Craniocervical Instability (CCI) Diagnosis: Supine MRI vs Upright MRI

[Hip]

Dr Vicenç Gilete and Dr Fraser Henderson both prefer patients to send in upright MRI or CT scans, which include the head in a full range of positions: neutral (head level), flexion (head looking down), extension (looking up) and rotational (looking left and right).

By contrast, Dr Paolo Bolognese prefers patients to send in standard supine MRI or CT scans, which normally only offer views of the head in the neutral position. Only occasionally does Dr Bolognese also request upright MRI scans with the full range of positions, as an additional test (he says this in his video).

Because of this, depending which of these neurosurgeons you send your scans to, you may receive a positive craniocervical instability (CCI) diagnosis from one, but a negative diagnosis from another. If you are considering fusion surgery for CCI, which is no small consideration, this situation is a little confusing. So I wanted to look into this issue in this thread.



One of the differences between neutral head position supine scans, versus upright scans with the full range of head positions, is that you cannot measure the translational BAI on the supine.

This is because the translational BAI is calculated as the change in BAI value, as the head moves between flexion, neutral and extension positions in the MRI or CT scan.

In CCI diagnosis, there are a number of important measurements of skull and spinal bone positions, and the translational BAI is an important measurement in cases of known ligamentous laxity (people with weak or lax ligaments).

In the CSF Craniocervical Instability Colloquium, San Francisco 2013, a consensus statement was arrived, in which the experts agreed that the primary measurements used for diagnosing CCI should be:

• Clivo-axial angle (CXA)
• Grabb-Oakes measurement
• Basion-axial interval (BAI)
• Translational BAI — used if there is ligamentous instability

The consensus statement says the CXA is ideally measured with the patient in a moderate flexion head position; but if this is not available, then neutral position will suffice in most circumstances.

The table below shows the normal, borderline and pathological ranges for these measurements a few other relevant measurements:

Reference Ranges for CCI and Chiari Measurements

The translational BAI and translational BDI are the maximum change in BAI and BDI as the head is moved from flexion to extension positions. The dynamic BDI is the change in BDI value as an upward force (of up to 35 lbs) is exerted on the head via invasive cervical traction.

More info in the MEpedia CCI article.

Spoiler: Reference Sources

References for Data in Above Table

CXA: Bolognese 2015 video at 14:28. Also Henderson 2016, Henderson 2018, Henderson 2018 video at 8:17.
Grabb-Oakes: Bolognese 2018 video at 49:30, 53:47 and 54:46. Also Henderson 2019.
BAI and BDI: Henderson 2018 video at 10:34. Also Harris study. Also Rojas et al.
BDI: Henderson 2016.
Translational BAI: Henderson 2019.
Translational BDI:
Dynamic BDI: Bolognese 2018 video at 58:31. Pathological range for dynamic BDI communicated by Dr Bolognese to an ME/CFS patient.
Dens Over Chamberlain: here. Also Gilete 2017 video at 1:16 (see written sheet of paper).
Cerebellar Tonsil Ectopia: here. Also here.
See also the consensus statement on CCI.

Results from my CCI survey indicate that a pathological translational BAI is common in ME/CFS patients with CCI, with 46% having this pathology. Next most common is a pathological CXA, found in 38%. Then a pathological Grabb-Oakes is found in 16%, and a pathological BAI in 9%.

16% of CCI positive ME/CFS patients had their odontoid peg 3 mm or more above Chamberlain's line. 21% had a retroflexed odontoid. 12% were diagnosed with Chiari.

So it seems that a pathological translational BAI is a common issue in ME/CFS. But if Dr Bolognese is not normally taking the translational BAI into consideration, he may diagnose many ME/CFS patients as negative for CCI, whereas presumably Dr Gilete and Dr Henderson may diagnose these patients positive for CCI.



As to who is right and who is wrong, well there may not be a clear distinction.

Dr Bolognese in this 2018 video at 31:19 says that there are over 20 different measurements that apply to the craniocervical junction, and originally nobody was entirely clear on which ones to employ for CCI diagnosis. Then 10 years ago, it was whittled down to 14 measurements, and then more recently reduced to just 2, according to Dr Bolognese: the CXA and Grabb-Oakes.

Though later in the same video at 54:20 Dr Bolognese also talks about the importance of the dynamic BDI for detecting CCI with vertical instability (cranial settling). He says whereas the CXA and Grabb-Oakes are good for detecting horizontal instability, they are not very sensitive for vertical instability, and if the patient's major problem is vertical instability, you need the dynamic BDI to detect that.

