Tuesday, November 15, 2016
Asleep DBS
Walk into room
Anesthesia does their thing
Retape eyes with tegaderm
Turn bed 90 degrees so head is facing ceretom
Pin with doro frame (radiolucent head holder) - shoulders may get in the way - pad shoulders with foam
Fix Mayfield on inner side of head holder
Bottom of Mayfield has to be angled back towards patient's feet - if its angled forward, it will hit CT scanner and you won't be able to get entire head in scan
Shave head
Prep with chloraprep then chlorhexidine then alcohol
Put in 5 fidicuial screws - 2 frontal, 1 left posterior (above/behind L ear), 2 right posterior (one by bregma, one above/behind R ear), roughly in a circle - bias screw 3 (posterior L screw) medial to allow for tunneling
CT scan
Put suction-cup registration device on head, non-sterile registration of fiducial screws with blue thing
Use stealth to determine optimal entry points
Mark entry holes with pen. Draw upside down U-shaped incision lines approx 1 cm radius around point
Prep again
Hand drill to mark entry points
Drape - body drape, 4 drapes in square on head - include fiducial screws in field. Ioban.
Incision
dissection between galea and periostium around incision and in flap. Or subperiosteal. depends on who you ask.
Hemostasis
On L side, dissect posteriorly to form pocket
Burr holes
Hemostasis
Screw in stimlock - do not strip screws - screws on both side, slit pointing down
Screw in base - two screws that are closer together approx 10 o'clock and 2 o'clock, the one pointing down at 6 o'clock. do not strip screws. Base has to be tight, all things are registered off base.
Make sure clip fits
Attach dome
Attach other registration device to left dome.
put steatlh probe into dome, find optimal trajectory and depth. tell scrub tech so they can set the cannula.
Attach nexdrive micropositioner
Then put in cannula
Figure out where it will enter dura. Bipolar dura at that point. a lot. Cruciate incision with 11-blade
Attempt to pass cannula - if any resistance, bipolar + 11 blade again. you need to get through Pia.
push in cannula to pre-determined depth into center hole
remove stylet
Place extender
Pass electrodes
remove cannula
apply clip to base of electrode
mark electrode where its coming out of the clip
Tisseal into burr hole
CT scan - merge with preop MRI with plan. deviation of <2mm is acceptable. More requires electrode replacement
Remove everything
close - do not hit the electrode with your needle. electrode is very sensitive and fragile.
DBS generator
pocket - incision 1 finger breath below clavicle approx 5-6 cm long. Cut down until you get deep to fascia into fat layer but above muscle, make a pocket deep enough to fit finger to knuckle.
ear incision - far down enough to be over the curve of skull (sucks to tunnel over the curve), but high up enough to access leads. bovie to bone. Look up to find leads, flip them out. use heavy scissors to start tunneling process to get through neck fascia, open them, then pull them back out. LEFT lead has tie on it. Use tunneler to get to pocket. go above the clavicle. anesthesia gets upset if you tunnel into subclavian artery. screw extension leads onto tunneler and pull back to incision (L and R extension leads are the same).
connect extension leads to head leads-- white = right, blue = left. Starting with left, remove plastic coverings off head lead, wipe lead with blood, put raytech below for clean field, put new clear plastic covering thingy on (skinny side first) connect extension lead on; when you screw the leads on, you must hold the lead by contact 1 tightly (i.e. provide counter-torque) otherwise you risk breaking the electrode. if you break the electrode, you're fired. screw to 1 click of tightness. Repeat with right electrode - only use white covering instead of clear.
connect leads to generator (for medtronic, left is contacts 1-7, right is 8-15, left goes in front, right in back, front is the brand name side). screw to 3 clicks of tightness. tuck generator into pocket, rep checks impedance. very high impedance = suboptimal contact between electrode contacts. very low = concern for short circuit i.e. electrode breakage. If all goes well, pull the electrodes straight. close pocket and head.
DBS generator replacement
Knife through skin, then continue with knife while feeling with your fingers. the only way you screw this up is by cutting the leads. cut towards the generator. Cut through capsule with knife, make sure to turn the corners around both edges. Pull out generator with kocher. unscrew/remove leads one at a time! front/left lead first. correction, the other way you screw this up is by mixing up L and R leads. Connect leads to new generator first left, then right - put the leads all the way in, make sure good contact, screw to 3 clicks of tightness. put generator back into pocket, brand name side up. Test generator to check impedances. If good, anchor with non-absorable suture if needed, close.
Walk into room
Anesthesia does their thing
Retape eyes with tegaderm
Turn bed 90 degrees so head is facing ceretom
Pin with doro frame (radiolucent head holder) - shoulders may get in the way - pad shoulders with foam
Fix Mayfield on inner side of head holder
Bottom of Mayfield has to be angled back towards patient's feet - if its angled forward, it will hit CT scanner and you won't be able to get entire head in scan
Shave head
Prep with chloraprep then chlorhexidine then alcohol
Put in 5 fidicuial screws - 2 frontal, 1 left posterior (above/behind L ear), 2 right posterior (one by bregma, one above/behind R ear), roughly in a circle - bias screw 3 (posterior L screw) medial to allow for tunneling
CT scan
Put suction-cup registration device on head, non-sterile registration of fiducial screws with blue thing
Use stealth to determine optimal entry points
Mark entry holes with pen. Draw upside down U-shaped incision lines approx 1 cm radius around point
Prep again
Hand drill to mark entry points
Drape - body drape, 4 drapes in square on head - include fiducial screws in field. Ioban.
Incision
dissection between galea and periostium around incision and in flap. Or subperiosteal. depends on who you ask.
Hemostasis
On L side, dissect posteriorly to form pocket
Burr holes
Hemostasis
Screw in stimlock - do not strip screws - screws on both side, slit pointing down
Screw in base - two screws that are closer together approx 10 o'clock and 2 o'clock, the one pointing down at 6 o'clock. do not strip screws. Base has to be tight, all things are registered off base.
Make sure clip fits
Attach dome
Attach other registration device to left dome.
put steatlh probe into dome, find optimal trajectory and depth. tell scrub tech so they can set the cannula.
Attach nexdrive micropositioner
Then put in cannula
Figure out where it will enter dura. Bipolar dura at that point. a lot. Cruciate incision with 11-blade
Attempt to pass cannula - if any resistance, bipolar + 11 blade again. you need to get through Pia.
push in cannula to pre-determined depth into center hole
remove stylet
Place extender
Pass electrodes
remove cannula
apply clip to base of electrode
mark electrode where its coming out of the clip
Tisseal into burr hole
CT scan - merge with preop MRI with plan. deviation of <2mm is acceptable. More requires electrode replacement
Remove everything
close - do not hit the electrode with your needle. electrode is very sensitive and fragile.
DBS generator
pocket - incision 1 finger breath below clavicle approx 5-6 cm long. Cut down until you get deep to fascia into fat layer but above muscle, make a pocket deep enough to fit finger to knuckle.
ear incision - far down enough to be over the curve of skull (sucks to tunnel over the curve), but high up enough to access leads. bovie to bone. Look up to find leads, flip them out. use heavy scissors to start tunneling process to get through neck fascia, open them, then pull them back out. LEFT lead has tie on it. Use tunneler to get to pocket. go above the clavicle. anesthesia gets upset if you tunnel into subclavian artery. screw extension leads onto tunneler and pull back to incision (L and R extension leads are the same).
connect extension leads to head leads-- white = right, blue = left. Starting with left, remove plastic coverings off head lead, wipe lead with blood, put raytech below for clean field, put new clear plastic covering thingy on (skinny side first) connect extension lead on; when you screw the leads on, you must hold the lead by contact 1 tightly (i.e. provide counter-torque) otherwise you risk breaking the electrode. if you break the electrode, you're fired. screw to 1 click of tightness. Repeat with right electrode - only use white covering instead of clear.
connect leads to generator (for medtronic, left is contacts 1-7, right is 8-15, left goes in front, right in back, front is the brand name side). screw to 3 clicks of tightness. tuck generator into pocket, rep checks impedance. very high impedance = suboptimal contact between electrode contacts. very low = concern for short circuit i.e. electrode breakage. If all goes well, pull the electrodes straight. close pocket and head.
