False-Positive and False-Negative Results of Motor Evoked Potential Monitoring During Surgery for Intramedullary Spinal Cord Tumors

False-Positive and False-Negative Results of Motor Evoked Potential Monitoring During Surgery for... Abstract BACKGROUND Motor evoked potential (MEP) recording is used as a method to monitor integrity of the motor system during surgery for intramedullary tumors (IMTs). Reliable sensitivity of the monitoring in predicting functional deterioration has been reported. However, we observed false positives and false negatives in our experience of 250 surgeries of IMTs. OBJECTIVE To delineate specificity and sensitivity of MEP monitoring and to elucidate its limitations and usefulness. METHODS From 2008 to 2011, 58 patients underwent 62 surgeries for IMTs. MEP monitoring was performed in 59 operations using transcranial electrical stimulation. Correlation with changes in muscle strength and locomotion was analyzed. A group undergoing clipping for unruptured aneurysms was compared for elicitation of MEP. RESULTS Of 212 muscles monitored in the 59 operations, MEP was recorded in 150 (71%). Positive MEP warnings, defined as amplitude decrease below 20% of the initial level, occurred in 37 muscles, but 22 of these (59%) did not have postoperative weakness (false positive). Positive predictive value was limited to 0.41. Of 113 muscles with no MEP warnings, 8 muscles developed postoperative weakness (false negative, 7%). Negative predictive value was 0.93. MEP responses were not elicited in 58 muscles (27%). By contrast, during clipping for unruptured aneurysms, MEP was recorded in 216 of 222 muscles (96%). CONCLUSION MEP monitoring has a limitation in predicting postoperative weakness in surgery for IMTs. False-positive and false-negative indices were abundant, with sensitivity and specificity of 0.65 and 0.83 in predicting postoperative weakness. Intraoperative neurophysiological monitoring, Neurosurgical procedures, Sensitivity and specificity, Spinal cord neoplasms ABBREVIATIONS ABBREVIATIONS CI confidence interval IMT intramedullary spinal cord tumor MEP motor evoked potential MRC Medical Research Council NCSS neurosurgical cervical spine scale SEP somatosensory evoked potential UIA unruptured intracranial aneurysm Surgery for intramedullary spinal cord tumors (IMTs) carries risks of damaging the long tracts and the gray matters in the spinal cord. It is judicious to have means to evaluate the integrity of the neural function during surgery. To warn against risks of postoperative neurological deficits, intraoperative neurophysiological monitoring is conducted. Somatosensory evoked potentials (SEPs) have been recorded by averaging potentials in the sensory cortex from the scalp electrode after electrical stimulation of the median or the posterior tibial nerves. It can assess the function of the sensory tract,1 but the role of SEP monitoring during surgery for IMTs is limited because SEPs often disappear at the onset of separation/opening of the posterior median sulcus.2 In addition, the value of SEP monitoring is further limited due to its inability to detect deterioration in the motor tract.3-5 Transcranial electrical motor evoked potentials (MEPs) are obtained by electrical stimulation of the cerebral cortex. Evoked muscle action potentials are recorded in the upper and lower extremities, assessing the integrity of the motor system.6-9 Unlike SEP, summation averaging isnot necessary in MEP recording, and motor tract function can be assessed without delay. Many authors report usefulness of MEP monitoring during surgery for IMTs.10-12 Some authors have recommended abandoning tumor resection when the MEP waveform deteriorates.13 One hundred percent sensitivity of MEP monitoring to detect postoperative motor deficits has been reported in surgery for IMTs.11 However, to date, there is no evidence for efficacy of intraoperative MEP monitoring in preventing postoperative neurological deficits in a series of surgery for IMTs. McCormick reported his results of total removal of ependymomas without using MEP monitoring, in which 3 out of 23 patients (13%) had worsening of neurological symptoms.4 By contrast, in a recently published orthopedic study of 46 cases of ependymoma, with MEP deterioration used as an indicator to discontinue resection, neurological worsening occurred in as many as 15 patients (33%). Total resection could be achieved only in 30 out of the 46 cases in the series.13 We have had false-positive and false-negative results with MEP monitoring while removing IMTs. In our practice, MEP monitoring is not reliable either in terms of specificity or sensitivity, and the claim of 100% sensitivity appears unrealistic. The present study was undertaken to review our own experiences with 62 consecutive surgeries on IMTs, to delineate realistic predictive values as well as specificity and sensitivity of MEP monitoring. METHODS During the period from July 1, 2008 to December 31, 2011, 58 patients underwent 62 surgeries for IMTs at the hospital the authors serve. Medical charts and the recordings of MEPs in all patients were reviewed retrospectively. Sevoflurane-based anesthesia was used in 60 operations, and propofol-based anesthesia was used in 2. Sevoflurane was maintained lower than 3% during surgery. Muscle relaxant (vecuronium) was administered only for the induction and not thereafter. Transcranial electrical MEPs recorded from the muscles were monitored in 59 operations. In 2 patients with pre-existing severe motor deficits, MEP monitoring was abandoned because it did not yield responses. In 1 patient, MEP monitoring was converted to D-wave monitoring, in which the potential is recorded on the surface of the cord, as no responses were elicited in the muscles. The following data were analyzed: incidence of adequate MEP responses, intraoperative changes in MEP waveform, and the changes in muscle power as observed on the day following surgery compared with preoperative status (assessed using Medical Research Council [MRC] grades).14 We set the criteria for positive MEP warning as the amplitude decrease below 20% of the initial control amplitude. When the amplitude diminished transiently and then recovered above the level, the findings were not included in the positive warnings. Ambulation was assessed using the neurosurgical cervical spine scale (NCSS) at discharge (Table 1).15 The NCSS is an official grading system defined by the Japanese Society of Spinal Surgery in 1992; it is a simple and practical grading system for expressing neurological status and has been widely used.16-19 It allocates 5 points each to the upper and lower extremity motor function and 4 points to sensory function or pain. TABLE 1. NCSS Scoring for Lower Extremity Motor Function Total disability (score 1)  Chair bound or bedridden  Severe disability (score 2)  Needs support in walking on flat, and unable to ascend or descend stairways.  Moderate disability (score 3)  Difficulty in walking on flat, and needs support in ascending or descending stairways.  Mild disability (score 4)  No difficulty in walking on flat, but mild difficulty in ascending or descending stairways.  Normal (score 5)  Normal walking, with or without abnormal reflexes.  Total disability (score 1)  Chair bound or bedridden  Severe disability (score 2)  Needs support in walking on flat, and unable to ascend or descend stairways.  Moderate disability (score 3)  Difficulty in walking on flat, and needs support in ascending or descending stairways.  Mild disability (score 4)  No difficulty in walking on flat, but mild difficulty in ascending or descending stairways.  Normal (score 5)  Normal walking, with or without abnormal reflexes.  View Large Transcranial electric stimulation was delivered through corkscrew-type subdermal electrodes placed in the C4 and C5 positions of the 10-20 international electroencephalography system. The stimulation was given as a train of 5 pulses with a duration of 200 μs and stimulation intervals of 2 ms at 200 to 620 V using Multipulse Stimulator D185 (Digitimer Ltd, Hertfordshire, United Kingdom). The MEP was recorded using needle electrodes placed in the muscles innervated by the spinal segments at the level of and distal to the lesion, using either NeuroPack M1 (Nihon Kohden, Tokyo, Japan), or Synax 1100 (NEC Corporation, Tokyo, Japan). To evaluate adverse effect of presence of IMT on elicitation of MEP, we compared the incidence of recordable MEP with a group with unruptured intracranial aneurysms (UIAs) undergoing clipping in the same period. Of 111 cases of UIAs, 108 had intraoperative MEP monitoring. Anesthesia was maintained using sevoflurane in 106 cases and propofol in 2 cases. The anesthetic protocols used were identical to those used in the IMT group. MEP recording was attempted in 222 muscles in the UIA group. To assess the overall sensitivity and specificity of the monitoring and diagnostic value in correlation to the outcome, the following indices were calculated: (1) positive predictive value as the ratio of the number of true positives divided by the number of positives (true and false positives combined), (2) negative predictive value as the ratio of true negatives to the number of negatives (true and false negatives combined), (3) sensitivity as the ratio of number of true positives divided by the summation of true positives and false negatives, and (4) specificity as the ratio of number of true negatives to the summation of true negatives and false positives (Table 2). All patients gave informed consent to this study, which was approved by the institutional review board. TABLE 2. Assessment of Validity of MEP Findings for Predicting Postoperative Weakness   Postoperative weakness  No postoperative weakness  MEP warning: positive  True positive  False positive  MEP warning: negative  False negative  True negative    Postoperative weakness  No postoperative weakness  MEP warning: positive  True positive  False positive  MEP warning: negative  False negative  True negative  FN, false negative; FP, false positive; TN, true negative; TP, true positive. “True positive” means the MEP warning indicated disturbance of motor function and the patient developed a postoperative weakness. “False positive” means the MEP warning indicated disturbance of motor function but the patient did not develop a postoperative weakness. “True negative” means the MEP did not indicate motor dysfunction and the patient did not develop a postoperative weakness. “False negative” means the MEP did not indicate disturbed motor function but the patient developed postoperative weakness. The following properties are calculated to assess validity of the test: Sensitivity = TP/(TP + FN); Specificity = TN/(FP + TN); positive predictive value = TP/(TP + FP); negative predictive value = TN/(TN + FN). View Large RESULTS Intramedullary Spinal Cord Tumors and Lesions The patient group consisted of 30 women and 28 men, age ranging from 11 to 77, with a mean of 48. Four patients had 2 surgeries because of multiple or long lesions. Six patients had tumors in the high cervical level (C1-C2 vertebrae), 27 in the midcervical to cervicothoracic spine (C3-T1 vertebrae) corresponding to the cervical enlargement. Nine tumors were in the thoracic spine (T2-T10 vertebrae) and 1 in the lower thoracic spine (T11-T12 vertebrae) arising from the lumbar enlargement. In some patients, the tumors extended beyond the boundaries; 6 tumors extended from the high cervical to the cervicothoracic level, 1 tumor from the high cervical to the thoracic spine, 4 tumors from the midcervical to the thoracic level, 3 tumors the lumbar enlargement to the conus medullaris, and 1 tumor involved the entire spinal cord (Table 3). The histological diagnoses were as follows: cavernous malformation in 15 patients, ependymoma in 14, hemangioblastoma in 5 (6 operations), subependymoma in 3 (6 operations), pilocytic astrocytoma in 3, fibrillary astrocytoma in 1, anaplastic astrocytoma in 4, glioblastoma in 2, and myelitis in 5 patients. In addition, the series included single cases with ganglioglioma, gangliocytoma, lipoma, schwannoma, malignant peripheral nerve sheath tumor, and sarcoidosis (Table 3). Total removal was achieved in cases with ependymomas, hemangioblastoma, and cavernous malformation in the present series. TABLE 3. The Intramedullary Tumors Classified by Location and Histological Diagnosis Level:     C1-C2 (high cervical)  6   C3-T1 (cervical enlargement)  27   T2-T10 (thoracic)  9   T11-T12 (lumbar enlargement)  1   L1-L2 (conus medullaris)  0   C1-T1 (high cervical–cervical)  6   C1-T10 (high cervical–thoracic)  1   C3-T10 (cervical–thoracic)  4   T11-L2 (thoracic–lumbar enlargement)  3   Holocord (high cervical–conus medullaris)  1  Histology:     Cavernous/venous malformation  15   Ependymoma  14   Subependymoma  3 (6 operations)   Hemangioblastoma  5 (6 operations)   Pilocytic astrocytoma  3   Fibrillary astrocytoma  1   Anaplastic astrocytoma  4   Glioblastoma  2   Ganglioglioma  1   Gangliocytoma  1   Lipoma  1   Schwannoma  1   Malignant peripheral nerve sheath tumor  1   Myelitis  5   Sarcoidosis  1  Level:     C1-C2 (high cervical)  6   C3-T1 (cervical enlargement)  27   