Dr Bolognese does dynamic BDI testing in his office: applying invasive neck traction to the patient (using up to 35 lbs upward force on the head), and employing a fluoroscopy Xray machine to measure the change in BDI when this force is applied.

If the patient's symptoms greatly improve under traction, it helps establish a cranial settling CCI diagnosis. See this 2018 video at 58:31. Dr Bolognese will not offer corrective surgery for CCI unless a patient's symptoms demonstrably improve under invasive traction.

Dr Bolognese in his patient instructions also asks patients to report the effects of manual neck traction performed by a physical therapist (PT), or over-the-door traction, which are a basic test for vertical instability CCI. Again if symptoms greatly improve under traction, it helps establish a cranial settling CCI diagnosis. Note though that some patients may not experience any improvements in symptoms under manual neck traction, but do experience improvements under invasive traction. So invasive traction is the gold-standard traction test for CCI.

In the 2018 video at 1:06:45 Dr Bolognese say that CT scans can be used to detect rotational instability of the craniocervical joint.

So the main CCI measurements used by Dr Bolognese appear to be the CXA, Grabb-Oakes and dynamic BDI under invasive neck traction.

Dr Bolognese in this 2018 video at 37:02 explains why he prefers supine MRIs. One of the reasons is that upright MRIs are all 1 tesla, and have a lower image resolution than supine 3 tesla MRIs. He says you also get motion artifacts in uprights, because the head is not positioned in an immobilizing frame, and so moves, blurring the image. Furthermore he says flexion and extension are not standardized in terms of head position.



Dr Henderson in this 2018 video at 7:47 says that the consensus meeting in 2013 in San Fransisco, it was concluded that CCI was best detected by the CXA, Grabb-Oakes and the Harris measurement (the Harris usually refers to the BAI, but also to the BDI). In addition, at 11:05 Henderson says it is important to examine the translational Harris (ie, translational BAI).

So the main CCI measurements used by Dr Henderson appear to be the CXA, Grabb-Oakes, BAI, and translational BAI.

And Henderson says that upright MRIs with the full range of head positions are needed.

In this 2012 video at 15:47, Dr Henderson says that it is often important to use an upright MRI, because the weight of the head bears down on the spine differently in an upright, compared to a supine MRI where the patient is lying down. He gives the case of a patient whose neutral head position CXA was 141º (non-pathological) in a supine MRI, yet their neutral CXA was 133º (pathological) in an upright MRI. So without an upright MRI, this pathological neutral CXA would not have been detected.

In this 2018 video at 26:37, Dr Henderson talks about a patient whose cervical spine looked perfectly normal on a supine MRI, but with her head in the extension position it was clear she then manifested cervical spinal stenosis. So in this case, without looking at MRI scans in the extension head position, this spinal stenosis pathology would have been missed.



Dr Gilete seems to use the same measurements as Dr Henderson: in all his reports which forum members here have posted on this thread, you can see that Gilete usually measures the CXA, Grabb-Oakes, BAI, and translational BAI, and that he measures all of these in neutral, flexion and extension head positions.

And if you look at this 2017 video at 1:16 which a CCI patient recorded her appointment with Dr Gilete, you can actually read the patient's full report on his desk (see image below).

Dr Gilete's Report for a Patient With CCI

UPDATE: as of September 2019, Dr Gilete has also started offering the same invasive traction test as used by Dr Bolognese.



IN SUMMARY:

The advantages of upright MRI scans are:

• The weight of the head bears down on the spine as it does in your normal waking hours. This weight can change the values of the measurements, but you would not detect these values on a supine MRI.
• You can have flexion and extension head positions, which allows the neurosurgeon to determine the translational BAI and translational BDI measurements.
• Furthermore, with flexion and extension positions, you can check if conditions such as cervical spinal stenosis, which may not have appeared in the neutral head position, might manifest in flexion or extension.
• You can have rotational head positions, which allows you to check for atlantoaxial instability.

The advantages of supine MRI scans are:

• Supine MRIs are much cheaper, and much more commonplace, whereas places offering upright MRIs are rare.
• Upright MRIs are only 1 tesla, but with supine you can have 3 tesla, which provides a higher resolution image, so you get better measurements.
• With the head held still in a frame, you do not get motion blurring on a supine MRI, but you may on an upright MRI.

