Stereotactic radiosurgery for idiopathic trigeminal neuralgia

Transcription

1 J Neurosurg 97: , 2002 Stereotactic radiosurgery for idiopathic trigeminal neuralgia BRUCE E. POLLOCK, M.D., LOI K. PHUONG, M.D., DEBORAH A. GORMAN, R.N., ROBERT L. FOOTE, M.D., AND SCOTT L. STAFFORD, M.D. Departments of Neurological Surgery and Radiation Oncology, Mayo Clinic and Foundation, Rochester, Minnesota Object. Each year a greater number of patients with trigeminal neuralgia (TN) undergo radiosurgery, including a large number of patients who are candidates for microvascular decompression (MVD). Methods. The case characteristics and outcomes of 117 consecutive patients who underwent radiosurgery were retrieved from a prospectively maintained database. The mean patient age was 67.8 years; and the majority (58%) of patients had undergone surgery previously. The dependent variable for all analyses of facial pain was complete pain relief without medication (excellent outcome). Median follow-up duration was 26 months (range 1 48 months). The actuarial rate of achieving and maintaining an excellent outcome was 57% and 55% at 1 and 3 years, respectively, after radiosurgery. A greater percentage of patients who had not previously undergone surgery achieved and maintained excellent outcomes (67% at 1 and 3 years) than that of patients who had undergone prior surgery (51% and 47% at 1 and 3 years, respectively; relative risk [RR] = 1.77, 95% confidence interval [CI] , p = 0.04). New persistent trigeminal dysfunction was noted in 43 patients (37%). Tolerable numbness or paresthesias occurred in 29 patients (25%), whereas bothersome dysesthesias developed in 14 patients (12%). Only a radiation dose of 90 Gy correlated with new trigeminal deficits or dysesthesias (RR = 3.10, 95% CI , p 0.001). Excellent outcomes in patients with new trigeminal dysfunction were achieved and maintained at rates of 76% and 74% at 1 and 3 years, respectively, after radiosurgery, compared with respective rates of 46% and 42% in patients who did not experience postradiosurgery trigeminal dysfunction (RR = 4.53, 95% CI , p 0.01). Conclusions. Radiosurgical treatment provides complete pain relief for the majority of patients with idiopathic TN. There is a strong correlation between the development of new facial sensory loss and achievement and maintenance of pain relief after this procedure. Because the long-term results of radiosurgery still remain unknown, MVD should continue to be the primary operation for medically fit patients with TN. KEY WORDS trigeminal neuralgia facial pain stereotactic radiosurgery T RIGEMINAL neuralgia is a facial pain syndrome estimated to affect 15,000 new patients in the United States each year. 9,35 Tumors involving the trigeminal nerve, ganglion, or divisions account for 2 to 3% of cases of TN. 9 In patients with multiple sclerosis, TN often develops as a result of demyelinating plaques in the dorsal nerve root entry zone, descending trigeminal tracts, or brainstem nuclei. 13 Nevertheless, most patients suffer from idiopathic TN that is thought to arise from vascular compression and focal demyelination of the trigeminal root. 10,14 It is postulated that damaged nerve fibers become hyperexcitable and normal sensory impulses cause adjacent pain fibers to discharge, resulting in paroxysms of facial pain. Medical therapy eliminates or significantly reduces pain in approximately 75% of patients and is considered the treatment of choice for newonset TN. 12 Unfortunately, the relief provided by medical therapy generally decreases over time and many patients must stop taking their medications due to side effects, including imbalance, cognitive dysfunction, and hyponatremia. Drug intolerance is particularly prevalent in elderly patients with Abbreviations used in this paper: CI = confidence interval; MR = magnetic resonance; MVD = microvascular decompression; RR = relative risk; TN = trigeminal neuralgia. TN. A variety of operations have been used in the treatment of patients with medically intractable TN: MVD, 2,3,8 radiofrequency rhizotomy, 15,30 glycerol rhizotomy, 24,32 and balloon compression. 6,7 Aside from MVD, the remaining operations purposefully damage the trigeminal system and are therefore associated with various degrees of postoperative sensory loss. As an alternative to open surgical procedures, in 1951 Lars Leksell 20 performed radiosurgery in a patient with TN by using a modified orthovoltage x-ray tube. Recently, there has been renewed interest in this technique for the treatment of patients with medically unresponsive TN. 18,34 In this study, we reviewed our initial 4-year experience with radiosurgery in the treatment of TN. Clinical Material and Methods Patient Population Between April 1997 and August 2000, 137 radiosurgical procedures were performed at the Mayo Clinic in Rochester, Minnesota, in patients with nontumor-related TN. Six patients with multiple sclerosis were excluded from the present analysis. Eleven patients underwent repeated radiosurgery for recurrent or persistent TN. For these 11 patients, only the results of their initial operation were included in 347