T2-T10 (thoracic)  9   T11-T12 (lumbar enlargement)  1   L1-L2 (conus medullaris)  0   C1-T1 (high cervical–cervical)  6   C1-T10 (high cervical–thoracic)  1   C3-T10 (cervical–thoracic)  4   T11-L2 (thoracic–lumbar enlargement)  3   Holocord (high cervical–conus medullaris)  1  Histology:     Cavernous/venous malformation  15   Ependymoma  14   Subependymoma  3 (6 operations)   Hemangioblastoma  5 (6 operations)   Pilocytic astrocytoma  3   Fibrillary astrocytoma  1   Anaplastic astrocytoma  4   Glioblastoma  2   Ganglioglioma  1   Gangliocytoma  1   Lipoma  1   Schwannoma  1   Malignant peripheral nerve sheath tumor  1   Myelitis  5   Sarcoidosis  1  View Large Detection of MEP and Evaluation of Muscle Strength Electrodes for MEP recording were placed in 212 muscles during 59 surgeries performed on 55 patients. Monitoring was performed in 2 muscles in 12 surgeries and in 4 muscles in 47 surgeries. Reproducible MEPs were recorded in all of the monitored muscles in 29 surgeries (29/59 = 49%), and in some muscles in 23 (23/59 = 39%). Strength was assessed before and 1 d after surgery in 208 of the 212 monitored muscles. Muscle strength improved 1 d after surgery in 2 muscles, stayed unchanged in 173, and decreased in 33 (probability of worsening: 0.16, 95% confidence interval [CI]: 0.11-0.21). Patterns of MEP Alterations: Positive and Negative Warnings MEP potentials were obtained in 150 muscles (150/212 = 71%); MEP stayed stable during surgery in 84 muscles (84/150 = 56%), became recordable during surgery after initial absence in 11 (11/150 = 7%), and showed mild decrease during surgery (defined as the amplitude not below 20% of the control at the beginning of procedure) in 3. MEP temporarily decreased during surgery but recovered by the end of the procedure in 15 muscles (15/150 = 10%). MEP diminished during surgery (defined as the amplitude decrease below 20% of the initial amplitude) in 14 muscles (14/150 = 9%). MEP disappeared without recovery in 23 muscles (23/150 = 15%). We defined the findings of MEP monitoring as “positive” when there was disappearance or persistent decline below 20% of the control amplitude at the beginning of procedure, and “negative” when the amplitude stayed above 20% of the control, or recovered above 20% after transient diminution. MEP Changes and Muscle Strength—Predictive Values, Sensitivity, and Specificity of the Monitoring Of the 37 muscles in which MEP change was positive during operations, strength was unaffected in 22 (false positive) and worsened in 15 (true positive), yielding a positive predictive value of 0.41 (15/37, with 95% CI: 0.25-0.58). Of the 113 muscles in which MEP warning was negative (stable during surgery: 84, recovered from initial absence: 11, mild or transient changes: 3+15 = 18), strength remained intact in 105 (true negative) and worsened in 8 (false negative; Table 4). The numbers yielded negative predictive value of 0.93 (105/113, 95% CI: 0.87-0.97). TABLE 4. Correlation of MEP Findings and Postoperative Muscle Strength of the 208 Muscles in the 59 Operations for Intramedullary Tumors/Lesions.   Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  10  48  58  Negative MEP findings:         Stable  6  78  84   Appeared after initial absence  0  11  11   Slight change  0  3  3   Temporarily disappeared or declined to < 20% baseline amplitude with recovery at the end  2  13  15    Subtotal: negative MEP  8  105  113  Positive MEP warnings:   Declined to <20% baseline amplitude  5  9  14   Disappeared  10  13  23    Subtotal: positive MEP  15  22  37    Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  10  48  58  Negative MEP findings:         Stable  6  78  84   Appeared after initial absence  0  11  11   Slight change  0  3  3   Temporarily disappeared or declined to < 20% baseline amplitude with recovery at the end  2  13  15    Subtotal: negative MEP  8  105  113  Positive MEP warnings:   Declined to <20% baseline amplitude  5  9  14   Disappeared  10  13  23    Subtotal: positive MEP  15  22  37  Positive Predictive Value was 0.41 (=15/37) and Negative Predictive Value was 0.93 (=105/113) View Large In the 8 muscles with false-negative MEP monitoring, the postoperative weakness was mild or transient; strength diminished from MRC grade 5 to 4 in 5 muscles. One patient with C1 cavernous malformation developed weakness in the left brachioradialis (from grade 5 to grade 2), and it recovered in several days. One patient with a large C5/C6 cavernous malformation had postoperative weakness in the tibialis anterior from grade 5 to 2, and another patient with a C2 cavernous malformation had postoperative weakness in the brachioradialis from grade 4 to 3. Both patients showed substantial recovery in 1 wk. In 58 muscles, in which no MEPs were detected throughout the procedure, motor strength improved after surgery in 2 of them, stayed comparable in 46, and diminished in 10 (probability of worsening: 0.17, 95% CI: 0.09-0.29). MEP monitoring yielded true-positive warning in 15 muscles, false-positive warning in 22 muscles (false-positive rate 22/37 = 59%), true-negative warning in 105 muscles, false negative in 8 muscles (false-negative rate 8/113 = 7%). The overall sensitivity, the ratio of the number of true positives divided by the summation of true positives and false negatives, was 0.65 in predicting worsening of muscle strength. The overall specificity, the ratio of the number of true negatives divided by the summation of the true negatives and the false positives, was 0.83. MEP Findings and Muscle Strength in Cervical Tumors—Monitoring for Segmental Motor Output (Upper Extremity) and Long-Tract Conduction (Lower Extremity) The tumor involved the cervical enlargement in 39 patients. In these patients, MEP monitoring was attempted in 66 upper extremity muscles and 74 lower extremity muscles. Responses elicited in the upper extremity muscles reflect functional status of the segmental innervation at the level of the tumor. MEP responses and alterations in the lower extremity muscles reflect function of the long-tract passing through the level of the cervical tumor. In the upper extremity, MEP responses were elicited in 58 of the 66 muscles (88%). In correlation with the outcome of strength, the monitoring yielded false-positive warning in 6 muscles, false negative in 4, true positive in 5, and true-negative results in 44 muscles (Table 5). The numbers yielded a positive predictive value of 0.45 (5/11) and a negative predictive value of 0.92 (44/48) for upper extremity weakness. Sensitivity was 0.56 (5/[4+5]) and the specificity was 0.88 (44/[6+44]) in predicting postoperative motor deficit in the cervical segments. TABLE 5. MEP Findings and Postoperative Muscle Strength of the 66 Upper Extremity Muscles in the 39 patients with Tumors Involving the Cervical Enlargement.   Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  0  7  7  Negative MEP findings:         Stable  3  35  38   Appeared after initial absence  0  3  3   Slight change  0  3  3   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  1  3  4    Subtotal: negative MEP  4  44  48  Positive MEP warnings:         Declined to <20% baseline amplitude  2  1  3   Disappeared  3  5  8    Subtotal: positive MEP  5  6  11    Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  0  7  7  Negative MEP findings:         Stable  3  35  38   Appeared after initial absence  0  3  3   Slight change  0  3  3   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  1  3  4    Subtotal: negative MEP  4  44  48  Positive MEP warnings:         Declined to <20% baseline amplitude  2  1  3   Disappeared  3  5  8    Subtotal: positive MEP  5  6  11  Positive Predictive Value was 0.45 (=5/11), and Negative Predictive Value was 0.92 (=44/48) View Large In the lower extremity, MEP responses were elicited in 46 of the 74 muscles (62%). False-positive warning was observed in 7 muscles, false negative in 2, true positive in 8, and true-negative warning in 29 muscles (Table 6). Positive predictive value was 0.53 (8/15) for lower extremity weakness and negative predictive value was 0.94 (29/31). Sensitivity was 0.80 (8/[8+2]), and the specificity was 0.81 (29/[29+7]) for predicting postoperative motor deficit in the lower segments. TABLE 6. MEP Findings and Postoperative Muscle Strength of the 74 Lower Extremity Muscles in the 39 patients With Tumors Involving the Cervical Enlargement.   Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  4  24  28  Negative MEP findings:         Stable  2  19  21   Appeared after initial absence  0  7  7   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  0  3  3    Subtotal: negative MEP  2  29  31  Positive MEP warnings:         Declined to <20% baseline amplitude  2  3  5   Disappeared  6  4  10    Subtotal: positive MEP  8  7  15    Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  4  24  28  Negative MEP findings:         Stable  2  19  21   Appeared after initial absence  0  7  7   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  0  3  3    Subtotal: negative MEP  2  29  31  Positive MEP warnings:         Declined to <20% baseline amplitude  2  3  5   Disappeared  6  4  10    Subtotal: positive MEP  8  7  15  Positive Predictive Value was 0.53 (=8/15) and Negative Predictive Value was 0.94 (=29/31) View Large Outcome in Locomotion Locomotion capability was assessed in the 55 patients using the NCSS scored before surgery, at discharge, and at the 1-yr postoperative visit. Locomotion score of NCSS at discharge was better than its preoperative score in 3 patients. Locomotion scores at discharge were unchanged in 42 patients (71%) and worse in 14 patients (24%). Locomotion 1 yr after surgery was assessed in 48 patients; it was better than preoperative status in 16 patients (33%), unchanged in 27 (56%), and worse in 5 (10%). Comparison was not feasible in 11 patients, 3 of whom underwent reoperation within 1 yr after surgery and 8 of whom did not return for the 1-yr visit. Of the 14 patients whose locomotion score was worse at discharge, it was better 1 yr after surgery as compared to preoperative status in 3 patients, was comparable to the preoperative status in 5, stayed worse in 3, and was unknown in 3. Of the 42 patients whose locomotion was unchanged at discharge, it improved in 1 yr after surgery in 10 patients (24%), stayed unchanged in 22, and worsened in 2. The 3 patients with improved locomotion at discharge showed progressive improvement in the score at 1 yr after surgery. MEP Warnings and Locomotion—False Positives and Negatives There were 18 patients who had positive MEP warnings in the lower extremities. Only 4 of them showed actual worsening of locomotion score (NCSS) at the time of discharge (true positive 4/18 = 22%, false positive 14/18 = 78%). Twenty-two patients did not have any warning of MEP changes, but 3 of them had locomotion worse than preoperative status (false negative 3/22 = 14%, true negative 19/22 = 86%). The results yielded positive predictive value of 0.22, negative predictive value of 0.86, sensitivity of 0.57, and specificity of 0.58. In 19 patients, MEPs were not recordable in any of the lower extremity muscles; however, only 7 of them had locomotion scores that were worse at discharge. Variation of MEP Elicitation in Different Muscles Incidence of recordable MEP for monitoring varied among the muscles. For brachioradialis, 53 out of 62 muscles (85%) had detectable MEPs, whereas in gastrocnemius, 54 of 86 muscles (63%) did have recordable MEPs. Occurrence of recordable MEP was significantly higher in the muscles of the upper extremities (88%, 95% CI: 0.79-0.95) than lower extremities (62%, 95% CI: 0.53-0.71; Table 7). TABLE 7. Incidence of Recordable MEP by Muscles in cases With Intramedullary Tumors/Lesions Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Biceps brachii  5  5  1.0 [NA]  Brachioradialis  62  53  0.85 [0.74-0.93]  Flexor carpi radialis  2  2  1.0 [NA]  Triceps brachii  7  7  1.0 [NA]  Thenar  2  2  1.0 [NA]   Upper extremity subtotal  78  69  0.88 [0.79-0.95]  Gastrocnemius  86  54  0.63 [0.52-0.73]  Quadriceps femoris  34  20  0.59 [0.41-0.75]  Tibialis anterior  10  7  0.70 [0.35-0.93]   Lower extremity subtotal  130  81  0.62 [0.53-0.71]   Total  208  150  0.72 [0.65-0.78]  Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Biceps brachii  5  5  1.0 [NA]  Brachioradialis  62  53  0.85 [0.74-0.93]  Flexor carpi radialis  2  2  1.0 [NA]  Triceps brachii  7  7  1.0 [NA]  Thenar  2  2  1.0 [NA]   Upper extremity subtotal  78  69  0.88 [0.79-0.95]  Gastrocnemius  86  54  0.63 [0.52-0.73]  Quadriceps femoris  34  20  0.59 [0.41-0.75]  Tibialis anterior  10  7  0.70 [0.35-0.93]   Lower extremity subtotal  130  81  0.62 [0.53-0.71]   Total  208  150  0.72 [0.65-0.78]  View Large MEP Elicitation and Preoperative Muscle Strength The incidence of measurable MEPs varied in relation to preoperative muscle weakness. Only 1 out of 4 muscles (25%) with preoperative muscle strength graded 1 or 2 showed reliable MEPs, while 29 out of 46 muscles (63%) with preoperative grade 4 and 117 out of 153 muscles (76%) with grade 5 did show reliable MEPs (Table 8). TABLE 8. Incidence of Measurable MEP and Preoperative Muscle Strength