Talk about “seeing more” – diagnostic imaging has been one of the areas of development in veterinary medicine that is almost unrecognisable from where we were just two decades ago. When the authors graduated, many practices were still hand-processing physical films, with the main advancement being the introduction of automated processors.
The move to digital radiography and the increasing adoption of ultrasound, followed by the advent of computed tomography (CT) and magnetic resonance imaging (MRI), has opened up a digital world of wonder and rapid and ongoing growth in insight, knowledge, understanding and refinement within diagnostic imaging. Artificial intelligence (AI) is now the single most important development in veterinary radiology – indeed, across the entire field of veterinary medicine – and it’s already here. The future really is now.
Advanced imaging utility
CT and MRI
It wasn’t too long ago that CT and MRI were restricted to the realm of the specialist centre, and their use in diagnostics was limited to the referral patient. Then came the advent of mobile units, opening their HGV doors to more locations. Now, with the increasing demand for advanced imaging to manage owner expectations and the desire to grow in-house expertise, knowledge and capability, more first-opinion practices are installing CT and MRI machines.
Companies developing and producing these machines are introducing increasingly affordable lease options based on usage, with attractive business prospects. The size and complexity of these machines are also decreasing, with standing CT and MRI options available for horses, negating some of the need for general anaesthetics and enabling more horses to be imaged as day patients. Practices with machines can act as “hubs” for neighbouring clinics to send patients for imaging without requiring referral. And, when married with teleradiology, a first-opinion vet can send an animal in their care for advanced imaging locally and receive a report from a diploma-holding radiologist within a day.
With this roll-out of advanced imaging equipment comes the need for greater education about the benefits and limitations of each machine and its modalities
With this roll-out of advanced imaging equipment comes the need for greater education about the benefits and limitations of each machine and its modalities to help decide which device to use. For example, helical CT and cone-beam CT should be used for specific and different clinical indications. We need to support informed decision making by practices wishing to add advanced imaging to their diagnostics portfolio. For this reason, VetCT offers free, independent advice on which machines best suit the needs of individual practices.
PET and ultrasound
Early adopters of positron emission tomography (PET) scans are already successfully using this technology in larger institutions, and although this modality is currently very expensive, it won’t be long before it’s more widely accessible. Ultrasound too is an area of great innovation, with new machines entering the market and increasing options for in-clinic cart-based machines and miniature mobile handsets for ambulatory use.
However, training is key for competence and confidence. This is another area where we can empower veterinary nurses to play an important role in diagnostics to benefit the patient, support the vet-led team and further career development.
You only need to look at your smartphone and the variety and volume of apps to be reminded how far beyond expectation software development has accelerated. With the increasing ease of sending high-quality image files and the rapid advancements in processing software, the post-acquisition phase of imaging is now an art form. Nowadays, three-dimensional reconstructions of complex vascular anomalies enable accurate surgical planning, and the printing of bespoke 3D implants mapped against complex fractures facilitates exact orthopaedic repairs.
Training is also benefitting from software advancement, with remote training options enabling upskilling and real-time CPD in the clinic through remote expert supervision for primary care practitioners on their own cases in practice. The recent development of ultrasound units that allow live viewing by remote third parties has increased the ease of providing remote guidance.
Whether it’s radiographic positioning, ultrasound or endoscopy, the ability to live stream a procedure creates a range of in-house training potential
Whether it’s radiographic positioning, ultrasound or endoscopy, the ability to live stream a procedure creates a range of in-house training potential, benefitting both patients and the veterinary team.
Due to the global shortage of radiologists and the ever-increasing demand for advanced imaging, in-house radiologists are an increasingly rare breed. For both work–life balance and variety, radiologists are increasingly adopting remote working for all or part of their reporting through teleradiology.
For the practitioner, gone are the days of physically mailing films to the local referral centre. Increased access to teams of radiologists around the clock means more reliable turnaround times and more efficient workflow for sending images and clinical histories. This way, every practice can have a radiologist “in-house”.
Far from reducing imaging interpretation skills in-house, radiologists are keen to support knowledge acquisition and diagnostic confidence by providing detailed, clinically relevant reports – another useful addition to “learning on the job”.
