How tech is reinventing healthcare
From deep learning to fighting health misinformation, tech is changing the healthcare for doctors and patients alike
“It’s time to get rid of the scars,” says Mark Slack, a gynaecologist and co-founder of robotic surgery startup CMR Surgical. Cambridge-based CMR Surgical was founded in 2014 – Slack remembers how, when the company started, surgeons wanted a robot that was suitable for all surgical disciplines: equivalent in cost to a straight-stick surgery, adaptable to any theatre, modular and quick to set up and take down. “There was nothing like it,” he says. “So what did we do? We had to build our own robot.” The result is the Versius, a surgical robot designed to help surgeons perform keyhole surgery. Today, the company employs around 500 people and, after raising more than $100 million (£75.2 million) in June 2018, it’s now valued at more than $1 billion (£759 million), making it one of the rare unicorns in medtech.
Keyhole surgery is broadly preferable to open surgery, Slack says. Overall, around 50 per cent of post-surgery complications are reduced by having keyhole surgery, rather than open. “If you have a small, minimal-access wound, almost none of the patients return to hospital,” he says. “If you have a large wound, about a fifth of patients are required to go back into theatre if they have a wound infection.”
The problem, according to Slack, is that fewer than 50 per cent of patients worldwide are able to take advantage of minimal-access surgery. There are many reasons: one, keyhole surgery is technically difficult and physically demanding. “It’s like trying to tap your head and rub your chest,” Slack says. “After many years of training minimal-access surgeons, I’ve realised that a lot of people struggle to actually master the fine art of advanced surgery.”
Robotics offers a potential solution. The Versius has multiple arms, each with seven degrees of freedom, allowing the surgeon to do procedures that aren’t feasible with standard laparoscopy.
It’s also equipped with 3D vision, giving surgeons augmented visualisation of the surgical field and allowing them to do trickier procedures.
Slack hopes that Versius will solve many of the problems encountered in surgery today. For instance, thousands of patients are harmed annually by post-surgical complications. “Minimal-access surgery will reduce complications by about 50 per cent,” he says. “And then if we could standardise surgery, this would reduce it further.”
Robotic surgery has further advantages over that performed by skilled humans: it provides a digital interface between patient and surgeon that can continually collect important data which would otherwise be lost: image analysis, user’s movements, telemetry from the robotic instruments and much more. “This gives us the ability to look at the surgeon’s hand movements,” Slack says.
“We already have data that shows us how to tell the difference between a novice user and an expert user. That means we can use it to identify surgeons who have picked up bad habits, and are maybe excessively using certain aspects of the robot, and therefore could adopt practices that would be better for the patient.”
CMR has also developed the Versius trainer, a simulator that surgeons can use to become familiar with the machine. An app, called MyVersius, stores videos that surgeons can then analyse to improve their surgical procedures. “That enables them to keep a digital logbook and to benchmark themselves against best practice,” Slack says. “It also means surgeons – if they get slightly bad results – can easily get back on track, because they can bring in others to help them and see why their results have deteriorated.”
Currently, only five to seven per cent of all surgeries are performed robotically. The Versius has already been used with more than 800 procedures and patients in multiple sites around the world, and has been adopted by four NHS hospitals and other sites in India and Europe. “Within the next five years, I believe you’re going to see an enormous acceleration of robotic surgery to the point where it will take over from manual laparoscopic procedures completely,” Slack says.
And surgical robots aren’t the only big technological breakthrough transforming healthcare. From Nobel-winning therapies to the fight against vaccination misinformation, here’s more of what we learned at WIRED Health: Tech 2020.
Using CRISPR to edit disease out of our DNA
Two weeks before she received the 2020 Nobel Prize in Chemistry, biochemist Jennifer Doudna spoke to WIRED Health from her house in California to talk about her discovery of gene editing technique CRISPR, which she compared to “molecular surgery”. “CRISPR is really a way to alter the DNA in cells and organisms in ways that allow precise correction of disease-causing mutations,” she says. “It also allows scientists to do all sorts of manipulations of genetic material in living cells and organisms. It gives them the ability to alter the details stored in the genome of any cell types and any organism, allowing us to both understand and manipulate that information.”
