Sunday, April 26, 2015

Intraparenchymal bleeds 

Type Distribution Hx Notes Management
Hypertensive Deep grey (putamen, thalamus), pons, cerebellum 

BG- 50%
Thal - 15%
BS/pons - 10-15%
Cb - 10%
Hypertension Pathophys: 
- vessel wall hypertrophy/ischemia/necrosis => vessel wall replaced with collagen 
- scar can fill lumen 
(lipohyalinosis) => ischemia
-charcot buchard aneurysmal weakening of scar  wall => bleed
- ABCD
- Control ICP (suspect high ICP if coma, edema/effacement, blood in vents, HTN/bradycard)
- INR < 1.5, plt > 100,000
- Control BP: goal CPP > 60, so MAP goal should be 60+ICP. If unknown ICP: MAP goal 90-100 (bc ICP>30-40 is unsurvivable anyways). If no ICP issues: goal SBP 120-160

Amyloid Lobar (esp parietal, occip), NOT deep grey. GRE shows microhemorrhages in lobar distribution.   Old (>75), not hypertensive. If young, look for other pathology  Amyloid deposits in small vessels - unknown why deep vessels spared.  Don’t anticoagulate - very high risk of rebleed 
DAI  Grey-white junction, corpus callosum, dorsal brainstem  High speed trauma Punctate bleeds Supportive 
Contusion Surface of convexity, anterior frontal, base of temporal lobes, coup-contrecoup (i.e occipital-frontal)   Trauma  - Pathophys: Brain scrapes against bumpy bone (temporal, frontal)
- Trauma bleeds: mixed density, boggy tissue, edema.  

Venous thrombosis Temporal (esp post temp), b/l BG/thalamus, parasagittal UC/IBD, OCPs, peripartum, HRT, kids with viral GI  - UC/IBD: hypercoag, + dehydration 2/2 diarrhea
- Kids: dehydration 2/2 viral sickness
- Anticoagulate, even though there is blood in the head: Heparin gtt bridge to coumadin 
- Look for clot (MRV/CTV) 
Tumor Anywhere Hx of cancer  - Mets (chorio, thyroid, RCC, breast, lung, melanoma) 
- GBM 
depends on tumor type 

Aneurysm 99% of the time, you will see SAH  aneurysmal SAH story  In isolated IPH, aneurysm is very unlikely, so no indication for immediate angio to look for one.  Aneurysms should be fixed immediately - AVMs and cavernomas should wait 
AVM/AVF/cavernoma “weird” location - non arterial or venous distribution: i.e. along tent (dural AVF), superficial lobar Age < 50, not hypertensive  Cavs cant be seen on angio; need MRI in 1-2 mos after blood clears  AVMs and Cavernomas should not be fixed immediately - should give brain time to metaphorically cool down. 

When to operate: 
- Relatively stronger indications: Cerebellar, rapidly deteriorating clinical status 

STICH I: 
- Lancet 2005;  data collected 1995-2003
- No difference in outcomes between early surgery and conservative treatment 
- Inclusion: spontaneous ICH > 2cm within 72 hrs, GCS > 5 
- Exclusion: secondary ICH (i.e. 2/2 aneurysm, AVM, tumor, trauma), bleed involving cerebellum or brainstem, severely disabled at baseline, couldn't undergo surgery within 24h of randomisation. 



Commentary/thoughts (from me) 
- the only group that seems to have any benefit are those with hematomas really close to the surface (<1cm), and even then, its just barely significant -- (OR 0.47-1.01) 
- 25% of the group randomized to conservative treatment ended up getting surgery-- like all neurosurgery trials, a lot of crossover to contend with. 
- people with GCS 5-8 who got surgery did much worse; in the context of 25% crossover, this finding may be confounded by a subset of people who are crashing and burning clinically and get emergent surgery as a heroic measure. In other words, this may be more indicative of poor slope of clinical function rather than a condemnation of surgery. Everyone who came in with GCS<8 in the trial ended up with an unfavorable outcome....
- the definition of a 'favorable' outcome was a function of prognosis - for older patients who came in with GCS<8, mRS 3 or less is considered favorable, which is fair but sad. 

Wednesday, April 22, 2015

Venous infarcts/hemorrhages 

General principles: 
- Venous thrombus leads to increased back pressure, which first manifests as increased vasogenic edema, and then as the process evolves to infarction, cytotoxic edema
-Arterial thrombus only leads to cytotoxic edema
- Venous infarcts are often midline, while arterial are typically unilateral
- Infarct/bleed/edema in temporal lobe or b/l thalamus: think venous (see below for image of distributions).
- Posterior temporal lobe pathology only => Labbe
- Arterial infarct/bleed will follow arterial distributions

Distributions of venous infarct/hemorrhage: 

{image source: radiologyassistant.nl}

Venous drainage patterns:
{image source: radiologyassistant.nl}

Reminder of venous anatomy:



Labbe: 

infarct 

bleed

cytotoxic & vasogenic edema 
{image source: radiologyassistant.nl}

Deep veins (i.e. galen/straight sinus, etc) :



L thalamus and R BG flair signal

bilateral BG flair signal 



{image source: radiologyassistant.nl}

Tuesday, April 21, 2015

MR Imaging (Basics)

General: 
- fat is always bright (except modified sequences like STIR) 
- bone is always dark
- marrow is somewhere in between

T1: 
- anatomy scans (resolution of normal anatomy good)
- things that are bright on T1: fat, subacute blood (=metHb), melanin, contrast. 
- fluid is dark on T1
- protip: T1 noncon is a good way to look for recent bleed (days to ~2-3 weeks) if you don't have GRE imaging.

T2
- Fluid is bright 
- FLAIR = T2 with CSF subtracted out.
- In cases of bleeds, intact RBC membrane acts as a local magnet, and will appear dark on T2.

