Trigeminal nerve integrated dose and pain outcome after gamma knife radiosurgery for trigeminal neuralgia

Abstract

Background

Gamma knife radiosurgery (GKRS) is an established treatment for trigeminal neuralgia. Identifying factors that influence outcome will help improve patients’ results.

Methods

We conducted a retrospective review of all patients treated with GKRS for trigeminal neuralgia at our institution from 2005 to 2010. Patients’ clinical features and treatment details were reviewed. Analysis was performed to identify predictors of response and recurrence.

Results

A hundred and forty five patients were treated. Mean follow up period was 24 months. At last follow up, 48 patients (33%) were pain free with no medications, and 48 patients (33%) were pain free maintained on medications. Twenty-eight patients (19%) had pain after the treatment but had significant reduction in their pain severity. Twenty-one patients (15%) did not have any significant pain reduction. Forty-four patients (30%) developed facial numbness. Recurrence occurred in 51 patients (35%). Post-treatment numbness was a predictor of good treatment response (OR 2.720, CI 1.193-6.200, p 0.0173). Higher integrated dose was a predictor of poor pain response to radiosurgery (OR 0.729, CI 0.566-0.940, p 0.0146). At an integrated dose value of 5.3 mJ or less, there was more than 50% chance of pain free outcome. Longer pain duration prior to treatment was the only independent predictor of increased recurrence risk (HR 1.038, 95%CI 1.001-1.075; p=0.0412).

Conclusions

Radiosurgery is an effective treatment modality for trigeminal neuralgia. Post treatment numbness is associated with good treatment response and higher integrated dose predicts poor outcome after radiosurgery for trigeminal neuralgia.

INTRODUCTION

Radiosurgery is one of several treatment modalities with proven efficacy in the management of trigeminal neuralgia [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. The relative lack of invasiveness makes it an attractive treatment modality in the elderly population and patients with multiple medical co-morbidities. Radiosurgery is a non-invasive tool, together with other ablative techniques such as glycerol rhizotomy, radiofrequency lesioning and balloon compression. These non-invasive modalities of treatment share a significant risk of recurrence of pain [2, 3, 4, 16, 17]. A specific disadvantage of radiosurgery is the latent period between the treatment and the clinical effect. Despite the large number of reports of radiosurgery for trigeminal neuralgia, there is a significant variation in patient selection, treatment planning techniques, and definition of predictors of good response. With better understanding of the impact of different treatment-variables on outcome, treatment can be altered to achieve better pain control.

It has been difficult to find an objective method of assessment of the effect of Gamma Knife Radiosurgery (GKRS) on the trigeminal nerve, although a recent study from our group highlights a promising role for diffusion tensor imaging in this process [18]. Clinically, the dose in GKRS is described based on the maximum dose delivered to the nerve, which is known to correlate with the clinical outcome [6, 19]. Some clinical reports have indicated that outcome can be also influenced by the difference in the dose delivery rate as measured by the integrated dose despite similar maximum dose [20, 21]. The integrated dose is the total energy deposited per unit volume of tissue. It is calculated as a summation of dose within the volume. It increases when the rate of radiation delivery is slower even if the prescribed dose is unchanged. Using plugs slows the dose delivery and increases the integrated dose.

In this paper, we are studying the hypothesis that differences in the integrated dose will influence pain outcome despite using similar maximum prescription dose.

METHODS

Patient Population

We conducted a retrospective review of all patients with trigeminal neuralgia who were treated with gamma knife radiosurgery at our institution from the time the unit was established in 2005 until 2010. All included patients had to have at least 6 months of follow up. One hundred and forty-five patients met our inclusion criteria. Eighty-five patients (59%) were female. Mean age of the patients was 67 years (range 30-89 years). Seventy-three patients (50%) had right-sided pain and 69 patients (48%) had left-sided pain. Three patients (2%) had bilateral pain. The most common pain distribution was V3 and V2-V3 (each occurred in 27% of the patients). Mean duration of pain by the time the patients were treated at our centre was 9 years (range 1-50 years). Sixteen patients (11%) described a component of atypical pain. Sixteen patients (11%) had trigeminal neuralgia secondary to multiple sclerosis (MS). Six patients (4%) had previous stereotactic radiosurgery (their only previous invasive treatment). Forty-two patients (29%) have undergone previous percutaneous rhizotomy and five patients (3%) have undergone previous microvascular decompression. Twenty patients (14%) had undergone more than one invasive treatment modality previously. The patients’ clinical features are summarized in Table 1.

Table 1.