The development of CRISPR originated, according to Doudna, in a “very obscure” research area dedicated to the study of how bacteria fight viral infection. In bacteria, Doudna explained, CRISPR-Cas9 is a defence system that provides immunity. “They do it by creating a small molecular recording of a virus, using the viral DNA,” she continued. “The cell stores a piece of viral DNA in the bacterial DNA, a bacterial chromosome called the CRISPR locus. And that stored DNA provides a template for making molecules of RNA that can in turn direct proteins called CRISPR associated proteins to find and destroy DNA that has a matching sequence. The cell creates a genetic vaccination card which allows it to remember, in a molecular sense, viruses that have infected it in the past and provide protection, if they show up in the future.”
According to Doudna, what emerged from that research was the realisation that CRISPR had the ability to create precise changes to DNA. “That was really the beginning of the CRISPR genome editing era,” Doudna says. “We now have a technology that is readily available to scientists worldwide.”
One of its current applications is in sickle cell disease, a disease caused by a single base-pair mutation in the genome that encodes a protein for haemoglobin. In patients with the mutation, this causes the formation of sickle-shaped red blood cells, which tend to block blood flow. Recently, several companies have initiated clinical trials using CRISPR to treat sickle cell disease. “One patient has received CRISPR therapy for sickle cell disease, and she has been effectively cured of her disease,” Doudna says. “That’s led to a lot of excitement in the field, and thoughts about how we can now expand this technology to treat others that have this or other types of rare genetic disease.” Doudna thinks that other diseases, such as muscular dystrophy, will soon also benefit from the application of CRISPR. “I suspect in five years, we’ll be talking about clinical trials and maybe real results,” she says.
The technology, however, remains experimental and expensive, and won’t become part of anyone’s typical healthcare options just yet – but Doudna remained optimistic that this may happen one day. “I would love to see this become a standard of care for certain diseases and, of course, to be affordable as well. Safety is essential, and there’s cautious optimism based on early clinical trial results,” Doudna explains.
Recently, scientists have been using CRISPR to develop new forms of disease diagnostics. According to Doudna, certain classes of CRISPR proteins have the ability to recognise a target sequence that matches the sequence of the guide RNA. “These systems can be used to detect viruses.” Doudna says. She also believes that in the coming months we will see CRISPR-based, point-of-care diagnostics that will allow for a rapid detection of Covid-19. From basic research in cells, to the development of new drugs and antibiotics, to the cultivation of nutritious, disease-resistant crops, Doudna believes that, “the possibilities are extraordinary. It’s really an exciting time to be working in this field.”
The race to develop an effective Covid-19 vaccine
In October 2020, more than 190 Covid-19 vaccines were in development, with around 40 being tested in humans. At the forefront of that race are Massachusetts-based Moderna and the German company BioNTech. Tal Zaks, the chief medical officer of Moderna, offered his perspective on the rapid pace: “Unfortunately, there’s high transmission, and this is the paradox of vaccine development,” he says. “The worse it is out there in terms of transmission, the quicker we’ll know whether a vaccine works.”
This is one of the reasons there’s no vaccine against other coronaviruses such as SARS and MERS: not enough people were infected. That’s not the case with Covid-19. “If nobody was getting infected, we would never know if a vaccine works,” Zaks added. “The more cases and the more transmission we have, the sooner we’ll be able to prove that these vaccines are indeed safe and effective.”
BioNTech’s vaccine underwent phase three clinical trials in more than 44,000 participants; Moderna, in collaboration with the US National Institutes of Health (NIH), was running its phrase three trials with 30,000 participants. As WIRED went to press, the results came in: BioNTech had achieved a 90 per cent efficacy rate; Moderna reported almost 95 per cent efficacy. Both companies are pioneers in messenger RNA therapeutics, which uses a molecule that can encode genetic information. “Scientists can take part of a virus that is important to induce a strong, effective immune response, encrypt that with messenger RNA, and deliver it in a way that can be taken up by human cells,” says Ugur Sahin, co-founder and CEO of BioNTech. Those cells, as a result, produce the vaccine. This method is, according to Sahin, a very precise way of inducing immune response in patients.
