Blog

A canine ultrasound protocol for GP vets

We do a lot of ultrasound training for vets in practice with all levels of experience and, in our role as peripatetic ultrasonographers, we get to see up close and personal how GP vets use ultrasound themselves.  Over the years I’ve gradually changed how I feel about how to approach teaching diagnostic ultrasound.  

In the beginning, I essentially used to talk people through the same systematic approach that I use.  With time I’ve realised that maybe I was underestimating the experience required to do some of this stuff effectively.  We work with over 200 practices on a day to day basis; and many of them have their own GP vets who do a significant amount of sonography -and many of them are really excellent vets in my humble estimation.  And yet, it’s really common that our assessment of a patient will be very different from theirs.  

In particular, guts, pancreas and abdominal vessels are very challenging generally.  And there are also many subtleties in the examination of other organs that can be missed or misinterpreted.  When I started doing full-time peripatetic sonography after 25 years of scanning as a GP vet I found that my learning curve was very steep indeed. I’ve resigned myself to accepting that it’s just not possible for anyone to acquire the experience and subtle technical skills to fully and accurately assess patients in a time scale of months or even a few years.

FAST scan techniques: which represent the purest form of point-of-care ultrasound (POCUS) for triage of urgent cases is obviously something that it’s highly desirable for GP vets to be able to do themselves.  Beyond this, there’s a strong case to say that, in situations where owners want optimal care, an examination by a ‘professional’ sonographer is frequently the best option for many patients. Obviously no-one wants to burn up clients’ resources unnecessarily but I’d argue that many pet owners are enormously attached to their cats and dogs: in any significant illness there’s a strong case for gathering as much information as can be gained non-invasively. 

However, I accept that clients can’t always afford a dedicated sonographer.  Furthermore, GP vets understandably and laudably want to be able to develop their skills. In my judgment there is undeniably a place between POCUS and dedicated sonographer examination to be filled by GP vets performing diagnostic ultrasound (DxUS).  This is an attempt to put together a protocol to maximise utility in this situation.

Some of the things I am factoring in:

• Apart from very limited examinations of musculoskeletal or eye structures, it’s important for each patient to have a global assessment of thorax and abdomen.  It’s just inadequate in my view to omit either.  If you’re going to go to the trouble of scanning a patient then you really have to try to assess heart, lungs, thorax and abdomen to get the whole picture.
• Keep it quick and relatively simple.  Each of these 8 views should take a few minutes at most: you should be done inside 30 minutes.
• Save video clips (ideally 10 seconds or so each time) for subsequent review and for sharing with colleagues either locally or remotely for a second opinion.  It’s the biggest danger with diagnostic ultrasound: that an interpretation based on a brief view by a single clinician is used as the basis for major decisions.  We should try hard to avoid misdiagnosis by sharing and reviewing images.  None of us know what we don’t know.
• It’s nice to be able to scan standing patients.  It’s less stressful for the animals, quicker and easier for the assistant holding.
• Remember: the more there is a big, obvious lesion, the exponentially greater the chances of you missing something important somewhere else!. The moment your eye lights on a ‘diagnosis’, challenge yourself doubly to assess the rest of the patient thoroughly.
• Although you can do broadly the same thing for cats, I’m going to make this post dog-specific.  The differences are limited but important enough to justify a separate article.

Protocol:

This is going to sound a bit long-winded because I’m going to discuss some of the issues involved as we go along -but it takes less time to do this protocol than to read about it.  I will add a video shortly.

Assuming the patient is stable, clip fur over the abdomen and lower part of both sides of the chest. Or you can use spirit if the coat isn’t heavy.

Start with the patient standing and perform lung ultrasound on both sides of the chest. It’s sensible to find out whether the patient is likely to have respiratory compromise before wrestling him/her into lateral recumbency.  I prefer a linear probe but micro-convex is also OK.  Align the plane of scanning  with the intercostal spaces and simply sweep the probe from one IC space to the next from caudal costal margin right into the axilla.  Save a video clip from each side. Note the presence/absence of glide sign, any pleural fluid, lung lesions, other mass lesions. If there is no effusion then you only need a depth setting of a few cm to view the pleura and any subpleural lesions. If there is an effusion then increase the depth to include deeper structures which may then be visible.