2 B. E. Pollock, et al. TABLE 1 Characteristics of 117 patients who underwent radiosurgery for TN Characteristic Value male/female ratio 54:63 age (yrs) mean 67.8 range duration of pain (yrs) mean 10.8 range patients who underwent surgery previously (%) 58 prior surgery (no.) mean 0.9 range 0 4 type of prior surgery (% of patients) MVD 27 glycerol rhizotomy 12 radiofrequency rhizotomy 24 balloon compression 10 peripheral neurectomy 1 % of patients w/ preop trigeminal dysfunction 42 % of patients w/ atypical pain 7 this study. Three patients participated in less than 1 year of follow-up review after the procedure (two patients died of unrelated causes and the other was lost to follow up). Thus, the outcomes of single-session radiosurgery for 117 patients with medically intractable idiopathic TN were available for analysis. The characteristics of the patients are outlined in Table 1. All patients described episodic shocklike pain involving one or more divisions of the trigeminal nerve. Additionally, eight patients (7%) also experienced atypical features such as burning or constant pain. Facial pain was experienced on the right side in 71 patients (61%) and on the left side in 46 patients (39%). Three patients (2%) experienced pain in V1, 27 (23%) in V2, 17 (15%) in V3, 18 (14%) in V1 and V2, 41 (35%) in V2 and V3, and 11 (9%) in all three trigeminal divisions. Overall, 32 patients (27%) experienced pain involving the ophthalmic branch of the trigeminal nerve. The majority of patients (58%) had previously undergone one or more operations. Among these 68 patients there was a mean of 1.6 operations including radiofrequency rhizotomy (37 patients), MVD (35 patients), balloon microcompression (17 patients), glycerol rhizotomy (17 patients), and peripheral neurectomy (one patient). In addition, 49 patients (42%) underwent a total of 109 peripheral ethanol blocks before they underwent radiosurgery. 11 Radiation Dosimetry All radiosurgical procedures were performed with the aid of a gamma knife (Leksell gamma knife model B; Elekta Instruments, Norcross, GA). A local anesthetic was administered to the patient, a stereotactic headframe (Leksell model G; Elekta Instruments) was placed on the patient s head, and stereotactic MR imaging was performed to identify the trigeminal nerve. The MR sequence used for dose planning was an axial postgadolinium T 1 -weighted three-dimensional volume acquisition in which matrices were divided into 1-mm slices. The dose plans consisted of a single 4-mm isocenter of radiation in 92 patients (79%). The median length of nerve within the 50% isodose line was 5.9 mm (range mm). The maximum radiation dose was 70 Gy in 27 patients (23%), 80 Gy in 14 patients (12%), and 90 Gy in 51 patients (44%). Although the cisternal nerve length varied among patients, in patients who received the higher radiation doses the isocenter was generally placed more distally on the nerve to decrease radiation exposure to the brainstem. Thus, the median isodose line touching the brainstem was 40% in patients who received 70 Gy, compared with the 20% isodose line in patients who received either 80 or 90 Gy. The remaining 25 patients (21%) were treated with two 4-mm isocenters placed along the length of the nerve. Twenty-four patients treated with two isocenters received a maximum radiation dose of 70 Gy and one patient was treated with 80 Gy. The median length of nerve within the 50% isodose line was 9.4 mm (range mm). Dose calculations throughout this interval used an output factor of 0.8 for the 4-mm collimator. Follow-Up Review and Statistical Analysis Patients were discharged on the day of the procedure and told to continue their usage of preoperative medications until facial pain stopped. At that point, patients were instructed to decrease their medications, as tolerated, over a several-week interval. Facial pain outcomes were defined as excellent (no pain and no medications), good (no pain and reduced level of