Dosimetric Analysis Using Automated Tools of Trigeminal Neuralgia Treatment with Stereotactic Radiosurgery.

PURPOSE/OBJECTIVE(S): Symptoms from trigeminal neuralgia (TGN) can be significantly debilitating. For treatment of TGN with stereotactic radiosurgery (SRS), there is some data that suggests dose and location of treatment along the trigeminal nerve (TN) pathway may be associated with improved outcomes; however, recurrence rates are widely variable. Limited data exists to evaluate dosimetric parameters of SRS for TGN in relation to patient outcome. This study evaluated the hypothesis that specific characteristics of high dose volume coverage of the TN was associated with outcomes in the treatment of TGN with SRS.

MATERIALS/METHODS: Patients who underwent a non-invasive stereotactic radiosurgery instrument for the treatment of TGN at a single institution from 2008-2020 were retrospectively reviewed. Patients were selected for treatment with SRS after consensus from neurosurgery and radiation oncology teams. Patient must have undergone MR FIESTA sequence imaging for treatment planning and had at least 3mo TGN follow up after SRS. Using commercially available software, an automated script was developed to extract detailed characteristics of isodose line structures (IDL) created from the delivered radiation treatment plan. Variables evaluated included voxel volume, mean/max of the HU signal, kurtosis/skewness (a measure of non-spheroidicity) of the IDL, and HU integral of the 140-195% IDL structures, all of which could serve as possible surrogates to the degree and shape of TN pathway dosimetric coverage. Patient reported pain was graded prior and post treatment using Barrow Neurologic Institute pain scale (BNI). Post SRS BNI score I-III was considered significant symptom improvement while BNI score IV-V was considered no benefit. Univariate and multivariable analysis was performed using the cox proportional hazard model.

RESULTS: Of the 73 patients included, 52 (71%) were female and 21 (29%) were male, and median age at time of treatment was 73 years (60-79 years). Pretreatment BNI scores were BNI III in 9 (12%) patients, BNI IV in 42 (58%) patients, and BNI V in 22 (30%) patients. After SRS, 62 (85%) patients experienced significant symptom relief (BNI I-III). Median time to symptom relief was 2mo (IQR 1-4mo) with a median duration of 12mo (2-28mo). Median time to treatment failure was 12mo (IQR 2-28mo). On multivariable analysis evaluating IDL characteristics associated with achieving symptom relief, 195% IDL kurtosis was associated with increased likelihood of achieving symptom relief (HR 1.94, p<.001), while and 180% IDL voxel size was associated with reduced likelihood of achieving symptom relief (HR 0.99, p<.001).

CONCLUSION: In the treatment of TGN with SRS, these initial results suggests that high dose IDL structure characteristics, specifically the size and shape features of the dose distributions, may be associated with likelihood of achieving symptom relief. Future directions include analysis of a richer set of dosimetric features in a larger cohort to further refine significant dosimetric parameters for treatment planning.