Blood on MR: 
Time course Substrate T1 T2
Hyperacute Intracellular oxyhemoglobin Iso Iso-br
Acute (hrs-days) Intracellular deoxyhemoglobin Iso Dark
Early subacute (days-wks) Intraceullar methemoglobin Bright  Dark
Late subacute  Extracellular methemoglobin Bright Bright
Remote  Ferritin/hemosiderin Dark Dark

DWI 
- restriction in mobility of water molecules (brownian motion) 
- restricted diffusion shows up bright = cytotoxic edema 
- ischemia leads to failure of membrane ion pumps; cells swell with fluid; water that used to be in the extracellular space is now in the cell, which limits its motion 
- most common cause of cytotoxic edema is ischemia, but there are other pathologies: necrosis (abscess, tumor center, radiation damage), epidermoid tumor, acute demyelination, acute encephalitis, CJD, tumors with high nuclear:cytoplasmic ratio, elevated protein or subacute blood
- T2 shine through: DWI is a T2-derived study, lesions that are bright on T2 (ie tumors) will be bright on DWI as well; thus we use the the apparent diffusion coefficient (calculated from T2) to showcase intrinsically T2-bright things
- theoretically, DWI changes appear as early as 6 mins after ischemic insult and as late as 2 weeks after; imaging changes will be apparent on ADC (dark) before they are apparent on DWI
- ADC-dark lesions will eventually become bright over time as the cells complete necrosis

Gradient recalled echo/susceptibility weighted imaging 
- takes advantage of paramagnetic qualities of deoxyhemoglobin and hemosiderin to detect blood products
- most sensitive imaging modality for blood 

Contrast
- When to use: refine diagnosis, see small images. 
- Things with no BBB that enhance: Pituitatry, hypothalamus, pineal gland, choroid plexus. 

MR patterns: 
- Dural-based extraaxial mass: meningioma, dural met, granulomatous disease (sarcoid) 
- Periventricular ischemia: small vessel disease, watershed infarct 

Monday, April 20, 2015

All about CSF 

WBCs
- Normal WBC in adult is <5, in neonates < 20
- 99% of subjects with bacterial meningitis will have WBC > 100, 87% will have > 1000
- <100 = think viral meningitis 
- DDx for elevated WBC in CSF - seizure, ICH, tumor, inflammatory CNS dz 
- A traumatic tap will increase WBC in CSF at a ratio of 500-1000:1 (RBC to WBC). Correction factor can be applied to determine "true" WBC. Shortcut: if the RBC:WBC ratio in the CSF is <<500-1000:1, it's probably infected 
- Normal CSF: 70% lymphs, 30% monos 

Organisms (more relevant for standardized tests than for clinical practice) 
- Gram stain positive in 60-80% of untreated bacterial meningitis, 40-60% of partially treated
- Sensitivity of gram stain >90% for Staph and Strep, ~50% for listeria. 
- Sensitivity of india ink for cryptococcus ~50% 
- Sensitivity of first acid-fast stain ~33% for TB, increased to >85% if 4 smears are done 
- Dx toxo via giemsa or wright stain 

Protein 
- Newborn should be <150, adults~15-45 (varies by lab) 
- Can be elevated in bleeds, inflammatory disease (MS, GBS), tumors, infections/abscesses
- Can be falsely elevated in traumatic tap - to correct, subtract 1 mg/dL protein for every 1000 RBC
- Table of pathology: (from Fishman RA. Cerebrospinal fluid in diseases of the nervous system. 2d ed. Philadelphia: Saunders, 1992) 
 
 
Average: mg per dL (g /L)
Range: mg per dL (g / L)
Bacterial meningitis
418 (4.18)
21 to 2220 (0.21 to 22.2)
Brain tumor
115 (1.15)
15 to 1920 (0.15 to 19.2)
Brain abscess
69 (0.69)
16 to 288 (0.16 to 2.88)
Aseptic meningitis
77 (0.77)
11 to 400 (0.11 to 4.0)
Multiple sclerosis
43 (0.43)
13 to 133 (0.13 to 1.33)
Cerebral hemorrhage
270 (2.7)
19 to 2110 (0.19 to 21.1)
Epilepsy
31 (0.31)
7 to 200 (0.07 to 2.0)
Acute ETOH
32 (0.32)
13 to 88 (0.13 to 0.88)
Neurosyphilis
68 (0.68)
15 to 4200 (0.15 to 42.0)
Glucose
- CSF glucose should be ~ 2/3 of serum glucose, although the ratio decreases as the serum glucoses increases; regardless of serum glucose, CSF glucose should almost always be <300 
- DDx decreased CSF glucose: infection, bleed, inflammation, hypoglycemia 

Viral Meningitis 
- >85% enteroviruses (echo, coxsackie) 
- Oral fecal or respiratory
- Affect people in the warmer months-  summer or early fall 
- Children < 3 mos = generally coxsackie B 
- Coxsackie B in pregnant women - harmless to mom, can cause serious cardiac infection to fetus in perinatal period 

CSF Color 
- CSF protein >150 or RBC>100,000 (traumatic tap) or serum bili > 10-15 can cause xanthochromia 
- Yellow: blood breakdown products, hyperbilirubinemia 
- Orange: blood breakdown pdts, excessive carotenoid ingestion 
- Pink: blood breakdown 
- Green: hyperbili, pus 
- Brown: meningeal melanomatosis 
- Xanthochromia can be seen as early as 2-4 hrs and will linger for weeks; in a suspected SAH, most people advise waiting 12 hrs after the suspected bleed to LP to give sufficient time for xanthochromia to develop 
- Neonates will often have xanthochromia in CSF 2/2 elevated bilirubin and protein