medications), fair (significantly less pain and fewer medications required; the patient would repeat the procedure to achieve this degree of pain relief), and poor (no significant change in pain or medication requirement). Patients were routinely contacted at 1 month, 3 months, and then yearly after the operation to discuss the degree of pain relief, current medications, and the development of any new trigeminal dysfunction. Postoperative trigeminal deficits were classified as numbness, paresthesias, or dysesthesias. Any additional surgeries were documented. All available follow-up information was placed into a prospectively maintained computer database for later analysis. Patient outcomes were censored at the last follow-up examination (80 patients) or at the time of additional surgery (37 patients). The overall median patient follow-up review after radiosurgery was 26 months (range 1 48 months): the median follow up of patients not requiring further surgery was 29 months (range months). Facial pain outcomes were calculated using the Kaplan-Meier productlimit method. 16 The dependent variable for all facial pain analyses was complete pain relief without medications (excellent outcome). A univariate analysis of the following variables was performed with regard to facial pain outcomes and postoperative trigeminal dysfunction: age, sex, pain duration, pain location (single or multiple divisions), prior surgery (yes or no), number of prior surgeries, preoperative trigeminal dysfunction, atypical features, number of isocenters, length of irradiated nerve, and radiation dose (70 Gy or 70 Gy). Twenty-one patients who only received peripheral alcohol blocks before radiosurgery were included in the no prior surgery group. In addition, facial pain outcomes were analyzed to determine if there was any correlation with the development of postoperative trigeminal dysfunction. Parametric variables were compared using the Student t-test and nonparametric variables were compared using the chi-square test. All factors with probability values 348

3 Radiosurgery for trigeminal neuralgia FIG. 1. Actuarial plot showing the percentage of 117 patients with TN in whom excellent outcomes (no pain and no medications) were achieved and maintained after stereotactic radiosurgery. of 0.15 or lower were entered into a stepwise multivariate analysis, which was performed using the Cox proportional hazards model. FIG. 2. Actuarial plots showing the percentages of patients in whom an excellent outcome (no pain and no medications) was achieved and maintained among those who did not undergo prior surgery (solid line, 49 patients) and those who did (dotted line, 68 patients; p = 0.04). Results Facial Pain Outcomes Patient outcomes were excellent in 69 patients (59%), good in 19 (16%), fair in 12 (10%), and poor in 17 (14%). Overall, 75% of patients were free from pain at some point after radiosurgery. The median time to complete pain relief was 3 weeks (range 1 day 20 weeks). Eleven (16%) of 69 patients with excellent outcomes experienced recurrent pain at a median of 8 months (range 2 30 months); nine (47%) of 19 patients with good outcomes complained of recurrent pain at a median of 5 months (range 4 25 months). Thus, in 68 patients (58%) either excellent (58 patients [50%]) or good (10 patients [8%]) facial pain outcomes were still maintained at the last follow-up examination. Excellent outcomes were achieved and maintained in 57% and 55% of patients at 1 and 3 years, respectively, after radiosurgery (Fig. 1). Excellent or good outcomes were achieved and maintained in 65% and 55% of patients at 1 and 3 years, respectively, after radiosurgery. Thirty-seven patients (32%) underwent additional surgery at a median of 8 months after radiosurgery (range 1 40 months). The surgeries included MVD in eight patients, glycerol rhizotomy in 17, and radiofrequency rhizotomy and balloon compression in one patient each. Ten patients underwent repeated radiosurgery for their recurrent facial pain. Excellent facial pain outcomes for these additional surgeries by type of surgery included MVD (six of eight patients), glycerol rhizotomy (12 of 17 patients), radiofrequency rhizotomy (one of one patient), balloon microcompression (0 of one patient), and repeated radiosurgery (eight of 10 patients). Overall, 27 (73%) of 37 patients achieved excellent outcomes in response to additional surgery after failed radiosurgical procedures. A univariate analysis of patient and treatment parameters found the following three factors at or near statistical significance