“The beauty here is we actually don’t use the virus,” Zaks says. “We’ve never had the virus in our labs, we don’t need it.” All that’s required is the genetic information
RNA vaccines are quicker to produce, more precise and very flexible as a technology. “The infrastructure required is relatively small and quick, which means in the manufacturing space, you have a tremendous agility that usual technologies don’t,” Zaks asserted. Vaccines typically take decades to develop, and Zaks and Sahin explained that their fast development cycles came from doing different stages of vaccine development in parallel, rather than sequentially. “We were planning for phase three and vaccine manufacturing even before knowing if the vaccine works,” Sahin says.
Various pharmaceutical companies have also adopted new standards of data sharing, transparency and collaboration, which have accelerated the process as well. “I only have two competitors here: the virus and the clock. The world needs more than just one company to succeed,” Zaks says. Sahin agreed: “It’s really important to see how people collaborated: Moderna teamed up with the NIH, we teamed up with Pfizer, AstraZeneca teamed up with the Oxford University. We have the strongest transparency. People see the data coming in almost in real time.”
The results so far have been encouraging. For instance, based on preliminary observations of patients’ immune responses, Moderna expects to deliver a vaccine with an efficacy above 60 per cent. “That’s good enough. And the reason is, if you get enough people immune, then you will stop transmission, and you will ultimately stop disease,” Zaks says.
When asked about the political pressures to deliver a vaccine as quickly as possible, however, Zaks was adamant that safety and efficacy would never be sacrificed in the process. “This is really personal, we all see what’s happening out there, we all either have had relatives or colleagues affected,” he says. “And we all live in a very different world these days. Half of my team has never set foot in the building, and I’ve yet to meet them and shake their hand. Personally, I have the luxury of working remotely, but that also means I can’t go and visit my mom who’s 80 years old, because she lives in a different country. This is the biggest pressure for my team and everybody in this domain: it’s that sense of responsibility that we can make a difference.”
The testing kit that could enable an era of personalised medicine
“I spent years of my academic life trying to shrink things to the size of a mobile phone,” says Chris Toumazou, CEO of medtech startup DnaNudge, of his early career as an electronics engineer specialising in mobile phone technology. “My honeymoon in the medical arena began when we made one of the world’s first cochlear implants for children who were born deaf. It was clear that if you apply a fraction of this technology to health care, you can make major innovations.”
Toumazou has made a career inventing wearable devices that could monitor vital signs and genetic tests to find if a drug would be metabolised by the patient. “That whole area frustrated me, because prevention was never as important as cure,” Toumazou says. “There were never business models for prevention… I thought, ‘I need to empower the consumer, to democratise DNA testing. I need people to look after their own health’. So rather than personalised health care, let’s make health personal.”
Toumazou decided to focus on a chronic disease epidemic that affects more than 60 per cent of the population: obesity and associated conditions such as diabetes and hypertension. In partnership with Cambridge University, he created a microchip that could detect genetic predispositions to type 2 diabetes, to obesity, to hypertension, to cholesterol – all associated with nutrition.
In 2015, he co-founded DnaNudge, and set up a store in Covent Garden. The company made the DnaBand, a wearable device which enabled customers to scan products at supermarkets and receive information on whether they were compatible with one’s biology. It was franchised to supermarkets such as Waitrose, Sainsbury’s and Tesco. “We’ve got every single macronutrient related to every product that you can find in all the supermarkets in the UK,” Toumazou says. “What this means is that you can effectively shop with your DNA.”
The technology worked in conjunction with a cartridge, called the DnaCartridge, and the NudgeBox, a shoe box-sized testing kit. “On this cartridge, we’ve effectively got this array. And on this array, we’ve got all these little wells. And in the wells, we’re putting the bait, the detectors of genetic errors in the genome,” he explained. The test consists of a simple cheek swab, which is then inserted into the cartridge and placed into the box. There the DNA is extracted from the sample and analysed.
Then, in March 2020, when Covid-19 took hold in the UK, Toumazou considered how his technology might help fight the virus. “I thought, ‘rather than looking at genetic errors, why don’t we look at the RNA (which carries information on the synthesis of proteins) of the virus?’,” he says. “All we had to do was replace the bait for the detection of Covid-19.”
