Hello and welcome to this lecture on innovations in equine orthopaedic imaging, and we're gonna be asking what's new in this field. So it's safe to say that in the last 5 years, there have been huge innovations in equine orthopaedic diagnostic imaging and that this technology has increased almost exponentially. We have new equipment, new techniques, and with these come a much greater knowledge of conditions that we already knew of.
And also a realisation of the existence of some that we didn't actually fully appreciate. Thanks to these advances, we also have a much greater retrospective understanding of the results of standard imaging techniques that we can use to our advantage. In addition to the mach the machinery, human knowledge has increased massively, and there's been the arrival and influence of artificial intelligence.
In this lecture, we're going to discuss some of these factors as relates to each of the different imaging modalities, but it's not intended to be a comprehensive review of the image systems available or of the relevant literature. More, it's really a pick of the clinically applicable techniques and papers that might be relevant to the equine practitioner. So for the learning objectives of today's sessions, we aim for you to gain an appreciation of new technologies, to know which diagnostic imaging techniques the equine orthopaedic world is currently talking about, and also what the options are that you can give to your clients.
We want you to be able to understand the advantages and disadvantages of each technology, and we aim to highlight new knowledge that has emerged or has been consolidated from these tech technological improvements and give you a realistic idea of which clinical questions it's possible to answer with these advances. So let's start with computed tomography. There have been huge and exciting advances in technology and practical applications of CT in the equine industry in the last 5 years.
New units now exist with a variety of options. Instead of just the traditional helical multi-slice scanners used in equine work previously, cone beam systems are now also available. Comb beam and fan beam multi-slice imaging of the limbs of horses is also possible in standing sedated horses rather than just horses under general anaesthesia, which was of course the norm just five years ago.
In addition, dynamic imaging of the cervical spine has been trialled on cadaver cases with the hope of expanding this technology to living horses, as is already the case in some small animal patients. So thinking about the standing CT first. The development of the cone beam unit for standing CT by the Hallmark Company has allowed CT of the lower limbs to become affordable and a safer choice for many equine clinics.
These units are much cheaper than conventional conventional helical fan beam CT units, and they also negate the risk of general anaesthesia. The scan time is still quick, but the horse is not enclosed in a complete ring or bulky equipment at any point during the procedure, which is the case with some of the units. The cons of these units are the imaging resolution is not quite as good as in multi-slice helical imaging.
And it is a little bit more prone to artefacts. However, multiple studies have provided firm evidence for the usefulness of these standing cone beam CT units in the field of diagnostic imaging. In particular, in the situation of pre-purchase situations, or when fracture is suspected, and there's a need for cross-sectional imaging without GA, then these kind of units are imperative.
So this is one of the papers that that looks at the comparison of cone beam and fanbeam CT and also compares the images found with low field MR. In this particular study, 35 limbs from 10 horses underwent cone beam CT, fan beam CT MRI, and macroscopic examination. These are all cadaver limbs.
The CT and MR images were examined for the presence of palmar ostochondrial disease, and imaging details of paler ostochondral disease and measurements of the dimensions of the disease and areas were measured. The imaging diagnosis, details and measurements were compared with macroscopic examination between modalities. In all, 48 pomerostochondral lesions were seen over 70 condyles.
And comparing macroscopic examination, the sensitivity and specificity of diagnoses were 95.8 and 63.6% for fanbeam CT 85.4 and 81.8 for cone beam CT, and 69% for MRI.
Intermodality agreement on diagnosis was moderate between the cone beam and the fanbeam CT. And there was agreement on imaging details between both cone beam and beam CT. Especially for subchondral bone irregularity.
Macroscopic porostochondrial disease was strongly correlated with MRI and moderately correlated with fan beam CT when the width was measured. The main limitations of this study reported was that the effect of motion artefact in live horses could not be assessed because this was a cadaver study. Nevertheless, the authors concluded that all imaging modalities, the fan beam, the cone beam, and the MR were able to detect the palma ostochondral lesions.
But they did concede that all of the modalities underestimated lesion size. The CT systems were more sensitive than the than the MR, but the differing patterns of signal intensity on the MR allowed additional changes associated with the Palmer endochondral Palma osteochondral disease that could not be identified on CT. Overall, the image features observed with cone beam and with fanbeam CT were reported to be similar.
Moving on to the advances in multi-slice helical fan beam, large bore CT units available. The image above shows the system that we use at the University of Liverpool Eine Hospital. And this is a multi-slice helical largeable scanner that we have had in place since 2018.
