Warts precision medicine got to do with it?

Today we are still a long way from precision medicine. I visited my GP this week, which I don’t do very often, so was interested to see how technology is coming along in primary healthcare. I was able to book my appointment online, and check-in at reception using a touch screen computer instead of queuing to speak to a receptionist. There was even a computerised board announcing which room to go to – all very nice efficiencies that no doubt save some staff costs.

But the consultation with the doctor was pretty much the same as I imagine it has been for a hundreds of years. He had a quick look at a small but ugly growth on my abdomen, asked a few questions (which relied on my very fallible memory of how long I had it, whether it had changed recently, etc) and said it looks like a wart. To quote him accurately, “we’ll assume it’s a wart until it’s something more serious.” Given the tools and time available there’s nothing else he could assume – in 99.9% of the cases he’s probably correct – but that 1 in a thousand times when it turns out to be something more serious this is a wasted opportunity of early diagnosis. The same scenario no doubt plays out hundreds of times a day in that one small medical practice for all sorts of complaints – the vast majority of symptoms of a common cold or viral infection will be exactly that.

But what if it my growth, or someone else’s nasty cough, does turn out to be a cancer? Not that I think it is, I’m just using it as a hypothetical case study. By the time it’s grown and metastasised the tumours may be too numerous to fight. And certainly any attempt to treat it will be lengthy and costly. Compare that to the cost ff it was found early and treated immediately.

precision medicine spot the cancer

I’m not a doctor, but after reviewing dozens of images online of different types of skin infection I can see that rarely do two photos of the same condition look the same. There seems to be as much variation within a disease as between them. So in this modern technological age we shouldn’t be relying on highly-educated guesswork, we should have precision medicine that can accurately determine the root cause of symptoms. I’d hope by now that a quick swab, or a pin prick of blood, could be taken and within minutes it’s analysed to determine if there are antibodies or blood markers of serious disease.

We are getting close to this – recent news has included a sample-free laser test for malaria with results in 20 seconds, a single test that identifies all past virus infections and most recently a urine test that detects unique protein signatures of pancreatic cancer. It will probably start slowly, with quicker and cheaper tests for the most common diseases coming first. Then as technology prices reduce rarer conditions will also be able to be tested – just in case.

Eventually, in some utopian medical future, these tests won’t even have to wait until you have symptoms and decide to visit a doctor. Your daily routine will include a simple test carried out at home before breakfast to check for any developing problems before you are even aware of them. It may be even further in the future that we can cure any disease we find, but we will always have a better chance of doing that if we detect it in its very early stages.

Virtual Bodies are Accelerating Research and Diagnostics

Technology accelerates rapidly because it’s growing exponentially, but does this apply to medical technology? One place it certainly does is in virtual organisms – 3D digital models of cells, individual organs and even whole bodies.

Virtual Bodies are Accelerating Research and Diagnostics

Lots of medical research is initially based on animal models – that is, studying non-human organisms such as worms, mice or monkeys, to see what affect drugs, gene therapy and physical procedures have on them. One advantage of these models is ethics  – people are less concerned if 1000 worms die in the course of science than a single human, though of course there are many that would prefer to ban animal testing especially on our closest relatives such as other primates. Another advantage is timescales, and this is particularly important when looking at longevity treatments. Even if we had a drug that we were confident did no harm, and signed up 1000 willing human volunteers to test it, it would take decades before we would know if it was an effective anti-aging treatment. However, using worms with lifespans of weeks, or mice who only live for 1-2 years, the effectiveness can be seen quickly, enabling different types and dosages of drugs to be tested before considering human trials. But as well as the animal rights considerations, this still takes years of research and a lot of manual (and expensive) handling of the creatures under test.

So what if we could create a virtual model of an entire worm, mouse or even human. If this was accurate enough to respond to physical and chemical factors, taking into account the complex interactions within the body, we could test as many different drug candidates as we could feed into the computer. And as computing power continues to grow exponentially we can then turn up the speed and see the results overnight that even with worms might take several weeks.

We’re already getting close to this scenario. Research teams around the world are working on individual organs and also linking these up together to demonstrate how the entire body would react. Here are just a few examples:

  • University of California-Davis School of Medicine’s ion channel based heart model predicted adverse effects of 2 drugs used to treat abnormal heart rhythm.
  • The mechanics of the complex geometry of the skeletal system has been modelled by the University of Jyväskylä  to determine how different exercises induce bone strain and strain rates and to research the causes of degenerative arthritis.
  • Living Heart Project has developed a physics-based digital 3D model of the heart that can be used to virtually test new physical devices and aid heart disease research, and allows surgeons to walk inside a massive heart projection to really understand how the organ works.
  • Virtual worm brain (OpenWorm project) simulates all the connections between the c. elegans worm’s 302 neurons and is able to control a Lego robot without a single line of code.
  • The EU’s Human Brain Project has developed a simplified virtual mouse brain mapped to different parts of a virtual mouse body, including spinal cord, whiskers, eyes and skin.
  • Virtual Physiological Human programme aims to create a computer simulated replica of the human body (“in silico”).

This last project also is also applicable at the treatment end of healthcare. Once a detailed and accurate virtual model of a “standard” human has been developed, this could then be configured with the physiological parameters to match an individual. Their personal data could be input into their virtual avatar to predict how their specific body would react to drugs and other treatments. Already, for example, the University of Pittsburgh has modelled the complex interaction of multiple inflammation markers in blood enabling trauma patients’ risk of multi-organ dysfunction to be calculated and appropriate intensive care allocated.

Personalised medicine is mainly trying to predict how individuals will respond to pharmaceuticals based on their genes – which given the efficacy of most drugs is better than blindly working down a list of potential treatments; but what if you could then try that drug in a personalised virtual body and see how it really reacted given a multitude of individual factors? That could save time finding the best available option, save money in the wasted time of healthcare professionals and drug costs, and most importantly save lives.

Links to research mentioned in this blog post are available on the Digital Modeling page.

Laboratory Automation – only the white coats remain

If you think, like I did until a few minutes ago, that blood tests at hospitals went into a big room full of people with white coats and pipette tubes you’d be wrong… apart from the white coats!

Here’s a great case study video from Siemens showing their Aptio Automation system working at the NHS Tayside Ninewells Hospital, Dundee, UK. It combines lots of types of testing (e.g. hematology, immumolgoy, coagulation) with all the test tubes loaded onto trays and then in they go. The operators (in the white coats still) tell the system the tests required and then it’s over to the machines. The appropriate diagnostics are all done without human intervention and then stored away, with the ability to recall them for additional tests or retests.

At Ninewells they’re currently handling nearly 2,000 tubes per hour at peak times – and when that emergency sample comes in they load it onto the tray the same as any other but the tube is then fast tracked – in the video you see it literally overtaking others.

Siemens Aptio Automation system

The clinical director of the hospital says turnaround times reduced by about 25% which must have been from an already automated system as it just wouldn’t be physically possible to prepare, test and store that many sample manually in anywhere near the current processing time.

OK so it’s not yet the pin prick of blood with instant diagnosis which I’m waiting for but it’s a good start – and as always, once it’s a technology there’s nothing stopping it becoming smaller, faster and cheaper.

Video available on YouTube: Laboratory Automation Improving Patient Testing