Baseline patients’ characteristics

Feature	Number (%)
Gender
Male	60 (41%)
Female	85 (59%)
Pain distribution
V 1	9 (6%)
V 2	21 (14.5%)
V 3	39 (27%)
V 1, 2	21 (14.5%)
V 2, 3	39 (27%)
V 1, 3	2 (1%)
V 1, 2, 3	14 (10%)
Laterality
Right	73 (50%)
Left	69 (48%)
Bilateral	3 (2%)
TN type
Idiopathic TN	129 (89%)
MS-related	16 (11%)
Previous invasive treatment
None	72 (50%)
Radiosurgery	6 (4%)
Rhizotomy	42 (29%)
MVD	5 (3%)
Multiple treatment modalities	20 (14%)
Atypical pain component
Yes	16 (11%)
No	129 (89%)

Radiosurgery Technique

Patients were treated with the Leksell Gamma Knife 4C (Elektra, Stockholm, Sweden). The treating team consisted of a neurosurgeon, radiation oncologist, medical physicist and radiation therapist. Planning was based on an MRI and CT obtained after applying the stereotactic frame. Patients received 80 Gy prescribed to a single isocenter targeting the cisternal segment of the trigeminal nerve. Patients who had undergone previous radiosurgery for trigeminal neuralgia received 50 Gy. A single 4-mm collimator was used for radiation delivery. If necessary, collimator plugs were used to ensure that the dose to the brainstem did not exceed 15 Gy to a volume of 1mm³.

Outcome and Statistical Analysis

Patients’ health records were reviewed for basic clinical and demographic features. Treatment details were obtained from the gamma knife unit database including the radiation dose and number of isocenters. Integrated dose was obtained from the Leksell Gamma Knife software. Follow up information was obtained by conducting telephone interviews where the patients were asked to report their current clinical status. Health records were also reviewed to document the patient’s last follow up clinic visit.

Treatment response was graded using the following four categories, in order of decreasing magnitude of response: no pain and no medications, no pain yet maintained on medications, persistent pain with significant improvement after radiosurgery, and finally no worthwhile improvement (see Table 2).

Table 2.

Pain outcome categories

category 1	No pain, no medications
Category 2	No pain with medications
Category 3	Pain reasonably controlled with medications
Category 4	No improvement from the treatment

Statistical analysis was performed using SAS 9.2 for Windows (SAS Institute; Cary, NC). The treatment outcome was dichotomized into good, consisting of those patients with no pain either with or without medications, and poor, comprising those patients with persistent pain after treatment. Pearson’s chi-square, and Fisher’s exact test were used to test for association of categorical variables. Univariate logistic regression analysis was used to identify predictors of treatment response. Predictors of recurrence were similarly identified, with patients reporting no response to treatment excluded from this analysis. Variables found to associate with recurrence in univariate analysis were entered into a multivariable Cox proportional hazards model to determine the independent hazard of recurrence for each predictor. In all analyses, statistical significance was set at p=0.05.

RESULTS

Pain Response

One hundred and forty-five patients met our inclusion criteria. Mean follow up period was 24 months (range 6-56 months). At last follow up, forty-eight patients (33%) were pain free with no medications. Forty-eight patients (33%) were pain free while still taking pain medications. Twenty-eight patients (19%) had pain after the treatment with significant reduction in their pain severity. Twenty-one patients (15%) did not have any significant pain reduction. Mean interval between treatment and response was 6 weeks (range 1 day- 6 months). The pain outcome after radiosurgery is shown in Table 3. Pain free status (outcome category 1 and 2) was achieved in 63% of the patients who had prior radiosurgery treatment at another institution compared to 67% of patients with no previous radiosurgery (p value 0.915).

Table 3.

Pain outcome at last follow up

Pain category	Number of patients (%)
Category 1	48 (33%)
Category 2	48 (33%)
Category 3	28 (19%)
Category 4	21 (15%)

Multivariate regression analysis was used to identify clinical and treatment-related factors that are associated with good response to radiosurgery. Two factors were significant predictors of pain outcome. Patients who received a higher integrated dose were less likely to be pain free with or without medications, (OR 0.729, CI 0.566-0.940, p 0.0146) (see below). Post-treatment numbness was associated with good pain outcome (OR 2.720, CI 1.193-6.200, p = 0.0173)(See Table 4). There was a trend for poor pain control in patients with multiple sclerosis-related trigeminal neuralgia but it did not reach statistical significance (OR 0.339, CI 0.096 1.193, p 0.0919).

Table 4.