This unit allows for standing examinations of the head and upper cervical spine to see two in most horses. But does require general anaesthesia for the unicular and axial skeletons. The 90 centimetre ball is fantastic.
It it allows for a large portion of the body of most horses to be imaged. So the vast majority of horses, we can do pretty much all of the body apart from the mid part of the thoracic spine, which just isn't accessible whichever way you try and put the horse into the machine. Various similar units are also up and running worldwide, and all have their advantages and disadvantages.
It's safe to say though that all have contributed greatly to the advancement of knowledge in equine orthopaedic and diagnostic imaging. So this image shows our unit again with the pelvis of a large 162 hand horse being scanned and fitting comfortably into that large diameter, bore. The main difficulty really with this is in the initial positioning, which can be a little bit fiddly and also a little bit exciting if the horse is maintained on total intravenous anaesthesia, which is the case in this animal.
With this positioning, once the horse is in position, then we can. Very clearly assess the femoral joints, the pelvis, the sacred iliac joints, the lumbar spine, and we can image these in relatively high detail. In order to image the stifles or upper forelimbs, the horse, however, has to be repositioned into lateral recumbency.
So I just want to play a video at this point, and this shows the pelvic and lumbar spine scanning process speeded up, which does make it look incredibly easy and stress free. But I can assure you that that is not the case. I think we speeded this up quite a lot.
And you can see that the quality of images obtained is very good, even in large horses. The bottom left image is the dorsal plane view of both cox femoral joints, with contrast, contrast agent injected into the left cox femoral joint, which is the one displayed on the right hand side of the screen. The top right image is the 3D volumetric reconstruction of of this same study.
I think you'd agree, I mean, these are really beautiful images that we could only have dreamed about 5 years or so, ago. So other multi-slice helical fan beam CT systems are also available, and these now allow for the distal limbs from the tarpus, Tarsus, and carpus to be imaged with the horse standing and then weight bearing. Again, a huge leap forward in technology.
The system shown on the slide in front of us is the CliR system, and this is the one at the University of Zurich. This system incorporates the same large ball CT helical scanner that we saw in the earlier slides, but uses a bespoke platform and gantry to greatly increase the areas of the horse that could be imaged while the horse is standing and sedated. The same system can then of course be used to image the pelvis, the lumbar spines, the sacred iliac joints, the cocc femoral joints, and the upper limbs, just as in the previous system.
A further multi-slice fan beam system. It's available for use in standing sedated horses, and this system is unique in presenting a fan beam weight bearing scanning protocol, which is different to the one that we saw previously, which is non-weight bearing. So this gives the advantage of fan beam and standing CT, but it does result in a slightly cumbersome scanning area at the point of actual image acquisition.
Although the speed of the scan is extremely quick. So actually, the practicalities and the limitations of the Of the practicalities of the horse standing in an enclosed ring are actually really quite minimal still because of the speed of the scan. So what have we learned from all these fantastic units?
Well, we've learned an awful lot, and it's impossible to go through all of the evidence-based medicine and literature that has come out from the advances in CT in the last 5 years or so. But I have picked out a few studies that I think really illustrate where we're going with this technology. So this study from Ogden and colleagues, looked at the technique, image quality and anatomical variation in 56 clinical cases of courses where the cadal spine and pelvis were were imaged.
Studies were carried out in a large bore helical CT machine, some in our clinic, and some in the, the Netherlands, and some in a further clinic in England, and the horses were all under general anaesthesia. Horses were all older than 6 months, and images were only examined from horses that were alive at the time of imaging. So postmortem examinations were were excluded.
The weight of the horses ranged from 85 to 680 kilogrammes, so really large horses going into these machines. The general anaesthesia time was reported to range from 10 to 60 minutes, which again is a really good time to be imaging these big animals in. The authors reported that there were no adverse events either before, during, or after the procedures.
Images in all horses were deemed to be of diagnostic quality. And there was quite a large range of anatomical variations reported. And these included the location of the diverging into spinous space.
The presence of spina bifida in the lumbar and sacral spine. The shape of the last lumbar vertebrae and the location of the in transverse joints in terms of where they were present and the degree of fusion or modelling that was found. The main limitations were those of image noise and beam hardening artefact.