 

[Yuno]

Great summary @Hip.
I came across this image comparing upright and supine mri and the resulting change of measurements. Don’t remember exactly where I found it though.

 

[JenB]

Thanks @Hip. This is great!

I thought pathological for translational BAI was > 2mm. Can you improve this page and site sources? https://www.me-pedia.org/wiki/Craniocervical_instability

 

[Yuno]

Thanks @Hip. This is great!

I thought pathological for translational BAI was > 2mm. Can you improve this page and site sources?

In these links pathological is also deemed to >2 though, if iam not mistaken.
https://link.springer.com/article/10.1007/s10143-018-01070-4


https://www.ehlers-danlos.com/pdf/2...8Baltimore-Diagnosis-of-headache-in-EDS-S.pdf

 

[leela]

Great summary @Hip.
I came across this image comparing upright and supine mri and the resulting change of measurements. Don’t remember exactly where I found it though.

Geepers, this looks like yet another reason why we feel better lying down.

 

[Hip]

I thought pathological for translational BAI was > 2mm. Can you improve this page and site sources? https://www.me-pedia.org/wiki/Craniocervical_instability

In these links pathological is also deemed to >2 though, if iam not mistaken.
https://link.springer.com/article/10.1007/s10143-018-01070-4

Yes > 2 mm is considered pathological for the translational BAI, and < 1 mm is the range for normal, according to this 2019 paper by Henderson.

But that paper also says that the criteria for corrective surgery is when the translational BAI is greater than 4 mm (see the section entitled "inclusion criteria for occipital-cervical fusion stabilization surgery").

So that's why in my table I put 4 mm as the threshold for pathological, and placed the range 1 to 4 as borderline.

Does that sound right, or do you think I should use 2 mm as the threshold for translational BAI pathology?

I believe when these surgeons give the threshold value for a pathological measurement, it means that if you are beyond the threshold, that indicates you are a possible candidate for surgery, or at least a candidate for conservative non-surgical treatments.



I will have a go at updating the MEpedia CCI page with these figures.

 

[Hip]

I came across this image comparing upright and supine mri and the resulting change of measurements. Don’t remember exactly where I found it though.

That image can be found in the Henderson 2012 video at 15:54. As you say, it shows how the clivo-axial angle changes considerably, and becomes pathological, when the patient is upright, compared to lying down.

 

[JenB]

The 2013 consensus document say 2mm, which is why I have that on MEpedia. Based on reports I’ve seen, that also seems to be the cutoff Gilete is using.

It’s possible the 2019 paper represents a more recent consensus or that it’s just Henderon’s Ideas. The table on MEpedia encapsulates divergences in opinion so it could say something like “> 2mm (some surgeons set the cutoff at 4mm)” or similar.

 

[Alvin2]

@Hip
Thanks for deep diving into this, i no longer have the functioning to do so but this makes a lot of sense

 

[Malea]

Great summary @Hip.
I came across this image comparing upright and supine mri and the resulting change of measurements. Don’t remember exactly where I found it though.View attachment 33428

A thought:
Something came to my eye regarding this image for the first time.
From reading on the fb groups in the last weeks I learned that a lot of us have pannus formations. (It’s the white thing above the odontoid) The pannus formation can compress the brainstem, which I think I can see in the right image. When it compresses the brainstem it lowers the CXA. So what I was told from many on fb, that the CXA measured in supine images should involve the panus formation. If this would have been done in the left image, I assume the CXA maybe would already have been pathological even in supine.

My example:
I asked a very knowledgeable patient to measure my CXA in my supine images. I have pannus formation. Without involving the pannus formation the CXA was around 144. when the pannus was involved it was around 135.


I have no idea how the different neurosurgeons value pannus formations and if they involve them in the CXA. Also, not all pannus formations seem to compress the brainstem.

 

[Hip]

It’s possible the 2019 paper represents a more recent consensus or that it’s just Henderon’s Ideas. The table on MEpedia encapsulates divergences in opinion so it could say something like “> 2mm (some surgeons set the cutoff at 4mm)” or similar.

That's sounds like a good idea.

In the 2019 paper, Henderson says that the normal translational BAI is always < 1 mm, but in their criteria for prescribing fusion surgery, "the authors allowed 3 mm for error", which then gets you to the > 4 mm threshold for pathological and recommended for surgery.

So it sounds like the study authors are adding this error margin to make sure that surgery is justified.