Multivariate regression analysis of predictors of pain free status after treatment

Variable	OR	CI	P value
Age	1.006	0.969-1.044	0.7693
Pain duration	1.011	0.934-1.093	0.7914
Gender	0.966	0.353-2.648	0.9470
Laterality	0.707	0.262 1.906	0.4935
Atypical	1.067	0.224-5.075	0.9354
Integrated dose	0.729	0.566-0.940	0.0146
Post-treatment numbness	2.720	1.193-6.200	0.0173
Previous radiosurgery	0.918	0.191-4.415	0.9150
MS-related trigeminal neuralgia	0.339	0.096 1.193	0.0919

Integrated Dose

The mean and median integrated doses to the trigeminal nerve were 3.2mJ and 3.1 mJ, respectively (range 0.5-7.6 and standard deviation 1.4). A higher integrated dose predicted poor pain outcome (OR 0.729, CI 0.566-0.940, p 0.0146). Based on the linear regression model between the integrated dose and pain free outcome, at integrated dose of 5.3 mJ or below, there is more than 50% chance of pain free outcome. Among the 145 patients treated, the integrated dose was available by the Gamma Knife software in 135 patients. Only 4 patients received an integrated dose above 5.3mJ. In those four patients, one patient achieved category 2, two patients achieved category 3 and one patient had category 4 outcome. There was no significant difference in pain outcome in those four patients compared to patients who had an integrated dose below 5.3mJ.

Side Effects

Four patients (3%) reported side effects within 24 hours after treatment. Two patients complained of significant headache, one patient complained of fatigue and one patient complained of dizziness. No patient required admission because of these complaints.

New post-treatment numbness in the trigeminal nerve distribution was reported in 44 patients (30%). One of the four patients with integrated dose above 5.3 mJ developed post-treatment facial numbness.

Recurrence

Recurrence occurred in 51 patients (35%). Mean time to recurrence was 10 months from the date of treatment. Twenty-one patients had another procedure during the follow up period. Twelve patients had another radiosurgery treatment and 4 of them (33%) became pain free (with or without medications). Four patients had rhizotomy and 5 had microvascular decompression. Pain duration prior to treatment was the only independent predictor of recurrence (HR 1.038, 95%CI 1.001-1.075; p=0.0412). Patients with longer pain duration before the treatment were at higher risk of recurrence.

DISCUSSION

Treatment Response

The efficacy of gamma knife radiosurgery in the treatment of trigeminal neuralgia has been shown in many clinical reports [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. Our study focuses on the identification of different treatment variables that influence treatment effects and in turn allow modification and assessment of these parameters to optimize patients’ outcome. We categorized our outcome as pain free group (further subdivided whether on pain medications or not). The third group was patients with worthwhile improvement, corresponding roughly to group 3B in the Barrow Neurological Institute Pain Scale [22]. And lastly, patients with no worthwhile response (see Table 2). Although the BNI report has been frequently used, we tried to use a classification system that is fairly simple and reproducible and would allow results comparison across different reports with different classification systems. Any classification system such as these will have limitations, and comparison between different classification systems needs to be carefully done.

In our treated patients, 66% became pain free at last follow-up (33% were pain free without medications) while 85% had significant improvement in their pain. Fifteen percent did not respond or did not have a significant decrease in their pain after treatment. We had a new facial numbness rate of 30%. Reported percentage of patients after radiosurgery for trigeminal neuralgia who have no to minimum pain or have pain that is controlled with medications (good response by our criteria and BNI score of roughly 1-3A) varies between 34%-73% [2, 3, 7]. Adequate pain control was achieved in 85% of our patients. The rate of adequate pain control one year from treatment (BNI score 1-3B) in two large series of at least 450 patients that were published recently were 80 and 75% [3, 23]. In these two large series the rate of trigeminal sensory deficit was 11% and 35%, respectively.

Using regression analysis, two factors predicted treatment outcome. The first factor was post-treatment numbness. Treatment-induced facial numbness has been associated with good pain outcome in previous reports [24, 25, 26, 27]. Furthermore, the literature suggests that the rate of these two variables correlate with the maximum dose with a suggestion that a dose between 80-90 Gy is ideal [7, 25, 26]. Pollock [26] reviewed published series of radiosurgery for trigeminal neuralgia. He found a significant correlation between the dose and both pain outcome and post-treatment facial sensory deficit. The question that should be answered is what is the dose that would achieve good pain outcome with acceptable rate of sensory deficit. In a series of patients treated with 90 Gy, Pollock [25] reported pain free rate of 61% at mean follow up of 14 months with 54% risk of trigeminal nerve numbness. Nicol [12] reported another series of 90 Gy dose with 73% pain free rate at median follow up of 14 months and sensory deficit rate of 17%. Another series of 90 Gy reported pain free percentage of 56% with 25% risk of trigeminal dysfunction [28]. In summary of the literature, it appears that clinical series where patients received a maximum dose of 90 Gy show pain control and trigeminal deficit rates that overlap with the rates in series where patients received 80 Gy (see Table 5). A randomized controlled trial or a comprehensive meta-analysis that adjusts for all other variables might be needed before a conclusion is made about the ideal dose between 80 and 90 Gy.