The result was COVID Nudge, a rapid, lab-free Covid-19 test. “The RNA effectively gets sliced, it gets buffered,” explained Toumazou .”Your RNA gets extracted through a membrane. The viral RNA is like fish, and you need to amplify it so it bites the bait.” That process is known as PCR – polymerase chain reaction – and the test delivers a result in almost an hour. “This is a decentralised technology. It doesn’t need a laboratory,” he says.
Toumazou and his team took the COVID Nudge into hospitals to validate its accuracy, testing it on 400 patients at St Mary’s Hospital and the Chelsea and Westminster Hospital in London, and the John Radcliffe Hospital in Oxford. In a paper published in The Lancet, it showed that the test has an average sensitivity of 95 per cent and a specificity close to 100 per cent. The Health Secretary, Matt Hancock, praised DnaNudge as “life-saving” and the UK government ordered 5.8 million tests to be deployed throughout the UK, Ireland and Scotland. “We bulldozed ourselves into the NHS,” he says. “We positioned ourselves in maternity wards, cancer wards, elective surgery wards and in mental-health wards.”
After Covid-19, Toumazou believes that DnaNudge will find applications in areas such as the detection of sexually transmitted diseases and oncology. “It’s unfortunate that it took such a chronic, awful virus, for us to get personalised decentralisation into hospitals,” Toumazou says. “The bottom line is that it’s taken a pandemic to really help an epidemic.”
Inoculating against a global misinformation epidemic
The Vaccine Confidence Project was founded in 2010 by Heidi Larson, an anthropologist at the London School of Hygiene and Tropical Medicine, to track the public’s trust in the effectiveness and the safety of vaccines. “In the past five years, we have developed a Confidence Index to get some measure of the messy emotional sentiment that we know has a tremendous effect on people’s trust and willingness to take vaccines,” Larson says.
Since 2015, the Project has conducted global surveys involving more than 280,000 participants in 149 countries. “One thing we’ve seen is that people’s feelings about vaccines have become far more volatile,” Larson says. “That’s a lot like political opinion polling. They used to be much more stable.” Currently, Europe is the continent where scepticism regarding vaccines is the highest in the world. In Lithuania, for instance, only 19 per cent of the population believes vaccines are safe. On the other end of the spectrum is Finland, with a confidence index of 66 per cent.
Around the world, countries dealing with geopolitical crises and religious extremism – such as Afghanistan, Azerbaijan, Nigeria, Pakistan and Serbia – consistently register very low indices of vaccine confidence. “These are countries where local tensions are influencing people’s trust in authorities, trust in who’s giving and managing immunisation programmes.” Larson says.
She mentions the case of Indonesia, where a drop in vaccine confidence was due to Muslim opposition to vaccines that included porcine gelatine, which is used as a stabiliser. “When religion comes up against scientific evidence, that’s a hard one,” she says. “You can’t really address religious beliefs with a scientific fact. They’re very different sentiments. So we need to find ways to humanise and make the reasons for accepting a vaccine more relevant.”
However, she has also observed that confidence in vaccines is increasing, partially due to the proactive efforts of health organisations around the globe. “I think that’s because there was a bit of a wake-up call to the world that actually, we are starting to lose confidence.”
The situation itself still remains highly uncertain, as there can be significant consequences if even only a small segment of the population rejects vaccination. “Sentiments which reflect negativity and disagreement about acceptance can be the tipping point phenomena”, she says. “You can have a good amount of the population going to take a vaccine, but that little difference – that swing vote, in political terms – can really be the deciding factor of getting to herd immunity or not.”
To Larson, it’s fundamental that public health campaigners don’t forget the emotional responses of the public. “As much as we spend time on the big numbers, we remind ourselves of the emotions of the public,” she says. For example, she mentioned a worldwide campaign in countries such as Denmark, Ireland, Columbia and Japan, in opposition to the HPV vaccine.
It was a movement that started after a number of young girls apparently had adverse reactions. Larson cautions that while many of these cases are still being investigated, the scientific data so far shows no link between the vaccine and any serious side effects. “The experience of the vaccination has created an ‘anxiety reaction’,” she says. “In many of these cases, we still need to deal with that and have empathy with the girls’ responses, and not just focus on the specific scientific evidence of whether the vaccine was related or not. We need to think about the person, about the girls and their mothers.”