And in larger horses, the images were of insufficient quality to interpret the soft tissues adequately. The authors ended up by concluding that CT of the cordial spine and pelvis in live horses with wide bore CT machines and modified patient infrastructure was safe and did indeed produce diagnostic images, at least for bony stretches. And these are some of the images that they presented.
So this slide illustrates several of the CT images from this paper. A and B are dorsal plane images centred at the sacred iliac joint. And they demonstrate unilateral fusion in a.
And bilateral fusion in B of the L6 and sacrum joints. A transverse image C is centred over the lumbar articular process joints, and this demonstrates asymmetry between the left and right sides and new bone formation and enlargements of the left articular process joints. The image in D shows a transverse image on the left and the dorsal plane image on the right, centred over the sacrum, showing multiple sites of spina bifida, which is, essentially a bilobed appearance to that dorsal spinous process, which is incompletely fused.
In this image from the same authors, there is partial free fusion of the L5, L6 into transverse joints and of and L6 and sacral joints. There's also a symmetry of these joints, and the shape of the L6 transverse process is asymmetrical between the left and the right hand sides. This paper reported similar findings in the variation of the lumbosacral joints to those recorded in the cadaver anatomical study by Stubson colleagues back in 2006.
So we know from this, cadaver study back in 2006 that the maximum dorsoventral movement in the lum sacral area is usually between L6 and S1. And this is represented by a larger divergence of the two spinous processes of the vertebrae where this occurs. However, in some horses, the maximal divergence of the spinal processes, and hence the greatest flexion, was found to be between L5 and L6.
This greatly alters the movement in this area. In the cadaver study previously mentioned, lumbosacral variations were found in 33% of horses. The variations that occurred in this study included 5 lumbar lumbar vertebrae instead of 6, but with the normal 5 sacral vertebrae.
In these cases, maximum dorsoventral motion occurred between L5 and S1. This variation was found in 8% of thoroughbreds and 16% of thoroughbred crosses, but in none of the standardbred horses examined. The conventional L6S5 formula, but with a L5 L6 maximal divergence was found to occur in 32% of thoroughbreds and 29% of thoroughbred crosses.
It's a huge variation in the anatomy of this region. Going back to the CT paper from Ogden and colleagues, they looked at the significance of these abnormalities and found that there was no statistically significant association between the location of the divergent in spinal space and an imaging diagnosis of sacroiliac joint osteoarthritis or ventral spondylosis. There was, however, statistically significant association between the location for the greatest divergence of the interspinous space and horses being diagnosed with intervertebral disc disease.
With a higher than expected number of horses, having a diagnosis of the intervertebral disc disease of the caudal lumbar and lumbar sacral spine when the divergence was located at both L6S1 and L5 L6 and when the divergence was located at L5 and between L5 and L6. So basically a double divergence present instead of a single L6 S1. And also when it was located, one space cranially at L5 and L6.
Three of the same authors then used the same cohort of horses to report on pathological findings. And these horses were presented for, predominantly for lameness, mainly to myself and my colleagues and poor performance. OI pathology was identified in 41 horses out of 56, and these included osteoarthritis of the sacred iliac joints, pathology of the intervertebral disc joints, pelvic fractures, osteoarthritis of the coopfemoral joints, ventral spondylosis of the vertebral bodies, and acetabular rib fracture in two horses.
Dislocation of the cups femoral joint and even a dorsal dermal sinus of the sacrum in one horse, which was quite quite bizarre finding. So this slide illustrates some of the findings in. The SI joint, and we've got left to the right hand side of the screen, and in A a transverse image, and B, a dorsal plane image centred on the sacred iliac joint.
In the same horse And we can see that we've got very markedly irregular sacred iliac joints outlined by the black arrowheads. In this sagittal image of the lumbar spine, although the image is somewhat grainy and has a lot of noise, due to the large mass of tissue over this area, a large ventral spondylosis of the lumbar vertebral bodies can be readily appreciated. The significance of all of these findings in ridden horses is currently unknown.
In dogs and humans, these are largely dismissed as incidental findings and to be expected in older animals. However, in horses that are still in athletic use and have to carry a rider, these are much more likely to be of greater relevance to clinical problems, and much further work is needed on this subject, which we can now excitingly do with all of the advances in image acquisition. We've also found during our clinical investigations that fractures to the bony prominences of the pelvis may not be quite so straightforward, a diagnosis or treatment as as we have previously perceived them to be.
This is a CT image and syntographic scan from a middle-aged girl doing with chronic Fy name lameness. We originally diagnosed a mildly displaced tubercoy fracture with graphic. Examination.
