Never mind the biosensors

There will undoubtedly be some significant advances in sensing and diagnosis over the next 15 years, but much of what is touted in the media simply doesn’t stack up from a clinical, regulatory or commercial perspective. The aim of this short article is to equip you with some additional insights to help you sift out the plausible from the impossible.

We’ll start by considering the frequency of measurement and whether a measurement is non-invasive or invasive, then we’ll look at implants and the feasibility of home testing before concluding with my thoughts about the future of diagnostics over the next 15 years.

How often?

Lots of measurements are only made once, or very infrequently. Whereas you don’t need to test a patient for their blood type more than once, a diabetic will need to measure their blood glucose several times a day over their entire life. These are obviously extreme examples, but it’s really important to be cognisant of both how often a measurement can be made and how often it needs to be made.

Technology development, miniaturisation and cost reduction over the last 20 years has made it possible to make some measurements on a frequent or even continuous basis. For example, it is now possible to attach sensors to a patient to measure their blood pressure, breathing rate, heart beat and blood oxygenation continuously. Over the last five years, it has been increasingly possible for these measurements to be processed, logged and uploaded to the cloud via smartphones.

An important attribute of these physiological measurements is that they can be made non-invasively (without breaking the skin) and using sensors which don’t get worn out or used up — they can keep taking measurements indefinitely. However, there’s a limit to what you can diagnose simply using physical measurements such as pulse rate, blood pressure and blood oxygenation. Whilst a smartphone connected to the cloud can do a great deal of signal processing and analysis, there still remains a fundamental limit to what you can detect using non-invasive sensors.

To stab or not to stab?

Some tests will always need access to the blood. For example, if you want to do a full blood count or check someone’s biochemistry then you have no choice but draw some blood to test. A lot of resource has been invested in trying to develop non-invasive biochemical measurements but unfortunately this is extraordinarily difficult to achieve. Pulse oximetry is the exception, since haemoglobin conveniently undergoes a large colour change when oxygenated and there’s lots of haemoglobin in blood. Frustratingly, most other analytes, such as glucose, are far less amenable to non-invasive monitoring. Billions of research dollars have been spent fruitlessly chasing the holy grail of non-invasive blood glucose measurement over the last 25 years and even the newest blood glucose sensors, such as Abbott’s new FreeStyle Libre Flash Glucose Monitoring System, still need to be placed under the skin.

What If we implant tiny sensors into the body?

As far back as 1987 the renowned biomedical futurologist Steven Spielberg brought us ‘Innerspace’ — portraying a future of implanted miniaturised devices, albeit with a fair degree of artistic license and a flagrant disregard for research ethics. Almost 30 years have passed and tremendous advances have been made in sensor technology and microfabrication, but implantable sensors still aren’t the answer to diagnosing most diseases.

If you just want to make a simple physical measurement then there’s a chance that you can do it with an implantable sensor. For example, CardioMEMS (now St. Jude Medical) has developed a miniaturised blood pressure sensor which is permanently implanted inside the pulmonary artery. The sensor is embedded inside a tiny glass brick — the size of a grain of rice, and delivered via right heart catheterisation and fixed to the inside wall of the pulmonary artery. The sensor is interrogated from outside the body and enables clinicians to regularly and accurately monitor pulmonary artery pressure and detect impending heart failure.

However, if you want to make a chemical or biological measurement then it’s exceedingly challenging to use an embedded sensor, and one of the biggest problems is that foreign material attracts clotting blood, gets attacked by the immune system and is eventually encapsulated by the body. This is OK if you’re measuring blood pressure but it’s a show stopper if you’re trying to make an optical measurement or if you need circulating blood to come into contact with your sensor. Clever coatings can minimise fouling but all surfaces will inevitably end up coated to some extent. The other big problem with embedded biosensors is that most sensors use biological molecules to recognise and bind to the target of interest. These biological molecules, such as antibodies or proteins, are inherently fragile. For example the antibodies attached to a pregnancy test strip are stable for two years in a dry package on the shelf in Walmart but won’t last more than a few hours inside your body. This means that the vast majority of biological and biochemical sensors can’t be implanted in the body.

‘Ok, but surely we’ll be doing lots of diagnostic tests at home in 2030?’