Get a right sided long axis 4 chamber view of the heart.  You can do this with the patient standing or in right lateral recumbency with the probe applied to the underside through a cut-out echo table.  Broadly-speaking find the heart close to the right axilla and rotate the probe to align the plane of scanning with the long axis of the heart until you can see 4 chambers.  Save a video clip covering several beats. It doesn’t have to be perfect alignment.  The aim is to check for pericardial effusion and to assess the overall proportions of atrial and ventricular chamber sizes, the shape of the atrial septum and the ventricular wall thickness on left and right. 

In dogs without cardiac issues:

• the atria should be roughly same size
• at end-diastole (freeze video and scroll to the last frame before the mitral leaflets shut) the left ventricle should be roughly three times the diameter of the right.
• at same time point the left ventricular free wall should be roughly twice the thickness of the right ventricular free wall.
• at the same time point the left ventricular lumen should be between 3.5-5.0 x left ventricular free wall thickness.

Callipers can be used to measure these things if in doubt. 

Make a subjective assessment of systolic function. Human studies show that, with a little practice, subjective assessment of systolic function (eyeballing it) can be as meaningful as technical measurements. On the whole, I believe that if a patient’s systolic function looks fine then it probably is fine.  If it looks poor then it probably is poor and if it looks borderline then no number of measurements will change that.  It’s a widely held misconception that ‘dilated cardiomyopathy’ is a condition diagnosed by echocardiography.  In reality, there is no specific set of figures which confirm ‘DCM’ in an individual patient -because DCM isn’t the only cause of myocardial compromise.  Far from it.  If there is poor systolic function then our job is to try to find a treatable primary cause (diet, tachycardia-induced cardiomyopathy, PDA, hypoadenocorticism, sepsis etc etc) rather than to measure more things on an echocardiogram. 

Rotate the probe 90 degrees clockwise (as seen from the sonographer’s position) to get a right short axis view of the heart with the silhouette of the heart as round as possible. Fan the probe to take the plane of scanning up to the level of the aorta and left atrium.  Now, this view is widely touted by all and sundry as ‘the view’ to assess left atrial size.  My personal experience, borne out pretty well by published data, is that it’s very difficult to achieve good repeatability even for experienced echocardiographers.  

https://onlinelibrary.wiley.com/doi/abs/10.1046/j.0140-7783.2003.00543.x

…and not surprisingly because the left atrium is a complicated 3D shape: small changes of angle can significantly affect measurements.  The widely-used cut off of 1.6:1 for LA:Ao ratio does not factor in a grey area to allow for intra-observer variability of at least 10%.  As with other measurements an obsession with generating numbers can greatly extend the length of time taken to perform the examination.  OK, if we’re trying to distinguish between dogs with mitral valve disease who might benefit from pimobendan and those who probably won’t then we have actual evidence (the EPIC study) based on actual specific measurements.  To take those measurements and make that distinction reliably, in my judgement, you probably need an experienced sonographer.  If that’s not achievable, to be brutally honest, I’m not sure that a GP vet is any worse just basing the decision on grade of murmur as opposed to echo findings.

If, on the other hand, you are dealing with a patient with clinical signs then, rather than measurements, I prefer GP vets to look at the shape of the atrial septum (in both short axis and long axis) looking for bowing to one side or another.  And to subjectively assess the shape of the left auricle for dilation.

Amongst other issues, a focus on teaching vets to rely on LA:Ao completely overlooks the possibility that there might actually be right atrial enlargement.  Atrial septal shape tells us something wider than LA:Ao. From a right long axis view the atrial septum may appear slanted in normal dogs depending on the angle of scanning.  Bowing is different.

These two views between them, assessed subjectively allow a broad overview of cardiac function: identifying changes in systolic function, volume overload or pressure overload of the right or left heart

I believe that, when assessing an individual patient, experienced veterinary cardiologists use measurements to varying degrees but they all rely fundamentally on their subjective judgement of the heart.  Apart from patients with intermittent arrhythmias, a normal-looking heart is unlikely to be responsible for clinical signs.  For example, all other things being equal in a dyspnoeic patient, left-sided volume overload (chamber dilation) indicates that left-sided congestive failure is the probable cause whereas right-sided changes imply primary lung pathology.