DBS generator replacement
Knife through skin, then continue with knife while feeling with your fingers. the only way you screw this up is by cutting the leads. cut towards the generator. Cut through capsule with knife, make sure to turn the corners around both edges. Pull out generator with kocher. unscrew/remove leads one at a time! front/left lead first. correction, the other way you screw this up is by mixing up L and R leads. Connect leads to new generator first left, then right - put the leads all the way in, make sure good contact, screw to 3 clicks of tightness. put generator back into pocket, brand name side up. Test generator to check impedances. If good, anchor with non-absorable suture if needed, close.
Friday, November 11, 2016
ACDF
Head towards anesthesia. towel roll under shoulders horizontally. head taped down - straight (important bc you don't want to fuse them turned, also helps identify midlines structures during surgery). arms tucked.
Mark midline cartilagenous structures, SCM, sternal notch. Use fluoro to localize your levels
- FYi - hyoid is approx C3, thyroid cartilage is approx C4-5, cricoid cartilage is approx C6
Small horizontal incision for 1-2 levels, big horizontal incision for 3 level, at 4+ levels you probably need CEA style vertical incision.
Dissect through skin to fat - once you approach platysma, put scissors underneath to elevate it and cut through it horizontally with bovie. Then dissect rostral and caudal in a sub-platysmal plane.
Find plane medial to SCM - and go straight through to to vertebrae. if you need to cut omohyoid, you can. SCM and carotid should be lateral, esophagus and trachea should be medial. If it doesn't dissect easily, you are in the wrong plane. Put in retractors (we often use clowards).
Use kittners to get all the fascia off the spine. Mountains are disks, valleys are vertebrae. Localize your operative level - recommend caspar pin into one of the vertebrae you are fusing; some people do needle into disc but this can theoretically poke a hole in the wrong annulus (does this increase risk of future disk hernation?) and then you have to pull it out and mark that level somehow and your mark could come off or it could slip out etc. To avoid adjacent segment disease, expose only the operative levels (ie. if you are doing 4-6 ACDF, expose only the bottom of 4 and the top of 6, only enough to get your plate in.
Once your level is confirmed, use bovie to elevate longus coli off just to the uncovertebral joints - don't go more lateral than that - you risk injury to sympathetic chain under longus coli, you also risk injury to verts. Your cloward (or whatever) retractors should be flush all the way to the vertebrae while doing this - otherwise if your bovie slips and there is no retractor stopping it, you will run into the carotid/IJ laterally or the esophagus medially. That is bad.
Measure depth and put in your color-coded retractor system of choice - Trimline, phantom, etc. Put in the remaining caspar pins, make sure they are all midline and lined up neatly. Use knife to cut square into your disk. Don't cut too aggressively lateral, thats where the vert lives.
Use a combination of curette (straight and curved), pituitary, kerrison and drill to remove the disk. Use the straight curette against the endplates to prepare them to receive fusion graft. you know you are at PLL when you feel the posterior edge of the disks. if there are osteophytes at the posterior aspect, drill them down.
Taking of PLL is controversial - some people always do it, some people rarely do it, some people say do it only if there is significant central stenosis/myelopathy (i.e. may not be needed if you are only doing foraminotomy). If you take PLL, then use nerve hook to get under PLL and flick towards center to get it away; you are done only when the shiny pretty dura is visible the entire length of the disk.
Make sure to run nerve hook at either edge to ensure the foramina are open and free. if they are not, take a kerrison and bite stuff off until they are open.
Make sure your endplates are prepared! Free of cartilage. That will prevent fusion. However do not be too aggressive, otherwise if you have no endplate then your graft will subsist (i.e. sink into cancellous bone)
Put in your sizing tool - it should fit very snug. determine appropriate size of graft. Then put in graft. it is our humble opinion that PEEK cages with tiny-ass bone insert fail more often than all-bone allografts. Whatever. something to ensure fusion. certainly do not leave in an all-metal/all-PEEK construct with no bone/fusion material. That is just silly. Hammer it in; ideal graft placement is relatively anterior, but fully flush at the anterior edge.
Plate over the system - screws should angle towards the center of each vertebrae (i.e. at bottom edge of bone screws will be angled inferior and medial.
Two view XR at the end.
Thoughts:
- A fast surgeon can complete a 1 level ACDF in 45 mins and a 2 level in 90. This is the goal to shoot for.
- if you ever do skip-level ACDFs (i.e you fuse 4-5 and 6-7), the intervening levels will fail and they will fail soon. additionally - if you fuse 5-6, because there is limited movement at 7-1 -- 6-7 may fail.
- when to do ACDF rather than PIF - when compression is mostly anterior rather than from ligamentum/uncovertebral hypertrophy, or when there is a kyphotic deformity you are trying to correct by jamming in a wedged graft
- if the disk is herniated such that it's in the middle of the vertebrae - consider corpectomy instead
- if there are many (i.e. entire subaxial cervical spine) levels of significant disk herniation causing bad cervical stenosis, and you don't want to do a 5 level ACDF, can consider big posterior lami and fusion instead. can also consider multi-level laminoplasty, either swing-door or french door. Risk for worsening of kyphosis, but typically not a problem unless >5 degrees of kyphosis already. Laminoplasty at C2 and C7 tend to cause a lot of pain but not so much in between (consider laminoplasty for C3-C6 + laminectomy C7 if needed). Can still do foraminotomy.
- if there are many (i.e. entire subaxial cervical spine) levels of significant disk hernation causing bad cervical stenosis, and the person is 90 years old or a medical train wreck and you want to get them off the table as fast as possible, then consider big posterior lami without fusion - although the risk of subsequent kyphotic deformity will be high
- if there is mostly uncovertebral disease and/or the patient is mostly having radiculopathy symptoms - consider posterior foraminotomy only.
Tuesday, November 8, 2016
original scan on bottom, new scan 1 year later on the top
- on original scan - L4-5 foraminal stenosis- you would expect an L4 radiculopathy since foraminal stenosis affects the exiting nerve root, and the L4 nerve root exist at L4-5.
- on new scan - now stenosis is at the lateral recess, you an see from the different slices that the traversing nerve roots get crushed by the disk. You would expect an L5 radic since the L5 nerve roots are passing by at this level.
Monday, September 19, 2016
Trauma Crani:
- after intubation, turn bed 90 degrees with the operative side facing away and the ET tube facing anesthesia.
- after intubation, turn bed 90 degrees with the operative side facing away and the ET tube facing anesthesia.
- position supine, head turned as much as you can - if patient is in c-collar or old, may need to bump shoulder in order to get enough posterolateral exposure of head
- shave generously, at least half head, sometimes whole head is easier.
- incision 1 cm off midline (plan incision 1 cm off midline, and then plan burr holes 1 cm from incision, that way you are 2 cm from sinus). Start anterior at the edge of hairline then travel back along the vertex - posterior limit is transverse sinus (and realistically, how far back you can reach on the head, which is a function of positioning). Travel laterally above the root of zygoma, then curve in front of ear - stay 1 cm away. Any closer to tragus and you risk violating EAM, any further and you risk facial nerve.
- use the knife to cut down to bone; over temporalis muscle, knife through skin to temporalis and then bovie through temporalis. If you hit STA, bipolar it until it stops bleeding.
- raney clips on scalp. not everyone likes these. some people like them edge to edge on scalp, some only like them where scalp is bleeding, some never use them at all. if you use them, make sure to include galea in clips.
- raney clips on scalp. not everyone likes these. some people like them edge to edge on scalp, some only like them where scalp is bleeding, some never use them at all. if you use them, make sure to include galea in clips.
- if not super emergent, consider stopping at pericranium and harvesting pericranial flap in case you get into frontal sinus. Also consider trying to save STA.... In a true crash crani, all you care about is speed; aim to get from skin incision to open dura in 15 mins, and to get out of the OR and into the ICU in less than hour.
- position periosteal at the root of the zygoma, scrape upwards to get temporalis muscle off in one piece quickly.
- retract flap using whatever system suits you; many like the towel clamps on flap and then some system to attach the towel clamps to
- Burr holes : keyhole, right above root of zygoma (as far down as you can get basically), posterior at least 1-2 cm above transverse sinus if not more, and then 1-2 up top - at least 2 cm off midline.
- remove remaining bone left by perforator drill bit, with straight curette, dissect off dura with woodson; some people will use other devices like penfield 1 or 3 to get additional dura off bone.
- turn the crani - start at the temporal bone. When turning on forehead, if you start at lateral edge of head and turn flap up towards forehead (rather than stating at the vertex and turning down) less risk of getting into frontal sinus.
- turn the crani - start at the temporal bone. When turning on forehead, if you start at lateral edge of head and turn flap up towards forehead (rather than stating at the vertex and turning down) less risk of getting into frontal sinus.