for achieving and maintaining an excellent outcome: younger patient age (p = 0.04), no previous surgery (p = 0.06), and longer length of the irradiated portion of the nerve (p = 0.15). Note that pain duration (p = 0.32), prior trigeminal deficits (p = 0.42), presence of atypical features (p = 1.00), and radiation dose (p = 0.53) did not correlate with excellent patient outcomes. A multivariate analysis demonstrated that patients with TN who had not undergone previous surgery had better outcomes after radiosurgery (RR = 1.77, 95% CI , p = 0.04); younger patient age (p = 0.22) and longer length of irradiated nerve (p = 0.53) were not associated with excellent facial pain outcomes. In 67% of patients who had not undergone prior surgery, excellent outcomes were achieved and maintained 1 and 3 years after radiosurgery, compared with 51% at 1 year and 46% at 3 years in patients who underwent surgery previously (Fig. 2). Complications of Radiosurgery New or worsened persistent trigeminal dysfunction developed after radiosurgery in 43 patients (37%). The median time to onset of these deficits was 8 months (range 2 20 months). Ten patients (8%) described areas of numbness, whereas 19 patients (16%) experienced tolerable paresthesias. Fourteen patients (12%) complained of significant dysesthesias that were bothersome. Three of these 14 patients also experienced corneal numbness, although visual loss caused by this complication was not observed. No patient experienced jaw weakness after the procedure. A radiation dose of 90 Gy correlated with postradiosurgery trigeminal deficits in a univariate analysis (RR = 3.10, 95% CI , p 0.001). Note that we detected no relation with the length of the irradiated nerve (p = 0.17) or preoperative trigeminal deficits (p = 0.64). Postradiosurgery trigeminal dysfunction strongly correlated with excellent facial pain outcomes. Excellent outcomes of 76% and 74% at 1 and 3 years, respectively, after radiosurgery were 349

4 B. E. Pollock, et al. FIG. 3. Actuarial plots showing the percentages of patients in whom excellent outcomes (no pain and no medications) were achieved and maintained among those with postradiosurgery trigeminal dysfunction (solid line, 43 patients) and those without (dotted line, 74 patients; p 0.01). achieved and maintained in patients with new trigeminal dysfunction, compared with rates of only 46% and 42%, respectively, in patients who did not experience postradiosurgery trigeminal dysfunction (RR = 4.53, 95% CI , p 0.01; Fig. 3). Two patients experienced other cranial nerve deficits following radiosurgery. One patient complained of ipsilateral hearing loss 1 week after radiosurgery; his hearing returned to baseline within several weeks. A second patient suffered ipsilateral facial weakness 2 months after radiosurgery; this patient s facial strength improved, but remained slightly abnormal 16 months after radiosurgery. Both patients became free from pain after the operation, making it almost impossible that a mistake in radiation administration could have occurred. Because the calculated radiation dose to the facial and vestibulocochlear nerves was lower than 2 Gy in both cases, it is unlikely that these new deficits represented radiation-related complications. Discussion The number of patients with TN who undergo stereotactic radiosurgery has increased dramatically over the past several years. Although 63 (8%) of the first 762 patients who underwent radiosurgery at the Karolinska Institute between 1968 and 1982 suffered from TN, the inability to visualize the trigeminal system clearly at that time limited the usefulness of this technique. 20 With the advent of computerized tomography and MR imaging, however, physicians are now able to obtain images of the trigeminal root for radiosurgery dose planning. Rand, et al., 27 reported on 12 patients with TN who underwent radiosurgery during which the maximum radiation doses ranged from 57.1 to 76 Gy. Eight of 11 patients experienced improvements in their facial pain, six of whom became free from pain. Encouraged by these early results, Kondziolka and colleagues 18 conducted a prospective multiinstitutional dose-escalation trial to characterize the results of this procedure further. Overall, 29 (58%) of 50 patients achieved freedom from pain and another 18 patients (36%) experienced more than a 50% reduction in pain. During a median follow-up period of 18 months, three patients (6%) experienced recurrent pain and three patients had facial paresthesias. Importantly, patients in whom the maximum radiation