Microvascular decompression (MVD) is one of the prime surgical approaches in intractable trigeminal neuralgia with an 80% success rate at one year after surgery. This presentation discusses the possibility of better patient selection based on advanced imaging techniques and discusses surgical strategies for those cases in which bona fide arterial compression is not found during exploration. Different forms of selective rhizotomy in such cases have not been subject to randomized studies. The approach to venous compression is also not standardized and subject to variations in venous anatomy. A personal view of surgical strategies in these situations is presented, based on personal experience and review of the literature. MVD for trigeminal neuralgia was popularized by Jannetta in the late sixties. He claimed that in the vast majority of cases vascular compression of the trigeminal root would be found, implying that a negative exploration was due to surgical inexperience [1]. There are certainly anatomical causes for an incomplete exposure of the trigeminal root such as an intervening petrosal vein complex or endostosis of the petrous bone that must be dealt with to allow for a complete inspection of the root. We are however aware of other etiologies of trigeminal neuralgia such as CPA mass lesions, bony distortions of the posterior fossa, and MS. Once the MRI was introduced a significant number of cases were found in which no vascular compression could be identified Moreover there are many cases of patients with trigeminal vascular contacts that do not have trigeminal neuralgia. The questions I should like to discuss are: 1. Can we improve the success rate of MVD for trigeminal neuralgia by better patient selection (approx. 80% at 1 year)? 2. Assuming a pathophysiological understanding of trigeminal neuralgia, how do we deal with cases without bona fide arterial root compression during exploration? 3. How should we relate to venous trigeminal root contacts? In 1994 Marshall Devor and I introduced the ignition hypothesis to explain the pathophysiology of trigeminal neuralgia [2]. The basic idea was that damage to the nerve root or ganglion could lead to spontaneous ectopic discharges from the trigeminal root sensory neurons. This activity would be synchronized and amplified by electrochemical cross-talk and after-discharge between damaged axons in the root. Biopsy specimen taken from compressed areas during MVD surgery showed a significant destruction and distortion of axonal myelin sheaths on EM. Biopsy specimens from roots without vascular compression were found to be virtually normal [3]. We speculated that in patients without vascular compression of the root, a different disease (post-infectious, vascular permeability, infarct) process is present in the ganglion. Additionally a genetic predisposition to this phenomenon may exist. The ignition hypothesis posits a similar pathophysiology in both situations. While patients with bona fide vascular compression will benefit from decompressing the mechanosensitive area of root damage, patients with ganglion pathology will need to undergo some form of rhizotomy or gangliotomy to cull hyperactive neurons and stop the pain [4]. We should therefore distinguish between these 2 groups of patients, those with significant vascular compression in which a MVD operation will have a high chance of success and those whose trigeminal neuralgia is from a different pathology and who can undergo a rhizolysis without a major surgical intervention. The question therefore arises what is the actual percentage of trigeminal neuralgia patients without bona fide vascular compression? I would not take outcome figures of MVD series in the literature as an indicator. With vigorous manipulation of the nerve root while looking for an offending vessel a partial rhizotomy effect may be achieved. Mere contact of an artery with a normal looking trigeminal root may be a red herring as well as an innocuous venous contact. In the literature we find a range of negative exploration from 3 to 20%. The higher percentage may be due to inclusion of vascular contacts that do not distort the nerve root [5-8]. Recent progress in MRI imaging has allowed us to substantially increase or diagnostic accuracy. A combination of 3-D TOF MRA with T2 DRIVE and phase contrast MRV allows for creating a 3-D model of the trigeminal root entry zone on a work station or computerized navigator with high sensitivity and specificity as correlated with actual surgical findings. According to Shimizu the sensitivity of this protocol (i.e. actually identifying a compressing vessel) varied from 84% for SCA to 50% for AICA, and 75% for veins. The specificity (i.e. correct identification of no compression) was 100% for SCA, 93% for AICA, and 79% for veins [9]. Sindou's group from Lyon had a sensitivity of 97% and a specificity of 100% using Ti 3D Gadolinium for the venous component. These figures are from using a 1.5 T machine [10]. Shimizu showed that a 3-T MRI will be able to pick up even small arterial branches not detected by the weaker magnet. In a recent paper by Lee et al [11], 29% of patients with type 1 trigeminal neuralgia had no vascular contact using advanced MRI techniques. I would argue that we need to rethink Jannetta's carte blanche for open explorations. If high quality 3-D MRI protocols fail to demonstrate a significant vascular contact, lesser invasive techniques of rhizolysis may give similar results to open exploration without the risks of posterior fossa surgery. What are the options when we are faced with a negative exploration of the trigeminal root? Techniques that have been described include: Arachnoidal dissection around the trigeminal root with 80% good results if significant adhesions found [6]; Partial sensory rhizotomy at the pontine entrance of the root [12]. With this technique a 17% failure rate was described at 1 year, with a 2.6% failure rate yearly therafter. Two-thirds of patients had a significant sensory deficit; Nerve root compression at the pons using a bipolar forceps [7], with a 43.2 % failure rate at 4 years and a 57% incidence of short term hypesthesia; Intraneural rhizolysis ("combing") [13]. With this technique a 18% failure rate and 17% post-operative numbness was reported. Nerve combing was introduced by Ma and Li into the English language literature in 2009 [14], having been first described in China in 1995. A microneedle is used to longitudinally split the trigeminal root into several fascicles. It has been suggested that abnormal excitatory interaxonal connections are thereby disrupted based on our ignition hypothesis that I previously described. I've used most of these techniques with varying degrees of success, applying knowledge of the somatotopy of the root as it relates to the clinical pain distribution. My feeling is that in all the discussed techniques, we are speaking about different methods of performing a partial rhizotomy and that there is not much to distinguish between them. The results are not dissimilar from percutaneous rhizotomy with a lower success rate than classic MVD, more recurrences over time, and occasionally troubling hypesthesia and dysesthesia Controlled comparative studies of these techniques are lacking and would be most welcome. The topic of venous compression is of special interest as even high quality MRI protocols may not be able to cull those cases where venous decompression will make a difference. Venous contact with the trigeminal root has been reported in 20-27% of explorations, usually together with arterial contacts [5, 15, 161. Venous compression only, most commonly from a transverse pontine vein, is present in 3.3-6.5% of explorations [16]. When encountering a venous compression we are faced with a dilemma Is it safe to coagulate and cut the vein? This will depend on the surgeon's instinct. Are there other potential collateral veins in the vicinity? Can the vein be safely mobilized instead? Matsushima [16] recommends decompression only if a single large vein is in sight. If there is a complex of veins in the vicinity, the offending vein may be safely cut. If the vein indents the root and has caused greyish discoloration it is the likely culprit of the neuralgia. Sindou reported a 67% success rate by coagulating and cutting the vein in this scenario [5]. Lee and Jannetta 151 reported a 25% failure rate of vein coagulation and cutting at 1 year. On reexploration they found regrowth of venules surrounding the root, requiring a partial rhizolysis. Personally I do add one of the rhizolysis procedures when not convinced of significant compressive venous pathology. However just venous decompression may also provide a degree of traumatic root damage due to the increased root manipulations required to deal with the abutting veins.