Preoperative muscle strength  Muscles  Measurable MEP  Incidence of recordable MEP [95% CI]  1  1  0  0 [NA]  2  3  1  0.33 [0.01-0.91]  3  5  3  0.60 [0.15-0.95]  4  46  29  0.63 [0.48-0.77]  5  153  117  0.76 [0.69-0.83]  Preoperative muscle strength  Muscles  Measurable MEP  Incidence of recordable MEP [95% CI]  1  1  0  0 [NA]  2  3  1  0.33 [0.01-0.91]  3  5  3  0.60 [0.15-0.95]  4  46  29  0.63 [0.48-0.77]  5  153  117  0.76 [0.69-0.83]  View Large Method of General Anesthesia and MEP Elicitation of MEPs was influenced by anesthetic agents. Thirty patients were anesthetized with sevoflurane + air + remifentanil, and 23 patients were anesthetized using sevoflurane + nitrous oxide + fentanyl. Combinations of propofol + remifentanil, sevoflurane + air + fentanyl and sevoflurane + nitrous oxide + remifentanil were used in 2 patients each (Table 9). Statistical significance was not detected regarding the incidence of MEPs among the anesthetic protocols. Sevoflurane was maintained at 1% to 2% in 49 patients, and 3 (6.1%) did not have MEP responses in any muscles. It was maintained at 2.5% to 3% in 8 patients, and 4 of them did not show MEP responses (Table 10). TABLE 9. Incidence of Recordable MEP With Different Anesthetic Protocols. Combination of Sevoflurane, Air, and Remifentanil was Associated With High Chance of Obtaining Recordable MEP     Observed MEP  Anesthetic agent  n  All muscles  Some muscles  No muscle  Propofol + remifentanil  2  2  0  0  Sevoflurane + air + fentanyl  2  1  1  0  Sevoflurane + N2O + fentanyl  23  10  7  6  Sevoflurane + air + remifentanil  30  15  14  1  Sevoflurane + N2O + remifentanil  2  1  1  0      Observed MEP  Anesthetic agent  n  All muscles  Some muscles  No muscle  Propofol + remifentanil  2  2  0  0  Sevoflurane + air + fentanyl  2  1  1  0  Sevoflurane + N2O + fentanyl  23  10  7  6  Sevoflurane + air + remifentanil  30  15  14  1  Sevoflurane + N2O + remifentanil  2  1  1  0  View Large TABLE 10. Sevoflurane Maintenance Dose and MEP Response. Statistical Significance was not Detected Regarding the Incidence of MEPs Among the Anesthetic Protocols     Observed MEP  Sevoflurane (%)  n  All muscles  Some muscles  No muscles  1  23  13  9  1  1.5  18  8  8  2  2  8  5  3  0  2.5  6  0  3  3  3  2  1  0  1      Observed MEP  Sevoflurane (%)  n  All muscles  Some muscles  No muscles  1  23  13  9  1  1.5  18  8  8  2  2  8  5  3  0  2.5  6  0  3  3  3  2  1  0  1  View Large MEP Responses in Patients Undergoing Clipping for UIAs Intraoperative MEP monitoring was performed in 108 patients with unruptured aneurysms during clipping operation (73 women and 35 men, age ranging from 33 to 82 yr, mean age 62.8 yr). Electrodes were placed in 2 muscles in 99 patients, 3 muscles in 6, 1 muscle in 2, and 4 muscles in 1 patient. A total of 104 patients (96%) showed MEP responses in all of the monitored muscles. In 2 patients, no responses were obtained. In total, MEPs were elicited in 216 out of 222 muscles (97%) monitored. The responses were obtained in 84 of 85 upper extremity muscles (99%) and 132 of 137 lower extremity muscles (96%; Table 11). TABLE 11. Incidence of Recordable MEP in Patients With Unruptured Intracranial Aneurysms. MEP was Observed in Almost All Muscles Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Brachioradialis  83  82  0.99 [0.93-1.00]  Thenar  2  2  1.0 [NA]  Upper extremity total  85  84  0.99 [0.94-1.00]  Tibialis anterior  137  132  0.96 [0.92-0.99]  Total  222  216  0.97 [0.94-0.99]  Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Brachioradialis  83  82  0.99 [0.93-1.00]  Thenar  2  2  1.0 [NA]  Upper extremity total  85  84  0.99 [0.94-1.00]  Tibialis anterior  137  132  0.96 [0.92-0.99]  Total  222  216  0.97 [0.94-0.99]  View Large DISCUSSION Through the analysis of MEP monitoring in the 212 muscles during 59 operations for 55 patients with intramedullary tumors, we found a high false-positive rate. Of 37 muscles with positive MEP warning during operation, 22 muscles did not show deterioration in the strength. The false-positive rate was 59%. The positive predictive value was limited to 0.41. In the muscles with no warning during operation, 93% (105/113) remained unaffected in strength. The false-negative rate was 7% and the negative predictive value was 0.93. Previous studies on MEP monitoring during surgery for IMTs have reported fewer false-positive rates (9.6%-23.8%) and false-negative rates to be none or very rare.11,12,20,21 Our findings contradict these descriptions. This discrepancy in findings from us and others is partly explained by a difference in definition of positive MEP warnings and method of assessing postoperative weakness. Criteria for MEP warning was defined as 50% decrease in amplitude,20 or changes in either voltage stimulation, duration, or complexity of MEP waveform, or loss of detectable wave.11,12 Postoperative neurological function was assessed using the McCormick scale or similar functional grading,11,20,21 or MRC strength grading.12 Studies that reported 0% false-negative cases assessed neurological functions using functional scales such as the McCormick scale that divides neurological status into normal, slightly paretic, severely paretic, and plegic. Some studies did not specify timing of postoperative assessment.11,20,21 One study exclusively included patients with normal or slightly paretic preoperative neurological status, and found no patients suffering prolonged neurological deterioration. They reported 100% sensitivity, with no false negatives.11 The study is biased and may be misleading because appearance of MEP is affected by the preoperative strength as demonstrated in the present study, and the outcome is influenced by the preoperative strength.4,22 Our series included patients with various degrees of preoperative motor weakness. We assessed muscle strength in MRC grading on postoperative day 1 and classified mild or transient weakness as postoperative weakness. This may have contributed to more false negatives and lower sensitivity in our series than in the previous studies. As for the MEP threshold for significant decline, we chose 20%, because a 50% decrease was commonly observed without neurological sequelae, but we wanted to be warned before MEP became flatline. When the MEP threshold was experimentally changed to 50% decrease, the number of muscles with MEP warning increased to 45, of which postoperative weakness developed in 16, and those without MEP warning decreased to 105, of which weakness occurred in 7. False-positive results, that is, positive MEP warning but no postoperative weakness, increased to 29 from 22 with 20% threshold. This resulted in sensitivity of 0.70, specificity of 0.77, positive predictive value of 0.36, and negative predictive value of 0.93. In the cervical spine tumors, the sensitivity was lower for the upper extremity monitoring (0.56) than that for the lower extremity monitoring (0.80), due to larger false negatives for the upper extremity muscles, suggesting MEP output via the segment involved by the tumor is less susceptible to surgical intrusion. This could be due to mixed, overlapping innervation of the upper extremity over a few segments. For example, an alteration limited to 1 segment, such as C6 or C7 alone, may not be reflected as a significant alteration in the MEP recorded from brachioradialis. Chances for false-negative results are increased, and sensitivity (true positive/[true positive + false negative]) would decrease. The MEP amplitude change is a better or more straightforward indicator for disturbances in the downstream conduction in the corticospinal tract than for alteration in the segmental motor function. Thus, when monitoring in cases of cervical tumor, it may be useful to be aware of the variance in false-negative MEP warnings between the upper and lower extremity muscles. MEP with transcranial stimulation and recording without averaging is primarily a method to monitor integrity of the monosynaptic pathway of corticospinal neurons and anterior horn motor neurons.23 However, detailed neurophysiological and anatomic studies reveal that the upper neurons of the connection scheme occupy only 5% to 10% of the entire neuronal population in the pyramidal tract.8,24 The descending neurons from the motor cortex are known to have variable connection patterns to the anterior horn cell, with a great majority connecting via the interneurons and propriospinal neurons.25-28 These multisynaptic pathways also contribute to the motions of the extremities, but MEP does not reflect alterations in the functions of such pathways. Some authors advocated enhancing sensitivity by adjusting the threshold amplitude for positive warning,29 but we would not consider this to be an authentic measure, in view of the inherent variability of the MEP monitoring and limitations to represent the anatomic substrates participating in the voluntary movement. In our series, the overall incidence of obtaining MEP in the muscles monitored was 71%, and in the cases of cervical tumors, the chance of getting MEP in the upper extremity was 88% and in the lower extremity was 62%. This is in accordance with previous studies, which found that MEP is not reliably elicited in patients with intramedullary tumors.30,31 The variability in the responsiveness to the monitoring indicates an inherent limitation in the reliability as a test in the pathological condition. Sevoflurane, at higher doses, is known to affect MEPs adversely.32 We routinely maintained sevoflurane lower than 3% during surgery. At this maintenance dose, MEP was recorded in 97% of muscles in our patients undergoing craniotomy for UIAs, and 96% of the patients with UIAs had MEP responses. Therefore, we would not consider the anesthetic protocol to be the cause of the higher false-positive and -negative rates observed in our series. MEP monitoring did not predict outcome of gait function at the time of discharge or 1 yr after surgery. In fact, MEP warnings in the lower extremity muscles had positive predictive value of 0.22 in predicting gait dysfunction at the time of discharge. Sensitivity and specificity were limited to 0.57 and 0.58. This may be due to the frequent false-positive MEP warnings for locomotion (78%) or less reproducibility in eliciting MEP in the lower extremity muscles (62% as opposed to 88% for the upper extremity muscles). The latter view is shared by other authors.30,31 Normal locomotion requires coordination of motor and sensory function, which may involve the tectospinal, rubrospinal, reticulospinal, and vestibulospinal tracts in addition to the corticospinal tract and the dorsal column/medial lemniscus. MEPs mainly reflect corticospinal tract function,6,33 and may not correlate with ambulation as a whole. Concurrent monitoring of SEP enables assessment of dorsal column function,34-36 but its role in IMT surgeries is limited, because SEP often disappears at the beginning of opening the dorsomedian sulcus to reach the intramedullary pathology, and SEP change does not correlate with postoperative weakness or loss of the dorsal column function.36-39 Therefore, we use only MEP as the electrophysiological monitoring during surgery of intramedullary tumors. When we face positive MEP warning during surgery for IMTs, we check for the possibility of dysfunction caused by surgical maneuvers, and then check for depth of anesthesia, alterations in body temperature, and blood pressure. When MEP decline occurred after rotating or retracting the spinal cord, possibly obstructing the blood flow to the cord, release of such tractions often results in recovery of MEPs. When MEP amplitude decreased during resection of the intramedullary tumor, we would suspend the maneuver while loosening the spinal cord rotation or pial retraction or by changing the place of microdissection from one part of the tumor to another. Often, these measures resulted in recovery of the amplitude. We did not administer steroids. We usually pursued total removal of the tumor as far as there was a good plane of dissection and as far as the thickness of the adjoining spinal cord parenchyma, typically the lateral funiculus, was maintained. Total resection of benign tumors such as ependymoma can accomplish a cure, and, when executed with utmost precision, it carries relatively small risks of debilitating neurological deficits.22 Premature discontinuation of removal following MEP warning, as has been described by an orthopedic group,13 would result in regrowth and progressive neurological worsening. A second surgery, as to be warranted, would be complicated with adhesive changes, posing more of a technical challenge and carrying greater risk. Tumor histology is an important factor in achieving total resection.40 Therefore, we value appropriate intraoperative histological diagnosis and the feasibility of dissection with a plane. They are vital for decision making regarding the extent of the resection of the tumor. It is certainly essential to preserve the anatomic integrity of the spinal cord tissue and its vascular supplies with utmost care, and this often requires maximum magnification for the surgical microscope. As a result, we achieved total removal in all cases of