However, we failed to appreciate that the fracture fragment had also been rotated, as well as displaced by the pull of the strong muscular attachments onto the tubercoccy. And this horse represented months after the original diagnosis with ongoing hi lameness. A CT scan then identified the true extent of the injury, and we can see that the ventral portion has been rotated backwards as well as displaced ventrally.
Using CT conditions of the coccofemoral joints were also readily identified. These are city images of three horses demonstrating pathology in the cox femoral region. Left is to the right in transverse and dorsal plane images.
And transverse a and dorsal plane B images centred over the femoral head in the same pony demonstrate a cranially displaced and dislocated femoral head as shown by the black arrowheads. Note there is new bone formation along the left ilium at the dislocated femoral head. And the transverse image say and dorsal plane images D.
Are centred over the left coccy femoral joint in the same horse also demonstrating an a acetabular rim fracture which has contributed to the dislocation of the femoral head. The transverse image E and the dorsal plane image F centred over the cot femoral joint in the same pony demonstrate the marked secondary osteoarthritis that has ensued from the two. Do other injuries.
Moving on to the limbs, contrast CT orthography has become increasingly utilised. This is certainly not a particularly new concept. However, the use has dramatically increased in the last 5 years, due mainly to the increased availability of the GA units and also the standard units that we talked about at the start of the lecture.
With increased use of these units, of course comes better understanding of the range of pathologies that exist in lame and non lame horses. And this paper from 2011 really beautifully illustrated the delineation that you can achieve it intraticular ligaments and meniscus the stifle joint using contrast arthrography. These are some of the images that we have obtained in our own unit using contrast orthography.
And the left image illustrates the cranial cruciate ligament and the right image delineates the caudal cruciate ligament. This, recent paper published in vet vet surgery describes 4 cases with called cruciate ligament vulsion fragments, and these were diagnosed on the CT examination. They were then confirmed and removed with arthroscopic treatment.
So the authors concluded that CTE and arthroscopic evaluation of the causal medial femo tibial joint, was valid. And the the CT examination confirmed the diagnosis and allowed evaluation of the stifle joint for comorbid comorbid comorbidities. A cranial intercondular arthroscopic approach facilitated the removal of these cor cord or cruciate inser insertional lavulsion fragments, although they did acknowledge that the removal of the fragments was not always complete.
Moving down to the fetlock region. This recent paper in animals. Describes magnetic resonance imaging, computer tomography, and radiographic findings in the metacarpophalangeal joint of 40 non-lame thoroughbred gearings.
The aim of this particular study was to describe MRI. CT and radiographic findings in the metacarpophalangeal joint of non-named cerebralings. 40 thoroughbreds under a low field MR in a standing hallmark unit, fanbeam CT in a standing Calibra unit, and radiographic examinations of both metacarpophalangeal joints.
Images were then assessed assessed subjectively. And a hypo attenuating lesion of the sagittal ridge of the third metacarpal bone was identified in 33 of 80 limbs in CT reconstructions. Cone shaped mineralizations in the sagittal ridge were detected in MR images and in CT images, and mild hyperattenuation was also common in the trabecular bone, in the dorsoed and polateral metacarpal condyles in CT reconstructions.
A focal lesion in the subchondral bone was seen in the proximal phalanx and in the third metacarpal bone in some horses. A large vascular channels were detected in the metacarpal condyles in. 5657 out of 80 limbs and in the proximals bones in all limbs.
The authors therefore concluded that signs of bone remodelling can be seen in yearling thoroughbred pet locks, that sagittal ridge lesions were common. And that they are likely associated with osteochondrosis or other developmental osteochondral defects. Focal lesions in the subchondral bone of the third metacarpal bone and proximal phalanx.
Could also indicate developmental abnormalities or subtle subchondral bone injuries. This study really nicely illustrated the usefulness of that standing fanbeam CT unit, that we looked at, in the previous slides in live horses. Considering briefly, the cervical spine in orthopaedic imaging, then CT has allowed us to assess the cervical spine in a way that has just not been possible before.
Our own caseload has a high, high number of plain CT and CT myelographic examinations that just would not have been possible 10 years ago. We knew from pathology papers that there was a huge range of disease possible in spi or spine that may be contributing to ataxia and lameness and pain and poor performance. But it's only relatively recently that we We've been able to now confidently diagnose these conditions in our live patients, and therefore have the opportunity to try to help them and to do something about it.