dose was 70 Gy or greater had a significantly greater chance of complete pain relief compared with patients in whom the dose was less than 70 Gy. This work, combined with an increase in the number of radiosurgical centers, has made radiosurgery a widely accessible treatment option for patients with facial pain that is refractory to medications; in fact, in 1999, 1347 patients in the United States underwent radiosurgery for TN (Gamma Knife Treatment Statistics, Norcross, GA). Although the long-term results of this procedure remain unknown, practitioners at some centers now claim that radiosurgery is the safest and most effective form of surgery currently available for TN and recommend early radiosurgical treatment once a clinical diagnosis has been established. 5,31 In the current series, we found that 68 (58%) of 117 patients were free from pain at a median follow-up period of 26 months. Specifically, 57% and 55% of patients were pain free without medication 1 and 3 years, respectively, after radiosurgery. Patients who had not undergone surgery previously had significantly better outcomes than patients who had undergone prior operations. Excellent outcomes were achieved and maintained 1 and 3 years after radiosurgery in 67% of patients who did not undergo previous surgery, compared with 51% at 1 year and 46% at 3 years in patients who underwent prior surgery. These results are excellent if the characteristics of our patients are considered: the mean patient age was 67.8 years and 68 patients (58%) had undergone an average of 1.6 prior surgeries. Table 2 provides a comparison of our results and those of other published studies on radiosurgery for idiopathic TN. 5,18,21,23,28,29,34 In these series, excellent outcomes were achieved within the first several weeks after the procedure in 35% to 74% of patients. The primary drawbacks of most of these series has been the limited number of patients (range patients) and the short follow-up periods (range months). 5,18,23, 29,34 Factors that have been associated with improved results are normal preoperative facial sensation, 21 higher radiation doses, 18,23,28 increased radiation dose to the brainstem, 5 and absence of prior surgery. 4,21,34 Negative predictors for pain relief after radiosurgery have included patients in whom TN is related to multiple sclerosis 4,19 and patients with atypical pain features. 21,29,31 Consequently, our results are quite similar to those of earlier reports and emphasize that, similar to other operations for TN, radiosurgery appears to be more effective when used as a primary surgery than when used in patients who experience recurrent pain after undergoing failed operations. One of the frequently cited advantages of radiosurgery, compared with percutaneous surgical techniques for TN, is the low risk of new trigeminal dysfunction. The incidence of this complication after radiosurgery has been reported to be between 0 and 17%; 5,18,21,23,28,29,34 however, 37% of our patients sustained a permanent sensory disturbance that was described as bothersome and troubling by 14 patients (12%). This increased incidence probably relates to the radiation dosimetry in our cases. Unlike most reports on TN, in which radiation doses between 70 and 75 Gy were deliv- 350

5 Radiosurgery for trigeminal neuralgia ered, 5,18,21,29,34 44% of our patients were treated with a radiation dose of 90 Gy. Multivariate analysis showed that a maximum radiation dose of 90 Gy was associated with increased rates of trigeminal sensory disturbance. Young, et al., 33 recently presented the results of 435 patients who underwent radiosurgery for TN between 1991 and The rates of facial sensory loss per radiation dose were 12% at 70 Gy, 20% at 80 Gy, and 29% at 90 Gy. Conversely, authors of two series have reported the use of a maximum radiation dose of 90 Gy, which was accompanied by the development of new facial paresthesias or numbness in 0% and 17% of their patients. 23,28 None of the patients in those studies experienced bothersome dysesthesias after radiosurgery, compared with 12 (24%) of 51 patients in our study who received 90 Gy. 26 To the best of our ability we can detect little difference in our technique compared with those used by other groups, although it is possible that we actually delivered approximately 8% more radiation as a result of our continued use of 0.8 as the output factor for the 4-mm collimator. Nonetheless, based on our experience with 90 Gy, we now limit the dose of radiation to 80 Gy during radiosurgery for TN in the hope of achieving better facial pain outcomes than those reported with 70 Gy (approximately 40% excellent outcomes), 5,21,29 while allowing an acceptable incidence of new facial sensory loss. Besides increased radiation dose, preoperative sensory loss has also been correlated with new facial numbness after radiosurgery. 