ependymomas, hemangioblastomas, and cavernous malformations in the present series. Of 14 patients with ependymomas included in the present study, 21% had worsening of lower extremity function at discharge, but all patients recovered 1 yr after surgery to the same or better grade than before surgery. We could not obtain measurable MEPs in the lower extremities in 19 patients. The rate of MEP elicitation is lower in patients with weakness.32 There has been no recommendation as to surgical strategy for those without detectable baseline MEPs. However, gait function after surgery was the same or better than before surgery in 12 out of the 19 patients (64%) with no baseline potentials. Gait was worse at the time of discharge in 7 patients, of whom 1 improved better than before surgery and 2 recovered to the same grade as preoperative state at 1-yr follow-up. We believe that attempt of total removal of the intramedullary tumors is justifiable even when MEPs cannot be elicited, if the histology and the anatomic findings indicate feasibility of resection, in view of the temporary and limited nature of the postoperative weakness in the majority of such cases.22 CONCLUSION MEP monitoring is by no means the perfect test in predicting functional outcome, with frequent false-positive and false-negative warnings, and with limited sensitivity and specificity. It should be utilized as an aid to detect intraoperative alterations in the motor signal conduction, and as one source of supportive information together with other findings, such as plane of cleavage, thickness of the remaining cord parenchyma, and histological diagnosis, in executing safe and effective resection of IMT. Disclosures Kazushige Itoki has received funding from the National Institutes of Health, Wellcome Trust/COAF, Howard Hughes Medical Institute, Australian Science Fund, Bill & Melinda Gates Foundation, World Bank, Research Councils UK, and the Department of Neurosurgery, Dokkyo University School of Medicine. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Part of the material was presented in the 12th Annual Meeting of the Japanese Society of Intraoperative Imaging on July 7, 2012, in Tsukuba, Japan, as an oral presentation. REFERENCES 1. Nash CL Jr, Lorig RA, Schatzinger LA, Brown RH. Spinal cord monitoring during operative treatment of the spine. Clin Orthop Relat Res . 1977; 126: 100- 105. 2. Whittle IR, Johnston IH, Besser M. Recording of spinal somatosensory evoked potentials for intraoperative spinal cord monitoring. J Neurosurg . 1986; 64( 4): 601- 612. Google Scholar CrossRef Search ADS PubMed  3. Kearse LA Jr, Lopez-Bresnahan M, McPeck K, Tambe V. Loss of somatosensory evoked potentials during intramedullary spinal cord surgery predicts postoperative neurologic deficits in motor function. J Clin Anesth . 1993; 5( 5): 392- 398. Google Scholar CrossRef Search ADS PubMed  4. McCormick PC, Torres R, Post KD, Stein BM. Intramedullary ependymoma of the spinal cord. J Neurosurg . 1990; 72( 4): 523- 532. Google Scholar CrossRef Search ADS PubMed  5. Seyal M, Mull B. Mechanisms of signal change during intraoperative somatosensory evoked potential monitoring of the spinal cord. J Clin Neurophysiol . 2002; 19( 5): 409- 415. Google Scholar CrossRef Search ADS PubMed  6. Merton PA, Morton HB. Stimulation of the cerebral cortex in the intact human subject. Nature . 1980; 285( 5762): 227. Google Scholar CrossRef Search ADS PubMed  7. Levy WJ Jr. Clinical experience with motor and cerebellar evoked potential monitoring. Neurosurgery  1987; 20 (1): 169- 182. Google Scholar PubMed  8. Taniguchi M, Cedzich C, Schramm J. Modification of cortical stimulation for motor evoked potentials under general anesthesia: technical description. Neurosurgery . 1993; 32( 2): 219- 226. Google Scholar CrossRef Search ADS PubMed  9. Calancie B, Harris W, Broton JG, Alexeeva N, Green BA. "Threshold-level" multipulse transcranial electrical stimulation of motor cortex for intraoperative monitoring of spinal motor tracts: description of method and comparison to somatosensory evoked potential monitoring. J Neurosurg . 1998; 88( 3): 457- 470. Google Scholar CrossRef Search ADS PubMed  10. Morota N, Deletis V, Constantini S, Kofler M, Cohen H, Epstein FJ. The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery . 1997; 41( 6): 1327- 1336. Google Scholar CrossRef Search ADS PubMed  11. Kothbauer KF, Deletis V, Epstein FJ. Motor-evoked potential monitoring for intramedullary spinal cord tumor surgery: correlation of clinical and neurophysiological data in a series of 100 consecutive procedures. Neurosurg Focus . 1998; 4( 5): e1. Google Scholar CrossRef Search ADS PubMed  12. Quiñones-Hinojosa A, Lyon R, Zada G et al.   Changes in transcranial motor evoked potentials during intramedullary spinal cord tumor resection correlate with postoperative motor function. Neurosurgery . 2005; 56( 5): 982- 993. Google Scholar PubMed  13. Matsuyama Y, Sakai Y, Katayama Y et al.   Surgical results of intramedullary spinal cord tumor with spinal cord monitoring to guide extent of resection. J Neurosurg Spine . 2009; 10( 5): 404- 413. Google Scholar CrossRef Search ADS PubMed  14. The Guarantors of Brain. Aids to the Examination of the Peripheral Nervous System . 5th ed. London, UK: Saunders Elsevier; 2010. 15. Kadoya S. Grading and scoring system for neurological function in degenerative cervical spine disease—neurosurgical cervical spine scale. Neurol Med Chir (Tokyo) . 1992; 32( 1): 40- 41. Google Scholar CrossRef Search ADS PubMed  16. Kim P, Wakai S, Matsuo S, Moriyama T, Kirino T. Bisegmental cervical interbody fusion using hydroxyapatite implants: surgical results and long-term observation in 70 cases. J Neurosurg . 1998; 88( 1): 21- 27. Google Scholar CrossRef Search ADS PubMed  17. Koyanagi I, Iwasaki Y, Hida K, Imamura H, Abe H. Magnetic resonance imaging findings in ossification of the posterior longitudinal ligament of the cervical spine. J Neurosurg . 1998; 88( 2): 247- 254. Google Scholar CrossRef Search ADS PubMed  18. Hida K, Iwasaki Y, Yano S, Akino M, Seki T. Long-term follow-up results in patients with cervical disk disease treated by cervical anterior fusion using titanium cage implants. Neurol Med Chir (Tokyo) . 2008; 48( 10): 440- 446. Google Scholar CrossRef Search ADS PubMed  19. Bucciero A, Zorzi T, Piscopo GA. Peek cage-assisted anterior cervical discectomy and fusion at four levels: clinical and radiographic results. J Neurosurg Sci . 2008; 52( 2): 37- 40. Google Scholar PubMed  20. Zentner J. Noninvasive motor evoked potential monitoring during neurosurgical operations on the spinal cord. Neurosurgery . 1989; 24( 5): 709- 712. Google Scholar CrossRef Search ADS PubMed  21. Costa P, Peretta P, Faccani G. Relevance of intraoperative D wave in spine and spinal cord surgeries. Eur Spine J . 2013; 22( 4): 840- 848. Google Scholar CrossRef Search ADS PubMed  22. Klekamp J. Treatment of intramedullary tumors: analysis of surgical morbidity and long-term results. J Neurosurg Spine . 2013; 19( 1): 12- 26. Google Scholar CrossRef Search ADS PubMed  23. Pechstein U, Cedzich C, Nadstawek J, Schramm J. Transcranial high-frequency repetitive electrical stimulation for recording myogenic motor evoked potentials with the patient under general anesthesia. Neurosurgery . 1996; 39( 2): 335- 343. Google Scholar CrossRef Search ADS PubMed  24. York DH. Review of descending motor pathways involved with transcranial stimulation. Neurosurgery . 1987; 20( 1): 70- 73. Google Scholar CrossRef Search ADS PubMed  25. Ralston DD, Ralston HJ 3rd. The terminations of corticospinal tract axons in the macaque monkey. J Comp Neurol . 1985; 242( 3): 325- 337. Google Scholar CrossRef Search ADS PubMed  26. Nathan PW, Smith M, Deacon P. Vestibulospinal, reticulospinal and descending propriospinal nerve fibres in man. Brain . 1996; 119( pt 6): 1809- 1833. Google Scholar CrossRef Search ADS PubMed  27. Chakrabarty S, Shulman B, Martin JH. Activity-dependent codevelopment of the corticospinal system and target interneurons in the cervical spinal cord. J Neurosci . 2009; 29( 27): 8816- 8827. Google Scholar CrossRef Search ADS PubMed  28. Martin JH. Neuroanatomy: Text and Atlas . 4th ed. New York: McGraw-Hill; 2012. 29. Muramoto A, Imagama S, Ito Z et al.   The cutoff amplitude of transcranial motor evoked potentials for transient postoperative motor deficits in intramedullary spinal cord tumor surgery. Spine (Phila Pa 1976) . 2014; 39( 18): E1086- E1094. Google Scholar CrossRef Search ADS PubMed  30. Rajshekhar V, Velayutham P, Joseph M, Babu KS. Factors predicting the feasibility of monitoring lower-limb muscle motor evoked potentials in patients undergoing excision of spinal cord tumors. J Neurosurg Spine . 2011; 14( 6): 748- 753. Google Scholar CrossRef Search ADS PubMed  31. Chen X, Sterio D, Ming X et al.   Success rate of motor evoked potentials for intraoperative neurophysiologic monitoring: effects of age, lesion location, and preoperative neurologic deficits. J Clin Neurophysiol . 2007; 24( 3): 281- 285. Google Scholar CrossRef Search ADS PubMed  32. Hayashi H, Kawaguchi M, Abe R et al.   Evaluation of the applicability of sevoflurane during post-tetanic myogenic motor evoked potential monitoring in patients undergoing spinal surgery. J Anesth . 2009; 23( 2): 175- 181. Google Scholar CrossRef Search ADS PubMed  33. Malhotra NR, Shaffrey CI. Intraoperative electrophysiological monitoring in spine surgery. Spine (Phila Pa 1976) . 2010; 35( 25): 2167- 2179. Google Scholar CrossRef Search ADS PubMed  34. Nagle KJ, Emerson RG, Adams DC et al.   Intraoperative monitoring of motor evoked potentials: a review of 116 cases. Neurology . 1996; 47( 4): 999- 1004. Google Scholar CrossRef Search ADS PubMed  35. Costa P, Bruno A, Bonzanino M et al.   Somatosensory- and motor-evoked potential monitoring during spine and spinal cord surgery. Spinal Cord . 2007; 45( 1): 86- 91. Google Scholar CrossRef Search ADS PubMed  36. Hyun SJ, Rhim SC. Combined motor and somatosensory evoked potential monitoring for intramedullary spinal cord tumor surgery: correlation of clinical and neurophysiological data in 17 consecutive procedures. Br J Neurosurg . 2009; 23( 4): 393- 400. Google Scholar CrossRef Search ADS PubMed  37. Kothbauer K, Deletis V, Epstein FJ. Intraoperative spinal cord monitoring for intramedullary surgery: an essential adjunct. Pediatr Neurosurg . 1997; 26( 5): 247- 254. Google Scholar CrossRef Search ADS PubMed  38. Sandalcioglu IE, Gasser T, Asgari S et al.   Functional outcome after surgical treatment of intramedullary spinal cord tumors: experience with 78 patients. Spinal Cord . 2005; 43( 1): 34- 41. Google Scholar CrossRef Search ADS PubMed  39. Sala F, Palandri G, Basso E et al.   Motor evoked potential monitoring improves outcome after surgery for intramedullary spinal cord tumors: a historical control study. Neurosurgery . 2006; 58( 6): 1129- 1143. Google Scholar CrossRef Search ADS PubMed  40. Karikari IO, Nimjee SM, Hodges TR et al.   Impact of tumor histology on resectability and neurological outcome in primary intramedullary spinal cord tumors: a single-center experience with 102 patients. Neurosurgery  2011; 68( 1): 188- 197. Google Scholar CrossRef Search ADS PubMed  Acknowledgment We thank Ms. Kayoko Iwata for her technical assistance in the intraoperative MEP monitoring throughout the series. COMMENT The authors reported their experience of MEP monitoring in 59 operations of intramedullary tumors. Evoked responses were recorded from 212 muscles of the upper and lower extremities following transcranial electrical stimulation. MEP was recorded in 71% of muscles. Decreased MEP amplitude below 20% of the initial level was observed in 37 muscles, yet 22 of them (59%) did not show postoperative motor weakness of the monitored muscles. The authors stressed the limitation of using MEP in predicting postoperative motor weakness in surgery of intramedullary tumors. The study was carefully designed to clarify the relationship of decreased amplitude of MEP and postoperative weakness of the corresponding muscle. Surgery of intramedullary tumors has a higher chance to insult the limited areas of functional pathways and gray mater neurons in the spinal cord. Muscle movement of extremities is mediated by multiple pathways, not limited to the corticospinal tracts. The authors correctly pointed out the mechanisms underlying false results and issues of intraoperative monitoring. This is an important study for neurosurgeons to understand “the real world” of using MEPs elicited from extremity muscles for intramedullary spinal cord tumor surgeries. Izumi Koyanagi Sapporo, Japan Copyright © 2017 by the Congress of Neurological Surgeons http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Operative Neurosurgery Oxford University Press

False-Positive and False-Negative Results of Motor Evoked Potential Monitoring During Surgery for Intramedullary Spinal Cord Tumors

Loading next page...
 
/lp/ou_press/false-positive-and-false-negative-results-of-motor-evoked-potential-tJUIZylyeh
Publisher
Congress of Neurological Surgeons
Copyright
Copyright © 2017 by the Congress of Neurological Surgeons
ISSN
2332-4252
eISSN
2332-4260
D.O.I.
