For the original paper click the title: Limitations of Intraoperative Transcranial Electrical Stimulation-Motor Evoked Potential Monitoring During Brain Tumor Resection Adjacent to the Primary Motor Cortex.
When running TcMEPs, if you stimulate at a high intensity the current or voltage can spread from the cortex deeper in the brain, additionally stimulating the descending corticospinal pathway (see original paper here). Thus, you can get TcMEP responses in the muscles that come from stimulating the descending corticospinal tract even though the cortex or tract above the stimulation may have sustained injury (either surgical injury, perforating vessel occlusion or subcortical infarct).
In the worst-case scenario, you have a false negative; the TcMEP responses were consistently present the entire surgery, but the patient wakes up with a hemispheric motor deficit. This paper reports such a scenario in 3 patients undergoing craniotomy and tumor resection in the precentral motor gyrus.
All three patients were having tumor resection in the vicinity of the motor cortex. The authors used the NIM Eclipse and stimulated at C3/C4 with 6 pulse/trains at 350-450 volts. They recorded contralateral deltoid, thenar, tibialis anterior and abductor hallucis. Their criteria for a change was 50% amplitude reduction and 10% latency (see comments below). There was no anesthesia listed.
Under these methods these three patients maintained their TcMEP responses throughout the surgery but woke up with hemispheric motor deficits.
Instead of relying solely on TcMEP, direct cortical stimulation (DCS-MEP), which involves directly stimulating the motor cortex with a grid, should be used on cases where the cortex and descending cortical spinal pathway in the brain are at risk.
It is easy to criticize this paper on several grounds, but it takes a certain level of bravery to publish a paper showing that you have erred. And they make the valid point that there are few articles published out there that explicitly discusses this TcMEP limitation. Yes, we are all taught that this can happen, but when you go to the literature, there are very few paper that tackle this specifically with real world consequences (patients waking up with deficits).
One recent paper also reported (buried in their results) a 2.5% of false negatives when monitoring TcMEP in aneurysms. Like the paper being reviewed here they failed to take into account stimulation spread. And here is paper that had a false negative from too high of stim (30mA) when using DCS-MEP. So it is happening.
Reduction of Stimulation
The authors use stimulation intensity (350-450V) and do not document any measures to reduce that stimulation. It may have been that the authors were using this level of stimulation as their standard practice without titrating down the stimulation to threshold level or more likely the anesthesia used (which they do not report)was such that they needed higher stimulation levels to get a response. When relying on motor evoked potentials the preferred anesthesia is TIVA.
The authors used a C3/C4 stimulating montage. One of the reasons this montage works so well is that the electrodes are far apart forcing the current to spread from one cortex to the other, passing through the deeper mesial (leg) regions. As we want to minimize the spread in brain surgeries a better stimulating montages would be Cz-C3/4 or Cz-C1/2. However, if your electrodes are not situated just over the homunculus or are too close together (i.e. current shunting through the skull) you may end up having to stimulate at higher intensities to get a response using these montages. This would negate your efforts to reduce stimulation spread. If this is the case it is critical you work with anesthesia to lower the anesthetic agents as much as possible to give you a good response at the lowest stimulation intensities, you can get.
Their alarm criteria of a 50% decrease in TcMEPs with a 10% shift in latency is a common alarm criteria for DCS-MEPs, though the latency shift portion has been dropped as latency shifts do not occur in the absence of significant amplitude attenuation. This criterion is at odds with the spinal TcMEP criteria which have more stringent alarm criteria so as to avoid the common false positives that result from anesthetic fade and other physiological variables. However, the probability of motor deficit in many of these surgeries is significantly higher than in spinal surgeries and so a more sensitive alarm criteria may be appropriate.
It would be fascinating to have craniotomy case like the one they describe in this paper in which the DCS-MEPs and TcMEPs were lost and then the TcMEP stimulation was systematically increased to see at what stage the TcMEP responses returned (i.e. current spread to the undamaged descending corticospinal pathway). However, modeling studies have shown that current spread can be affected by individual sulcal and skull patterns making generalization challenging. Modeling and animal studies would also give us some insight into this
As was recommended by Szelenyi et al, when the surgery is in the brain it is imperative you minimize your TcMEP stimulation intensity, use more focal montages and whenever possible use DCS-MEP as your primary motor modality.
What is the highest you would stimulate on a cortical or brainstem surgery?
Do you prefer to use voltage or current for TcMEP and why?
Let us know in the comments below.