The images here show us moderate mineralization in the intervertebral discs of C4 C5 and C5 C6 and marked enlargement of the articular process joints of C6 and C7 and subsequent narrowing of the intervertebral nerve foramina. Dynamic CT is an area of increasing interest in both small and large animals and particularly in respect to the cervical spine. And if we think back to the previous slide, we we've got a good idea of what happens to the neck and extension now, but that doesn't really help us when our patients are moving and the neck is inflexion.
So hence the need for development of protocols for looking at dynamic CT. This paper had specimens of 12 warm blood horses, so these were cadaver specimens, and they used a custom made motorised testing device used to position and manipulate the neck and perform dynamic TD and 3D CT imaging. Images were obtained with a 320 detector ACT scanner and a solid state detector design that allowed image acquisition of volumetric axial length without moving the CT couch.
Dynamic videos were required. And divided into 4 phases of of movement. The authors found that medial translation displacement of the articular process joint surface was significantly greater inflexion.
Then when the horse was in. Tension movement They then concluded that the study provided a first step in the investigation of the potential of dynamic 3D 3D CT in veterinary medicine. This is a technique that has only just begun to be explored.
And leaves much room for it for refinement prior to its induction in routine practise. It is being used a little bit in small animal imaging, and particularly in in spike or Tor malformation in, in bigger dogs in in Dobermans, but still really only in the research setting. So this is an exciting, exciting development and hopefully one that we can refine to a clinical point of view to look at really what happens to And the articular process joints and to the subarachnoid space and the spinal cord, when the neck is inflexion, extension and lateral movement, that we're just not able to see currently with the setup of CT machines.
OK, so let's leave CT behind and we'll move on to. Recent clinically relevantly advances in MR. GAMR of upper limbs is now possible in low field units.
We used to be able to perform. GAMR in high field units in the UK, but now those those units are, are no longer in action. There are some worldwide and in the states, but these low field units have become available in Europe more recently.
This is a 0.25 Tesla machine, and this allows for imaging up to and including the stifle and carpus, and also allows heads of small horses and folds to be imaged. And these are the kind of images that we can expect from these machines.
So I think you'll agree that even though it's low field, we still have very good high quality images that can be of diagnostic quality. So these authors from Munich in Germany and Pulse of Pennsylvania in the US described their technique and observations in 76 cases of stifle MR using the Assote 0.25 Tesla rotating Grand A MR system.
There was an MR customised table and compatible anaesthetic equipment. All horses had a positive response to diagnostic analgesia of the stifle joint without full explanation of lameness on conventional imaging. In all horses, the stifle examinations were successful and complete, and typical MR lesions found included bone marrow lesions, osteocyst-like lesions, cruciate ligament desmopathy, meniscal tearing, or a combination of these injuries.
Retrospective reviewing confirmed initial radiographic and ultrasonographic images had failed to identify these lesions. Surgically, accessible lesions were confirmed on arthroscopy, if exploration was indicated. MR was useful to estimate the extent of cruciate and meniscal pathology as well as bone marrow lesions, and bone cystic-like lesions were more thoroughly investigated.
The protocol allowed for routine stifle MR independent of breed, age, or gender, and based on the results, the authors concluded that low field and stifle MR is safe and can delineate bone and soft tissue pathology, allowing a better understanding of stifle pathology and prognosis. In the world of standing MR there have also been huge advances. Firstly, in motion correction software for upper limbs, in standing MR hallmark units, and that's with the introduction of INA technology.
INA enhances the capabilities of motion correction technology and has been designed by Hallmark specifically for veterinary applications. INA significantly reduces the impact of limb movements during image acquisition and helps to maintain diagnostic integrity. It allows much better imaging of higher joints, particularly in fastpin echo and stir sequences.
So these are an example of the improvement image quality in the upper limb that we can get in standing courses using this technique. This is the same horse imaged in our clinic in October 2024. And we can see that the image on the left is a T2 FSC transverse MRI image.
So the image that we would have, traditionally used to look at the upper limb. And this is the same limb and same area scanned in the same session using the T2 FSC transverse IA protocol, the new protocol. And I think you'll agree that the image on the right hand side is dramatically improved in image quality and definition due to the reduction of that motion artefact.
So really exciting times in the world of standing MR imaging. Staying with the standing MR, their new state of the art, artificial intelligent algorithms are also being developed that effectively remove noise from images and ensure that the resolution and contrast remain intact. This innovation is going to allow for significant reduction in scan times and improve patient care.