10.1093/ons/opx113
Publisher site
See Article on Publisher Site

Abstract

Abstract BACKGROUND Motor evoked potential (MEP) recording is used as a method to monitor integrity of the motor system during surgery for intramedullary tumors (IMTs). Reliable sensitivity of the monitoring in predicting functional deterioration has been reported. However, we observed false positives and false negatives in our experience of 250 surgeries of IMTs. OBJECTIVE To delineate specificity and sensitivity of MEP monitoring and to elucidate its limitations and usefulness. METHODS From 2008 to 2011, 58 patients underwent 62 surgeries for IMTs. MEP monitoring was performed in 59 operations using transcranial electrical stimulation. Correlation with changes in muscle strength and locomotion was analyzed. A group undergoing clipping for unruptured aneurysms was compared for elicitation of MEP. RESULTS Of 212 muscles monitored in the 59 operations, MEP was recorded in 150 (71%). Positive MEP warnings, defined as amplitude decrease below 20% of the initial level, occurred in 37 muscles, but 22 of these (59%) did not have postoperative weakness (false positive). Positive predictive value was limited to 0.41. Of 113 muscles with no MEP warnings, 8 muscles developed postoperative weakness (false negative, 7%). Negative predictive value was 0.93. MEP responses were not elicited in 58 muscles (27%). By contrast, during clipping for unruptured aneurysms, MEP was recorded in 216 of 222 muscles (96%). CONCLUSION MEP monitoring has a limitation in predicting postoperative weakness in surgery for IMTs. False-positive and false-negative indices were abundant, with sensitivity and specificity of 0.65 and 0.83 in predicting postoperative weakness. Intraoperative neurophysiological monitoring, Neurosurgical procedures, Sensitivity and specificity, Spinal cord neoplasms ABBREVIATIONS ABBREVIATIONS CI confidence interval IMT intramedullary spinal cord tumor MEP motor evoked potential MRC Medical Research Council NCSS neurosurgical cervical spine scale SEP somatosensory evoked potential UIA unruptured intracranial aneurysm Surgery for intramedullary spinal cord tumors (IMTs) carries risks of damaging the long tracts and the gray matters in the spinal cord. It is judicious to have means to evaluate the integrity of the neural function during surgery. To warn against risks of postoperative neurological deficits, intraoperative neurophysiological monitoring is conducted. Somatosensory evoked potentials (SEPs) have been recorded by averaging potentials in the sensory cortex from the scalp electrode after electrical stimulation of the median or the posterior tibial nerves. It can assess the function of the sensory tract,1 but the role of SEP monitoring during surgery for IMTs is limited because SEPs often disappear at the onset of separation/opening of the posterior median sulcus.2 In addition, the value of SEP monitoring is further limited due to its inability to detect deterioration in the motor tract.3-5 Transcranial electrical motor evoked potentials (MEPs) are obtained by electrical stimulation of the cerebral cortex. Evoked muscle action potentials are recorded in the upper and lower extremities, assessing the integrity of the motor system.6-9 Unlike SEP, summation averaging isnot necessary in MEP recording, and motor tract function can be assessed without delay. Many authors report usefulness of MEP monitoring during surgery for IMTs.10-12 Some authors have recommended abandoning tumor resection when the MEP waveform deteriorates.13 One hundred percent sensitivity of MEP monitoring to detect postoperative motor deficits has been reported in surgery for IMTs.11 However, to date, there is no evidence for efficacy of intraoperative MEP monitoring in preventing postoperative neurological deficits in a series of surgery for IMTs. McCormick reported his results of total removal of ependymomas without using MEP monitoring, in which 3 out of 23 patients (13%) had worsening of neurological symptoms.4 By contrast, in a recently published orthopedic study of 46 cases of ependymoma, with MEP deterioration used as an indicator to discontinue resection, neurological worsening occurred in as many as 15 patients (33%). Total resection could be achieved only in 30 out of the 46 cases in the series.13 We have had false-positive and false-negative results with MEP monitoring while removing IMTs. In our practice, MEP monitoring is not reliable either in terms of specificity or sensitivity, and the claim of 100% sensitivity appears unrealistic. The present study was undertaken to review our own experiences with 62 consecutive surgeries on IMTs, to delineate realistic predictive values as well as specificity and sensitivity of MEP monitoring. METHODS During the period from July 1, 2008 to December 31, 2011, 58 patients underwent 62 surgeries for IMTs at the hospital the authors serve. Medical charts and the recordings of MEPs in all patients were reviewed retrospectively. Sevoflurane-based anesthesia was used in 60 operations, and propofol-based anesthesia was used in 2. Sevoflurane was maintained lower than 3% during surgery. Muscle relaxant (vecuronium) was administered only for the induction and not thereafter. Transcranial electrical MEPs recorded from the muscles were monitored in 59 operations. In 2 patients with pre-existing severe motor deficits, MEP monitoring was abandoned because it did not yield responses. In 1 patient, MEP monitoring was converted to D-wave monitoring, in which the potential is recorded on the surface of the cord, as no responses were elicited in the muscles. The following data were analyzed: incidence of adequate MEP responses, intraoperative changes in MEP waveform, and the changes in muscle power as observed on the day following surgery compared with preoperative status (assessed using Medical Research Council [MRC] grades).14 We set the criteria for positive MEP warning as the amplitude decrease below 20% of the initial control amplitude. When the amplitude diminished transiently and then recovered above the level, the findings were not included in the positive warnings. Ambulation was assessed using the neurosurgical cervical spine scale (NCSS) at discharge (Table 1).15 The NCSS is an official grading system defined by the Japanese Society of Spinal Surgery in 1992; it is a simple and practical grading system for expressing neurological status and has been widely used.16-19 It allocates 5 points each to the upper and lower extremity motor function and 4 points to sensory function or pain. TABLE 1. NCSS Scoring for Lower Extremity Motor Function Total disability (score 1)  Chair bound or bedridden  Severe disability (score 2)  Needs support in walking on flat, and unable to ascend or descend stairways.  Moderate disability (score 3)  Difficulty in walking on flat, and needs support in ascending or descending stairways.  Mild disability (score 4)  No difficulty in walking on flat, but mild difficulty in ascending or descending stairways.  Normal (score 5)  Normal walking, with or without abnormal reflexes.  Total disability (score 1)  Chair bound or bedridden  Severe disability (score 2)  Needs support in walking on flat, and unable to ascend or descend stairways.  Moderate disability (score 3)  Difficulty in walking on flat, and needs support in ascending or descending stairways.  Mild disability (score 4)  No difficulty in walking on flat, but mild difficulty in ascending or descending stairways.  Normal (score 5)  Normal walking, with or without abnormal reflexes.  View Large Transcranial electric stimulation was delivered through corkscrew-type subdermal electrodes placed in the C4 and C5 positions of the 10-20 international electroencephalography system. The stimulation was given as a train of 5 pulses with a duration of 200 μs and stimulation intervals of 2 ms at 200 to 620 V using Multipulse Stimulator D185 (Digitimer Ltd, Hertfordshire, United Kingdom). The MEP was recorded using needle electrodes placed in the muscles innervated by the spinal segments at the level of and distal to the lesion, using either NeuroPack M1 (Nihon Kohden, Tokyo, Japan), or Synax 1100 (NEC Corporation, Tokyo, Japan). To evaluate adverse effect of presence of IMT on elicitation of MEP, we compared the incidence of recordable MEP with a group with unruptured intracranial aneurysms (UIAs) undergoing clipping in the same period. Of 111 cases of UIAs, 108 had intraoperative MEP monitoring. Anesthesia was maintained using sevoflurane in 106 cases and propofol in 2 cases. The anesthetic protocols used were identical to those used in the IMT group. MEP recording was attempted in 222 muscles in the UIA group. To assess the overall sensitivity and specificity of the monitoring and diagnostic value in correlation to the outcome, the following indices were calculated: (1) positive predictive value as the ratio of the number of true positives divided by the number of positives (true and false positives combined), (2) negative predictive value as the ratio of true negatives to the number of negatives (true and false negatives combined), (3) sensitivity as the ratio of number of true positives divided by the summation of true positives and false negatives, and (4) specificity as the ratio of number of true negatives to the summation of true negatives and false positives (Table 2). All patients gave informed consent to this study, which was approved by the institutional review board. TABLE 2. Assessment of Validity of MEP Findings for Predicting Postoperative Weakness   Postoperative weakness  No postoperative weakness  MEP warning: positive  True positive  False positive  MEP warning: negative  False negative  True negative    Postoperative weakness  No postoperative weakness  MEP warning: positive  True positive  False positive  MEP warning: negative  False negative  True negative  FN, false negative; FP, false positive; TN, true negative; TP, true positive. “True positive” means the MEP warning indicated disturbance of motor function and the patient developed a postoperative weakness. “False positive” means the MEP warning indicated disturbance of motor function but the patient did not develop a postoperative weakness. “True negative” means the MEP did not indicate motor dysfunction and the patient did not develop a postoperative weakness. “False negative” means the MEP did not indicate disturbed motor function but the patient developed postoperative weakness. The following properties are calculated to assess validity of the test: Sensitivity = TP/(TP + FN); Specificity = TN/(FP + TN); positive predictive value = TP/(TP + FP); negative predictive value = TN/(TN + FN). View Large RESULTS Intramedullary Spinal Cord Tumors and Lesions The patient group consisted of 30 women and 28 men, age ranging from 11 to 77, with a mean of 48. Four patients had 2 surgeries because of multiple or long lesions. Six patients had tumors in the high cervical level (C1-C2 vertebrae), 27 in the midcervical to cervicothoracic spine (C3-T1 vertebrae) corresponding to the cervical enlargement. Nine tumors were in the thoracic spine (T2-T10 vertebrae) and 1 in the lower thoracic spine (T11-T12 vertebrae) arising from the lumbar enlargement. In some patients, the tumors extended beyond the boundaries; 6 tumors extended from the high cervical to the cervicothoracic level, 1 tumor from the high cervical to the thoracic spine, 4 tumors from the midcervical to the thoracic level, 3 tumors the lumbar enlargement to the conus medullaris, and 1 tumor involved the entire spinal cord (Table 3). The histological diagnoses were as follows: cavernous malformation in 15 patients, ependymoma in 14, hemangioblastoma in 5 (6 operations), subependymoma in 3 (6 operations), pilocytic astrocytoma in 3, fibrillary astrocytoma in 1, anaplastic astrocytoma in 4, glioblastoma in 2, and myelitis in 5 patients. In addition, the series included single cases with ganglioglioma, gangliocytoma, lipoma, schwannoma, malignant peripheral nerve sheath tumor, and sarcoidosis (Table 3). Total removal was achieved in cases with ependymomas, hemangioblastoma, and cavernous malformation in the present series. TABLE 3. The Intramedullary Tumors Classified by Location and Histological Diagnosis Level:     C1-C2 (high cervical)  6   C3-T1 (cervical enlargement)  27   T2-T10 (thoracic)  9   T11-T12 (lumbar enlargement)  1   L1-L2 (conus medullaris)  0   C1-T1 (high cervical–cervical)  6   C1-T10 (high cervical–thoracic)  1   C3-T10 (cervical–thoracic)  4   T11-L2 (thoracic–lumbar enlargement)  3   Holocord (high cervical–conus medullaris)  1  Histology:     Cavernous/venous malformation  15   Ependymoma  14   Subependymoma  3 (6 operations)   Hemangioblastoma  5 (6 operations)   Pilocytic astrocytoma  3   Fibrillary astrocytoma  1   Anaplastic astrocytoma  4   Glioblastoma  2   Ganglioglioma  1   Gangliocytoma  1   Lipoma  1   Schwannoma  1   Malignant peripheral nerve sheath tumor  1   Myelitis  5   Sarcoidosis  1  Level:     C1-C2 (high cervical)  6   C3-T1 (cervical enlargement)  27   T2-T10 (thoracic)  9   T11-T12 (lumbar enlargement)  1   