Table 5.

Summary of outcome of reported series

Series	Dose (Gy)	Pain outcome	Trigeminal deficit (%)
Current series	80	66% pain-free at 24 months	30%
Knodziolka [3]	80	46% BNI 1-3B at 5 years	11%
Verheul [23]	80	58% BNI 1-3B at 5 years	35%
Massager [14]	90	59% pain-free at 42 months	38%
Nicol [12]	90	73% pain-free at 14 months	17%
Smith [28]	90	56% pain-free at 23 months	25%
Pollock [25]	90	61% pain-free at 14 months	54%

Although we standardized the maximum dose at 80 Gy for primary cases, we tried to analyze other treatment factors for their impact on outcome. During treatment planning, the radiation beams are contoured to avoid excess radiation to critical neural structures. This is achieved with collimator plugs that block selected sources of radiation beams. This results in slowing the dose-delivery rate without changing the maximum radiation dose to the trigeminal nerve. This can potentially alter the biological efficacy of the radiation on the nerve. Massager et al [20] found that shielding which prolongs treatment duration and increases radiation integrated dose results in an increased biological effect of radiation as reflected by higher rate of pain relief and post-treatment facial numbness. Flickinger et al [21] reported the outcome in a group of patients who were randomized to either one or two isocenters with the same maximum dose for both subgroups. Pain control was similar between the two groups but facial numbness was more in the group that was treated with two isocenters. Higher rates of both pain control and facial numbness suggest an increase in the biological effect of radiation with higher integrated dose. In our series, an increase in the integrated dose was associated with poor pain control (OR 0.729, CI 0.566-0.940, p 0.0146). With more plugs, the integrated dose increases and the time needed to deliver the same maximum dose increases, thereby affecting the dose rate. Because only four patients had integrated dose above 5.3 mJ, there was no significant difference in pain control or posttreatment numbness between those four patients and the rest of the cohort. Our finding suggests slower dose delivery decreases the biological efficacy of the radiation. From our linear regression, at an integrated dose of 5.3 mJ or less, there was more than 50% chance of painfree outcome. We do not suggest that radiosurgeons should use this as an upper limit for the integrated dose. Instead, we used it to refer to the 50% point on the linear regression. We think our finding should raise the question of the importance of dose delivery rate on pain outcome which can be confirmed by future reports.

Recurrence

Fifty-one patients (35%) reported recurrent pain over a mean follow up of 24 months. Mean duration between treatment and the occurrence of recurrent pain was 10 months. Longer pain duration prior to treatment was associated with higher risk of recurrence (HR 1.038, 95%CI 1.001-1.075; p=0.0412). Twelve patients had another radiosurgery treatment and 4 of them (33%) achieved pain free status (with or without medications).

Loss of treatment benefit with time is a well-established phenomenon in patients receiving radiosurgery for trigeminal neuralgia [2, 3, 4, 16, 17]. Kondziolka found the chance of having adequate pain control (BNI score of 1-3B) to be 80% at one year and only 30% at 10 years [3]. Despite this decline, the 3-year efficacy of this treatment is reported to be 41%-71% [3. 10, 23]. Despite this loss of benefit with time, radiosurgery remains a valuable option in elderly population with medical co-morbidities.

Pain free outcome was achieved in 33% of re-treated cases. This is in concordance with the literature that suggests repeat radiosurgery for trigeminal neuralgia is associated with some efficacy [11, 29, 30].

Study Limitations

Any clinical study, particularly one that deals with the subjective sensation of pain, is subject to significant limitations. With the variability in the methods of results reporting in the literature, comparison with other series is challenging and, associated with limitations. Our report is a retrospective review with the inherent bias of this methodology. The relatively short-term follow up of our patients may underestimate our recurrence rate.

CONCLUSIONS

This report supports existing literature about radiosurgery efficacy in trigeminal neuralgia. Post-treatment numbness predicts better pain outcome. Slower treatment delivery may decrease treatment efficacy. Despite significant time-dependent decline in pain control, radiosurgery is a valuable treatment option for this condition.

ABBREVIATIONS

CI:

Confidence interval

CT:

computed tomography

GKRS:

Gamma knife radiosurgery Gy: Gray

HR:

Hazard ratio

mJ:

Milli Joule

MRI:

Magnetic resonance imaging

MS:

Multiple sclerosis

OR:

Odds ratio