It's hoped that the combining the AI with the INA motion correction will end up reducing scan times by 50% or more without compromising quality, and this is being rolled out across Hallmark standing MR sites in 2025. OK, let's look at a different kind of technology further. So, positron emission tomography is a nuclear medicine imaging technique that has recently become available for use in the horse.
Compared with planar centigraphy, commonly used in equine bone scans, PET has the advantage of providing cross-sectional imaging, resulting in more specific localization of lesions. Studies in human medicine have demonstrated that the functional information obtained with sodium fluoride PET scanning. Complements the structural information provided by CT or MR so we get functional rather than structural information.
And that has led to identification of lesions that are not readily apparent using other imaging techniques, providing valuable information in determining the activity of lesions. Pilot studies have demonstrated that the sodium fluoride pet identifies stress remodelling lesions in the fetlocks of thoroughbred racehorses and that the sodium fluoride uptake was identified most commonly in the palma metacarpal condyles and in the proximal pet sesamoid bones at site where catastrophic breakdowns then occurred. Some of these lesions were not recognised with imaging, other imaging modalities.
These regions of uptake corresponded on histopathology to areas of increased vascularity and increased osteoblastic activity. So this represents an early stage of the lesion prior to the occurrence of microscopic damage. The identification of these early lesions is considered paramount in the prevention of catastrophic breakdown in racehorses.
Sodium fluoride pet is also useful in the assessment of lameness localised to the equine foot. This is illustrated in the ability to recognise chondrosesamoidian ligament enthusiopathy using. Fluoride, fluoro dioxide, glu glucose, FDG.
And this has been reported for the assessment of deep digital flexor tendinopathy. The main limitation of equine pet has been the requirement for general anaesthesia in the past. And this has limited the ability of the technique as a screening tool or for following up scans.
To address this issue, a PET scanner with a horizontally orientated, freely open openable ring of detectors made by mile pet and long mile veterinary imaging has been designed. This is allowed for ease of positioning and safe release if the horse was to move during the scan. And this is the system that we can see on the slide in front of us.
Mile pet from Longm veterinary pet imaging is the world's first dedicated equine positron emission tomography scanner. And this enables safe imaging of the equine limb without anaesthesia in a standing sedated horse. The device is fairly compact, so there are no additional construction costs, and any rooms such as the centigraphy area can be quickly turned into a pet imaging suite.
It's adjustable in height, the ring opens for safety, and it can also be freely integrated into multimodality fusion software. So when it's combined with CT MR and radiography, it can be fused with pet to confirm lesion localization, given the correct software. The radio pharmaceuticals used typically are sodium fluoride, which is the bone tracer, and then in the chemical form, 18 fluoro deoxyglu glucose and FFDG, which is the soft tracer and soft tissue tracer.
So this paper reported on a prospective preclinical experimental study, and the goals of this were to validate the safety of this scanner, to assess image quality and to optimise scanning protocols. 6 research horses were imaged 3 times, and 6 horses in active race training were also imaged once under standing sedation. Multiple scans of both front fetlocks were obtained with different scan durations and axial fields of view.
A total of 94 FET scans were attempted and 90 provided images of diagnostic value. Radio tracer uptake was the main factor affecting imaging quality. While motion did not represent a major issue.
Scan duration of field of view also affected image quality. 8 specific lesions were identified on the pet images from anaesthetized horses and were also independently recognised on the pet images obtained on standing horses. This study val validated the safety and practicality of PET scanning specifically designed to image the distal limb in standing horses.
They found that proper preparation of horses similar to bone centigraphy was important for image quality. And a 4 minute scan with 12 centimetre field of view was considered optimal for clinical fetlock imaging. This technology has now been made commercially available.
So these are some of the examples of the images that we might expect from this this system. And we can see here a standing foot scan, combined with MR fusion to give us some structural information as well. So this case shows a 10 year old thoroughbred gelding with chronic left for limb lameness that resolved with paler digital nerve block.
We've got sagittal left 4 ft images and this marked focal and sodium fluoride uptake in the navicular bone corresponding to an area of osseous lysis on the distal aspect of the flexor cortex. We can see then that on the sagittal MR images that this corresponds to an area of hyperintensity and the fusion of the pet and the MR images and gives us that nice clarification of where the lesion is identified. This is a further case of dual tracer scan, so we're using both sodium fluoride and the FDG tracer.