- if blood starts coming out at copious volume when you are near the vertex, it may just be a venous lake in bone; apply bone wax to bone edges; if it stops, its not the sinus. If it is the sinus, do not panic, put a big gelfoam over it and wait for it to stop. Do not use small gelfoam for this, there's a chance it'll get sucked into the sinus and cause an obstruction.
- dissect dura off bone, remove bone flap
- if you caused a durotomy with the footplate or with the burr holes, reasonable to make a c-shaped dural opening. Alternatively, if you think there is a good chance you'll be able to do dural re-closure/bone replacement immediately, also reasonable to make c-shaped incision as it is easier to close.
- if brain is very swollen and you are aiming for maximum decompression, open dura in stellate manner.
- if brain is swollen, make sure to extend the craniotomy flush to the floor of the middle fossa - if this is not done, the temporal lobe will not be decompressed and the patient can still have temporal lobe swelling and uncal herniation even after decompression. use rongeur in biting, not tearing or twisting motions.
- if brain is swollen, make sure to extend the craniotomy flush to the floor of the middle fossa - if this is not done, the temporal lobe will not be decompressed and the patient can still have temporal lobe swelling and uncal herniation even after decompression. use rongeur in biting, not tearing or twisting motions.
- evacuate hematoma.
- if you are doing a loculated/membraney subdural, do not strip the membrane. Peeling off membrane takes off a layer or two of cortical cells and predisposes people to seizures/status.
- inspect dural edge and bone edge for bleeding, acquire hemostasis.
- replace bone flap if you think appropriate.
- always leave a drain. always leave a drain. always leave a drain. curl it around so the entire flap is included within the area that is being drained. the point is you can't get perfect hemostasis because you don't want to apply so much electrocautery to the flap so that it won't heal, but if you leave without a drain or place it suboptimally people can get really bad subgaleal hematomas. and a subgaleal hematoma in a person with no bone = epidural hematoma.
- close the galea. remember the words of harvey cushing.
- staples to skin for speed
- if you are doing a loculated/membraney subdural, do not strip the membrane. Peeling off membrane takes off a layer or two of cortical cells and predisposes people to seizures/status.
- inspect dural edge and bone edge for bleeding, acquire hemostasis.
- replace bone flap if you think appropriate.
- always leave a drain. always leave a drain. always leave a drain. curl it around so the entire flap is included within the area that is being drained. the point is you can't get perfect hemostasis because you don't want to apply so much electrocautery to the flap so that it won't heal, but if you leave without a drain or place it suboptimally people can get really bad subgaleal hematomas. and a subgaleal hematoma in a person with no bone = epidural hematoma.
- close the galea. remember the words of harvey cushing.
- staples to skin for speed
Sunday, August 28, 2016
ICP management for dummies
So you get a page that the ICPs are high. What do you do?
Step 1: Do not panic. If you panic, everyone panics.
Step 2: See the patient. Evaluate. Think. Are they sitting upright, talking to you, stating they feel great with the ICP monitor reading some absurd number? It might be wrong. Re-zero, troubleshoot, scan if you need to. Alternatively, some people live at ICPs of 50-70 - think of all the pseduotumor patients. Just because the number is high does not necessarily mean something bad is happening; the reason why elevated ICPs are bad is because of concern for decreased cerebral perfusion. If someone is neurologically intact, then they are clearly perfusing their brain. You can sit on high-grade SAH patients with ICPs in the 60s as long as they are awake and talking to you and without deficit; However, if someone is posturing with a blown pupil, then something probably has to be done about it.
Step 3: ABC - Airway, breathing, circulation first, always. Are they intubated? Do they need to be? Are they hemodynamically stable? If they are not protecting their airway, secure the airway first. If they are unstably hypotensive, push fluids/pressors/blood as appropriate - because if you send an unstable patient to CT and they code in the scanner, you're retarded
Step 4: ICP algorithm
- First thing, get the head of bed up, ensure neck is not excessively extended, or worse, flexed. This sounds really small but ensuring great venous drainage can buy you a lot of ICP lowering benefit. The only thing that should stop you from doing this is (very) unstable T/L spine fractures - which is why you should radiographically clear spines on any TBI as soon as possible, in anticipation of this occurring in the future, so you're not hurridly looking for total spine scans when your patient is herniating. I was once told by a team that they could not sit up a patient with very elevated ICPs because of straight-leg precautions after angio. Weigh the risk of groin hematoma vs rapid death from herniation, realize why this is absurd, and get the head of bed up. In the very least, you can do reverse-T.
- Second thing, assuming they are intubated, which people with ICP monitors often are, increase resp rate. Learn how to work your hospital ventilators so you know how to do this. This move does not buy you long-term ICP relief, but it gets you a rapid lowering that buys you time to think and plan your next moves. Reiteration: this is not a viable option for ICP management in any time-course longer than a few hours. You will get rebound effects, and after a while you will get worsening of ischemia from lower oxygen delivery to brain. A second thought: look at the rest of the parameters of your vent while you're there - if the PEEP is absurdly high, consider lowering it to improve venous drainage. Preventing herniation takes precedence over ARDSnet protocols or whatever else. Get an ABG to confirm no CO2 retention but don't wait for the results before you change the rate.
- Third, hyperosmolar therapy. Mannitol vs hypertonic saline. They both have rebound effects - the brain parenchymal osms will reset, and then you will get rebound edema. This happens much faster with mannitol than with hypertonic saline; mannitol buys you a few hours, with hypertonics you get about 24-48 hours. This is why we only use mannitol when we are about to crash someone to the OR for decompression, or if you can't get your serum osms up with hypertonic saline alone for whatever reason. Otherwise, I will always use 3% first. FYI a 250cc bag of 3% pushed at max speed delivers the same salt load as a 25cc bolus of 23% saline and you don't have to deal with an ICU team getting antsy about central lines. FYI#2 never allow central line placement to delay OR in someone who needs it stat; if its really needed, anesthesia can do it under the drapes while you operate. How much hypertonics can you give? I start getting nervous when Na>160 and Osm>320. At that point, if its not working, its not working and you should try something else.
NB: problematically, if the BBB is compromised, the hypertonics will go directly into that tissue and draw fluid in and actually worsen edema. This isn't a problem if your ICP issues are from a traumatic acute subdural, but it is a big problem if they are from a big ischemic stroke -- all the mannitol will go into the bad side and swell and crush the good side even more.
- Fourth, sedation. I like boluses of propofol, which is often rapid and very effective at lowering ICPs, as long as blood pressures will tolerate. From subjective experience, propofol has more ICP lowering effect than other things, like versed or precedex, but I have no objective evidence to support that. Pushing sedation of whatever agent of your choice is a fast and effective way of both lowering ICPs and (theoretically) of lowering cerebral oxygen demand. This is titratable - you can go from a few propofol boluses to get someone through a procedure or an event, all the way to sedating someone into burst suppression for recalcitrant ICP issues. If you're going to burst suppress, pentobarb >> extremely high dose propofol (i.e. 200mcg/kg/hr).
- Fifth, paralytics. When all else fails, you can put people on drips of non-depolarizing paralytics; this is especially helpful for people who are bucking the vent or shivering etc and you think this movement is contributing to ICP elevation. NB: most people prefer cis drips for this because of its predictable pharmacokinetics, whereas for roc or vec, the elimination is non-linear is if you put people on a drip of roc or vec for several days, it can take them > 24-48 hours to wake up and start moving, *even if* their train of fours show 4/4 twitches, so give people generous amount of time before you start brain death testing. You can have up to 80% residual neuromuscular blockade and have 4/4 twitches. Needless to say, do not remove your ICP monitor until they wake up, because you won't have an exam to follow.
Side notes:
- figure in early whether or not people will be surgical candidates. In people who are not surgical candidate, I would use exclusively salt, sedation, and paralytics and only use hyperventilation transiently to get someone through an ICP spike; I wouldn't use mannitol unless I couldn't get their osms high enough with Na alone.
- in people with EVDs, draining off CSF (or lowering the drain) can be an excellent option to temporarily lower ICPs, especially if there is a hydrocephalus component. If you get a dramatic ICP response to the drainage of a relatively small amount of CSF, that tells worlds about the ventricular compliance and about where you may be sitting on the pressure-volume curve.
So you get a page that the ICPs are high. What do you do?
Step 1: Do not panic. If you panic, everyone panics.