L1-L2 (conus medullaris)  0   C1-T1 (high cervical–cervical)  6   C1-T10 (high cervical–thoracic)  1   C3-T10 (cervical–thoracic)  4   T11-L2 (thoracic–lumbar enlargement)  3   Holocord (high cervical–conus medullaris)  1  Histology:     Cavernous/venous malformation  15   Ependymoma  14   Subependymoma  3 (6 operations)   Hemangioblastoma  5 (6 operations)   Pilocytic astrocytoma  3   Fibrillary astrocytoma  1   Anaplastic astrocytoma  4   Glioblastoma  2   Ganglioglioma  1   Gangliocytoma  1   Lipoma  1   Schwannoma  1   Malignant peripheral nerve sheath tumor  1   Myelitis  5   Sarcoidosis  1  View Large Detection of MEP and Evaluation of Muscle Strength Electrodes for MEP recording were placed in 212 muscles during 59 surgeries performed on 55 patients. Monitoring was performed in 2 muscles in 12 surgeries and in 4 muscles in 47 surgeries. Reproducible MEPs were recorded in all of the monitored muscles in 29 surgeries (29/59 = 49%), and in some muscles in 23 (23/59 = 39%). Strength was assessed before and 1 d after surgery in 208 of the 212 monitored muscles. Muscle strength improved 1 d after surgery in 2 muscles, stayed unchanged in 173, and decreased in 33 (probability of worsening: 0.16, 95% confidence interval [CI]: 0.11-0.21). Patterns of MEP Alterations: Positive and Negative Warnings MEP potentials were obtained in 150 muscles (150/212 = 71%); MEP stayed stable during surgery in 84 muscles (84/150 = 56%), became recordable during surgery after initial absence in 11 (11/150 = 7%), and showed mild decrease during surgery (defined as the amplitude not below 20% of the control at the beginning of procedure) in 3. MEP temporarily decreased during surgery but recovered by the end of the procedure in 15 muscles (15/150 = 10%). MEP diminished during surgery (defined as the amplitude decrease below 20% of the initial amplitude) in 14 muscles (14/150 = 9%). MEP disappeared without recovery in 23 muscles (23/150 = 15%). We defined the findings of MEP monitoring as “positive” when there was disappearance or persistent decline below 20% of the control amplitude at the beginning of procedure, and “negative” when the amplitude stayed above 20% of the control, or recovered above 20% after transient diminution. MEP Changes and Muscle Strength—Predictive Values, Sensitivity, and Specificity of the Monitoring Of the 37 muscles in which MEP change was positive during operations, strength was unaffected in 22 (false positive) and worsened in 15 (true positive), yielding a positive predictive value of 0.41 (15/37, with 95% CI: 0.25-0.58). Of the 113 muscles in which MEP warning was negative (stable during surgery: 84, recovered from initial absence: 11, mild or transient changes: 3+15 = 18), strength remained intact in 105 (true negative) and worsened in 8 (false negative; Table 4). The numbers yielded negative predictive value of 0.93 (105/113, 95% CI: 0.87-0.97). TABLE 4. Correlation of MEP Findings and Postoperative Muscle Strength of the 208 Muscles in the 59 Operations for Intramedullary Tumors/Lesions.   Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  10  48  58  Negative MEP findings:         Stable  6  78  84   Appeared after initial absence  0  11  11   Slight change  0  3  3   Temporarily disappeared or declined to < 20% baseline amplitude with recovery at the end  2  13  15    Subtotal: negative MEP  8  105  113  Positive MEP warnings:   Declined to <20% baseline amplitude  5  9  14   Disappeared  10  13  23    Subtotal: positive MEP  15  22  37    Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  10  48  58  Negative MEP findings:         Stable  6  78  84   Appeared after initial absence  0  11  11   Slight change  0  3  3   Temporarily disappeared or declined to < 20% baseline amplitude with recovery at the end  2  13  15    Subtotal: negative MEP  8  105  113  Positive MEP warnings:   Declined to <20% baseline amplitude  5  9  14   Disappeared  10  13  23    Subtotal: positive MEP  15  22  37  Positive Predictive Value was 0.41 (=15/37) and Negative Predictive Value was 0.93 (=105/113) View Large In the 8 muscles with false-negative MEP monitoring, the postoperative weakness was mild or transient; strength diminished from MRC grade 5 to 4 in 5 muscles. One patient with C1 cavernous malformation developed weakness in the left brachioradialis (from grade 5 to grade 2), and it recovered in several days. One patient with a large C5/C6 cavernous malformation had postoperative weakness in the tibialis anterior from grade 5 to 2, and another patient with a C2 cavernous malformation had postoperative weakness in the brachioradialis from grade 4 to 3. Both patients showed substantial recovery in 1 wk. In 58 muscles, in which no MEPs were detected throughout the procedure, motor strength improved after surgery in 2 of them, stayed comparable in 46, and diminished in 10 (probability of worsening: 0.17, 95% CI: 0.09-0.29). MEP monitoring yielded true-positive warning in 15 muscles, false-positive warning in 22 muscles (false-positive rate 22/37 = 59%), true-negative warning in 105 muscles, false negative in 8 muscles (false-negative rate 8/113 = 7%). The overall sensitivity, the ratio of the number of true positives divided by the summation of true positives and false negatives, was 0.65 in predicting worsening of muscle strength. The overall specificity, the ratio of the number of true negatives divided by the summation of the true negatives and the false positives, was 0.83. MEP Findings and Muscle Strength in Cervical Tumors—Monitoring for Segmental Motor Output (Upper Extremity) and Long-Tract Conduction (Lower Extremity) The tumor involved the cervical enlargement in 39 patients. In these patients, MEP monitoring was attempted in 66 upper extremity muscles and 74 lower extremity muscles. Responses elicited in the upper extremity muscles reflect functional status of the segmental innervation at the level of the tumor. MEP responses and alterations in the lower extremity muscles reflect function of the long-tract passing through the level of the cervical tumor. In the upper extremity, MEP responses were elicited in 58 of the 66 muscles (88%). In correlation with the outcome of strength, the monitoring yielded false-positive warning in 6 muscles, false negative in 4, true positive in 5, and true-negative results in 44 muscles (Table 5). The numbers yielded a positive predictive value of 0.45 (5/11) and a negative predictive value of 0.92 (44/48) for upper extremity weakness. Sensitivity was 0.56 (5/[4+5]) and the specificity was 0.88 (44/[6+44]) in predicting postoperative motor deficit in the cervical segments. TABLE 5. MEP Findings and Postoperative Muscle Strength of the 66 Upper Extremity Muscles in the 39 patients with Tumors Involving the Cervical Enlargement.   Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  0  7  7  Negative MEP findings:         Stable  3  35  38   Appeared after initial absence  0  3  3   Slight change  0  3  3   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  1  3  4    Subtotal: negative MEP  4  44  48  Positive MEP warnings:         Declined to <20% baseline amplitude  2  1  3   Disappeared  3  5  8    Subtotal: positive MEP  5  6  11    Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  0  7  7  Negative MEP findings:         Stable  3  35  38   Appeared after initial absence  0  3  3   Slight change  0  3  3   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  1  3  4    Subtotal: negative MEP  4  44  48  Positive MEP warnings:         Declined to <20% baseline amplitude  2  1  3   Disappeared  3  5  8    Subtotal: positive MEP  5  6  11  Positive Predictive Value was 0.45 (=5/11), and Negative Predictive Value was 0.92 (=44/48) View Large In the lower extremity, MEP responses were elicited in 46 of the 74 muscles (62%). False-positive warning was observed in 7 muscles, false negative in 2, true positive in 8, and true-negative warning in 29 muscles (Table 6). Positive predictive value was 0.53 (8/15) for lower extremity weakness and negative predictive value was 0.94 (29/31). Sensitivity was 0.80 (8/[8+2]), and the specificity was 0.81 (29/[29+7]) for predicting postoperative motor deficit in the lower segments. TABLE 6. MEP Findings and Postoperative Muscle Strength of the 74 Lower Extremity Muscles in the 39 patients With Tumors Involving the Cervical Enlargement.   Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  4  24  28  Negative MEP findings:         Stable  2  19  21   Appeared after initial absence  0  7  7   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  0  3  3    Subtotal: negative MEP  2  29  31  Positive MEP warnings:         Declined to <20% baseline amplitude  2  3  5   Disappeared  6  4  10    Subtotal: positive MEP  8  7  15    Muscle strength  MEP findings  Worse  Unchanged/improved  Sum  No response  4  24  28  Negative MEP findings:         Stable  2  19  21   Appeared after initial absence  0  7  7   Temporarily disappeared or declined to <20% baseline amplitude with recovery at the end  0  3  3    Subtotal: negative MEP  2  29  31  Positive MEP warnings:         Declined to <20% baseline amplitude  2  3  5   Disappeared  6  4  10    Subtotal: positive MEP  8  7  15  Positive Predictive Value was 0.53 (=8/15) and Negative Predictive Value was 0.94 (=29/31) View Large Outcome in Locomotion Locomotion capability was assessed in the 55 patients using the NCSS scored before surgery, at discharge, and at the 1-yr postoperative visit. Locomotion score of NCSS at discharge was better than its preoperative score in 3 patients. Locomotion scores at discharge were unchanged in 42 patients (71%) and worse in 14 patients (24%). Locomotion 1 yr after surgery was assessed in 48 patients; it was better than preoperative status in 16 patients (33%), unchanged in 27 (56%), and worse in 5 (10%). Comparison was not feasible in 11 patients, 3 of whom underwent reoperation within 1 yr after surgery and 8 of whom did not return for the 1-yr visit. Of the 14 patients whose locomotion score was worse at discharge, it was better 1 yr after surgery as compared to preoperative status in 3 patients, was comparable to the preoperative status in 5, stayed worse in 3, and was unknown in 3. Of the 42 patients whose locomotion was unchanged at discharge, it improved in 1 yr after surgery in 10 patients (24%), stayed unchanged in 22, and worsened in 2. The 3 patients with improved locomotion at discharge showed progressive improvement in the score at 1 yr after surgery. MEP Warnings and Locomotion—False Positives and Negatives There were 18 patients who had positive MEP warnings in the lower extremities. Only 4 of them showed actual worsening of locomotion score (NCSS) at the time of discharge (true positive 4/18 = 22%, false positive 14/18 = 78%). Twenty-two patients did not have any warning of MEP changes, but 3 of them had locomotion worse than preoperative status (false negative 3/22 = 14%, true negative 19/22 = 86%). The results yielded positive predictive value of 0.22, negative predictive value of 0.86, sensitivity of 0.57, and specificity of 0.58. In 19 patients, MEPs were not recordable in any of the lower extremity muscles; however, only 7 of them had locomotion scores that were worse at discharge. Variation of MEP Elicitation in Different Muscles Incidence of recordable MEP for monitoring varied among the muscles. For brachioradialis, 53 out of 62 muscles (85%) had detectable MEPs, whereas in gastrocnemius, 54 of 86 muscles (63%) did have recordable MEPs. Occurrence of recordable MEP was significantly higher in the muscles of the upper extremities (88%, 95% CI: 0.79-0.95) than lower extremities (62%, 95% CI: 0.53-0.71; Table 7). TABLE 7. Incidence of Recordable MEP by Muscles in cases With Intramedullary Tumors/Lesions Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Biceps brachii  5  5  1.0 [NA]  Brachioradialis  62  53  0.85 [0.74-0.93]  Flexor carpi radialis  2  2  1.0 [NA]  Triceps brachii  7  7  1.0 [NA]  Thenar  2  2  1.0 [NA]   Upper extremity subtotal  78  69  0.88 [0.79-0.95]  Gastrocnemius  86  54  0.63 [0.52-0.73]  Quadriceps femoris  34  20  0.59 [0.41-0.75]  Tibialis anterior  10  7  0.70 [0.35-0.93]   Lower extremity subtotal  130  81  0.62 [0.53-0.71]   Total  208  150  0.72 [0.65-0.78]  Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Biceps brachii  5  5  1.0 [NA]  Brachioradialis  62  53  0.85 [0.74-0.93]  Flexor carpi radialis  2  2  1.0 [NA]  Triceps brachii  7  7  1.0 [NA]  Thenar  2  2  1.0 [NA]   Upper extremity subtotal  78  69  0.88 [0.79-0.95]  Gastrocnemius  86  54  0.63 [0.52-0.73]  Quadriceps femoris  34  20  0.59 [0.41-0.75]  Tibialis anterior  10  7  0.70 [0.35-0.93]   Lower extremity subtotal  130  81  0.62 [0.53-0.71]   Total  208  150  0.72 [0.65-0.78]  View Large MEP Elicitation and Preoperative Muscle Strength The incidence of measurable MEPs varied in relation to preoperative muscle weakness. Only 1 out of 4 muscles (25%) with preoperative muscle strength graded 1 or 2 showed reliable MEPs, while 29 out of 46 muscles (63%) with preoperative grade 4 and 117 out of 153 muscles (76%) with grade 5 did show