And we've got an 11 year old Lisanna Gilding. With a several month history of progressive left and lameness, this blocks to an abaxial nerve block with no significant findings on radiographs. And we can see on the transverse left foot images, in addition to the sodium fluoride uptake at distoedial P2, there is FDG uptake in the collateral ligament of the distal interphalangeal joint, showing that we have both enthusopathy and active demopathy of the collateral ligament insertion.
This case is one of a 9 year old warm blood hunter jumper, and this was 4 out of 5 left hand limb lame, acutely lame, 20 minutes into a ride, positive to left eye inflexion, mild improvement in our actual nerve block, but resolution with perineal tibial nerve block. On the bone bone centigraphy, there was no significant abnormality. But the PET scan showed that there was a marked focal increase of sodium fluoride uptake in the distal tibia.
The pet in the CT fused MPR images showed focal increased sodium fluoride uptake localised at the subchondral bone of the medial groove of the distal tibia. The diagnosis was was the subchondral injury of the distal tibia. And then Finding for the PET cases, a use of positron of PET scanning and MR to identify central tarsal bone necrosis.
So this was presented for septic arthritis of the tars joint and the tarsal metatarsal joint. It had various synovial flushes but remained lame. Given the clinical deterioration, then advanced imaging, using sodium fluoride, PET scanning and MR were performed.
The images revealed a loss of blood supply and severe sclerosis of part of the central tarsal bone and potential devitalized bone. Histopathology confirmed necrosis and sclerosis in the dorsal medial aspect of the central tarsal bone. And then, areas of necrosis.
So this report highlighted the value in the use of PET scanning and MR for the early identification of osteonecrosis. OK, so, what about the more conventional imaging modalities such as ultrasound? Have they been left behind with all of these innovations?
Well, no, not really. We've had some really good advances in these as well, both in the hospital setting. And in the field setting in the form of point of care ultrasound machines.
So ultrasound guided injection techniques have enjoyed a revival in recent years, spurred on by some really nice papers describing this techniques and by CPD providers teaching at clinicians how to do these. These images show us a cadaver, a 3D model of the equine number spine. These images of courtesy of Doctor David Stack here at the University of Liverpool, and we can see that we've got a illustration of how to perform an ultrasound guided injection of the lumbar.
Particular forts and joints. And with a little bit of practise, this technique is readily achievable in the field with both a higher spec machine, but also with smaller handheld machines that we're going to talk about on the next few slides. So what is pocus?
Well, pocus is real-time instant ultrasound, that is available, patient side. It differs from traditional sound in a variety of ways. In the traditional ultrasound, assesses an anatomical region uses predefined parameters and measurements to provide a diagnosis.
Whereas point of care ultrasound assesses one part of the body at a particular time to answer a very specific question in the context of a physical exam and patient history. For example, in human medicine, does this patient with lower back pain and fever have kidney stones? In our patients, does this patient with respiratory difficulty have free pleural fluids?
In traditional ultrasound, then. The expertise have been complex and images are particularly in the human field and small animals have been interpreted by radiologists or cardiologists with significant ultrasound experience. Whereas pocus is more user friendly and can be performed by any pocus trained clinician.
Traditionally, ultrasound can also take hours or days to receive the results, whereas with point of care ultrasound, clinicians get clinically useful results in real time and can use that to increase the accuracy of their bedside assessment, or in our case stable or field side. Instead of the patient travelling to the machine, the machine also usually travels to the patient. But actually that's what we've all been doing for years.
So really, point of care ultrasound has come a full cycle for us that that's what we've always been doing with equine cases. We've moved a little bit towards higher spec and hospital based. Machines and reporting systems in recent years.
But actually there's been a real move back towards point of care ultrasound as well, because it's quicker, easier to use, and there've been huge advances in the technology. So the main difference is that point of ultrasound is not a comprehensive examination. Rather, it's an extension of the clinical exam.
In human medicine, it's become synonymous with more portable and cheaper equipment and the helltown units that can be used at the bedside. For eine vets in particular, stableside, handheld machines are, ideal. So many different handheld units have now become available, most of these connect to Apple or Android devices, and are really user friendly.
Some even are wireless. One unit, in particular, is pretty robust. It doesn't have piezoelectric crystals in the probe, but rather utilises a novel micro microchip technology and that's the butterfly unit in the left hand top picture.