Step 2: See the patient. Evaluate. Think. Are they sitting upright, talking to you, stating they feel great with the ICP monitor reading some absurd number? It might be wrong. Re-zero, troubleshoot, scan if you need to. Alternatively, some people live at ICPs of 50-70 - think of all the pseduotumor patients. Just because the number is high does not necessarily mean something bad is happening; the reason why elevated ICPs are bad is because of concern for decreased cerebral perfusion. If someone is neurologically intact, then they are clearly perfusing their brain. You can sit on high-grade SAH patients with ICPs in the 60s as long as they are awake and talking to you and without deficit; However, if someone is posturing with a blown pupil, then something probably has to be done about it.
Step 3: ABC - Airway, breathing, circulation first, always. Are they intubated? Do they need to be? Are they hemodynamically stable? If they are not protecting their airway, secure the airway first. If they are unstably hypotensive, push fluids/pressors/blood as appropriate - because if you send an unstable patient to CT and they code in the scanner, you're retarded
Step 4: ICP algorithm
- First thing, get the head of bed up, ensure neck is not excessively extended, or worse, flexed. This sounds really small but ensuring great venous drainage can buy you a lot of ICP lowering benefit. The only thing that should stop you from doing this is (very) unstable T/L spine fractures - which is why you should radiographically clear spines on any TBI as soon as possible, in anticipation of this occurring in the future, so you're not hurridly looking for total spine scans when your patient is herniating. I was once told by a team that they could not sit up a patient with very elevated ICPs because of straight-leg precautions after angio. Weigh the risk of groin hematoma vs rapid death from herniation, realize why this is absurd, and get the head of bed up. In the very least, you can do reverse-T.
- Second thing, assuming they are intubated, which people with ICP monitors often are, increase resp rate. Learn how to work your hospital ventilators so you know how to do this. This move does not buy you long-term ICP relief, but it gets you a rapid lowering that buys you time to think and plan your next moves. Reiteration: this is not a viable option for ICP management in any time-course longer than a few hours. You will get rebound effects, and after a while you will get worsening of ischemia from lower oxygen delivery to brain. A second thought: look at the rest of the parameters of your vent while you're there - if the PEEP is absurdly high, consider lowering it to improve venous drainage. Preventing herniation takes precedence over ARDSnet protocols or whatever else. Get an ABG to confirm no CO2 retention but don't wait for the results before you change the rate.
- Third, hyperosmolar therapy. Mannitol vs hypertonic saline. They both have rebound effects - the brain parenchymal osms will reset, and then you will get rebound edema. This happens much faster with mannitol than with hypertonic saline; mannitol buys you a few hours, with hypertonics you get about 24-48 hours. This is why we only use mannitol when we are about to crash someone to the OR for decompression, or if you can't get your serum osms up with hypertonic saline alone for whatever reason. Otherwise, I will always use 3% first. FYI a 250cc bag of 3% pushed at max speed delivers the same salt load as a 25cc bolus of 23% saline and you don't have to deal with an ICU team getting antsy about central lines. FYI#2 never allow central line placement to delay OR in someone who needs it stat; if its really needed, anesthesia can do it under the drapes while you operate. How much hypertonics can you give? I start getting nervous when Na>160 and Osm>320. At that point, if its not working, its not working and you should try something else.
NB: problematically, if the BBB is compromised, the hypertonics will go directly into that tissue and draw fluid in and actually worsen edema. This isn't a problem if your ICP issues are from a traumatic acute subdural, but it is a big problem if they are from a big ischemic stroke -- all the mannitol will go into the bad side and swell and crush the good side even more.
- Fourth, sedation. I like boluses of propofol, which is often rapid and very effective at lowering ICPs, as long as blood pressures will tolerate. From subjective experience, propofol has more ICP lowering effect than other things, like versed or precedex, but I have no objective evidence to support that. Pushing sedation of whatever agent of your choice is a fast and effective way of both lowering ICPs and (theoretically) of lowering cerebral oxygen demand. This is titratable - you can go from a few propofol boluses to get someone through a procedure or an event, all the way to sedating someone into burst suppression for recalcitrant ICP issues. If you're going to burst suppress, pentobarb >> extremely high dose propofol (i.e. 200mcg/kg/hr).
- Fifth, paralytics. When all else fails, you can put people on drips of non-depolarizing paralytics; this is especially helpful for people who are bucking the vent or shivering etc and you think this movement is contributing to ICP elevation. NB: most people prefer cis drips for this because of its predictable pharmacokinetics, whereas for roc or vec, the elimination is non-linear is if you put people on a drip of roc or vec for several days, it can take them > 24-48 hours to wake up and start moving, *even if* their train of fours show 4/4 twitches, so give people generous amount of time before you start brain death testing. You can have up to 80% residual neuromuscular blockade and have 4/4 twitches. Needless to say, do not remove your ICP monitor until they wake up, because you won't have an exam to follow.
Side notes:
- figure in early whether or not people will be surgical candidates. In people who are not surgical candidate, I would use exclusively salt, sedation, and paralytics and only use hyperventilation transiently to get someone through an ICP spike; I wouldn't use mannitol unless I couldn't get their osms high enough with Na alone.
- in people with EVDs, draining off CSF (or lowering the drain) can be an excellent option to temporarily lower ICPs, especially if there is a hydrocephalus component. If you get a dramatic ICP response to the drainage of a relatively small amount of CSF, that tells worlds about the ventricular compliance and about where you may be sitting on the pressure-volume curve.
Thursday, August 18, 2016
Open lumbar spine surgery - Part I
Types of tables:
- Jackson with OSI frame: standard setup that we use - has table with the rails and the pads for chest, hips, thighs. can adjust level of rods at each end so you get extra T or reverse-T - if you're doing say an L5-S1 fusion and the sacral slope is really steep, you can get feet up so you're working horizontally, etc etc. When putting people on OSI, put chest pad on upper chest- not directly against chin, not too low for women. Some people think you should straighten knees relative to hips so you avoid hypothetical risk of fusing people in hip flexion position, however in real life this is rarely a problem.
- Jackson table with Wilson frame: forced flexion over the arc. Gets you better access to the disc, which is useful if you're doing a discectomy, however can give you a false sense of security about how much decompression you have achieved. Like you feel nerve roots, you think they are decompressed because they are in forced flexion, but then when back goes straight everything closes off and then you're not decompressed anymore. Also as a note - if you're gonna use a Wilson frame, better to do Jackson table because its' lower - Wilson frame is high, unless you're really tall you're gonna be on 4 steps.
- Jackson table w Axis frame - can bend in half. Most useful for XLIF/DLIF when you position patient laterally, really opens up space between iliac crest and and ribs.
- Jackson flattop - useful vs standard table because its radiolucent and also no giant pedicle in the middle- so if you're doing an ALIF you have an easier time getting C-arm under table.
Arm positioning :
- up for middle T/L - make sure no hyperextension of arms back or too anterior, keep elbows soft, 90 degrees; protect against axillary nerve injury
- down/wrapped for C and upper T
Setting up the room:
- head towards anesthesia, scrub nurse at legs, surgeon on each side of back.
- assuming the pathology doesn't have laterality, choose the side of patient that doesn't have the base of the C-arm on it, so you're not moving your steps/foot pedals around every time the C-arm moves.
- when you move c-arm out of the way, move it towards anesthesia/patient head - because your scrub nurse is by patient's legs and you don't want the c-arm between you
General principles
- Look at CT scan before every case; look at anatomy - planning on lami, make sure there is actually lamina at every level. Some people have big facets, and there may be no lamina to take at a given level, or some atypical scoliosis, such that you have to drill at an angle. if you meant to do a lami, without a fusion, and you violate the facet capsule or worse drill facet when you didn't mean to, then you're hosed because you've destabilized that level.
- determine pedicle sizes on scan, pick out screw size and length beforehand, write them on the board
Process:
- localize level first with c-arm.
- mark incision
- prep and drape wide
- 15 blade through skin
- buzz through fat with impunity
- put finger in and feel for the spinous processes - directly above spinous process is fascia. every time. every time. that's where it is. lateral to spinous process will be bumps for paraspinous muscle.
- take cobb and scrape last bit of fat off fascia. In everyone except super old frail people with terrible fascia, you can scrape and you will not accidentally break through as long as you're not jabbing aggressively
- cut on either side of spinous process. Take cobb, put raytech on it, and scrape down along spinous process. you should be sub-periosteal. if you are not, you will be in muscle - it will bleed like crazy and the patient will have a lot more postop pain. alternative to cobb is bovie down bone (but be careful directly lateral to lamina- there lie nerve roots). problem with bovie is it will cause paraspinal muscles to tighten and fight you - option to ask anesthesia for short-acting muscle relaxant. if someone has pacemaker or otherwise can't tolerate electrocautery, then you have to do the bipolar and cut thing which sucks.