reliable MEPs (Table 8). TABLE 8. Incidence of Measurable MEP and Preoperative Muscle Strength Preoperative muscle strength  Muscles  Measurable MEP  Incidence of recordable MEP [95% CI]  1  1  0  0 [NA]  2  3  1  0.33 [0.01-0.91]  3  5  3  0.60 [0.15-0.95]  4  46  29  0.63 [0.48-0.77]  5  153  117  0.76 [0.69-0.83]  Preoperative muscle strength  Muscles  Measurable MEP  Incidence of recordable MEP [95% CI]  1  1  0  0 [NA]  2  3  1  0.33 [0.01-0.91]  3  5  3  0.60 [0.15-0.95]  4  46  29  0.63 [0.48-0.77]  5  153  117  0.76 [0.69-0.83]  View Large Method of General Anesthesia and MEP Elicitation of MEPs was influenced by anesthetic agents. Thirty patients were anesthetized with sevoflurane + air + remifentanil, and 23 patients were anesthetized using sevoflurane + nitrous oxide + fentanyl. Combinations of propofol + remifentanil, sevoflurane + air + fentanyl and sevoflurane + nitrous oxide + remifentanil were used in 2 patients each (Table 9). Statistical significance was not detected regarding the incidence of MEPs among the anesthetic protocols. Sevoflurane was maintained at 1% to 2% in 49 patients, and 3 (6.1%) did not have MEP responses in any muscles. It was maintained at 2.5% to 3% in 8 patients, and 4 of them did not show MEP responses (Table 10). TABLE 9. Incidence of Recordable MEP With Different Anesthetic Protocols. Combination of Sevoflurane, Air, and Remifentanil was Associated With High Chance of Obtaining Recordable MEP     Observed MEP  Anesthetic agent  n  All muscles  Some muscles  No muscle  Propofol + remifentanil  2  2  0  0  Sevoflurane + air + fentanyl  2  1  1  0  Sevoflurane + N2O + fentanyl  23  10  7  6  Sevoflurane + air + remifentanil  30  15  14  1  Sevoflurane + N2O + remifentanil  2  1  1  0      Observed MEP  Anesthetic agent  n  All muscles  Some muscles  No muscle  Propofol + remifentanil  2  2  0  0  Sevoflurane + air + fentanyl  2  1  1  0  Sevoflurane + N2O + fentanyl  23  10  7  6  Sevoflurane + air + remifentanil  30  15  14  1  Sevoflurane + N2O + remifentanil  2  1  1  0  View Large TABLE 10. Sevoflurane Maintenance Dose and MEP Response. Statistical Significance was not Detected Regarding the Incidence of MEPs Among the Anesthetic Protocols     Observed MEP  Sevoflurane (%)  n  All muscles  Some muscles  No muscles  1  23  13  9  1  1.5  18  8  8  2  2  8  5  3  0  2.5  6  0  3  3  3  2  1  0  1      Observed MEP  Sevoflurane (%)  n  All muscles  Some muscles  No muscles  1  23  13  9  1  1.5  18  8  8  2  2  8  5  3  0  2.5  6  0  3  3  3  2  1  0  1  View Large MEP Responses in Patients Undergoing Clipping for UIAs Intraoperative MEP monitoring was performed in 108 patients with unruptured aneurysms during clipping operation (73 women and 35 men, age ranging from 33 to 82 yr, mean age 62.8 yr). Electrodes were placed in 2 muscles in 99 patients, 3 muscles in 6, 1 muscle in 2, and 4 muscles in 1 patient. A total of 104 patients (96%) showed MEP responses in all of the monitored muscles. In 2 patients, no responses were obtained. In total, MEPs were elicited in 216 out of 222 muscles (97%) monitored. The responses were obtained in 84 of 85 upper extremity muscles (99%) and 132 of 137 lower extremity muscles (96%; Table 11). TABLE 11. Incidence of Recordable MEP in Patients With Unruptured Intracranial Aneurysms. MEP was Observed in Almost All Muscles Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Brachioradialis  83  82  0.99 [0.93-1.00]  Thenar  2  2  1.0 [NA]  Upper extremity total  85  84  0.99 [0.94-1.00]  Tibialis anterior  137  132  0.96 [0.92-0.99]  Total  222  216  0.97 [0.94-0.99]  Muscle  n  Recordable MEP  Incidence of recordable MEP [95% CI]  Brachioradialis  83  82  0.99 [0.93-1.00]  Thenar  2  2  1.0 [NA]  Upper extremity total  85  84  0.99 [0.94-1.00]  Tibialis anterior  137  132  0.96 [0.92-0.99]  Total  222  216  0.97 [0.94-0.99]  View Large DISCUSSION Through the analysis of MEP monitoring in the 212 muscles during 59 operations for 55 patients with intramedullary tumors, we found a high false-positive rate. Of 37 muscles with positive MEP warning during operation, 22 muscles did not show deterioration in the strength. The false-positive rate was 59%. The positive predictive value was limited to 0.41. In the muscles with no warning during operation, 93% (105/113) remained unaffected in strength. The false-negative rate was 7% and the negative predictive value was 0.93. Previous studies on MEP monitoring during surgery for IMTs have reported fewer false-positive rates (9.6%-23.8%) and false-negative rates to be none or very rare.11,12,20,21 Our findings contradict these descriptions. This discrepancy in findings from us and others is partly explained by a difference in definition of positive MEP warnings and method of assessing postoperative weakness. Criteria for MEP warning was defined as 50% decrease in amplitude,20 or changes in either voltage stimulation, duration, or complexity of MEP waveform, or loss of detectable wave.11,12 Postoperative neurological function was assessed using the McCormick scale or similar functional grading,11,20,21 or MRC strength grading.12 Studies that reported 0% false-negative cases assessed neurological functions using functional scales such as the McCormick scale that divides neurological status into normal, slightly paretic, severely paretic, and plegic. Some studies did not specify timing of postoperative assessment.11,20,21 One study exclusively included patients with normal or slightly paretic preoperative neurological status, and found no patients suffering prolonged neurological deterioration. They reported 100% sensitivity, with no false negatives.11 The study is biased and may be misleading because appearance of MEP is affected by the preoperative strength as demonstrated in the present study, and the outcome is influenced by the preoperative strength.4,22 Our series included patients with various degrees of preoperative motor weakness. We assessed muscle strength in MRC grading on postoperative day 1 and classified mild or transient weakness as postoperative weakness. This may have contributed to more false negatives and lower sensitivity in our series than in the previous studies. As for the MEP threshold for significant decline, we chose 20%, because a 50% decrease was commonly observed without neurological sequelae, but we wanted to be warned before MEP became flatline. When the MEP threshold was experimentally changed to 50% decrease, the number of muscles with MEP warning increased to 45, of which postoperative weakness developed in 16, and those without MEP warning decreased to 105, of which weakness occurred in 7. False-positive results, that is, positive MEP warning but no postoperative weakness, increased to 29 from 22 with 20% threshold. This resulted in sensitivity of 0.70, specificity of 0.77, positive predictive value of 0.36, and negative predictive value of 0.93. In the cervical spine tumors, the sensitivity was lower for the upper extremity monitoring (0.56) than that for the lower extremity monitoring (0.80), due to larger false negatives for the upper extremity muscles, suggesting MEP output via the segment involved by the tumor is less susceptible to surgical intrusion. This could be due to mixed, overlapping innervation of the upper extremity over a few segments. For example, an alteration limited to 1 segment, such as C6 or C7 alone, may not be reflected as a significant alteration in the MEP recorded from brachioradialis. Chances for false-negative results are increased, and sensitivity (true positive/[true positive + false negative]) would decrease. The MEP amplitude change is a better or more straightforward indicator for disturbances in the downstream conduction in the corticospinal tract than for alteration in the segmental motor function. Thus, when monitoring in cases of cervical tumor, it may be useful to be aware of the variance in false-negative MEP warnings between the upper and lower extremity muscles. MEP with transcranial stimulation and recording without averaging is primarily a method to monitor integrity of the monosynaptic pathway of corticospinal neurons and anterior horn motor neurons.23 However, detailed neurophysiological and anatomic studies reveal that the upper neurons of the connection scheme occupy only 5% to 10% of the entire neuronal population in the pyramidal tract.8,24 The descending neurons from the motor cortex are known to have variable connection patterns to the anterior horn cell, with a great majority connecting via the interneurons and propriospinal neurons.25-28 These multisynaptic pathways also contribute to the motions of the extremities, but MEP does not reflect alterations in the functions of such pathways. Some authors advocated enhancing sensitivity by adjusting the threshold amplitude for positive warning,29 but we would not consider this to be an authentic measure, in view of the inherent variability of the MEP monitoring and limitations to represent the anatomic substrates participating in the voluntary movement. In our series, the overall incidence of obtaining MEP in the muscles monitored was 71%, and in the cases of cervical tumors, the chance of getting MEP in the upper extremity was 88% and in the lower extremity was 62%. This is in accordance with previous studies, which found that MEP is not reliably elicited in patients with intramedullary tumors.30,31 The variability in the responsiveness to the monitoring indicates an inherent limitation in the reliability as a test in the pathological condition. Sevoflurane, at higher doses, is known to affect MEPs adversely.32 We routinely maintained sevoflurane lower than 3% during surgery. At this maintenance dose, MEP was recorded in 97% of muscles in our patients undergoing craniotomy for UIAs, and 96% of the patients with UIAs had MEP responses. Therefore, we would not consider the anesthetic protocol to be the cause of the higher false-positive and -negative rates observed in our series. MEP monitoring did not predict outcome of gait function at the time of discharge or 1 yr after surgery. In fact, MEP warnings in the lower extremity muscles had positive predictive value of 0.22 in predicting gait dysfunction at the time of discharge. Sensitivity and specificity were limited to 0.57 and 0.58. This may be due to the frequent false-positive MEP warnings for locomotion (78%) or less reproducibility in eliciting MEP in the lower extremity muscles (62% as opposed to 88% for the upper extremity muscles). The latter view is shared by other authors.30,31 Normal locomotion requires coordination of motor and sensory function, which may involve the tectospinal, rubrospinal, reticulospinal, and vestibulospinal tracts in addition to the corticospinal tract and the dorsal column/medial lemniscus. MEPs mainly reflect corticospinal tract function,6,33 and may not correlate with ambulation as a whole. Concurrent monitoring of SEP enables assessment of dorsal column function,34-36 but its role in IMT surgeries is limited, because SEP often disappears at the beginning of opening the dorsomedian sulcus to reach the intramedullary pathology, and SEP change does not correlate with postoperative weakness or loss of the dorsal column function.36-39 Therefore, we use only MEP as the electrophysiological monitoring during surgery of intramedullary tumors. When we face positive MEP warning during surgery for IMTs, we check for the possibility of dysfunction caused by surgical maneuvers, and then check for depth of anesthesia, alterations in body temperature, and blood pressure. When MEP decline occurred after rotating or retracting the spinal cord, possibly obstructing the blood flow to the cord, release of such tractions often results in recovery of MEPs. When MEP amplitude decreased during resection of the intramedullary tumor, we would suspend the maneuver while loosening the spinal cord rotation or pial retraction or by changing the place of microdissection from one part of the tumor to another. Often, these measures resulted in recovery of the amplitude. We did not administer steroids. We usually pursued total removal of the tumor as far as there was a good plane of dissection and as far as the thickness of the adjoining spinal cord parenchyma, typically the lateral funiculus, was maintained. Total resection of benign tumors such as ependymoma can accomplish a cure, and, when executed with utmost precision, it carries relatively small risks of debilitating neurological deficits.22 Premature discontinuation of removal following MEP warning, as has been described by an orthopedic group,13 would result in regrowth and progressive neurological worsening. A second surgery, as to be warranted, would be complicated with adhesive changes, posing more of a technical challenge and carrying greater risk. Tumor histology is an important factor in achieving total resection.40 Therefore, we value appropriate intraoperative histological diagnosis and the feasibility of dissection with a plane. They are vital for decision making regarding the extent of the resection of the tumor. It is certainly essential to preserve the anatomic