Unless you are going to use Doppler (colour, PW, CW) on a weekly basis then, at best, I’m not sure it’s going to be reliable enough to add anything and, at worst, it’ll be misleading and you’ll end up spending so much time on it that you’ll miss the bigger picture. Do less Doppler, take more temperatures and spend more time thinking is my advice!

For abdominal views I’m keen for people to concentrate on looking for free fluid, abnormally hyperechoic abdominal fat and to save video of the major organs that can be used to guide further work up -especially with advice from colleagues. Use a microconvex probe as a general rule.

Sub-xiphisternal view: apply the probe to the abdominal wall immediately caudal to the xiphisternum (in the ventral midline obviously) with the plane of scanning aligned in a longitudinal plane.  With conventional set-up, when the mark on the probe is cranial, the left of the screen will be cranial.  Angle forward until you can see the diaphragm and adjust the depth to accommodate its complete curve.  Fan the probe (keep the same contact point, swing the tail of the probe) slowly to left and right to view the right and left sides of the cranial abdomen.  Save a video clip of this.  You should capture falciform fat, liver, gallbladder, stomach and spleen. 

From the same contact point, rotate the probe 90 degrees with the mark now to the patient’s right side.  Now fan the probe to view more cranially and more caudally.  Again, save a video clip.

Left kidney area: apply the probe high on the left flank immediately behind the last rib.  You may need to hunt around a little to find the left kidney. Generally try to keep the probe aligned so that you are scanning in a longitudinal plane. Fan the probe dorsally and ventrally.  Save some video of all organs which you can see in this area (hopefully, spleen, kidney, aorta).

Right kidney area: repeat the process high on the right flank.  In deeper-chested dogs you will often need to move the probe onto intercostal spaces to see the kidney and the area of the pylorus and right pancreas. Save video.

Mid-abdomen: apply the probe to the umbilicus aligned in longitudinal plane.  Fan from left to right and save video.  Mainly to look at the appearance of gut and mesentery.

Bladder area: apply the probe to the side of the abdomen in the inguinal area in longitudinal plane.  Locate the fluid-filled bladder, adjust to include bladder neck and proximal urethra and then slide the contact point dorsally to view colon, uterus (if present) and major vessels.  Save video.

That’s it!

So, yes, you would be right in thinking that this abdominal part is largely a modification of the AFAST protocol with which many GP vets will be familiar.  Just add on a mid-abdominal view, do a little more looking round and save video of everything.

In dogs with acute abdominal signs the most important thing is to be conscious of noting any locally hyperechoic abdominal fat which may help localise the problem.

As with cardiac Doppler, I’d argue that unless you are scanning several times a week it’s going to be largely unrewarding to look at the finer point of adrenals, pancreas and portal vein.  The biggest problem people experience with canine adrenals is not the technicalities of scanning them, it’s that they don’t spend enough time looking at the patient’s skin changes.  There are no changes in the canine pancreas which are specific for pancreatitis: as a basis for further investigation and treatment it’s enough to know that abdominal fat in the area of the pancreas looks inflamed.

Obviously, this post can’t really be a complete guide to the ultrasound of the dog.   There’s enormous natural variation and subtlety which is way beyond the scope of what we’re talking about here.  The crux of the protocol for me is to get some video of all areas of the abdomen and thorax.  Perfection of technique and interpretation is never possible.  Comprehensive survey of the thorax and abdomen with a view to picking up major changes certainly is achievable.  You need to set out to acquire a specific set of images: deal with the finer points of interpretation later as much as possible.

If there are subtle changes in the area of the right pancreas then it may well be that no-one can interpret those with confidence.  Don’t beat yourself up if you’re not sure if your patient has pancreatitis and don’t spend 25 minutes prodding at it.  But if you fail to notice a 4cm patch of consolidation in the right lung then that’s clearly a suboptimal outcome. Hey, we’ve all been there!