- your cobb will then land on lamina. scrape ROSTRAL and LATERAL - next you will land on a valley of bone which is the superior aspect of pars, between the superior facet and the lamina. then scrape your cobb more rostral and more lateral - you will go over a bump, this is the facet capsule, and then you will land on the transverse process. If you're only doing a lami, you don't need to see TP but if you're doing a fusion you will need to expose a lot of the TP.
- THE PEDICLE IS ROSTRAL TO THE SPINOUS PROCESS. once you find the spinous process, you have to go rostral to find pedicle. this is especially true in thoracic spine - if you try to take shortcuts and aren't thoughtful, and you go for the pedicle directly next to the spinous process, you will put the screw in the wrong level.
- the anatomy is as such: from rostral to caudal superior facet, TP/pedicle, lamina, inferior facet, with the pars representing the bridge of bone connecting superior to inferior facet
NB
- if you're doing a lami, you cannot drill all the way lateral - if you take pars, you destabilize
- you cannot bovie with impunity directly lateral to the spinous process/lamina - that's where the nerve roots come out. (if patient jumps and they are not paralyzed, its because your electrocautery hit a nerve root). this is why we don't paralyze spine cases.
- there is often a large artery next to pars - if it starts bleeding like crazy, don't dive after it with electrocautery, as you might hit nerve roots.
- if someone has giant, super hypertrophied facets, you might need to drill off the inferior aspect of facet to access entry point for pedicle screw - be careful - you can only drill approx 1/3 of facet joint before you destabilize
Thoughts on cervical decompression
- if most of someone's compression is coming from disk, you can do ACDF
- if there is a hypertrophied ligamentum flavum or facet arthropathy, you can get a lot more decompression of spinal cord if you come from the back
Types of tables:
- Jackson with OSI frame: standard setup that we use - has table with the rails and the pads for chest, hips, thighs. can adjust level of rods at each end so you get extra T or reverse-T - if you're doing say an L5-S1 fusion and the sacral slope is really steep, you can get feet up so you're working horizontally, etc etc. When putting people on OSI, put chest pad on upper chest- not directly against chin, not too low for women. Some people think you should straighten knees relative to hips so you avoid hypothetical risk of fusing people in hip flexion position, however in real life this is rarely a problem.
- Jackson table with Wilson frame: forced flexion over the arc. Gets you better access to the disc, which is useful if you're doing a discectomy, however can give you a false sense of security about how much decompression you have achieved. Like you feel nerve roots, you think they are decompressed because they are in forced flexion, but then when back goes straight everything closes off and then you're not decompressed anymore. Also as a note - if you're gonna use a Wilson frame, better to do Jackson table because its' lower - Wilson frame is high, unless you're really tall you're gonna be on 4 steps.
- Jackson table w Axis frame - can bend in half. Most useful for XLIF/DLIF when you position patient laterally, really opens up space between iliac crest and and ribs.
- Jackson flattop - useful vs standard table because its radiolucent and also no giant pedicle in the middle- so if you're doing an ALIF you have an easier time getting C-arm under table.
Arm positioning :
- up for middle T/L - make sure no hyperextension of arms back or too anterior, keep elbows soft, 90 degrees; protect against axillary nerve injury
- down/wrapped for C and upper T
Setting up the room:
- head towards anesthesia, scrub nurse at legs, surgeon on each side of back.
- assuming the pathology doesn't have laterality, choose the side of patient that doesn't have the base of the C-arm on it, so you're not moving your steps/foot pedals around every time the C-arm moves.
- when you move c-arm out of the way, move it towards anesthesia/patient head - because your scrub nurse is by patient's legs and you don't want the c-arm between you
General principles
- Look at CT scan before every case; look at anatomy - planning on lami, make sure there is actually lamina at every level. Some people have big facets, and there may be no lamina to take at a given level, or some atypical scoliosis, such that you have to drill at an angle. if you meant to do a lami, without a fusion, and you violate the facet capsule or worse drill facet when you didn't mean to, then you're hosed because you've destabilized that level.
- determine pedicle sizes on scan, pick out screw size and length beforehand, write them on the board
Process:
- localize level first with c-arm.
- mark incision
- prep and drape wide
- 15 blade through skin
- buzz through fat with impunity
- put finger in and feel for the spinous processes - directly above spinous process is fascia. every time. every time. that's where it is. lateral to spinous process will be bumps for paraspinous muscle.
- take cobb and scrape last bit of fat off fascia. In everyone except super old frail people with terrible fascia, you can scrape and you will not accidentally break through as long as you're not jabbing aggressively
- cut on either side of spinous process. Take cobb, put raytech on it, and scrape down along spinous process. you should be sub-periosteal. if you are not, you will be in muscle - it will bleed like crazy and the patient will have a lot more postop pain. alternative to cobb is bovie down bone (but be careful directly lateral to lamina- there lie nerve roots). problem with bovie is it will cause paraspinal muscles to tighten and fight you - option to ask anesthesia for short-acting muscle relaxant. if someone has pacemaker or otherwise can't tolerate electrocautery, then you have to do the bipolar and cut thing which sucks.
- your cobb will then land on lamina. scrape ROSTRAL and LATERAL - next you will land on a valley of bone which is the superior aspect of pars, between the superior facet and the lamina. then scrape your cobb more rostral and more lateral - you will go over a bump, this is the facet capsule, and then you will land on the transverse process. If you're only doing a lami, you don't need to see TP but if you're doing a fusion you will need to expose a lot of the TP.
- THE PEDICLE IS ROSTRAL TO THE SPINOUS PROCESS. once you find the spinous process, you have to go rostral to find pedicle. this is especially true in thoracic spine - if you try to take shortcuts and aren't thoughtful, and you go for the pedicle directly next to the spinous process, you will put the screw in the wrong level.
- the anatomy is as such: from rostral to caudal superior facet, TP/pedicle, lamina, inferior facet, with the pars representing the bridge of bone connecting superior to inferior facet
NB
- if you're doing a lami, you cannot drill all the way lateral - if you take pars, you destabilize
- you cannot bovie with impunity directly lateral to the spinous process/lamina - that's where the nerve roots come out. (if patient jumps and they are not paralyzed, its because your electrocautery hit a nerve root). this is why we don't paralyze spine cases.
- there is often a large artery next to pars - if it starts bleeding like crazy, don't dive after it with electrocautery, as you might hit nerve roots.
- if someone has giant, super hypertrophied facets, you might need to drill off the inferior aspect of facet to access entry point for pedicle screw - be careful - you can only drill approx 1/3 of facet joint before you destabilize
Thoughts on cervical decompression
- if most of someone's compression is coming from disk, you can do ACDF
- if there is a hypertrophied ligamentum flavum or facet arthropathy, you can get a lot more decompression of spinal cord if you come from the back
Wednesday, August 17, 2016
Cranioplasty
- just like for a decompressive hemicrani - when you're making a giant trauma flap incision, you have to think about positioning such that you can reach the back of the head/most posterior aspect of incision. If someone has excellent neck mobility, you can just put them flat and turn head all the way over. If someone doesn't have good neck mobility (ie. c-collar, really old, arthritis, contractures) then you put a bump under the shoulder to get you access to back of head. similar principle for shunt - shoulder bump to straighten out neck for optimal tunneling.
- putting someone on a horse-shoe means that you can reach more posterior around their head more easily - think about your hand/wrist position relative to head vs them being flat on a table, and having the table block you
- shave around the incision area only
- cover eyes well with tegaderm, and then again with the 1000 drape, put drape as low across brow as you can to get the biggest field, but make sure to shield eyes - chlorhexadine is very caustic to corneas
- stuff xeroform into the ears only if you intend to use a chlorhexadine prep - it is ototoxic. if you are painting betadine or using alcohol only this step is skiappable
- prep and drape wide - you will have to tunnel a drain out of your flap
- feel bone edges under incision - if you have bone under incision you can cut all the way down. if you do not, you have to be careful - go thru skin with knife and then carefully bovie/dissect because its scalp - dura - brain. you will definitely not have any bone over the squamous temporal bone, because if you did a good decompression originally, you put a burr hole right at the pterion/over root of zygoma, and you kerrison'd all the way flush to the floor of the middle fossa for good temporal lobe decompression.
