Hello Readers,
For us number/science-y people, there is nothing more frustrating than unavoidable errors. And that is exactly what plaguing this project right now. I must accept that there is unavoidable error at every step of the operation. Every single step. So (partially in an attempt to vent some of this frustration) this post will detail some of the errors that have manifested itself in my research.
MRI Scan distortions: No matter how good you are at building an MRI scanner (which is a completely common skill that should be in every handyman's repertoire), it is nearly impossible to create a homogenous magnetic field. The inhomogeneity of the magnetic field means the MRI scans are often distorted. That is, if you scanned a perfect circle, a perfect circle may not show up on the monitor. Maybe the circle will be somewhat elongated or maybe one side of the circle is more flat than others. This is extremely not great when trying to target a small deep brain structure because, well, you may spend twenty minutes painstakingly centering your cursor on the STN (a common Parkinson's patient DBS target) but in reality, that STN may be a millimeter off. The operation hasn't even begun and you've already got error.
Distortion is more severe in higher resolution MRIs, such as 3T MRIs. A 3T MRI would have much more distortion than a 1.5T MRI. You trade high quality images for more image distortion. CT scans have no distortion, but CT scans also cannot visualize any of the soft tissues. We really can't have it all.
You see, millimeter accuracy is needed for successful DBS surgery largely because of the small STN size, but even more importantly because the STN is close (1-3mm) to the anatomic structures such as the red nucleus, posterior limb of the internal capsule, medial lemniscus, and substantia nigra, the stimulation of which may result in adverse reactions.
The only good thing to come out of this mess is that MRI distortions occur most significantly on the edges of the image, and the middle (where these structures are) are less likely to be subject to significant distortion.
Image Fusing: Pre-op and post op scans must be fused together in order to properly estimate errors in electrode placement. Image fusing is a frustrating thing to do because in order for some parts of the image to line up other parts will not. And that just creates (you guessed it) more error. This error can present itself also in millimeter amounts, but measurement estimations can differ from person to person.
Brain Shift: You can't control this one. Brain shift occurs naturally because your brain is floating around in a hard shell (skull) so naturally if you tilt your head, it will move slightly toward the direction of gravity. I wrote two rather long articles on brain shifts (which you can read here and here) so I won't elaborate much on it now. What I will say is one way to avoid the brain shifting too much is imaging the patient in the same position as the surgery will be done, so at least the brain won't shift due to patient movement.
Some DBS centers have tried to drain all the CSF out of the brain before doing the operation and imaging so no brain shift can occur (theoretically). The brain sinks to the bottom of the skull during the operation, but once the CSF fills the skull again (and the brain returns to its original position) the electrode will shift in the brain (another error).
In general, there is nothing we can do to avoid these errors (these aren't all the errors either). I sincerely hope reading this post hasn't resulting in any stressful hair loss. I'm sad to report the same cannot be applied to me.
More on errors in a future post!
Holly
For us number/science-y people, there is nothing more frustrating than unavoidable errors. And that is exactly what plaguing this project right now. I must accept that there is unavoidable error at every step of the operation. Every single step. So (partially in an attempt to vent some of this frustration) this post will detail some of the errors that have manifested itself in my research.
MRI Scan distortions: No matter how good you are at building an MRI scanner (which is a completely common skill that should be in every handyman's repertoire), it is nearly impossible to create a homogenous magnetic field. The inhomogeneity of the magnetic field means the MRI scans are often distorted. That is, if you scanned a perfect circle, a perfect circle may not show up on the monitor. Maybe the circle will be somewhat elongated or maybe one side of the circle is more flat than others. This is extremely not great when trying to target a small deep brain structure because, well, you may spend twenty minutes painstakingly centering your cursor on the STN (a common Parkinson's patient DBS target) but in reality, that STN may be a millimeter off. The operation hasn't even begun and you've already got error.
Distortion is more severe in higher resolution MRIs, such as 3T MRIs. A 3T MRI would have much more distortion than a 1.5T MRI. You trade high quality images for more image distortion. CT scans have no distortion, but CT scans also cannot visualize any of the soft tissues. We really can't have it all.
 |
| CT Scan on the left, MRI on the right. You see how the bone on these scans don't match up? That screams ERROR to me. |
You see, millimeter accuracy is needed for successful DBS surgery largely because of the small STN size, but even more importantly because the STN is close (1-3mm) to the anatomic structures such as the red nucleus, posterior limb of the internal capsule, medial lemniscus, and substantia nigra, the stimulation of which may result in adverse reactions.
The only good thing to come out of this mess is that MRI distortions occur most significantly on the edges of the image, and the middle (where these structures are) are less likely to be subject to significant distortion.
Image Fusing: Pre-op and post op scans must be fused together in order to properly estimate errors in electrode placement. Image fusing is a frustrating thing to do because in order for some parts of the image to line up other parts will not. And that just creates (you guessed it) more error. This error can present itself also in millimeter amounts, but measurement estimations can differ from person to person.
Brain Shift: You can't control this one. Brain shift occurs naturally because your brain is floating around in a hard shell (skull) so naturally if you tilt your head, it will move slightly toward the direction of gravity. I wrote two rather long articles on brain shifts (which you can read here and here) so I won't elaborate much on it now. What I will say is one way to avoid the brain shifting too much is imaging the patient in the same position as the surgery will be done, so at least the brain won't shift due to patient movement.
Some DBS centers have tried to drain all the CSF out of the brain before doing the operation and imaging so no brain shift can occur (theoretically). The brain sinks to the bottom of the skull during the operation, but once the CSF fills the skull again (and the brain returns to its original position) the electrode will shift in the brain (another error).
In general, there is nothing we can do to avoid these errors (these aren't all the errors either). I sincerely hope reading this post hasn't resulting in any stressful hair loss. I'm sad to report the same cannot be applied to me.
More on errors in a future post!
Holly
Post a Comment