Trigeminal neuralgia is a severe and debilitating pain disorder for which many patients become medically refractory. Stereotactic radiosurgery is the least invasive surgical option for patients with trigeminal neuralgia.

      The purpose of this study was to evaluate longer-term outcomes and the durability of pain relief after radiosurgery, and the evolution of the technique over time.

      Over a 19 year interval, gamma knife stereotactic radiosurgery (GKSR) was performed in 1,000 patients with trigeminal neuralgia. In the majority of patients MRI stereotactic guidance was used to deliver a single 4mm isocenter and a maximum dose of 80 Gy. We evaluated detailed outcomes in a subset of 503 medically refractory patients and stratified outcomes according to a validated rating scale.  These patients were evaluated up to 2009 to allow for longer term outcomes.  Patients were evaluated up to 16 years after GKSR; 107 patients had over five years of follow-up.

      The median patient age was 70 years (range,16–99 years). 605 were female and 395 were male.  The average duration of trigeminal neuralgia pain was seven years (max, 50).  

      In our comprehensive review of the first 503 patients, 93% had pain relief using the Barrow Neurological Institute (BNI) scale. Favorable outcomes (BNI scores I-IIIb) were achieved in 75% at 1 year, 59% at 3 years, 43% at 5 years and 29% at 10 years. A faster initial pain response was seen in patients with typical trigeminal neuralgia (p=0.014), patients without prior surgery (p=0.014), and patients with pain < 3 years (p=0.012). Two hundred forty three patients (51%) reported some recurrent pain 3 to 144 months after initial relief (median, 48 months). Fifty-three patients (10.5%) developed new or increased subjective facial sensory symptoms and one developed deafferentation pain; in 17 this resolved. Patients who developed sensory loss had better long-term pain control (78% at five years).

      The main limitation of this study is its retrospective nature and lack of a specific control group.  We are continuing to collect outcomes data on all subjects.

      Gamma knife radiosurgery was an effective minimal-access approach for medically refractory trigeminal neuralgia. Some degree of facial sensory loss may improve long term pain relief and usually is not bothersome. Pain recurrence can be managed with repeat radiosurgery or a different procedure.

      This study establishes clinical outcomes for trigeminal neuralgia radiosurgery so that patients and doctors can compare radiosurgery to other medical or surgical treatments. Because the dataset is large, this data will be of value for the design of randomized trials.

      Medical therapy is the mainstay of trigeminal neuralgia (TN) management but patients frequently require surgical intervention due to persistent pain or treatment-related side effects. Available surgeries include microvascular decompression via craniotomy or percutaneous rhizotomies such as radiofrequency thermooagulation, balloon microcompression or glycerol rhizotomy.¹⁷ These procedures have variable but definite rates of risk and pain recurrence and clinicians often choose between them based on these outcomes.

      Gamma Knife radiosurgery (GKSR) is a minimally invasive surgical approach to create a lesional effect with tumors, blood vessels, or nerve fibers (axons). It can be used to manage trigeminal neuralgia via precise targeting of the trigeminal nerve anterior to the brainstem. Lars Leksell first performed radiosurgery using a prototype guiding device linked to an x-ray tube.^13,14 Since then, modern radiosurgery devices have been used for this indication. It is important to understand the value of radiosurgery for this pain disorder.^{3,5,7,9,10,15,16,19-23,25}

      Our hypothesis was that through retrospective evaluation of a large patient series who had this procedure, we could determine not only how well it worked but determine an accurate morbidity profile.

      Thus, we evaluated the effectiveness of GKSR over time, how quickly relief was achieved, whether relief was durable, and determined factors associated with success or complications to help clinicians and patients choose amongst various options.

      Prior reports found that radiosurgery was well tolerated and could be performed on an outpatient basis, that approximately 80% of patients had some form of pain relief and that side effects were related to dose and treatment technique. ^{3,5,7,10,15,16,19-23,25}

      The main limitation of our study method was that data was evaluated retrospectively (although collected prospectively).

      This study was approved by the University of Pittsburgh Institutional Review Board.

      The work described in this report was funded by a grant to Dr. Kano from the Osaka Medical Research Foundation for Incurable Diseases.

      Between 1992 and 2011, 1,000 GKSR procedures were performed for trigeminal neuralgia patients including 156 repeat GKSR procedures. The median age was 70 years (range 16–99 years).

      In this complete series, there were 396 men and 604 women.

      All patients had longstanding pain refractory to medical management using agents such as carbamazepine, phenytoin, baclofen, or gabapentin. Within the cohort of 503 patients with idiopathic TGN suitable for evaluation, the median symptom duration was 84 months (range 3–444 months). Pain was distributed in the V2 and V3 distributions of the trigeminal nerve (56%), followed by V3 (15%), V2 (12%), V1 and V2 (12%), all trigeminal distributions (2.0%) and only V1 plus V3 (1%).  Surgery had been performed previously in 215 (42%) patients and multiple surgeries had been performed in 57%. GKSR was the first surgery in 288 (58%) of patients.  The subset of 503 patients suitable for outcomes analysis were part of a cohort evaluated up to 2009¹².  Those managed after 2009 were censored from the outcomes analysis but were included in our study of the technique.  The characteristics of these patients were similar to the entire clinical experience.