- for temporalis muscle, carefully dissect it off the dura - if you don't dissect between temporalis muscle and scalp, you can have a better chance of not causing a frontalis palsy. this is also a good place to find the bone edge and begin to develop the plane between periostium and dura.
- if you are lucky, there will be a good plane between periostium and dura, and you can follow that plane all the way around - you are not done until you see all bone edges. Put screws into the bone flap, push it flush all the way against anterior aspect of bone. if temporalis muscle is large and healthy, you can just close (fascia on fascia) anterior temporalis muscle against posterior. If its kind of bad looking, consider putting in a mesh over the cranial defect where temporal bone used to be, to buttress it and prevent a hollowing defect later on.
- if you are not lucky, the whole thing will be scarred down and socked in and you just have to create a plane - find the bone, do not violate dura.
- leave a subgaleal drain, use hemovac (same width of drain all the way around vs JP which is a wider drain; HMV hurts less coming out, creates a smaller hole to close when removed). Always leave drain because it will bleed a lot and you will not be able to use aggressive electrocautery because otherwise it won't heal well.
- tunnel your drain before you close, tunnel it outside of your flap becusae it will heal better. do not sew in drain until you finish closing, otherwise you'll shift the position of the drain in head and where you tunnel it out may not be as sterile as the rest of your field so do that last.
- close galea with 3-0 vicryl with C-23 needle
- close skin with either absorbable sutures or staples - for healthy people who will heal well, choose absorbable - it looks nicer, plus you don't have to hurt people 2 weeks later when you take them out. If there is any question about whether someone will heal an incision though, use staples - you can hypothetically leave staples in forever. also they are faster.
- if you are struggling to close, either because scalp is really scarred or tight, consider nylon sutures with horizontal mattress- you can use it to pull the scalp incrementally closer together
- just like for a decompressive hemicrani - when you're making a giant trauma flap incision, you have to think about positioning such that you can reach the back of the head/most posterior aspect of incision. If someone has excellent neck mobility, you can just put them flat and turn head all the way over. If someone doesn't have good neck mobility (ie. c-collar, really old, arthritis, contractures) then you put a bump under the shoulder to get you access to back of head. similar principle for shunt - shoulder bump to straighten out neck for optimal tunneling.
- putting someone on a horse-shoe means that you can reach more posterior around their head more easily - think about your hand/wrist position relative to head vs them being flat on a table, and having the table block you
- shave around the incision area only
- cover eyes well with tegaderm, and then again with the 1000 drape, put drape as low across brow as you can to get the biggest field, but make sure to shield eyes - chlorhexadine is very caustic to corneas
- stuff xeroform into the ears only if you intend to use a chlorhexadine prep - it is ototoxic. if you are painting betadine or using alcohol only this step is skiappable
- prep and drape wide - you will have to tunnel a drain out of your flap
- feel bone edges under incision - if you have bone under incision you can cut all the way down. if you do not, you have to be careful - go thru skin with knife and then carefully bovie/dissect because its scalp - dura - brain. you will definitely not have any bone over the squamous temporal bone, because if you did a good decompression originally, you put a burr hole right at the pterion/over root of zygoma, and you kerrison'd all the way flush to the floor of the middle fossa for good temporal lobe decompression.
- for temporalis muscle, carefully dissect it off the dura - if you don't dissect between temporalis muscle and scalp, you can have a better chance of not causing a frontalis palsy. this is also a good place to find the bone edge and begin to develop the plane between periostium and dura.
- if you are lucky, there will be a good plane between periostium and dura, and you can follow that plane all the way around - you are not done until you see all bone edges. Put screws into the bone flap, push it flush all the way against anterior aspect of bone. if temporalis muscle is large and healthy, you can just close (fascia on fascia) anterior temporalis muscle against posterior. If its kind of bad looking, consider putting in a mesh over the cranial defect where temporal bone used to be, to buttress it and prevent a hollowing defect later on.
- if you are not lucky, the whole thing will be scarred down and socked in and you just have to create a plane - find the bone, do not violate dura.
- leave a subgaleal drain, use hemovac (same width of drain all the way around vs JP which is a wider drain; HMV hurts less coming out, creates a smaller hole to close when removed). Always leave drain because it will bleed a lot and you will not be able to use aggressive electrocautery because otherwise it won't heal well.
- tunnel your drain before you close, tunnel it outside of your flap becusae it will heal better. do not sew in drain until you finish closing, otherwise you'll shift the position of the drain in head and where you tunnel it out may not be as sterile as the rest of your field so do that last.
- close galea with 3-0 vicryl with C-23 needle
- close skin with either absorbable sutures or staples - for healthy people who will heal well, choose absorbable - it looks nicer, plus you don't have to hurt people 2 weeks later when you take them out. If there is any question about whether someone will heal an incision though, use staples - you can hypothetically leave staples in forever. also they are faster.
- if you are struggling to close, either because scalp is really scarred or tight, consider nylon sutures with horizontal mattress- you can use it to pull the scalp incrementally closer together
Saturday, June 25, 2016
Differential Diagnosis of Intra-Axial Brain Tumor based on appearance in Adults
Intra-axial: 75% astrocytic tumor or mets
Tumors that invade the corpus callosum and cross midline
- GBM (rarely has leptomeningeal spread)
- Lymphoma (can have leptomeningeal spread, typically homogenously enhancing but in immunocompromised hosts can appear ring-enhancing or heterogenous)
- Mets (can have leptomeningeal enhancement, often cystic)
Multiple lesions
- usually mets
- multifocal GBM - rarer but exists
- gliomatosis cerebri
- lymphoma can appear as multiple lesions
- CNS metatstases - i.e. medulloblastoma, ependymoma, oligodendro
Genetic diseases that cause multiple tumors :
- NF1 (optic glioma, astrocytoma)
- NF2 (meningiomas, schwannoma, ependymoma of brain and spinal cord)
- Tuberous Sclerosis (SEGA, ependymoma)
- VHL (cerebellar/retinal/spinal cord hemangioblastomas, endolymphatic sac tumor)
Cortical-based
Oligodendro
- often cortical/subcortical, and are shown on MRI "extending all the way to the cortex"
- classically frontal but can be in any lobe.
- 70-90% calcified.
- usually T1 dark, T2 bright (except calcified areas - which will show up as T2 dark/T2* blooming
- 50% will enhance with Gad - usually heterogenously. Gad enhancement is not a reliable indicator of grade.
- Typically do not restrict on DWI. Often older people (40-50s+)
Ganglioglioma
- mix of glial and neuronal cells (if the glial component de-differentiates it turns into a GBM; if the neuronal component de-differentiates it turns into a neuroblastoma).
- 45% of the time it will appear as a cyst with mural nodule in a cortical area, but it can be very variable in appearance--- simple cyst with small mural nodule, complex cyst with large, heterogenous mural nodule, solid tumor only.
- The mural nodule has variable enhancement - sometimes enhances vividly
- not much edema.
- Usually affects children and young adults.
DNET:
- arise from cortical or deep grey matter.
- Predilection for temporal lobes.
- often associated with cortical dysplasia
- often cause intractable seizures.
- May be cyst with mural nodule.
- Enhances 20-30% of the time.
- T1 isodense, T2 bright with sometimes "bubbly" appearance
- not much peritumoral edema.
- Rarely grow in size over time, excellent oncologic prognosis but often removed because of the intractable seizures they cause - seizures often stop when the tumor is resected.
- Typically affects children and young adults.
Of note: in clinical practice, DNETs and GG often appear very similar radiographically, are both T1-dark, T2-bright tumors that rarely enhance (DNET can sometimes appear as ring-enhancing but not commonly, GG are a little more likely to be cystic with mural nodule, etc), with predilections for the temporal lobe, cause seizures, and have excellent seizure freedom outcomes with resection. See this retrospective series from Turkey with 52 patients.
PXA:
- cortical tumors with cystic component and vivid enhancement with Gad.
- 98% supratentorial. Mostly temporal.
- Often low grade/slow growing without much edema.
- T1 iso/hypointense. 50-60% cyst with mural nodule that enhances strongly. T2 - cyst often looks different than CSF due to proteinacious content.
- Avascular on angio despite vidid enhancement.
- Often affects kids and young adults.