Advances in our knowledge base
As the use of advanced cross-sectional imaging increases across practices and species, knowledge of imaging pathology has increased in both breadth and depth. For example, the increasing number of CT scans of rabbits reported by VetCT has revealed insights into abdominal pathology, including rabbits showing liver lobe torsion, appendicitis (Di Girolamo et al., 2021) and gastric dilatation and volvulus. All of these are conditions that may be much more common than currently thought and where surgery is likely to be urgently indicated. Performing an exploratory laparotomy on a sick rabbit is challenging, and having clear guidance about the pathology and the appropriate surgical correction required preoperatively will help us to optimise the prognosis for these patients.
Although conventional radiographs and ultrasound are far from obsolete and remain the affordable option […], the more we learn from cross-sectional imaging the more aware we are of the limitations
But it’s not just rabbits: although conventional radiographs and ultrasound are far from obsolete and remain the affordable option and the exam of choice in many cases, the more we learn from cross-sectional imaging the more aware we are of the limitations. This leads to more appropriate case selection for each diagnostic imaging modality, which, in turn, increases efficiency, reduces cost and improves patient outcomes by decreasing the frequency of suboptimal choice of imaging procedures.
Yes, it’s already here! Science fiction is now science fact when it comes to AI in everything from medical diagnostics to the dawn of “automated killer robots”, as Professor Stuart Russell warns us in his BBC Radio 4 lecture (Russell, 2021). AI development has rapidly outpaced global consensus over regulation, ethics and even terminology for “killer robots” – and so it has for AI in veterinary medicine.
There are huge potential benefits of AI in radiology as a speciality. As the VetCT position statement (Johnson and Labruyère, 2022) highlights: “There is a limited pool of veterinary radiologists and an ever-increasing need for expert interpretation of radiographs and more advanced diagnostic imaging modalities. This presents a huge opportunity for the development of AI and related technologies to better address demand, save time and potentially improve clinical knowledge and outcomes.”
Imaging lends itself to data-driven algorithms as, since the inception of digital radiography, imaging data has accumulated in vast global online data banks held by universities, practices and teleradiology companies. It therefore behoves veterinary radiology as a speciality to lead the way with both research and guiding the ethical and legal implications of AI that can be adapted and applied in other areas of medicine and surgery.
However, despite its enormous potential for good, the development and commercialisation of AI tools are outpacing research, publication, regulation and education across the profession globally. The responsibility – as ever – lies with the individual practitioner to apply and interpret medical tools (including AI) to benefit the welfare of the animals in their care. In the absence of clear research underpinning evidence-based guidance on best practice for the application of these technologies, we’re asking our colleagues to use their sight while blindfolded.
Despite its enormous potential for good, the development and commercialisation of AI tools are outpacing research, publication, regulation and education across the profession globally
There is a real risk of undermining trust in AI and new technologies if they are deployed without the appropriate evidence base in research and governance, especially at a time when the profession needs all the support and efficiency savings it can get from technological advancement. There is a moral responsibility – in the absence of regulatory incentives – for individuals, organisations and companies developing AI to employ high standards of self-regulation to safeguard both the practitioner and animal welfare.
Typically, human health AI advancements outpace veterinary ones, which is unsurprising given the stakes at play with humankind and the available funds. However, regulation is much lighter in the veterinary industry, and patient data is much easier to access.
There is an enormous opportunity for collaboration between animal and human health to develop and trial imaging modalities and AI algorithms that speed the development of diagnostic tools in both fields. This makes diagnostic imaging another area where a two-way flow of information and expertise would benefit both human and animal health enormously.
The future of diagnostic imaging is now; what remains uncertain is how well we steer it. To realise the full potential of new diagnostic insights, collaboration, pooling data, sharing knowledge and self-regulation are essential.
Our teams are benefitting from increased diagnostic confidence, improved patient outcomes and upskilling – all of which feed into job satisfaction and, ultimately, retention
What is more certain is that we are seeing more clearly than ever, and more of the animals in our care are benefitting from the rapidly increasing accessibility of advanced imaging modalities. Our teams are benefitting from increased diagnostic confidence, improved patient outcomes and upskilling – all of which feed into job satisfaction and, ultimately, retention. It is possibly the most exciting point in the field of veterinary radiology since the inception of the speciality.