The value of ‘pattern recognition’

01 January 2015, at 12:00am

Dr DAVID WILLIAMS continues his series with a discussion of the link between astronomy and ophthalmology and how we can so easily recognise stars and conditions which are common…

IT is not just the considerable nip in the air while I take my evening perambulation that tells me winter is here.

Over the Eastern horizon strides the hunter with his three-starred belt and nebulous sword. Orion disappeared in the early spring and only now makes his appearance again in the winter sky. Orion, perhaps the best known of the constellations. I use him at the beginning of my ophthalmology lectures to the vet students in Cambridge.

“What is this?” I ask them, putting him up as one of my first slides and, of course, everyone can reply with the correct answer.

“How do you know?” I enquire. Well it’s obvious, everyone says, we’ve seen it so often, it’s just there imprinted in our minds. But what relevance has astronomy got to ophthalmology?

Recognising the signs

Looking at the night sky, just like looking at animal eyes, is a matter of pattern recognition. See if you can recognise this one: a miotic pupil, a drooping upper eyelid and a third eyelid drawn half way across the face of the cornea. And all these three in a five-year-old Golden Retriever.

Easy: it’s a classic Horner’s syndrome. Did you get it?

Here’s another one: a dark brown circle resting in the bottom of the anterior chamber of a black Labrador. That’s an iris cyst, almost for certain. Next, here’s another black circular mass, this time in the lower eyelid of a 10-year-old Persian cat who also seems to have a couple of similar if smaller dark circular masses in her upper lip. These are apocrine hidrocystomas, benign tumours of the sweat glands of the lip and lid, also known as cysts of Moll.

So much of ophthalmology is simple pattern recognition. I guess that one might say the same of much of veterinary medicine. The thin, scraggy, ageing but still overly active cat is going to be hyperthyroid isn’t it, while the thin, scraggy but lethargic and polydypsic cat is far more likely to be in chronic renal failure.

We know that because we see those all the time. Maybe the Horner’s syndrome is common enough to be readily recognised but the iris cyst and the aprocine hidrocystoma may be less easy to spot.

The trouble is that, for me, when these cases are referred to me it takes a few seconds for me to recognise the pattern, while the referring vet may never have seen it before. Then my job is to explain to the owner why it was less than a minute for me to see what the problem was where a group of vets hadn’t managed to come to a diagnosis in a couple of weeks.

They hadn’t recognised the pattern, never having seen it before, while I was familiar with it having been looking at eyes for 25 years! But how do you recognise a pattern if you haven’t seen it before?

A particular case brings that home to me with particular force. It must have been November 1997 and before my morning clinic at the vet school I popped into the library to scan the shelves for interesting new reading material.

There was the latest Veterinary Clinics of North America on small animal ophthalmology. Flicking through the pages I came across Dave Ramsey’s interesting chapter on orbital surgery with a weird case I had never seen before: a young, female Golden Retriever with eyes looking as if they were popping out on stalks, white sclera all the way around and strangely no protrusion of the third eyelid as one would normally see with a retrobulbar mass.

This was extraocular myositis, something I had never seen before. So imagine my surprise when my first case of the day was a presumed glaucoma in – you’ve guessed it – a young, female Golden retriever, the spitting image of the picture I had just seen in that VCNA issue!

Jumping with glee

It doesn’t take much to get me excited, as those of you who know me will recognise, and so jumping with glee I left the owner and the students (asking them to do a full clinical examination of course) while I rushed back to the library, grabbed the issue off the shelves and raced back to the dog and rather bemused owner to show her what looked for all the world like a picture of her pet already published in an American journal!

Now my point here is that had I not seen the picture in Ramsey’s article, I would have spent a long time scratching my head and a fair few diagnostic tests and imaging modalities before coming to the diagnosis. To have seen it once before, even if only in a research article, led me to the answer in a second or two.

How does that help in the cases coming into your clinics, you might ask? When I was an intern and resident with Dr Keith Barnett, I had the joy and privilege of spending two-and-ahalf years with him seeing every case that he saw, countless new cases a week which very quickly built up my mental library of ophthalmic patterns.

I realised in later years that the way to show students the hundreds of cases they need to see to be able to recognise the classic patterns in veterinary ophthalmology was to harness the power of the internet and produce a website on which the cases I see could be placed.

There are about 250 cases on the old archived site at http://davidlwilliams. and a growing number on the newer blog (, from dermoids to hypertensive retinopathy.

The only problem with the blog is that there really isn’t room for longer discussion of some of the cases. Maybe I’ll tell you a bit more about one we mentioned above, Horner’s syndrome.

Wikipedia will tell you that the oculosympathetic palsy is named after the Swiss ophthalmologist Johann Frederick Horner who first reported the syndrome in 1869. But much as I love and value Wikipedia, you can’t believe everything you read in it.

In actual fact the first case demonstrating the ocular signs of cervical sympathetic denervation was recognised by William Keen, an army surgeon in the American Civil war, five years before Horner saw his case.

Faced with an army recruit shot in the neck and with a miotic pupil and a drooping upper eyelid, Keen remembered an illustration in his old physiology textbook, John Call Dalton’s A Treatise on Human Physiology.

Dalton was remarkable in his day for basing his writing on experimental studies, and here he had transected the cervicosympathetic trunk and produced what we now know as Horner’s syndrome. Keen recalls the event in his textbook Gunshot Wounds and Other Injuries of the Nerves written several years later.

I’m not sure what the injured soldier thought to Keen’s exclamation, “You are Dalton’s cat!” but Keen had the diagnosis in an instant: a perfect example of pattern recognition even across the species boundary!