Contain T1-bright tissue that resembles fat
Lipoma
- always located in subarachnoid spaces - believed to be from maldevelopment of meninx primativa (subarachnoid precursor)
- pericallosal (can wrap around corpus callosum - associated with agensis of corpus callosum in 50% of cases) - 45%
- quadrigeminal cistern (associated with underdevelopment of inferior colliculus) - 35%
- suprasellar cistern - 15%
- CP angle - 10%
- sylvian fissure - 5%
Dermoid:
- Can be thought of as on the spectrum from epidermoid (only squamous epithelial tissue) to dermoid (ectodermal only) to teratomas (contain tissue from all 3 embryonic layers)
- They appear bright on T1 but there is actually no fat in them. They contain squamous epithelial tissue + ectodermal appendages like hair follicles, sweat glands, sebaceous glands - which secrete sebum which looks T1 bright (from cholesterol and other things). Technically no adipose tissue, because that is from mesenchymal tissue so it would technically be a teratoma if there was mesenchyme
- Often midline - suprasellar, subfrontal, cerebellar vermis
- They can rupture and cause a chemical meningitis - leptomeningeal enhancement
- Very rarely will transform into squamous carcinoma
Teratoma - contains tissue from all 3 embryonic layers.
- rare in general population - however 25-50% of fetal brain tumor
- usually pineal or suprasellar
- T1 bright from fat, can enhance
Sources:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3876643/
http://www.radiologyassistant.nl/en/p47f86aa182b3a/brain-tumor-systematic-approach.html
http://radiopaedia.org/articles/dysembryoplastic-neuroepithelial-tumour
http://radiopaedia.org/articles/ganglioglioma
http://radiopaedia.org/articles/pleomorphic-xanthoastrocytoma
http://radiopaedia.org/articles/intracranial-dermoid-cyst-1
http://radiopaedia.org/articles/gliomatosis-cerebri
http://radiopaedia.org/articles/leptomeningeal-enhancement
Intra-axial: 75% astrocytic tumor or mets
Tumors that invade the corpus callosum and cross midline
- GBM (rarely has leptomeningeal spread)
- Lymphoma (can have leptomeningeal spread, typically homogenously enhancing but in immunocompromised hosts can appear ring-enhancing or heterogenous)
- Mets (can have leptomeningeal enhancement, often cystic)
Multiple lesions
- usually mets
- multifocal GBM - rarer but exists
- gliomatosis cerebri
- lymphoma can appear as multiple lesions
- CNS metatstases - i.e. medulloblastoma, ependymoma, oligodendro
Genetic diseases that cause multiple tumors :
- NF1 (optic glioma, astrocytoma)
- NF2 (meningiomas, schwannoma, ependymoma of brain and spinal cord)
- Tuberous Sclerosis (SEGA, ependymoma)
- VHL (cerebellar/retinal/spinal cord hemangioblastomas, endolymphatic sac tumor)
Cortical-based
Oligodendro
- often cortical/subcortical, and are shown on MRI "extending all the way to the cortex"
- classically frontal but can be in any lobe.
- 70-90% calcified.
- usually T1 dark, T2 bright (except calcified areas - which will show up as T2 dark/T2* blooming
- 50% will enhance with Gad - usually heterogenously. Gad enhancement is not a reliable indicator of grade.
- Typically do not restrict on DWI. Often older people (40-50s+)
Ganglioglioma
- mix of glial and neuronal cells (if the glial component de-differentiates it turns into a GBM; if the neuronal component de-differentiates it turns into a neuroblastoma).
- 45% of the time it will appear as a cyst with mural nodule in a cortical area, but it can be very variable in appearance--- simple cyst with small mural nodule, complex cyst with large, heterogenous mural nodule, solid tumor only.
- The mural nodule has variable enhancement - sometimes enhances vividly
- not much edema.
- Usually affects children and young adults.
DNET:
- arise from cortical or deep grey matter.
- Predilection for temporal lobes.
- often associated with cortical dysplasia
- often cause intractable seizures.
- May be cyst with mural nodule.
- Enhances 20-30% of the time.
- T1 isodense, T2 bright with sometimes "bubbly" appearance
- not much peritumoral edema.
- Rarely grow in size over time, excellent oncologic prognosis but often removed because of the intractable seizures they cause - seizures often stop when the tumor is resected.
- Typically affects children and young adults.
Of note: in clinical practice, DNETs and GG often appear very similar radiographically, are both T1-dark, T2-bright tumors that rarely enhance (DNET can sometimes appear as ring-enhancing but not commonly, GG are a little more likely to be cystic with mural nodule, etc), with predilections for the temporal lobe, cause seizures, and have excellent seizure freedom outcomes with resection. See this retrospective series from Turkey with 52 patients.
PXA:
- cortical tumors with cystic component and vivid enhancement with Gad.
- 98% supratentorial. Mostly temporal.
- Often low grade/slow growing without much edema.
- T1 iso/hypointense. 50-60% cyst with mural nodule that enhances strongly. T2 - cyst often looks different than CSF due to proteinacious content.
- Avascular on angio despite vidid enhancement.
- Often affects kids and young adults.
Contain T1-bright tissue that resembles fat
Lipoma
- always located in subarachnoid spaces - believed to be from maldevelopment of meninx primativa (subarachnoid precursor)
- pericallosal (can wrap around corpus callosum - associated with agensis of corpus callosum in 50% of cases) - 45%
- quadrigeminal cistern (associated with underdevelopment of inferior colliculus) - 35%
- suprasellar cistern - 15%
- CP angle - 10%
- sylvian fissure - 5%
Dermoid:
- Can be thought of as on the spectrum from epidermoid (only squamous epithelial tissue) to dermoid (ectodermal only) to teratomas (contain tissue from all 3 embryonic layers)
- They appear bright on T1 but there is actually no fat in them. They contain squamous epithelial tissue + ectodermal appendages like hair follicles, sweat glands, sebaceous glands - which secrete sebum which looks T1 bright (from cholesterol and other things). Technically no adipose tissue, because that is from mesenchymal tissue so it would technically be a teratoma if there was mesenchyme
- Often midline - suprasellar, subfrontal, cerebellar vermis
- They can rupture and cause a chemical meningitis - leptomeningeal enhancement
- Very rarely will transform into squamous carcinoma
Teratoma - contains tissue from all 3 embryonic layers.
- rare in general population - however 25-50% of fetal brain tumor
- usually pineal or suprasellar
- T1 bright from fat, can enhance
Contain Calcifications
- Oligodendro - rare, but often calicfied
- Astrocytoma - common, but infrequently calcified
- Pinealcytoma - are not calcified in and of themselves, but contain the inherent calcifications of the pineal gland
- Craniopharyngioma - suprasellar
Cystic lesions with same intensity as CSF
- Arachnoid
- Neurenteric cyst
- Neurenteric cyst
- Enlarged virchow-robin space
Bright on T1
- most tumors are isodense or slightly hypodense on T1. T1 hyperintensity implies presence of one of the elements which are hyperintense on T1, which include the following:
- Fat -> Lipoma, Teratoma
- Cholesterol -> Craniopharyngioma, Dermoid, colloid cyst
- Melanin -> Melanoma met
- Subacute blood (intracellular or extracellular metHb) -> Hemorrhagic tumor - melanoma met, breast/lung (uncommonly bleed), follicular thyroid met, renal cell met, choriocarcinoma met. pituitary apoplexy
- Mineralization with paramagnetic divalent cations ( Ca, Copper, manganese )
- Proteinaceous fluid - neurentetic cyst
Dark on T2
- Most tumors are bright on T2 dude to high water content. Low T2 signal implies:
- hypercellularity -> PNET/medullo, lymphoma, mucinous adenoCA mets, high-cellularity parts of GBM (although GBM is usually T2 bright)
- calcifications -> oligo, astro, craniopharyngioma
- many flow voids -> hemangioblastoma
Contrast enhancement:
- Fundamentally based on the integrity of the BBB
Tumors that do not have a BBB will enhance vividly
- extra-axial masses like meningioma, schwannoma
- non-CNS tumors like lymphoma, metastatic lesions (breast, lung, etc)
- CNS tumors derived from tissue that does not have a BBB - pituitary, pineal (including germinoma and other pineal gland tumors), choroid plexus
Homogenous enhancement
- Mets
- Lymphoma (non immunocompromised)
- Meningioma - and its great mimics, hemangiopericytoma and dural based MALT lymphomas
- Schwannoma/neurofibroma
- Mural nodule of hemangioblastoma, JPA
- Germinoma, other pineal gland tumors
- Pituitary adenoma
- Ganglioglioma
Patchy Enhancement
- Mets
- GBM
- Non tumor things (like radiation necrosis)
Ring Enhancement
- MAGICALDR
- Mets
- Abscess
- GBM
- Ischemia (subacute stroke), Infection (neurocystercercosis, toxoplasma, TB, blasto/histo/crypto, nocardia, listeria)
- Contusion
- Alternative weird stuff (sarcoid, vasculitis, behcets)
- Lymphoma
- Demyelination (MS, ADEM)
- Radiation necrosis
Leptomeningeal Enhancement
- leptomeningeal carcinomatosis (breast, lung, melanoma, leukemia/lymphoma)
- from CNS tumors - GBM, medullo/PNET, ependymoma, choroid plexus carcinoma.
