Article Text
PDF
Abstract
OBJECTIVES Over the past few years many reports have shown that posteroventral pallidotomy is an effective method for treating advanced cases of Parkinson's disease. The main differences with earlier descriptions were the use of standardised evaluation with new high resolution MRI studies and of single cell microrecording which can electrophysiologically define the sensorimotor portion of the internal globus pallidus (GPi). The present study was performed on a consecutive series of 40 patients with Parkinson's disease who underwent posteroventral pallidotomy to determine localisation discrepancies between the ventriculography based theoretical and the electrophysiologically defined target for posteroventral pallidotomy.
METHODS The tentative location of the posteroventral GPi portion was defined according to the proportional Talairach system. Single cell recording was performed in all patients. The definitive target was chosen according to the feasibility of recording single cells with GPi cell features, including the presence of motor drive and correct identification of the internal capsule and of the optic tract by activity recording and microstimulation.
RESULTS In all 40 patients the electrophysiologically defined sensorimotor portion of the GPi was lesioned, with significantly improved cardinal Parkinson's disease symptoms as well as levodopa induced dyskinesias, without damage to the internal capsule or optic tract. Significant differences between the localisation of the ventriculography based theoretical versus electrophysiological target were found in depth (p<0.0008) and posteriority (p<0.04). No significant differences were found in laterality between both approaches. Difference ranges were 8 mm for laterality, 6.5 mm for depth, and 10 mm for posteriority.
CONCLUSIONS Electrophysiologically defined lesion of GPi for posteroventral pallidotomy, shown to be effective for treating Parkinson's disease, is located at a significantly different site from the ventriculography based theoretical target.
- pallidotomy
- Parkinson's disease
- microrecording
- globus pallidus
Statistics from Altmetric.com
Over the past few years many reports have shown that posteroventral pallidotomy is an effective method for treating advanced cases of Parkinson's disease.1-6 The main differences with earlier descriptions are the use of standardised evaluation methods such as the core assessment program for intracerebral transplantation (CAPIT),7 high resolution MRI targeting, and single cell microrecording within the striatum, which can electrophysiologically define the sensorimotor portion of the internal globus pallidus (GPi).
Commonly, GPi localisation has been performed indirectly, using the anterior commissure-anterior commissure (AC-PC) line as reference, whether determined by ventriculography8, CT9, or MRI.10 However, due to substantial individual AC-PC line variation, target coordinates are expressed as a range, rather than as exact values. In a survey of current practice of pallidotomy within the United States, Favre et al 11 reported the results of a questionnaire sent to different centres, of which 50% performed the procedure with microrecording and 50% without, but neither data on the outcome nor on complications of the procedure were presented. Even though no conclusion can be drawn from such a study, it provoked the controversy over the usefulness and safety of microrecording. Those who perform microrecording argue that it remains the key to successful lesioning and the prevention of complications, helping not only with target site definition but also with the volume and shape of the lesion.12 Those who do not perform microrecording contend that it is no better than the use of macrostimulation to obtain a correctly located pallidotomy and that if it increases the risk of side effects then it should only be used for research purposes.13
The present study was carried out to determine localisation discrepancies between the ventriculography based theoretical and the electrophysiologically defined target within the sensorimotor portion of the GPi in a prospective series of patients who underwent posteroventral pallidotomy.
Methods
A consecutive series of 40 patients who underwent microguided posteroventral pallidotomy was included in the analysis. All patients met UKPDS-BB clinical criteria for idiopathic Parkinson's disease,14 had a positive response to levodopa, and presented motor fluctuations and marked dyskinesias.
Inclusion criteria for the posteroventral pallidotomy programme were basically as follows: lack of sustained benefit from the use of antiparkinsonian medication on activities of daily living (ADL) scores; unified Parkinson's disease rating scale (UPDRS) scores with bradykinesia and rigidity predominating over resting tremor; response to levodopa complicated with severe biphasic or interdose dyskinesias noticeably interfering with the patient's quality of life; markedly asymmetric symptoms; and Hoehn and Yahr score greater than III in the on stage. Patients were evaluated and followed up according to CAPIT. Basal scores were calculated by averaging two preoperative evaluations, and postoperative scores corresponded to 1 year follow up.
Mean (SD) UPDRS motor section and dyskinesia score values
For each case, anatomical and electrophysiological targets were calculated as below.