integrity of the spinal cord tissue and its vascular supplies with utmost care, and this often requires maximum magnification for the surgical microscope. As a result, we achieved total removal in all cases of ependymomas, hemangioblastomas, and cavernous malformations in the present series. Of 14 patients with ependymomas included in the present study, 21% had worsening of lower extremity function at discharge, but all patients recovered 1 yr after surgery to the same or better grade than before surgery. We could not obtain measurable MEPs in the lower extremities in 19 patients. The rate of MEP elicitation is lower in patients with weakness.32 There has been no recommendation as to surgical strategy for those without detectable baseline MEPs. However, gait function after surgery was the same or better than before surgery in 12 out of the 19 patients (64%) with no baseline potentials. Gait was worse at the time of discharge in 7 patients, of whom 1 improved better than before surgery and 2 recovered to the same grade as preoperative state at 1-yr follow-up. We believe that attempt of total removal of the intramedullary tumors is justifiable even when MEPs cannot be elicited, if the histology and the anatomic findings indicate feasibility of resection, in view of the temporary and limited nature of the postoperative weakness in the majority of such cases.22 CONCLUSION MEP monitoring is by no means the perfect test in predicting functional outcome, with frequent false-positive and false-negative warnings, and with limited sensitivity and specificity. It should be utilized as an aid to detect intraoperative alterations in the motor signal conduction, and as one source of supportive information together with other findings, such as plane of cleavage, thickness of the remaining cord parenchyma, and histological diagnosis, in executing safe and effective resection of IMT. Disclosures Kazushige Itoki has received funding from the National Institutes of Health, Wellcome Trust/COAF, Howard Hughes Medical Institute, Australian Science Fund, Bill & Melinda Gates Foundation, World Bank, Research Councils UK, and the Department of Neurosurgery, Dokkyo University School of Medicine. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Part of the material was presented in the 12th Annual Meeting of the Japanese Society of Intraoperative Imaging on July 7, 2012, in Tsukuba, Japan, as an oral presentation. REFERENCES 1. Nash CL Jr, Lorig RA, Schatzinger LA, Brown RH. Spinal cord monitoring during operative treatment of the spine. Clin Orthop Relat Res . 1977; 126: 100- 105. 2. Whittle IR, Johnston IH, Besser M. Recording of spinal somatosensory evoked potentials for intraoperative spinal cord monitoring. J Neurosurg . 1986; 64( 4): 601- 612. Google Scholar CrossRef Search ADS PubMed  3. Kearse LA Jr, Lopez-Bresnahan M, McPeck K, Tambe V. Loss of somatosensory evoked potentials during intramedullary spinal cord surgery predicts postoperative neurologic deficits in motor function. J Clin Anesth . 1993; 5( 5): 392- 398. Google Scholar CrossRef Search ADS PubMed  4. McCormick PC, Torres R, Post KD, Stein BM. Intramedullary ependymoma of the spinal cord. J Neurosurg . 1990; 72( 4): 523- 532. Google Scholar CrossRef Search ADS PubMed  5. Seyal M, Mull B. Mechanisms of signal change during intraoperative somatosensory evoked potential monitoring of the spinal cord. J Clin Neurophysiol . 2002; 19( 5): 409- 415. Google Scholar CrossRef Search ADS PubMed  6. Merton PA, Morton HB. Stimulation of the cerebral cortex in the intact human subject. Nature . 1980; 285( 5762): 227. Google Scholar CrossRef Search ADS PubMed  7. Levy WJ Jr. Clinical experience with motor and cerebellar evoked potential monitoring. Neurosurgery  1987; 20 (1): 169- 182. Google Scholar PubMed  8. Taniguchi M, Cedzich C, Schramm J. Modification of cortical stimulation for motor evoked potentials under general anesthesia: technical description. Neurosurgery . 1993; 32( 2): 219- 226. Google Scholar CrossRef Search ADS PubMed  9. Calancie B, Harris W, Broton JG, Alexeeva N, Green BA. "Threshold-level" multipulse transcranial electrical stimulation of motor cortex for intraoperative monitoring of spinal motor tracts: description of method and comparison to somatosensory evoked potential monitoring. J Neurosurg . 1998; 88( 3): 457- 470. Google Scholar CrossRef Search ADS PubMed  10. Morota N, Deletis V, Constantini S, Kofler M, Cohen H, Epstein FJ. The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery . 1997; 41( 6): 1327- 1336. Google Scholar CrossRef Search ADS PubMed  11. Kothbauer KF, Deletis V, Epstein FJ. Motor-evoked potential monitoring for intramedullary spinal cord tumor surgery: correlation of clinical and neurophysiological data in a series of 100 consecutive procedures. Neurosurg Focus . 1998; 4( 5): e1. Google Scholar CrossRef Search ADS PubMed  12. Quiñones-Hinojosa A, Lyon R, Zada G et al.   Changes in transcranial motor evoked potentials during intramedullary spinal cord tumor resection correlate with postoperative motor function. Neurosurgery . 2005; 56( 5): 982- 993. Google Scholar PubMed  13. Matsuyama Y, Sakai Y, Katayama Y et al.   Surgical results of intramedullary spinal cord tumor with spinal cord monitoring to guide extent of resection. J Neurosurg Spine . 2009; 10( 5): 404- 413. Google Scholar CrossRef Search ADS PubMed  14. The Guarantors of Brain. Aids to the Examination of the Peripheral Nervous System . 5th ed. London, UK: Saunders Elsevier; 2010. 15. Kadoya S. Grading and scoring system for neurological function in degenerative cervical spine disease—neurosurgical cervical spine scale. Neurol Med Chir (Tokyo) . 1992; 32( 1): 40- 41. Google Scholar CrossRef Search ADS PubMed  16. Kim P, Wakai S, Matsuo S, Moriyama T, Kirino T. Bisegmental cervical interbody fusion using hydroxyapatite implants: surgical results and long-term observation in 70 cases. J Neurosurg . 1998; 88( 1): 21- 27. Google Scholar CrossRef Search ADS PubMed  17. Koyanagi I, Iwasaki Y, Hida K, Imamura H, Abe H. Magnetic resonance imaging findings in ossification of the posterior longitudinal ligament of the cervical spine. J Neurosurg . 1998; 88( 2): 247- 254. Google Scholar CrossRef Search ADS PubMed  18. Hida K, Iwasaki Y, Yano S, Akino M, Seki T. Long-term follow-up results in patients with cervical disk disease treated by cervical anterior fusion using titanium cage implants. Neurol Med Chir (Tokyo) . 2008; 48( 10): 440- 446. Google Scholar CrossRef Search ADS PubMed  19. Bucciero A, Zorzi T, Piscopo GA. Peek cage-assisted anterior cervical discectomy and fusion at four levels: clinical and radiographic results. J Neurosurg Sci . 2008; 52( 2): 37- 40. Google Scholar PubMed  20. Zentner J. Noninvasive motor evoked potential monitoring during neurosurgical operations on the spinal cord. Neurosurgery . 1989; 24( 5): 709- 712. Google Scholar CrossRef Search ADS PubMed  21. Costa P, Peretta P, Faccani G. Relevance of intraoperative D wave in spine and spinal cord surgeries. Eur Spine J . 2013; 22( 4): 840- 848. Google Scholar CrossRef Search ADS PubMed  22. Klekamp J. Treatment of intramedullary tumors: analysis of surgical morbidity and long-term results. J Neurosurg Spine . 2013; 19( 1): 12- 26. Google Scholar CrossRef Search ADS PubMed  23. Pechstein U, Cedzich C, Nadstawek J, Schramm J. Transcranial high-frequency repetitive electrical stimulation for recording myogenic motor evoked potentials with the patient under general anesthesia. Neurosurgery . 1996; 39( 2): 335- 343. Google Scholar CrossRef Search ADS PubMed  24. York DH. Review of descending motor pathways involved with transcranial stimulation. Neurosurgery . 1987; 20( 1): 70- 73. Google Scholar CrossRef Search ADS PubMed  25. Ralston DD, Ralston HJ 3rd. The terminations of corticospinal tract axons in the macaque monkey. J Comp Neurol . 1985; 242( 3): 325- 337. Google Scholar CrossRef Search ADS PubMed  26. Nathan PW, Smith M, Deacon P. Vestibulospinal, reticulospinal and descending propriospinal nerve fibres in man. Brain . 1996; 119( pt 6): 1809- 1833. Google Scholar CrossRef Search ADS PubMed  27. Chakrabarty S, Shulman B, Martin JH. Activity-dependent codevelopment of the corticospinal system and target interneurons in the cervical spinal cord. J Neurosci . 2009; 29( 27): 8816- 8827. Google Scholar CrossRef Search ADS PubMed  28. Martin JH. Neuroanatomy: Text and Atlas . 4th ed. New York: McGraw-Hill; 2012. 29. Muramoto A, Imagama S, Ito Z et al.   The cutoff amplitude of transcranial motor evoked potentials for transient postoperative motor deficits in intramedullary spinal cord tumor surgery. Spine (Phila Pa 1976) . 2014; 39( 18): E1086- E1094. Google Scholar CrossRef Search ADS PubMed  30. Rajshekhar V, Velayutham P, Joseph M, Babu KS. Factors predicting the feasibility of monitoring lower-limb muscle motor evoked potentials in patients undergoing excision of spinal cord tumors. J Neurosurg Spine . 2011; 14( 6): 748- 753. Google Scholar CrossRef Search ADS PubMed  31. Chen X, Sterio D, Ming X et al.   Success rate of motor evoked potentials for intraoperative neurophysiologic monitoring: effects of age, lesion location, and preoperative neurologic deficits. J Clin Neurophysiol . 2007; 24( 3): 281- 285. Google Scholar CrossRef Search ADS PubMed  32. Hayashi H, Kawaguchi M, Abe R et al.   Evaluation of the applicability of sevoflurane during post-tetanic myogenic motor evoked potential monitoring in patients undergoing spinal surgery. J Anesth . 2009; 23( 2): 175- 181. Google Scholar CrossRef Search ADS PubMed  33. Malhotra NR, Shaffrey CI. Intraoperative electrophysiological monitoring in spine surgery. Spine (Phila Pa 1976) . 2010; 35( 25): 2167- 2179. Google Scholar CrossRef Search ADS PubMed  34. Nagle KJ, Emerson RG, Adams DC et al.   Intraoperative monitoring of motor evoked potentials: a review of 116 cases. Neurology . 1996; 47( 4): 999- 1004. Google Scholar CrossRef Search ADS PubMed  35. Costa P, Bruno A, Bonzanino M et al.   Somatosensory- and motor-evoked potential monitoring during spine and spinal cord surgery. Spinal Cord . 2007; 45( 1): 86- 91. Google Scholar CrossRef Search ADS PubMed  36. Hyun SJ, Rhim SC. Combined motor and somatosensory evoked potential monitoring for intramedullary spinal cord tumor surgery: correlation of clinical and neurophysiological data in 17 consecutive procedures. Br J Neurosurg . 2009; 23( 4): 393- 400. Google Scholar CrossRef Search ADS PubMed  37. Kothbauer K, Deletis V, Epstein FJ. Intraoperative spinal cord monitoring for intramedullary surgery: an essential adjunct. Pediatr Neurosurg . 1997; 26( 5): 247- 254. Google Scholar CrossRef Search ADS PubMed  38. Sandalcioglu IE, Gasser T, Asgari S et al.   Functional outcome after surgical treatment of intramedullary spinal cord tumors: experience with 78 patients. Spinal Cord . 2005; 43( 1): 34- 41. Google Scholar CrossRef Search ADS PubMed  39. Sala F, Palandri G, Basso E et al.   Motor evoked potential monitoring improves outcome after surgery for intramedullary spinal cord tumors: a historical control study. Neurosurgery . 2006; 58( 6): 1129- 1143. Google Scholar CrossRef Search ADS PubMed  40. Karikari IO, Nimjee SM, Hodges TR et al.   Impact of tumor histology on resectability and neurological outcome in primary intramedullary spinal cord tumors: a single-center experience with 102 patients. Neurosurgery  2011; 68( 1): 188- 197. Google Scholar CrossRef Search ADS PubMed  Acknowledgment We thank Ms. Kayoko Iwata for her technical assistance in the intraoperative MEP monitoring throughout the series. COMMENT The authors reported their experience of MEP monitoring in 59 operations of intramedullary tumors. Evoked responses were recorded from 212 muscles of the upper and lower extremities following transcranial electrical stimulation. MEP was recorded in 71% of muscles. Decreased MEP amplitude below 20% of the initial level was observed in 37 muscles, yet 22 of them (59%) did not show postoperative motor weakness of the monitored muscles. The authors stressed the limitation of using MEP in predicting postoperative motor weakness in surgery of intramedullary tumors. The study was carefully designed to clarify the relationship of decreased amplitude of MEP and postoperative weakness of the corresponding muscle. Surgery of intramedullary tumors has a higher chance to insult the limited areas of functional pathways and gray mater neurons in the spinal cord. Muscle movement of extremities is mediated by multiple pathways, not limited to the corticospinal tracts. The authors correctly pointed out the mechanisms underlying false results and issues of intraoperative monitoring. This is an important study for neurosurgeons to understand “the real world” of using MEPs elicited from extremity muscles for intramedullary spinal cord tumor surgeries. Izumi Koyanagi Sapporo, Japan Copyright © 2017 by the Congress of Neurological Surgeons

Journal

Operative NeurosurgeryOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from
Google Scholar,
PubMed
Create lists to
organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off