      All patients reported the typical features of trigeminal neuralgia as a sharp, lancinating or shooting electric-shock type of pain often triggered by facial activity or stimulation. Twenty-six (6%) noted additional features such as a more constant dull, aching, or burning pain. In addition, 103 (20.4%) had some sensory disturbance (usually paresthesias) preoperatively, and three patients (0.5%) had partial deafferentation pain caused by prior ablative surery.

      Radiosurgical Technique Different models of the Gamma Knife (Elekta, Stockholm Sweden)(models U, A , B, C, 4C, and Perfexion) were used during this study. A single 4-mm isocenter was used in 953 patients (95%) and two 4-mm isocenters were used in the rest. With a single isocenter, the target was 3 to 8 mm anterior from the junction of the trigeminal nerve and pons. We administered maximum doses of 60 (2%), 70 (4%), 75 (3%), 80 (88%), 85 (2%), and 90 Gy (2%).  The radiosurgical target initially was closer to the pons and over time moved 1-2 mm anterior but remained a posterior nerve target in the majority of patients. Because some patients had a short nerve segment or because of prior surgery had a nerve that was difficult to identify along the entire course, a mid-nerve target or an anterior target was chosen in individual patients.

      Clinical Follow-up We evaluated the degree of pain relief, the latency interval until pain relief, the need for further surgical procedures, the use of medications, and the development of new symptoms or signs. To evaluate the effectiveness of GKSR we calculated the time to initial response and the duration of pain relief using the Kaplan–Meier product -limit method. Pain outcome was scored using the Barrow Neurological Institute (BNI) scale, as shown in Data Share 2.  We defined BNI scores I to IIIb as representing favorable outcomes, whereas BNI scores of IV and V were defined as treatment failures. For patients who described facial sensory dysfunction, we also inquired about their quality of life related to trigeminal neuralgia and whether sensory symptoms bothered them. BNI scores were as follows: I (no pain and taking no medication); II (occasional pain but taking no medication); IIIa (no pain but taking medication); IIIb (pain improved but not adequately controlled with medication); IV (some pain, not adequately controlled with medication); V (no relief).  The first follow-up was at 3 months or earlier if the patient requested an evaluation due to poor pain control, and then requested yearly. Patients who became pain free were later contacted in 2009 for a late evaluation.

      Within the evaluation cohort of 503 patients, 83% completed follow-up. Forty-two patients did not have follow-up data and were excluded from the analysis.

      All serial follow-up information was obtained via direct contact with the patient or their referring physicians. This data was collected by neurosurgeons who had not participated in patient care using telephone interviews.

      The tests and scoring methods were validated.

      We correlated outcomes with various clinical parameters (including sex, age, duration of symptoms, presence or absence of preoperative sensory dysfunction, presence of atypical pain features, number of previous failed surgical procedures, and maximum dose), using a log rank test with p<.05 set as significant, a stepwise (forward conditional) multivariate analyses and the Cox proportional hazards model with p<.10 set as signficant.  Outcomes were calculated using actuarial statistics over a 16 year period. The median follow-up duration was 2 years (median 24, range 3–156 months).

      We did not use a power calculation as no control group was studied.

      P<.05 was chosen as the level of statistical significance for the log-rank test.

      Pain Relief From the cohort analyis of 503 patients, 449 (89%) responded to GKSR after the procedure at a median latency of one month. The median interval until pain relief (grade I to IIIb) was one month. The duration of pain relief after the initial response was evaluated using the product limit method of Kaplan and Meier. Pain relief (grade I, II, IIIa or IIIb) was achieved and maintained in 375 patients (75%) at 1 year, 297 patients (59%) at 2 year, 217 patients (43%) at 5 years and 143 patients (29%) at 10 years. At the final evaluation, 330 (66%) patients continued to have pain control (BNI grade I-IIIb). One hundred and forty-four patients (29%) achieved group I, 11 patients (2%) were group II, 120 patients (24%) were group IIIa, 55 patients (11%) were group IIIb, 126 patients (25%) were group IV, and 47 patients were group V pain status (9%). Pain eventually recurred in 240 patients at a median of 48 months (range, 3-144 months). Additional surgery was performed in 144 patients.