 

[Hip]

From reading on the fb groups in the last weeks I learned that a lot of us have pannus formations. (It’s the white thing above the odontoid) The pannus formation can compress the brainstem, which I think I can see in the right image. When it compresses the brainstem it lowers the CXA.

Dr Bolognese in this 2015 video at 11:33 says that the Grabb-Oakes line represents the sum total the tilt of the dens bone (aka odontoid) towards the brainstem, plus the thickness of the adjoining ligament (some diseases like rheumatoid arthritis can cause ligament swelling and thickening). If the tilt is significant and the ligament swollen, it impinges on the brainstem.

 

[dreampop]

Really nice and clear summary. Do they discuss these measurements in the healthy population?

 

[Hip]

Really nice and clear summary. Do they discuss these measurements in the healthy population?

The normal ranges in my table in the first post (the figures in green) indicate the measurements found in the healthy population. So if your measurements are within this normal range, then you have nothing to be worried about.

For example, when it comes to the BAI and the BDI, Harris performed a study on 400 normal subjects, and found that 80% have a BAI in the range of 0-6 mm, and 18% in the range 7-12 mm. Therefore in 98% of normal subjects the BAI was less than 12 mm. That's why ≥ 12 mm was chosen by Harris as the threshold for pathological.



Sometimes though the boundary between normal and pathological is less clearcut. For example, Chiari I malformation is diagnosed if the cerebellar tonsils protrude more than 3 to 5 mm below the opening at the bottom of the skull (that protrusion is called an ectopia).

But there are a lot of patients with an ectopia or more than 5 mm who show no symptoms. One study found that although all patients with a cerebellar tonsil ectopia of greater than 12 mm were symptomatic for Chiari, 30% of those with an ectopia between 5 and 10 mm had no symptoms.

So you could have an ectopia of say 8 mm on the MRI, which is beyond the normal 3 to 5 mm threshold for diagnosing Chiari, but still have no symptoms of Chiari. So with Chiari, the threshold for pathological is not clearcut.


To a lesser degree this is also the case with Grabb-Oakes: some neurosurgeons use ≥ 9 mm as the threshold, others think ≥ 8.5 mm is more appropriate.

 

[Hip]

Neck Traction Versus Cervical Collar

Neck traction and a wearing a cervical collar (Philadelphia collar) are two things your CCI neurosurgeon may suggest you try, to see if your CCI symptoms improve.

However, it's worth pointing out that these two approaches are intended for different forms of CCI:

• Neck traction can improve vertical instability CCI (cranial settling / basilar invagination)
• Cervical collar can improve horizontal instability CCI (where the skull slides horizontally over the spine as you tilt your head up and down, rather than pivoting on the spine as it should)

Neck traction can mean manual traction (where a neurosurgeon pulls the patient's head upwards with his hands), invasive traction (where a hoist is attached to the skull by screws), or home over-the-door neck traction devices, which are sold online. In this 2018 video at 1:03:25 Dr Bolognese talks about traction tests and the improvements in symptoms that result (improvements which help confirm the diagnosis of vertical instability CCI).

But CCI patients with horizontal instability do not respond to neck traction. Horizontal instability patients however may feel better with a cervical collar, which immobilizes the neck to prevent forwards and backwards horizontal movements of the skull across the top of the spine.

Bolognese says in this 2018 video at 1:00:55 that a cervical collar only stabilizes the head, it does not lift the head. So if a patient has horizontal instability, a cervical collar can make them feel better, but if they have vertical instability and cranial settling, then the collar will not help, because it does not raise the head like neck traction does.


More info in this post.



EDIT: contrary to what I said above, @jeff_w says that neck traction can improve the symptoms of all types of CCI: vertical, horizontal and rotational CCI.

 

[Gingergrrl]

@Hip Your last post was really interesting and I was not aware of that difference. What would it mean for someone like me? I have chronic, severe neck (and right arm) pain from a serious car accident and another injury, and I feel significant, but temporary, improvement from manual traction. But my supine MRI (looked at by Dr. B) showed that I do not have CCI/AAI.

Does it just mean that other types of neck injuries also benefit from this type of manual traction? I have not tried a neck brace except back in 2006 right after the car accident (when I was hit at high speed and my car flipped upside down) and I had whiplash. I've debated if some type of neck brace could help me now but from reading your post above, I suspect that a neck brace would be less helpful than traction?