21 It has been hypothesized that radiosurgery provides a high rate of pain relief with a low risk of facial numbness due to a subtotal or selective injury of afferent fibers in the trigeminal system. 18,21,28,34 We noted that postradiosurgery trigeminal dysfunction strongly correlated with facial pain outcomes; patients with new trigeminal deficits were more than 4.5 times more likely to have achieved and maintained excellent outcomes after radiosurgery compared with patients who did not experience postradiosurgery trigeminal dysfunction. The correlation between facial sensory loss and improved outcomes parallels the radiofrequency rhizotomy experience 30 and supports recent experimental data showing that radiosurgery at sufficient doses causes unselective destruction of both myelinated and unmyelinated sensory fibers. 17 Rogers, et al., 29 also found an association between new facial numbness and pain relief in their study. All five patients who experienced numbness had excellent outcomes compared with only 17 (35%) of 49 patients who did not experience new numbness. The results of repeated radiosurgery for TN provide more evidence linking postradiosurgery sensory loss and pain relief. Pollock, et al., 25 described 10 patients who underwent a second radiosurgery for recurrent facial pain at a median of 13 months after the initial procedure. Eight patients attained freedom from pain without medications and in each case some degree of trigeminal dysfunction developed. Conversely, the two patients without a new trigeminal deficit continued to experience significant pain and required further surgery. As a result, we believe that radiosurgery should be considered to be an ablative procedure whose long-term success depends on the production of some degree of trigeminal dysfunction. The ideal surgery for patients with medically unresponsive TN should quickly alleviate facial pain, allow patients to discontinue their medications, and preserve trigeminal function. Microvascular decompression, which is directed at the neurovascular compression believed to cause the pain TABLE 2 Published results of radiosurgery for the treatment of idiopathic TN* Outcomes (% of patients) No. of Follow Excellent Authors & Year Patients Up (mos) Excellent or Good series w/ crude rates of patient outcomes Young, et al., NS Régis, et al., NS 74 NS Brisman, et al., NS Nicol, et al., Rogers, et al., NS series w/ actuarial rates of patient outcomes Kondziolka, et al., NS Maesawa, et al., NS 59 present study * NS = not stated. Excellent outcome indicates a patient free from pain without medication; good outcome represents a patient free from pain while using reduced medication. Outcome at 2 years posttreatment. All patients in the series. Patients who received a radiation dose of at least 70 Gy. for the majority of patients with idiopathic TN, provides such an option to patients. 1,2 Barker, et al., 2 reported outcomes in 1185 patients with TN who underwent MVD performed by Dr. Peter J. Jannetta over a 20-year interval. Ten years after surgery, 70% of patients remained pain free without medications and in only 1% of patients did significant facial numbness develop. Although few surgeons will be able to replicate these impressive results, other series have documented the efficacy and safety of MVD for idiopathic TN. 1,3,8 By comparison, our finding that excellent outcomes were achieved and maintained 1 and 3 years after radiosurgery in 49 patients (67%) who did not undergo previous surgery appears quite similar to the results obtained using MVD. In our study, however, only 13 (27%) of these 49 patients continued to be free from pain without medications and without new trigeminal dysfunction as of the last follow-up examintion. Consequently, we believe that MVD should remain the primary operation for patients with TN if they are medically fit. Still, many patients with TN are elderly, suffer from significant comorbid medical conditions, or have recurrent facial pain after surgery. How does radiosurgery measure up to percutaneous procedures such as radiofrequency rhizotomy, glycerol rhizotomy, or balloon microcompression for this patient population? Based on our data and other reports with actuarial patient follow-up review, 21 approximately 55 to 60% of patients will be free from pain without medications 2 years after radiosurgery. In comparison, North, et al., 24 found that 60% of patients had excellent outcomes 2 years after glycerol rhizotomy as a first surgery, but only 25% of patients who had undergone prior surgery did not require either medications or further operations. Kanpolat and associates 15 similarly noted that 60% of patients experienced no pain following a single radiofrequency rhizotomy over the same interval. Brown 6 performed one or more percutaneous compressions in 183 patients and found that, although 93% of patients 351