TENTATIVE TARGET PLANNING (VENTRICULOGRAPHY BASED THEORETICAL TARGET)
Under local anaesthesia a Talairach frame was affixed at four points.15 Ventriculography was used to identify the AC and PC, the AC-PC line, and thalamic height. The proportional Talairach lines were established according to these parameters.4 The tentative target for the initial track was calculated 2–3 mm ahead of the midpoint on the AC-PC line, at a depth equivalent to one quarter of the thalamic height below the AC-PC line (4 to 6 mm) and 18–21 mm lateral to the midline.
MICRORECORDING DEFINITION OF TARGET (ELECTROPHYSIOLOGICAL TARGET)
Microrecording was made by means of a platinum/iridium (80%/20%) microelectrode with an impedance of 0.8–1.2 Megohms measured at 1000 Hz with glass insulation, sheathed inside a 26 gauge stainless steel tube which was itself insulated with polyamide tubing with an outer diameter of 0.625 mm (FHC 14TDS KM). A preamplifier (ARS-3D-1 remote probe, Atlanta, USA) was connected to a differential amplifier, impedance meter, and biphasic pulse generator (ARS-MDA4I, Atlanta, USA). The signal was amplified, isolated, and led to oscilloscopes, to a window discriminator, and to an audio equaliser. The window discriminator had different voltage levels to allow triggering pulses and to count and display firing frequencies. The signal was recorded in a high fidelity videocassette using an analog video/audio recording system (ARS-DC1, Atlanta, USA) for postprocessing analysis. On line recording and raster display were monitored together with an EMG and an accelerometer signal. The number of recording tracks performed at each operation depended on: (1) the ability to identify unequivocally the putamen, GPe, and GPi; (2) the presence of motor drive (enhanced phase response to proprioceptive stimuli in GPi neurons); (3) correct identification of the internal capsule by microstimulation; and (4) correct identification of the optic tract by microstimulation and activity recording after visual stimulation.
STATISTICAL ANALYSIS
Differences were evaluated by multivariate analysis of variance (MANOVA) with post hoc paired ttest. All p values are two tailed. The threshold for significance was p<0.05. In order not to underestimate the range of differences between the ventriculography based theoretical and electrophysiological targets due to the presence of negative and positive gaps, an Sqr (theoretical-electrophysiological)2 was applied to all differences and mean SD range of differences calculated on this value. Difference mode was also calculated to disclose a tendency for a positive or negative value.
Results
CLINICAL OUTCOME
In all 40 patients the electrophysiologically defined sensorimotor portion of the GPi was lesioned, with significant improvement in cardinal Parkinson's disease symptoms as well as in levodopa induced involuntary movements. Forty patients, 25 men and 15 women, completed 1 year postsurgical follow up after posteroventral pallidotomy treatment for Parkinson's disease. Mean age was 57.94 (SD 11.65) years, with a mean disease duration of 13.87 (SD 4.83) years. Mean Hoehn and Yahr score was 3.78 (SD 1.07) in the off state and 3.29 (SD 0.96) in the on state. Mean ADL score was 37.69 (SD 27.8) in off versus 50.25 (SD 24.3) in on. Preoperative and postoperative clinical evaluation scores are displayed in the table. No damage to the optic tract was seen in any patient. Morbidity included transient facial palsy in one patient, transient crural paresis in two, transient dysarthria in one, postoperative delirium in two, seizures in one at 2 months after surgery, postoperative pneumonia in one, subdural haematoma in one, and wound infection in one. Therefore, overall morbidity of the procedure was 14.81%. Although not reported in this series, only one patient out of 75 died from an intracerebral haematoma (1.3%).16
DIFFERENCES IN VENTRICULOGRAPHY BASED THEORETICAL VERSUS ELECTROPHYSIOLOGICAL TARGET
Significant differences between the localisation of the ventriculography based theoretical versus electrophysiological target were found in depth (y; F(1,39)=26.5; p<0. 0008) and anteroposterior locality (z;F(1,38)=4.11;p<0.04). No significant differences were found in laterality (x; F(1,38)=0.25; p<0.6) between both approaches. Difference ranges were 8 mm for laterality, 6.5 mm for depth, and 10 mm for posteriority. Modes for differences in z, y, and x were –2, 1, and 0 respectively (fig 1 A, B, and C).
Bars indicate differences(mm) between the ventriculography based theoretical and electrophysiological target for each patient. (A) Laterality from midline (x), (B) posteriority to AC line (y), and (C) depth from AC-PC line (z). In order not to underestimate the differences because of negative and positive values, Sqr (theoretical-electrophysiological)2 was applied as in the text. Differences in depth show a tendency of theoretical target estimation to be deeper, whereas differences in anteroposterior locality are highly variable, representing target variability rather than theoretical targeting error.