In the context of Covid-19, efforts to work with local groups to build confidence around a vaccine are even more urgent. “Some people in public health and immunisation says, ‘Well, fine, we have such a bad virus that’s having such a bad impact on the world, maybe it will change the mind of people who are agitating against vaccines’,” Larson says. Sadly though, she concludes, the exact opposite is occurring. “This isn’t just a misinformation problem.
This is a relationship problem.”
Deep learning at your doctor’s surgery
“After 30 years of reading cardiograms, I can never tell whether it’s from a man or woman, or the age of the person,” says Eric Topol, a cardiologist from Scripps Research in La Jolla, California. “A machine can detect if a person has anaemia, heart muscle dysfunction, or difficult diagnoses like amyloidosis or hypertrophic cardiomyopathy.”
Topol is excited not only about how machines are already better than experts at spotting problems, but how they can discover patterns that experts wouldn’t even notice. “In Japan, instead of relying on the human eyes of the gastroenterologist, they are using a machine to pick up polyps in real time, with machine vision, and detecting whether or not they could be cancerous and whether they should have a biopsy,” he says. “Machines will not replace physicians – but physicians making use of AI will soon replace those not using it.”
Still, these are early days for the application of AI in healthcare. Pearse Keane, a consultant ophthalmologist at Moorfields Eye Hospital in London, has been leading a collaboration between Moorfields and Google’s DeepMind Health. In 2018, it famously published a proof-of-concept paper in the journal Nature showing the first successful AI diagnosis for eye disease. “At the same time, it’s a little bit awkward, because the fact is we haven’t saved the sight of millions of people yet,” Keane says. “And we’re halfway through a journey to getting to that point. The algorithm that we’ve developed isn’t in clinical use at the minute, so we’re trying to implement this now.” Pearse calls that process “code to clinic”.
Keane mentioned the INSIGHT study, which is looking into eye disease and its link to other conditions such as diabetes and dementia. “We are using the eye as a window to the rest of the body,” Keane says. “This is an old idea, but in the past couple of years, with deep learning, we can now look at a retinal photograph and say: ‘This is a woman, she’s 58 years old, she’s not a smoker or a diabetic, her BMI is around 25, and her blood pressure is around 150 over 85’. Now, to me, that’s staggering.”
Of course, there are easier ways to tell if a person is a man or a woman instead of applying AI to an image of a retina. The INSIGHT study is analysing more than three million OCT (optical coherence tomography) scans from around 300,000 patients at Moorfields Hospital taken over a period of 40 years. “We now know, for every person at Moorfields who has had a retinal scan, who’s gone on to develop a heart attack, or who’s gone on to develop Alzheimer’s or other causes of dementia,” he says. “Why we’re excited is that we think that if we can get the appropriate data sets, and by using deep learning, then we can find much more in the back of the eye about the rest of the body’s health.”
AI will also impact patients, helping them to manage their health and self-diagnose health problems. In 2018, Apple developed the first US Food and Drug Administration-approved deep learning algorithm for consumers for atrial fibrillation detection through their Apple smartwatch. Other AI applications exist for self-diagnosis of urinary tract infections, skin cancers and ear infections.
“The application of AI for healthcare and medicine is not just about precision, but also accuracy,” says Topol. “We want to have a medicine that makes fewer mistakes, but that’s not all.” The most important aspect for him, is how AI can promote a stronger human connection between doctor and patient. “Medicine has eroded terribly. It’s a rush job.
We see patients in single digit numbers of minutes. And that’s not enough, you need the gift of time, which AI can give back so people don’t feel so pushed,” he says. What Topol wants is for clinicians to spend more time with patients. For instance, natural language processing can record and transcribe conversations, allowing doctors to spend more time looking at the patient rather than the keyboard. “In the next year this will be the standard,” he believes. “Rather than doctors being data clerks, they will be making eye contact with patients. There’s no algorithm for empathy. That’s a human characteristic that we have to nurture and get it back the way it used to be.”