6 B. E. Pollock, et al. were initially relieved of their pain, the crude recurrence rate at 55 months was 25%; no actuarial data were provided in that study. Thus, without controlling for the patient populations being treated, it appears that the general rate of pain relief after radiosurgery is roughly equivalent to those of published reports for other ablative procedures used to treat TN. Three important facts should be remembered, however, whenever one compares these different surgical techniques. 1) Although not specifically discussed in most papers, the long-term success of these surgeries depends on the production of some degree of trigeminal injury. For example, Taha and colleagues 30 found that the median pain-free interval in patients with either analgesia or dense hypalgesia after radiofrequency rhizotomy was more than 15 years, whereas it was only 32 months in patients with mild hypalgesia. In this regard, radiosurgery is no different. Thus, for optimal results these destructive techniques must balance opposing goals, that is, the production of an injury to trigeminal fibers that is sufficient to provide adequate pain relief without causing so much damage that bothersome dyesthesia or anesthesia dolorosa develops. 2) Although patients typically realize the benefits of percutaneous procedures within the first few hours after the operation, radiosurgery often requires several weeks or months for the therapeutic peak to be reached. Thus, for patients with severe pain who are having difficulty speaking or maintaining adequate hydration, radiosurgery should be considered a poor option. 3) Although the extent of radiation exposure produced by TN radiosurgery is quite low globally, the high doses used for these procedures can lead to significant injury to adjacent vasculature. 22 Over the past several years, the first author has performed posterior fossa explorations on eight patients after failed radiosurgery: in five of these patients a region of presumed radiation injury was noted on the superior cerebellar artery, at the site where the vessel contacted the trigeminal nerve. The median time from radiosurgery to posterior fossa surgery was 15 months (range 8 43 months), and the median maximum radiation dose was 90 Gy (range Gy). In fact, four of these five patients received either 90 Gy (three patients) or underwent two radiosurgical procedures (70 Gy and 80 Gy, respectively; one patient). Thus, vascular injury may primarily occur in the setting of high-dose TN radiosurgery, or repeated procedures. To date, no brainstem or cerebellar ischemic event has been described after TN radiosurgery. Yet, this potential complication remains a concern due to the delayed and ongoing nature of radiation injury to blood vessels and the low number of patients who have been followed for more than 5 years after the procedure. Conclusions Radiosurgery is an ablative procedure that provides most patients suffering from TN with durable pain relief if accompanied by the development of some degree of new trigeminal dysfunction. In our experience, a dose prescription of 90 Gy was associated with an unacceptably high rate of bothersome dysesthesias and should not be used for treatment of this disorder. Given published data, MVD should continue to be the procedure of choice for medically fit patients who suffer from idiopathic TN. References 1. Apfelbaum RI: Neurovascular decompression: the procedure of choice? Clin Neurosurg 46: , Barker FG II, Jannetta PJ, Bissonette DJ, et al: The long-term outcome of microvascular decompression for trigeminal neuralgia. N Engl J Med 334: , Bederson JB, Wilson CB: Evaluation of microvascular decompression and partial sensory rhizotomy in 252 cases of trigeminal neuralgia. J Neurosurg 71: , Brisman R: Gamma knife radiosurgery for primary management for trigeminal neuralgia. J Neurosurg 93 (Suppl 3): , Brisman R, Mooij R: Gamma knife radiosurgery for trigeminal