning on the day of treatment, they are now performed in advance, prior to the actual treatment. For analysis, we gathered plan data from 20 cases of acoustic neuroma and 25 cases of meningioma that underwent GK treatment prior to September 2020 (W1). Additionally, an equal number of cases after March 2021 (W2) were selected and exported for further assessment. We conducted a comparison based on reported plan parameters from the GK TPS. Additionally, we utilized the Gaussian Weighted Conformity Index (GWCI), which assesses the level of over- and under-coverage at each point on the target contour.
Results: For meningioma, W1 demonstrated a higher mean value of selectivity. On the other hand, for Acoustic Neuroma, W2 exhibited better selectivity and coverage. Two combinations of over(negative)- and under(positive)-coverage, (-1.5, 0.5) and (-1.5, 1.0), were considered for GWCI calculation based on the mean values of over- and under-coverage in the considered cases. The GWCI results showed a similar trend. W1 had slightly better GWCI mean values for meningioma, while W2 had slightly better GWCI for acoustic neuroma. However, there was no statistically significant difference observed across all the parameters.
Conclusion: We evaluated the effects of modifying the workflow on the plan quality for cases of meningioma and acoustic neuroma. Our findings indicate that the plan quality remains consistent, regardless of any modifications to the workflow. It is interesting to observe that even though the mean volume of meningioma for W1 is larger, the number of shots required was less than W2. Further analysis is necessary to determine the underlying cause. Table 1. The analyzed parameters mean and standard deviation were calculated for the workflows. There was no statistically significant difference between two workflows for the considered cases.

	Meningioma			Acoustic Neuroma
	W1	W2	p-value	W1	W2	p-value
Volume [cc]	2.90±2.52	2.75±1.75	0.814	2.03±2.23	1.98±1.59	0.934
# of shots	16.7±8.5	20.2±10.4	0.188	14.9±9.8	14.7±8.5	0.942
Selectivity	0.719±0.080	0.688±0.094	0.198	0.713±0.146	0.768±0.069	0.130
Coverage	0.995±0.004	0.995±0.012	0.992	0.995±0.005	0.997±0.002	0.155
GWCI(-1.5, 0.5)	0.919±0.030	0.917±0.039	0.807	0.950±0.022	0.957±0.015	0.246
GWCI(-1.5, 1.0)	0.933±0.027	0.926±0.035	0.452	0.959±0.019	0.964±0.014	0.269