- infectious - meningitis (bacterial, viral, crypto, TB)
- sarcoid
- post-LP (<5%), post-surgery/hemorrhage/trauma
Sources:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3876643/
http://www.radiologyassistant.nl/en/p47f86aa182b3a/brain-tumor-systematic-approach.html
http://radiopaedia.org/articles/dysembryoplastic-neuroepithelial-tumour
http://radiopaedia.org/articles/ganglioglioma
http://radiopaedia.org/articles/pleomorphic-xanthoastrocytoma
http://radiopaedia.org/articles/intracranial-dermoid-cyst-1
http://radiopaedia.org/articles/gliomatosis-cerebri
http://radiopaedia.org/articles/leptomeningeal-enhancement
Friday, June 24, 2016
Bright
|
Dark
|
|
T1
|
- contrast
- fat/cholesterol (lipoma, teratoma, dermoid, lipomatous ependymoma, cholesteatoma) ** to differentiate fat from other T1-bright lesions, look for chemical shift artifact (dark band at edge of fat on one side, light band at the other)
- melanin
- early subacute blood
(intracellular metHb 3-7 days) and late
subacute blood (extraceullar metHb, 1-4 weeks)
- subacute thrombus (i.e.
venous sinus)
- protein-rich fluid
(colloid, rathke’s cleft, ectopic posterior pituitary, craniopharyngioma –
also cholesterol and blood)
- minerals (microcalcifcations,
iron, manganese – hepatic encephalopathy, copper)
- laminar necrosis (from
global hypoxemia or immunosuppression – appears 2 weeks after insult) – may
appear as cortical ribbon
- slowly flowing fluid
|
- CSF
- Edema
- flow voids
- calcium (bone) –
although microcalcifications often appear as bright on T2 due to their
interaction of water molecules
|
T2
|
- Edema (from tumor, infection,
inflammation, ischemia, vasculitis, radiation-induced, chemo-induced, migraines, etc)
- late subacute blood (extracellular metHb)
- CSF (virchow robin spaces)
- Demyelination
- Myelinolysis - Degeneration |
- contrast
- acute blood (deoxy Hb)/early subactue blood (intracellular
metHb), chronic blood
(hemosiderin)
- melanin
- mucous/protein (i.e.
colloid/rathke cyst
- hypercellular tumors
(high nucleus to cytoplasm ratio – medullo, lymphoma, highest grade parts of
high grade gliomas) ** - tends to be dark but not black
- minerality (iron,
copper, calcium)
- flow voids/turbulent flow
– of blood and CSF (jets can appear
dark)
- air
- fibrous
tissue/bone
|
Diffusion
|
- T2 shine through
- acute ischemia (<2-3 weeks)
- highly cellular tumors (lymphoma, medullo/PNET,
meningioma, chordoma, germinoma, hemangiopericytoma, pinealblastoma, cortical
part of high-grade gliomas – cyst part typically does not restrict in
gliomas)
- abscess (cystic part restricts 2/2 pus)
- mucinous metastasis like breast or colon adenocarcinoma (mucus restricts)
- epidermoid
- prion disease (CJD, kuru – cortical ribboning)
- toxic (carbon monoxide, methanol, Wernicke, maple syrup
urine, gluteric aciduria, methyl malonic aciduria other inborn errors of
metabolism diseases)
- adrenoleukodystrophy
|
- old stroke (> 3 weeks)
- necrosis (i.e. core of GBM/mets)
|
GRE/SWI
|
Blooming
- air
- blood of any age
- mineralization
- inflammation
|
Source papers:
Monday, June 6, 2016
Sympathetic storming
Epidemiology
- Typically occurs in young patients with significant/diffuse brain injury - TBI/DAI, SAH, big IPH, etc.
- Archetypally a young male with bad DAI -- likely no true gender predilection but rather trauma tends to affect male > female. And perhaps age predilection because young people have a more robust sympathetic response, or maybe because the degree of neurological injury that is typically associated tends to be mortal in older adults, or maybe because high grade SAH or IPH or diffuse injury occurs more in middle age than late age
Pathophys
- Poorly understood
- Originally believed to be exclusively a function of deep white matter injury; however its also seen in bilateral/diffuse cortical injury
- Perhaps decrease in the dampening signals? exaggerated sympathetic response to all stim, instead of only to severe/noxious stim.
Clinical Presentation
- Paroxysmal bouts of tachycardia, hypertension, diaphoresis, fever, mydriasis
- Characteristically waxing/waning, rather than constant (i.e. alcohol withdrawal)
- Typically occurs 3-5 days after the initial injury, and resolves on the scale of days to weeks but can start as early as immediately after the injury and last for years
Treatment
- Very severe (i.e. uncontrollable blood pressures leading to problematic sequelae) - precedex gtt and/or esmolol gtt
- Less severe/transitioning off gtts/on the floor - clonidine, propanolol, gabapentin (especially useful for controlling storming that directly follows stim - like turning/bathing/etc)
- Some people believe that opiates like morphine are an integral part of treating storming, some people don't.
- You can always snow people into the ground with propofol or drips of benzos or narcotics, but it's an inelegant solution and some people believe that they are suboptimal ways of treating storming.
Epidemiology
- Typically occurs in young patients with significant/diffuse brain injury - TBI/DAI, SAH, big IPH, etc.
- Archetypally a young male with bad DAI -- likely no true gender predilection but rather trauma tends to affect male > female. And perhaps age predilection because young people have a more robust sympathetic response, or maybe because the degree of neurological injury that is typically associated tends to be mortal in older adults, or maybe because high grade SAH or IPH or diffuse injury occurs more in middle age than late age
Pathophys
- Poorly understood
- Originally believed to be exclusively a function of deep white matter injury; however its also seen in bilateral/diffuse cortical injury
- Perhaps decrease in the dampening signals? exaggerated sympathetic response to all stim, instead of only to severe/noxious stim.
Clinical Presentation
- Paroxysmal bouts of tachycardia, hypertension, diaphoresis, fever, mydriasis
- Characteristically waxing/waning, rather than constant (i.e. alcohol withdrawal)
- Typically occurs 3-5 days after the initial injury, and resolves on the scale of days to weeks but can start as early as immediately after the injury and last for years
Treatment
- Very severe (i.e. uncontrollable blood pressures leading to problematic sequelae) - precedex gtt and/or esmolol gtt
- Less severe/transitioning off gtts/on the floor - clonidine, propanolol, gabapentin (especially useful for controlling storming that directly follows stim - like turning/bathing/etc)
- Some people believe that opiates like morphine are an integral part of treating storming, some people don't.
- You can always snow people into the ground with propofol or drips of benzos or narcotics, but it's an inelegant solution and some people believe that they are suboptimal ways of treating storming.
Tuesday, May 31, 2016
Aging of stroke
|
Time
|
DWI
|
ADC
|
T2
|
CT
|
|
30 mins
|
First becomes visible -
bright
|
First starts to become
visible – dark – in animal stroke models ADC changes visible in < 5 mins
|
Invisible
|
invisible
|
|
6-8 hours
|
Clearly Bright
|
Clearly Dark
|
First starts to become
visible – bright (edema)
|
First start to become
visible - loss of grey-white, insular ribbon, etc
|
|
24 hours
|
Clearly Bright
|
Clearly Dark
|
Clearly bright
|
Clearly dark
|
|
1-4 days
|
Clearly Bright
|
Max darkness
|
Clearly bright
|
Clearly dark
|
|
7 days
|
Max brightness
|
Clearly dark
|
Clearly bright
|
Clearly dark
|
|
10-15 days
|
Signal starts to fade
|
Reverses from dark to
bright, sometimes becomes invisible
|
Clearly bright
|
Clearly dark
|
|
2-3 weeks
|
Signal fading/reversing
|
Becomes bright
|
Clearly bright
|
Clearly dark
|
|
>30 days
|
Dark
|
bright
|
Max brightness
|
Clearly dark
|
Subscribe to:
Posts (Atom)