All have varying battery life and practicalities, but all have given significant improvements in the availability of ultrasounds within both the stable and in the field and for the general equine practitioner, and really the quality of these images has improved drama dramatically in the last 5 years. So. I, I've tried all of these units that are on the screen, and all of them have advantages and disadvantages.
The butterfly has a really long batch life, which is great and really robust and perhaps is a little bit poor on image resolution, but certainly a very practical workhorse. And the Cri in the bottom right hand screen has superior, superior image resolution. But it does get quite hot when you're scanning, which is a little bit disconcerting, although reportedly not a problem, and to be expected, and, and it, but it does give us superior resolution.
So all have their advantages and disadvantages. The other things that these units offer are some form of image enhancement for the needle injections. So, as well as learning how to perform these imaging, These ultrasound guided techniques more readily, the technology, the AI allows us to see the needle more clearly.
And you can see this video on the Clarrius system showing just how well that needle lights up to guide in your guiding your ultrasound injection. I think any orthopaedic electron orthopaedic imaging and the advances made in the last few years wouldn't be complete without a quick mention of artificial intelligence. So we've talked about it a little bit already in terms of the software and the machinery, but what about in the actual thinking capability.
So, in human medical diagnostic, Artificial intelligence has been increasingly used. AI comes from improving human diagnostics or providing good quality diagnostic, diagnostics at lower cost. And this study analysed the diagnostic performance of a widely used AI radio radiology software versus 17 veterinary radiologists in interpreting canine and feline radiographs.
They aim to establish whether the performance of commonly used AI, The algorithms match the performance of typical radiologists and can therefore be reliably used. They also tried to identify, in which cases artificial intelligence was most effective. So they had 15 canine and feline radiographic studies in DCon format, and they were anonymized and 21 reported by 11 board certified veterinary radiologists and processed with 22 commercially and widely used AAA AI software.
Dedicated to small animal radiography. The AI software used a deep learning algorithm and returned a coded 24 abnormal or normal diagnosis for each finding in the study. The radiologists provided a written report.
And all reports were coded into categories matching the codes from the AI software and classified as normal or abnormal. The sensitivity and specificity and accuracy of each 27 radiologists and the AI software were calculated. The variance in agreement between each radiologist and the AO's software was measured to calculate the difference.
So AI matched the radiologists in accuracy and was more specific, but less sensitive than 30, . Than 30 human radiologists. AI did better than the medial radiologists overall in low and high ambiguity cases and in high ambiguity cases, AI accuracy remained high, although it was less effective at detecting certain lesions.
So this study confirmed AI. Its reliability, especially in low ambiguity ambiguity scenarios, and the conclusion was that AI performed almost as well as the best veterinary radiologist in in certain settings. However, its strengths lied more in conforming, in confirming normal rather than detecting abnormal.
And it also did not provide differential diagnosis. Therefore, the broader use of AI could be reliably increased but does require further human input. So deep learning in AI also involves the development of algorithms that simulate the problem solving capabilities of the human mind.
Sophisticated AI technology has received significant interest recently in veterinary work. And this review identified 39 primary research articles directly applying deep learning to animal disease detection or management, excluding non-primary research reviews and unrelated AI studies. The key findings from the current body of research highlighted that an increase in the utilisation of deep learning models across various diagnostic areas.
Over the past decade, radiographic imaging has emerged as the most impactful as regards to deep learning. Various studies have demonstrated notable success in the classification of primary thoracic lesions and cardiac disease from radiographs using these deep learning models compared to specialist veterinary benchmarks. While deep learning showed promise in veterinary diagnostics, several challenges remained.
These challenges were reported to range from the need for large and diverse data sets, the potential for interpretability issues, and the importance of consulting with experts throughout model development to ensure validity. The authors concluded that a thorough understanding of these considerations for the design and implementation of deep learning in veterinary medicine is imperative for driving future research, research and development efforts in the field. So potentially one day we are going to be replaced by deep learning in AI, but the method for it, the, message from this systematic review, which is very recent.
Was that that is still a little way off. So hopefully I'll get to retirement before it takes over. OK, so in summary, it's a really exciting time to be involved in equine orthopaedic imaging, CT and MRI great leaps forward, the introduction of PET scanning, really, increasing the depth of our knowledge further.
AI, should be of benefit, whether it gives us further challenges, whether it replaces us all, I think that remains to be seen. So thank you very much for listening and I hope that has been useful in considering just where we are where we are at with diagnostic orthopaedic imaging, currently.