      Sensory Dysfunction All procedures were successfully concluded without incident. Fifty-three patients (10.5%) later developed increased facial sensory dysfunction such as paresthesias or objective facial sensory loss, all within the first two years. In ten of these (19%) a later recurrence of pain was also found. The 1-, 3-, and 5-year rates for maintenance of pain relief in patients who noted sensory dysfunction were 91%, 82.0%, and 78%, respectively. We evaluated these 53 patients further to understand the effect of sensory dysfunction on their quality of life. Thirteen could not be contacted and five of these had died. In seventeen, the symptoms had resolved (16 described their quality of life to be improved after radiosurgery). Of 23 with persistent dysfunction, seven described it as “not bothersome”, ten as “mildly bothersome”, six as “severely bothersome”, and zero as “intolerable”. Nineteen of these said their quality of life was improved after GKSR and that sensory dysfunction was a good “trade-off” for pain relief. Four said their quality of life was worse. One patient (0.2%) who already had decreased facial sensation after a prior microvascular decompression developed deafferentation pain after 75 Gy GKSR with two isocenters. No other trigeminal deficit (such as motor function or  chewing) or other cranial neuropathy was found. Management of Recurrent Pain Because of recurrent pain, 87 patients underwent a second radiosurgery and 72 were suitable for this analysis. Although the techniques varies amongst centers, we targeted the nerve a few mm anterior to the first target to allow some overlap. This becomes simpler with a longer nerve length. The median interval between procedures was 30 months, and subsequent follow-up averaged two years. The average age was 78. The maximum dose at the second radiosurgery varied between 60 and 70 Gy. After the second GKSR, 68 patients (94%) had pain relief at some time. The actuarial rate for pain control was 90%, 73%, 63%, 56% at 1-, 3-, 5-, 10-years respectively. Eight patients (11%) developed new sensory dysfunction after the second radiosurgery which was permanent in six.

      Factors associated with earlier initial pain relief (grade I to IIIB) were absence of atypical trigeminal neuralgia features (p=0.014), no prior surgery (p=0.014), and pain for < 3 years before GKSR (p=0.01). A shorter pain duration before GKSR was not correlated with a grade I outcome(p =0.098). Factors associated with earlier pain recurrence were age < 65 years (univariate:p=0.012, multivariate: p=.06), patients with atypical pain features (univariate,p=0.02, multivariate, p=.04), > 3 surgeries univariate:p=0.03, multivariate:p=.095), and patients without any new facial sensory symptoms (multivariate: p<.0001).

      The goal of trigeminal neuralgia surgery is complete elimination of pain and need for medication. To achieve this, there is a balance between surgical risk, maintenance of normal nerve function, and the known rates for pain relief. Not all procedures relieve pain and not all patients choose to stop medication, particularly if they have been taking it for years. For patients who have failed one or more prior surgeries, the expectations for complete relief are reduced. Unfortunately, since pain can come and go, a patient may report that they have no pain at one assessment (grade I), only to have it recur soon afterwards (grade II or worse) but not report it. The role of radiosurgery in the management of medically refractory trigeminal neuralgia has evolved. Questions regarding treatment durability, long-term complications, and appropriate patient selection mandate that analyses of radiosurgical outcomes be available to compare with reports on other options.²⁶

      Pain Relief Other centers have found a latency interval to pain relief after radiosurgery of four to eight weeks.^10,12,18,28  We found that 89% of patients noted significant improvement at an average of one month. We found that patients with typical trigeminal neuralgia, patients who underwent GKSR as their initial surgical procedure, and patients who underwent earlier GKSR (< 3 years) after pain onset had faster pain relief (grade I to IIIb). By 12 months after GKSR, 11% of patients still had pain. We advocate repeat GKSR only if complete pain relief had been achieved initially, with subsequent recurrence.¹⁸ The majority of patients experience lasting and meaningful pain reduction with few complications after GKSR. In this series, 75% of patients achieved or maintained pain control (BNI grade I-IIIb), 59% had pain relief at 3 years, 43% maintained relief at 5 years, and 29% were still controlled with or without medications at 10 years. The best result, BNI grade I (pain free-off medication), was achieved in 28% of patients. Of those with a significant pain recurrence, 60% underwent further surgery and 40% were maintained on medication alone.  Clearly, pain control can lessen over time necessitating further medical or surgical management. In a prior analysis we did not find that the radiosurgical dose rate correlated with outcome.¹

      Sensory Dysfunction Post-rhizotomy paresthesias or numbness of varying degrees is observed in 6 to 70% of patients after thermal rhizotomy, glycerol rhizotomy, or balloon microcompression, depending on the technique and the goal of the surgeon. ^{3,4,6,7,10,18,24,27} In our study, sensory dysfunction was found in 11%. Only one patient (0.19%) developed deafferentation pain. This patient had undergone a prior microvascular decompression.  Only ten of 53 patients who later developed facial sensory dysfunction had pain recurrence after GKSR. The 1-, 3-, and 5-year rates for maintenance of pain relief in patients who later developed facial sensory symptoms were 91%, 82%, and 78%, respectively. These data suggest that patients who developed facial sensory symptoms had a reduced rate of recurrent pain (p<0.0001).

      Gamma knife radiosurgery is effective for trigeminal neuralgia and associated with low risk. This procedure maintains facial sensation in most patients, and can be performed regardless of age or medical condition. The use of a higher dose may provide better and longer-lasting outcomes although this remains to be seen. In patients with recurrent pain, a repeat GKSR procedure may reestablish pain control. However, these outcomes are not as good as those after a first microvascular decompression, ² and thus we continue to recommend consideration of that approach for younger patients suitable for invasive surgery. Future work should consider randomized trials between radiosurgery and medical therapy for patients who have failed medication at least once.