neuralgia: dose-volume histograms of the brainstem and trigeminal nerve. J Neurosurg 93 (Suppl 3): , Brown JA: Percutaneous treatment of trigeminal neuralgia: advances and problems. Clin Neurosurg 46: , Brown JA, McDaniel MD, Weaver MT: Percutaneous trigeminal nerve compression for treatment of trigeminal neuralgia: results in 50 patients. Neurosurgery 32: , Burchiel KJ, Clarke H, Haglund M, et al: Long-term efficacy of microvascular decompression in trigeminal neuralgia. J Neurosurg 69:35 38, Chang TMW, Cascino TL, Onofrio BM: Comprehensive study of diagnosis and treatment of trigeminal neuralgia secondary to tumors. Neurology 43: , Dandy WE: Concerning the cause of trigeminal neuralgia. Am J Surg 24: , Fardy MJ, Zakrzewska JM, Patton DW: Peripheral surgical techniques for the management of trigeminal neuralgia alcohol and glycerol injections. Acta Neurochir 129: , Fields HL: Treatment of trigeminal neuralgia. N Engl J Med 334: , Hutchins LG, Harnsberger HR, Jacobs JM, et al: Trigeminal neuralgia (tic douloureux): MR imaging assessment. Radiology 175: , Jannetta PJ: Arterial compression of the trigeminal nerve at the pons in patients with trigeminal neuralgia. J Neurosurg 26 (Suppl): , Kanpolat Y, Savas A, Bekar A, et al: Percutaneous controlled radiofrequency trigeminal rhizotomy for the treatment of idiopathic trigeminal neuralgia: 25-year experience in 1,600 patients. Neurosurgery 48: , Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53: , Kondziolka D, Lacomis D, Niranjan A, et al: Histological effects of trigeminal nerve radiosurgery in a primate model: implications for trigeminal neuralgia radiosurgery. Neurosurgery 46: , Kondziolka D, Lunsford LD, Flickinger JC, et al: Stereotactic radiosurgery for trigeminal neuralgia: a multiinstutional study using the gamma unit. J Neurosurg 84: , Kondziolka D, Perez B, Flickinger JC, et al: Gamma knife radiosurgery for trigeminal neuralgia: results and expectations. Arch Neurol 55: , Leksell L: Stereotactic radiosurgery. J Neurol Neurosurg Psychiatry 46: , Maesawa S, Salame C, Flickinger JC, et al: Clinical outcomes after stereotactic radiosurgery for idiopathic trigeminal neuralgia. J Neurosurg 94:14 20, Maher CO, Pollock BE: Radiation induced vascular injury after stereotactic radiosurgery for trigeminal neuralgia: case report. Surg Neurol 54: , Nicol B, Regine WF, Courtney C, et al: Gamma knife radiosurgery using 90 Gy for trigeminal neuralgia. J Neurosurg 93 (Suppl 3): , North RB, Kidd DH, Piantadosi S, et al: Percutaneous retrogasserian glycerol rhizotomy. Predictors of success and failure in 352

7 Radiosurgery for trigeminal neuralgia the treatment of trigeminal neuralgia. J Neurosurg 72: , Pollock BE, Foote RL, Stafford SL, et al: Results of repeated gamma knife radiosurgery for medically unresponsive trigeminal neuralgia. J Neurosurg 93 (Suppl 3): , Pollock BE, Phuong LK, Foote RL, et al: High-dose trigeminal neuralgia radiosurgery associated with increased risk of trigeminal nerve dysfunction. Neurosurgery 49:58 64, Rand RW, Jacques DB, Melbye RW, et al: Leksell Gamma Knife treatment of tic douloureux. Stereotact Funct Neurosurg 61 (Suppl 1):93 102, Régis J, Bartolomei F, Metellus P, et al: Radiosurgery for trigeminal neuralgia and epilepsy. Neurosurg Clin N Am 10: , Rogers CL, Shetter AG, Fiedler JA, et al: Gamma Knife radiosurgery for trigeminal neuralgia: the initial experience of The Barrow Neurological Institute. Int J Radiat Oncol Biol Phys 47: , Taha JM, Tew JM Jr, Buncher CR: A prospective 15-year follow up of 154 consecutive patients with trigeminal neuralgia treated by percutaneous stereotactic radiofrequency thermal rhizotomy. J Neurosurg 83: , Urgosik D, Vymazal J, Vladyka V, et al: Gamma knife treatment of trigeminal neuralgia: clinical and electrophysiological study. Stereotact Funct Neurosurg 70 (Suppl 1): , Young RF: Glycerol rhizolysis for treatment of trigeminal neuralgia. J Neurosurg 69:39 45, Young RF, Jacques DS, Mark R, et al: Gamma Knife radiosurgery for treatment of trigeminal neuralgia: long-term results. Neurosurgery 49: , 2001 (Abstract) 34. Young RF, Vermeulen SS, Grimm P, et al: Gamma Knife radiosurgery for treatment of trigeminal neuralgia: idiopathic and tumor related. Neurology 48: , Wepsic JG: Tic douloureux: etiology, refined treatment. N Engl J Med 288: , 1973 Manuscript received October 16, Accepted in final form April 29, Address reprint requests to: Bruce E. Pollock, M.D., Department of Neurological Surgery, Mayo Clinic, 200 First Street SW, Rochester, Minnesota pollock.bruce@mayo.edu. 353