In 30 patients (70%), none of the theoretical x, y, or z coordinates matched the electrophysiological coordinates. In five patients (12.5%), one theoretical coordinate matched the electrophysiological coordinate, and in four patients (10%), two coordinates. In only one patient did the theoretical target match the electrophysiological target (fig 2 A, B, and C).
(A) Raw coordinates in the x axis, (B) y axis, and (C) z axis for each patient, The x, y, and z theoretical coordinates matched the electrophysiological coordinates in only one out of 40 patients.
Discussion
This study highlights three main findings: (1) There is a significant individual variability in the spatial position of an atlas defined GPi target. (2) The electrophysiologically defined localisation for posteroventral pallidotomy, shown to be effective for treating Parkinson's disease, is located at a significantly different site than the ventriculography based theoretical target. (3) Differences in anteroposterior locality were highly variable, whereas in depth the theoretical target was commonly located deeper than the electrophysiological target. This may be interpreted as follows: whereas differences in anteroposterior locality are due to target variability, differences in depth are mainly attributable to a limitation of our theoretical calculation of the target according to the proportional Talairach system. These findings confirm the need for electrophysiological recording not only to improve target selection, but also to prevent optic tract damage that could otherwise have occurred in many of our patients when the ventriculography based target was calculated according to the Talairach proportional system.
There is general agreement within contemporary neurosurgery that the lesion in the posterior part of the GPi is the most effective in treating Parkinson's disease.1-6 Theoretically, the lesion should include the entire sensorimotor territory of the GPi, including the neurons that give rise to the ansa lenticularis and fasciculus lenticularis.17 Careful placement of the lesion in this area is mandatory due to the vital surrounding structures. Whether or not intraoperative microrecording is necessary to perform pallidotomy safely and effectively is still a matter of controversy, despite the excellent outcomes reported by several groups1-6 Opponents to microrecording argue that this technique increases the cost of the procedure and is labour intensive, difficult to perform well, and time consuming, while providing data of little practical use.18 19 Whether or not the use of microrecording results in a higher incidence of complications is a difficult issue to address, as it is troublesome to determine whether described complications occurred during microrecording or lesion performance. In our experience with 75 procedures, the single intracerebral haematoma occurred during lesion performance rather than during microrecording16 and none of the related complications could be attributed to microrecording. Currently, the average number of tracks in our group is three to four, but it has changed with time and our learning curve. Exceptionally, in one patient more than 8 tracks were necessary for target determination, thus prolonging surgery duration to over 5 hours and exposing the patient to a greater risk of complications and fatigue. However, to perform the lesion according to the ventriculography based theoretical target would probably have resulted in a poorer surgical outcome as this patient had a significant difference between the theoretical and electrophysiologically defined targets.
Errors in target selection by ventriculography20 and discrepancies between the electrophysiological and theoretical target when the second is calculated by ventriculography and the Talairach proportional system are greater than those found with an MRI based AC-PC calculation,21 meaning that high resolution MRI based targeting is more accurate. Our results are in agreement with those from Alterman et al 22showing that microrecording provides practical information for functional localisation of the target, altering the theoretical target position. We also agree with Alterman et althat microrecording should not be performed by inexperienced groups. At variance with these authors, so far we refrain from performing macrostimulation because of our confidence in GPi, optic tract and internal capsule microrecording and microstimulation.
On reviewing MRI studies of patients not included in the present series but referred to our centre with previously unsuccessful operations, we saw that patients with lesions that only partially involved GPi, or lesions spreading to the GPe, sustained less benefit than those in whom the lesion had been properly situated and this is in agreement with other previous reports.23 Viteket al published their results on microrecording guided pallidotomy in a series of patients followed up for 18 months that showed no beneficial effect wearing off.1 This is in agreement with the results published by Lang et al,5 but at variance with those presented by Samii et al in a 2 year follow up of posteroventral pallidotomy not guided by microrecording, who reported that the beneficial effects of surgery were not sustained.24 This suggests that performance of the lesion under microelectrode guidance could also have a beneficial effect on the long term effect of surgery. Unfortunately, the only way to resolve the usefulness of microrecording in posteroventral pallidotomy is by means of a prospective randomised double blind evaluation, which has not yet been carried out.