Companies such as CUE would have you believe that’s the case. This US start-up company claims to be ‘creating the most advanced consumer health product ever seen’ which will allow consumers to measure five parameters in their own home, including vitamin D, testosterone, fertility hormones and influenza. The CUE promotional video is very nicely produced and the industrial design is nothing less than you’d expect from a Californian tech company, but I suspect the underlying premise is flawed and won’t ‘open up an entirely new world of possibilities for understanding and improving our health’. I’m sure CUE will sell some devices to the wealthy-worried-well with a penchant for new technology, but I don’t expect this or other similar products to have an immediate and lasting impact on healthcare provision for a number of reasons.

Firstly, cost of goods will always be a challenge. Consumers are extremely cost sensitive and history suggests that the overheads of producing diagnostic tests to the standards required by regulators are very burdensome. Costs from QA, QC, calibration, in-process controls, regulatory compliance and process validation put enormous pressure on margins.

Secondly, home diagnostic products are often trying to address needs that don’t really exist. I’m not convinced that consumers will really continue to spend both time and money in order to monitor their vitamin D or testosterone levels once the novelty wears off. And if a diagnostic test tells me I have influenza then what can I do about it? I feel too ill to go to work anyway, my family can’t exactly move out, and there aren’t any antiviral therapies worth taking.

Thirdly, the implications of integrating home use diagnostics with existing health care services are not straightforward and will take time to be established. I’m not sure how my GP would react to a call from me saying: ‘Hey doc, I’ve just diagnosed myself with a cool new gadget, could you write me a prescription for these drugs that Wikipedia says I need?’.

Lastly, the idea of home use diagnostics isn’t new. The iSTAT point-of-care diagnostic system (acquired by Abbott in 2003) has been available for over 18 years, and it is naïve to think that the world’s big diagnostic companies haven’t stopped to consider the potential of home diagnostics market during the last 20 years; every one of them will have commissioned a marketing initiative or two with the aim of exploring the potential of home use diagnostics. On top of that, hundreds of small diagnostic start-up companies have claimed that their particular spin on point-of-care diagnostics is the key to unlocking this vast untapped market. Yet companies large and small have generally failed to find a new and significant convergence of unmet need, clinical practice, health economics and technology. Yes, there are some niche applications beyond blood glucose, urine dipsticks and pregnancy testing, but it rarely makes sense to do complex clinical diagnostics in the physician’s office rather than the hospital, or at home.

Don’t get me wrong, I’m full of admiration for the diagnostic innovators and entrepreneurs at companies such as CUE who are building on the knowledge gained in the last 20 years to extend our current and future capabilities. Even if they don’t achieve all of their goals they will certainly pave the way for future innovators. However, I do suspect a fair few CUE devices will end up gathering dust at the back of cupboards, along with the odd Nokia Communicator, DAT player and abandoned ice cream maker.

‘Ok Mr Sceptical, so what do you think we’ll be doing in 2030?’

My personal hunch is that we will see advances in sensing, microfluidics, signal transduction and MEMS technology having the greatest impact on testing in central laboratories, not in the home or doctor’s office. Central labs are where the vast majority of clinical diagnostic tests are currently performed, and these facilities can afford large automated systems, negotiate significant volume discounts on reagents, have QC/QA procedures in place and are already integrated with hospital information systems.

If new technologies can significantly reduce costs, through miniaturisation and multiplexing, then it may be economically viable to test the general population on a regular basis for a whole panel of biomarkers, detecting diseases earlier and allowing preventative action to be taken.

This is generally too expensive at the present time but as diagnostics costs fall and treatment costs rise (drugs, surgical intervention, long term care and so on) then we may reach a tipping point when routine screening becomes economically viable. For example, everyone over the age of 40 might give a blood sample every month and send it off to a central lab which will do the biochemical analysis, compare the data against previous readings, and report any concerns directly to the primary care physician.

An exciting consequence of this shift is the increased possibility of detecting those cancers which are undetectable using a one-off test but which can sometimes be spotted by monitoring a panel of biomarkers over an extended period of time. At the moment, these tests are too expensive and don’t have particularly high sensitivity and specificity, but if used as part of a monthly screen which is analysed by computer, these tests could potentially detect cancers and other serious diseases at an early stage, saving lives and saving money.

The future of diagnostics and monitoring is exciting. The challenges are higher for today’s innovators than for their predecessors, but today we have an unprecedented suite of emerging technologies at our disposal. I suspect that the individuals and companies that will enhance patient care the most will be those who understand their users, and which can harness new technology to create products which offer demonstrable and measurable economic benefits.

This article was taken from issue 7 of Insight magazine. Get your free copy of the latest issue here.