I did three mos of PT at the end of 2018 (where I did manual traction once a week) and also tried a traction machine (one where you lie flat, not the over the door kind which scares me ). I found the manual traction more helpful than the machine but I've thought about buying a similar machine to try at home since I'm no longer doing PT.

 

[Yuno]

@Hip
I’m wondering though, a pathological CXA means horizontal instability and deformity of the brainstem, so the cxa / deformity would be totally left unchanged by lifting the head?
So if Dr. B. sees a positive response to traction as a prerequisite to surgery, does it mean in his point of view surgery is only recommended in the cases of vertical instability, cranial settling and basilar invagination?
Because from my impression Dr. G seems to diagnose and recommend surgery a lot based on the translational BAI, which would also mean horizontal instability.

 

[Hip]

I feel significant, but temporary, improvement from manual traction. But my supine MRI (looked at by Dr. B) showed that I do not have CCI/AAI.

Does it just mean that other types of neck injuries also benefit from this type of manual traction?

Not really sure. This whole complex subject of CCI measurement and diagnosis is hard to grapple with. One thing that confuses me — and your question touches on this — is how Dr Bolognese diagnoses vertical instability CCI just from a supine MRI alone.

Looking at his YouTube videos, Dr Bolognese makes use of three main measurements: the CXA, Grabb-Oakes (both of which he says are more sensitive to detecting horizontal instability than vertical), and then the dynamic BDI (which is sensitive for vertical instability). The dynamic BDI is the change in BDI (measured in his office using an Xray fluoroscope) when a patient is placed under invasive neck traction (dynamic BDI is not to be confused with translation BDI; the latter is the change in BDI from flexion to extension head positions).

So as far as I understand it (which is not very far), the most important test Dr Bolognese employs for vertical instability, the dynamic BDI, is done if the patient comes to his office. But in his patient instructions, Dr B also asks patients to report the effects of manual neck traction performed by a PT, which is also a basic test for vertical instability CCI.


Then if we look at the measurements Dr Gilete uses — usually the CXA, Grabb-Oakes, BAI, translational BAI and the dens protrusion above Chamberlain's line — most of these are oriented to detecting horizontal instability, except for the dens protrusion which measures cranial settling, and the possible exception of the BAI (I am unclear about what the BAI actually measures in terms of pathologies).

I’m wondering though, a pathological CXA means horizontal instability and deformity of the brainstem, but the brainstem would be “relaxed” (getting back it’s healthy form) again, if the head was lifted by traction and a normal CXA was restored, wouldn’t it?
So wouldn’t traction also help In that case?

Although Dr Bolognese says that the CXA (and Grabb-Oakes) is more sensitive to horizontal instability, the CXA does have some sensitivity to vertical instability, and in fact Dr Henderson says in this 2012 video at 27:05 that CXA is a surrogate measurement of both basilar invagination (cranial settling) as well as brainstem deformity.

Cranial settling is where the tip of the dens (also called the odontoid peg or odontoid process) pokes into the skull, and it can then impinge on the brainstem.

If your pathological CXA is due to cranial settling, it will look a bit like the image at the right:

Source: here

So yes in this case, I imagine a CCI based on a pathological CXA due to basilar invagination (cranial settling) would improve under vertical neck traction.

But I am guessing that you may also get a pathological CXA for other reasons, such as the skull being positioned too far forward horizontally with respect to the top of the spine (not sure if this actually occurs though), and in this hypothetical case I don't think neck traction would help fix a pathological CXA.

 

[Sidny]

The images above are totally mind-blowing. I don’t see how if someone had a physical abnormality, as illustrated in the second picture, that they wouldn’t have significant debilitating symptoms.

Looks like basilar invagination (cranial settling) results in severe alteration of critical neurological structures.

Within just months of my primary infection and illness onset I could literally see (in the mirror) my skull position had changed, almost as if it was sinking into my neck and spine. I wonder what appropriate imaging would turn up in my case.

Thanks for your continued hard work on this subject @Hip

 

[Yuno]

Thank you @Hip . That makes sense.
What I still don’t really understand though is that Dr. B, taking a positive response to traction as a prerequisite, only seems to see surgery as a solution for vertical instability, not bothering much about horizontal issues.

Dr. G, in contrast, seems to diagnose and deem fit for surgery a lot on basis of the translational BAI, which I thought was about horizontal instability, so it would seem that he sees surgery as means to alleviate both kinds of instabilities, horizontal and vertical.
Or do I get it wrong?